A comparative study of three rapid diagnostic tests _RDTs_ for

Document Sample
A comparative study of three rapid diagnostic tests _RDTs_ for Powered By Docstoc
					A	
  comparative	
  study	
  of	
  three	
  rapid	
  diagnostic	
  tests	
  
(RDTs)	
  for	
  malaria	
  diagnosis	
  in	
  Oromia	
  Regional	
  
State,	
  Ethiopia	
  
	
  
	
  

	
  
	
  

April	
  2010	
  
	
  

	
  
	
  
	
  

	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                         	
  


Acknowledgments	
  
CareStart	
   and	
   ParaScreen	
   RDTs	
   were	
   generously	
   provided	
   free	
   of	
   charge	
   for	
   this	
   study	
   by	
  
AccessBio	
  USA	
  courtesy	
  of	
  New	
  Millennium	
  World	
  Import	
  and	
  Export	
  plc	
  and	
  Zephyr	
  Biomedicals	
  
India,	
   respectively.	
   ICT	
   Diagnostics	
   South	
   Africa	
   kindly	
   provided	
   reduced-­‐price	
   ICT	
   Combo	
   RDTs	
  
for	
  inclusion	
  in	
  this	
  evaluation.	
  

We	
   would	
   like	
   to	
   thank	
   Oromia	
   Regional	
   Health	
   Bureau	
   for	
   facilitating	
   this	
   study,	
   and	
   the	
  
Ethiopian	
  Health	
  and	
  Nutrition	
  Research	
  Institute	
  for	
  conducting	
  the	
  heat	
  stability	
  testing.	
  Jimma	
  
Zonal	
  Health	
  Department	
  provided	
  valuable	
  assistance	
  in	
  coordinating	
  the	
  field	
  work.	
  We	
  would	
  
also	
  like	
  to	
  thank	
  the	
  staff	
  and	
  patients	
  of	
  Asendabo,	
  Dimtu	
  and	
  Serbo	
  health	
  centers,	
  as	
  well	
  as	
  
Bulbul	
   clinic	
   and	
   Dogosso,	
   Bala	
   Waggo,	
   Waktola,	
   Gudeta	
   Bula,	
   Burka	
   Asendabo,	
   Kejelo,	
   Decha	
  
Nadhi	
  and	
  Kotecha	
  Gibe	
  health	
  posts.	
  

Helen	
   Counihan	
   (Malaria	
   Consortium	
   UK),	
   Richard	
   Reithinger	
   (PMI/Ethiopia),	
   Jan	
   Kolaczinksi	
  
(Malaria	
   Consortium	
   Africa),	
   Ruth	
   Ashton,	
   Takele	
   Kefyalew	
   and	
   Gezahegn	
   Tesfaye	
   (Malaria	
  
Consortium	
   Ethiopia)	
   were	
   involved	
   in	
   study	
   design	
   and	
   planning.	
   Implementation	
   and	
   data	
  
collection	
   was	
   led	
   by	
   TK	
   and	
   GT,	
   with	
   support	
   from	
   Oromia	
   Regional	
   Health	
   Bureau.	
   RA	
  
conducted	
  data	
  analysis.	
  RA	
  prepared	
  the	
  final	
  report	
  with	
  HC,	
  RR,	
  JK	
  and	
  GT.	
  
We	
   thank	
   Bonnie	
   Cundill	
   (London	
   School	
   of	
   Hygiene	
   and	
   Tropical	
   Medicine)	
   for	
   statistical	
   advice,	
  
and	
   Joe	
   Malone,	
   Hiwot	
   Teka	
   (PMI/Ethiopia)	
   and	
   Michael	
   Aidoo	
   (CDC/Atlanta)	
   for	
   valuable	
  
comments	
  on	
  this	
  report.	
  
The	
  work	
  described	
  in	
  this	
  report	
  was	
  supported	
  by	
  the	
  United	
  States	
  Agency	
  for	
  International	
  
Development	
  (Cooperative	
  Agreement	
  663-­‐A-­‐00-­‐09-­‐00404-­‐00).	
  




                                                                           2	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                  	
  


CONTENTS
Abbreviations ........................................................................................................................................ 5	
  
1.	
   Summary .........................................................................................................................................6	
  
2.	
   Introduction ...................................................................................................................................8	
  
      2.1.	
   Malaria	
  in	
  Ethiopia...................................................................................................................8	
  
      2.2.	
   Rapid	
  diagnostic	
  tests ............................................................................................................8	
  
      2.3.	
   Rationale	
  for	
  multi-­‐species	
  RDTs	
  in	
  Ethiopia.........................................................................9	
  
3.	
   Objectives..................................................................................................................................... 10	
  
      3.1.	
   General	
  objective .................................................................................................................. 10	
  
      3.2.	
   Specific	
  objectives ................................................................................................................ 10	
  
4.	
   Study	
  area ..................................................................................................................................... 11	
  
5.	
   Study	
  period..................................................................................................................................12	
  
6.	
   Methods ........................................................................................................................................12	
  
      6.1.	
   Study	
  design...........................................................................................................................12	
  
      6.2.	
   RDT	
  selection	
  for	
  evaluation ................................................................................................12	
  
      6.3.	
   Inclusion	
  and	
  exclusion	
  criteria ........................................................................................... 13	
  
      6.4.	
   Sample	
  size........................................................................................................................... 13	
  
      6.5.	
   Training ................................................................................................................................. 14	
  
      6.6.	
   Data	
  and	
  sample	
  collection ................................................................................................. 14	
  
      6.6.1.	
   Health	
  centers ................................................................................................................... 14	
  
      6.6.2.	
   Health	
  posts ...................................................................................................................... 15	
  
      6.6.3.	
   Quality	
  control	
  measures ................................................................................................. 15	
  
      6.7.	
   Heat	
  stability	
  assessment .................................................................................................... 16	
  
      6.8.	
   Analysis..................................................................................................................................17	
  
      6.8.1.	
   Data	
  entry ...........................................................................................................................17	
  
      6.8.2.	
   RDT	
  validity	
  and	
  reliability	
  compared	
  to	
  microscopy.......................................................17	
  
      6.8.3.	
   RDT	
  ease-­‐of-­‐use ................................................................................................................ 18	
  
      6.8.4.	
   RDT	
  heat	
  stability ............................................................................................................. 18	
  
7.	
   Ethical	
  considerations ................................................................................................................. 18	
  
8.	
   Results .......................................................................................................................................... 19	
  
      8.1.	
   RDT	
  transportation	
  and	
  storage	
  conditions........................................................................ 19	
  
      8.2.	
   Study	
  population	
  description .............................................................................................. 19	
  
      8.3.	
   Microscopy	
  results ............................................................................................................... 19	
  
      8.4.	
   Microscopy	
  vs.	
  RDT	
  results...................................................................................................21	
  
      8.5.	
   RDT	
  ease-­‐of-­‐use....................................................................................................................26	
  
      8.6.	
   Heat	
  stability ........................................................................................................................28	
  
9.	
   Discussion.....................................................................................................................................29	
  
      9.1.	
   RDT	
  performance	
  compared	
  to	
  microscopy .......................................................................29	
  
      9.2.	
   Findings	
  in	
  similar	
  studies ....................................................................................................30	
  
      9.3.	
   Suitability	
  of	
  RDT	
  for	
  use	
  at	
  health	
  posts............................................................................ 31	
  
      9.4.	
   Additional	
  factors	
  determining	
  FMOH	
  RDT	
  choice ............................................................ 32	
  
10.	
   Conclusions	
  and	
  recommendations.......................................................................................... 33	
  
   10.1.	
   Which	
  RDT	
  to	
  choose	
  for	
  Ethiopia ..................................................................................... 33	
  
   10.2.	
   Future	
  activities	
  after	
  selection	
  of	
  a	
  multi-­‐species	
  RDT	
  in	
  Ethiopia ................................. 33	
  
11.	
   References................................................................................................................................... 35	
  
12.	
   Appendices .................................................................................................................................38	
  

                                                                               3	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                	
  


    Heat	
  stability	
  protocol ....................................................................................................................38	
  
    Informed	
  consent............................................................................................................................ 41	
  
    Enrolment	
  questionnaire ................................................................................................................43	
  
    Laboratory	
  results	
  -­‐	
  RDT .................................................................................................................45	
  
    Laboratory	
  results	
  -­‐	
  Microscopy.................................................................................................... 46	
  
    RDT	
  ‘ease	
  of	
  use’	
  questionnaire .....................................................................................................47	
  
               	
  




                                                                              4	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                 	
  


Abbreviations	
  
ACT	
       	
          Artemisinin-­‐based	
  combination	
  treatment	
  
EHNRI	
     	
          Ethiopian	
  Health	
  and	
  Nutrition	
  Research	
  Institute	
  
FIND	
      	
          Foundation	
  for	
  Innovative	
  New	
  Diagnostics	
  
HEW	
       	
          Health	
  Extension	
  Worker	
  
HRP2	
      	
          Histidine-­‐rich	
  protein-­‐2	
  
LLIN	
      	
          Long-­‐lasting	
  insecticide-­‐treated	
  mosquito	
  net	
  
FMoH	
      	
          Federal	
  Ministry	
  of	
  Health	
  
NMCP	
      	
          National	
  Malaria	
  Control	
  Program	
  
NPV	
       	
          Negative	
  predictive	
  value	
  
ORHB	
      	
          Oromia	
  Regional	
  Health	
  Bureau	
  
pLDH	
      	
          Plasmodium	
  lactate	
  dehydrogenase	
  
PPV	
       	
          Positive	
  predictive	
  value	
  
RDT	
       	
          Rapid	
  diagnostic	
  test	
  
SD	
        	
          Standard	
  deviation	
  
WBC	
       	
          White	
  blood	
  cell	
  
WHO	
       	
          World	
  Health	
  Organization	
  




                                                                           5	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                         	
  


1.    Summary	
  
Malaria	
   transmission	
   in	
   Ethiopia	
   is	
   unstable	
   and	
   variable,	
   with	
   both	
   Plasmodium	
   falciparum	
   and	
   P.	
  
vivax	
   causing	
   malaria.	
   The	
   National	
   Malaria	
   Control	
   Program	
   has	
   expressed	
   a	
   commitment	
   to	
  
improving	
   parasitological	
   diagnosis	
   of	
   malaria	
   at	
   all	
   levels	
   of	
   the	
   health	
   system,	
   particularly	
  
community-­‐level	
  (peripheral	
  health	
  facilities,	
  health	
  posts).	
  Because	
  P.	
  falciparum	
  and	
  P.	
  vivax	
  are	
  
endemic	
  at	
  significant	
  levels,	
  there	
  is	
  a	
  need	
  to	
  identify	
  the	
  most	
  appropriate	
  multi-­‐species	
  rapid	
  
diagnostic	
   test	
   (RDT)	
   for	
   use	
   in	
   the	
   Ethiopian	
   context.	
   The	
   RDT	
   selected	
   for	
   use	
   in	
   peripheral	
  
health	
  facilities	
  should	
  be	
  able	
  to	
  detect	
  both	
  P.	
  falciparum	
  and	
  other	
  Plasmodium	
  species,	
  have	
  
high	
  sensitivity	
  and	
  specificity,	
  be	
  easy	
  to	
  use	
  and	
  stable	
  when	
  stored	
  at	
  ambient	
  temperature.	
  
A	
   quantitative	
   study	
   of	
   three	
   multi-­‐species	
   (pf-­‐HRP2	
   and	
   pan-­‐pLDH)	
   RDTs	
   (CareStart®,	
  
ParaScreen®	
   and	
   ICT	
   Combo®)	
   was	
   conducted	
   at	
   three	
   health	
   centers	
   in	
   Jimma	
   zone,	
   Oromia	
  
Regional	
   State,	
   Ethiopia	
   using	
   microscopy	
   as	
   the	
   ‘gold	
   standard’	
   to	
   determine	
   RDT	
   performance.	
  
Gold	
   standard	
   microscopy	
   comprised	
   two	
   blinded	
   readings	
   for	
   all	
   blood	
   films,	
   with	
   further	
  
validation	
   of	
   results	
   at	
   a	
   national	
   reference	
   laboratory.	
   Ease	
   of	
   use	
   was	
   assessed	
   by	
   collecting	
  
opinions	
   of	
   health	
   extension	
   workers	
   (HEW)	
   who	
   used	
   the	
   RDTs	
   in	
   their	
   work	
   at	
   community	
  
health	
  posts	
  after	
  receiving	
  basic	
  training.	
  In	
  addition,	
  heat	
  stability	
  of	
  the	
  RDTs	
  was	
  tested	
  in	
  a	
  
controlled	
   laboratory	
   setting	
   at	
   the	
   WHO/FIND	
   Malaria	
   RDT	
   Lot-­‐Testing	
   Laboratory	
   at	
   the	
  
Ethiopian	
  Health	
  and	
  Nutrition	
  Research	
  Institute	
  (EHNRI)	
  in	
  Addis	
  Ababa.	
  
Between	
   May	
   and	
   July	
   2009,	
   a	
   total	
   of	
   2400	
   patients	
   attending	
   the	
   three	
   health	
   centers	
   were	
  
enrolled	
   in	
   the	
   study;	
   complete	
   data	
   are	
   available	
   from	
   2383	
   study	
   participants.	
   Microscopy	
  
results	
   showed	
   that	
   23.2%	
   of	
   patients	
   had	
   malaria;	
   of	
   which	
   53.8%	
   were	
   caused	
   by	
  P.	
   falciparum	
  
and	
   44.4%	
   by	
   P.	
   vivax;	
   1.8%	
   of	
   confirmed	
   infections	
   were	
   mixed	
   P.	
   falciparum	
   and	
   P.	
   vivax	
  
infections.	
   The	
   sensitivity	
   of	
   all	
   three	
   RDTs	
   in	
   detecting	
   P.	
   falciparum	
   or	
   mixed	
   infection	
   was	
  
exactly	
   the	
   same:	
   85.3%	
   (95%	
   confidence	
   interval	
   80.9-­‐89.1).	
   Specificity	
   of	
   the	
   RDTs	
   was	
   also	
  
similar	
  for	
  detection	
  of	
  P.	
  falciparum	
  or	
  mixed	
  infection:	
  92.2%	
  for	
  ICT	
  Combo	
  (95%	
  CI	
  90.9-­‐93.3),	
  
91.9%	
   for	
   CareStart	
   (95%	
   CI	
   90.7-­‐83.1)	
   and	
   91.5%	
   for	
   ParaScreen	
   (95%	
   CI	
   90.2-­‐92.7).	
   For	
   detecting	
   P.	
  
vivax	
   infection,	
   the	
   sensitivity	
   of	
   both	
   CareStart	
   and	
   ICT	
   Combo	
   was	
   84.9%	
   (95%	
   CI	
   79.8-­‐89.1)	
  
while	
   ParaScreen	
   had	
   a	
   lower	
   sensitivity	
   of	
   82.4%	
   (95%	
   CI	
   77.1-­‐87.0,	
   p=0.02).	
   Specificity	
   of	
   all	
   RDTs	
  
in	
   detecting	
   P.	
   vivax	
   was	
   similar:	
   96.4%	
   for	
   CareStart,	
   96.0%	
   for	
   ICT	
   Combo	
   and	
   95.5%	
   for	
  
ParaScreen.	
  
HEWs	
   were	
   not	
   unanimous	
   in	
   their	
   preferred	
   RDT	
   test.	
   When	
   summarizing	
   scores	
   assigned	
   to	
  
various	
   aspects	
   of	
   each	
   RDT,	
   CareStart	
   had	
   the	
   highest	
   score,	
   followed	
   by	
   ParaScreen	
   and	
   ICT	
  
Combo.	
  Many	
  HEWs	
  commented	
  that	
  the	
  ‘lab	
  in	
  a	
  pack’	
  style	
  of	
  the	
  CareStart	
  RDT	
  test	
  was	
  very	
  
useful	
  for	
  their	
  work	
  in	
  the	
  community,	
  which	
  involves	
  moving	
  from	
  house	
  to	
  house.	
  Each	
  RDT	
  
has	
  a	
  different	
  device	
  used	
  to	
  collect	
  the	
  blood	
  sample	
  (i.e.	
  capillary	
  tube,	
  pipette	
  or	
  loop),	
  but	
  
HEWs	
  did	
  not	
  favor	
  any	
  device	
  over	
  another.	
  The	
  most	
  relevant	
  comment	
  from	
  HEWs	
  in	
  terms	
  of	
  
RDT	
  ease	
  of	
  use	
  was	
  the	
  presence	
  of	
  markings	
  on	
  the	
  RDT	
  cassette	
  to	
  assist	
  with	
  interpretation	
  
of	
   results.	
   ICT	
   Combo	
   has	
   no	
   individual	
   markings	
   for	
   the	
   Pan	
   and	
   Pf	
   bands,	
   while	
   CareStart	
   labels	
  
the	
  bands	
  with	
  numbers,	
  and	
  ParaScreen	
  has	
  the	
  most	
  unambiguous	
  markings	
  with	
  ‘Pf’	
  and	
  ‘Pan’	
  
labeling	
   the	
   bands.	
   The	
   lack	
   of	
   markings	
   on	
   ICT	
   Combo	
   makes	
   it	
   very	
   difficult	
   for	
   HEWs	
   to	
  
correctly	
  interpret	
  positive	
  results.	
  
ParaScreen	
   and	
   CareStart	
   HRP2	
   and	
   pLDH	
   bands	
   were	
   able	
   to	
   detect	
   high	
   and	
   low	
   intensity	
   P.	
  
falciparum	
  and	
  P.	
  vivax	
  infections	
  after	
  90	
  days	
  of	
  heat	
  stability	
  testing	
  at	
  35°C	
  and	
  45°C,	
  and	
  after	
  



                                                                                  6	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                     	
  


72	
   hours	
   at	
   60°C.	
   The	
   ICT	
   Combo	
   pLDH	
   band	
   failed	
   the	
   testing	
   procedure	
   after	
   60	
   days	
   at	
   35°C	
  
and	
   60	
   days	
   at	
   45°C,	
   when	
   P.	
   vivax	
   infection	
   of	
   200	
   parasites	
   per	
   µl	
   was	
   not	
   detected.	
  
Furthermore,	
   at	
   60°C	
   the	
   HRP2	
   band	
   of	
   ICT	
   Combo	
   failed	
   to	
   detect	
   high	
   intensity	
   P.	
   falciparum	
  
infection	
   after	
   12	
   hours,	
   and	
   the	
   pLDH	
   band	
   failed	
   to	
   detect	
   low	
   intensity	
   P.	
   vivax	
   when	
   tested	
  
after	
  72	
  hours	
  at	
  60°C.	
  All	
  three	
  RDT	
  tests	
  showed	
  weak	
  pLDH	
  bands	
  throughout	
  testing	
  with	
  P.	
  
vivax	
   at	
   200	
   parasites	
   per	
   µl	
   at	
   all	
   temperatures.	
   ICT	
   pLDH	
   bands	
   were	
   consistently	
   graded	
   as	
  
weaker	
   than	
   those	
   of	
   CareStart	
   and	
   ParaScreen	
   in	
   detecting	
   low	
   intensity	
   P.	
   vivax	
   at	
   all	
   three	
  
temperature	
  evaluations.	
  

When	
   comparing	
   our	
   results	
   to	
   other	
   findings,	
   particularly	
   the	
   recent	
   WHO/FIND	
   first	
   round	
   of	
  
RDT	
   product	
   evaluation,	
   CareStart	
   has	
   shown	
   consistently	
   good	
   performance	
   and	
   reliability,	
  
whilst	
   ParaScreen	
   showed	
   much	
   lower	
   sensitivity	
   in	
   the	
   WHO/FIND	
   testing	
   than	
   in	
   the	
   current	
  
study.	
   ICT	
   Combo	
   showed	
   higher	
   sensitivity	
   in	
   the	
   WHO/FIND	
   evaluations,	
   particularly	
   to	
   low	
  
density	
  P.	
  falciparum	
  infection.	
  
When	
   considering	
   the	
   performance	
   of	
   the	
   three	
   RDTs	
   compared	
   to	
   microscopy,	
   their	
   heat	
  
stability	
   and	
   ease	
   of	
   use	
   by	
   HEWs,	
   as	
   well	
   as	
   consistency	
   of	
   results	
   between	
   published	
  
evaluations,	
   we	
   believe	
   that	
   currently	
   CareStart	
   is	
   the	
   most	
   appropriate	
   RDT	
   for	
   the	
   Ethiopian	
  
context.	
  	
  




                                                                                7	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                    	
  


2.      Introduction	
  	
  
2.1. Malaria	
  in	
  Ethiopia	
  
Approximately	
  52	
  million	
  people	
  in	
  Ethiopia	
  are	
  considered	
  to	
  be	
  at	
  risk	
  of	
  malaria	
  [1].	
  The	
  disease	
  
is	
   reported	
   to	
   be	
   the	
   leading	
   cause	
   of	
   morbidity	
   and	
   mortality	
   accounting	
   for	
   about	
   12%	
   of	
   the	
  
total	
   outpatient	
   visits	
   and	
   9.9%	
   of	
   the	
   total	
   admissions	
   in	
   2007	
   [2].	
   	
   Approximately	
   4-­‐6	
   million	
  
clinical	
   malaria	
   cases	
   are	
   reported	
   annually	
   across	
   all	
   health	
   facilities	
   in	
   the	
   country.	
   The	
   actual	
  
number	
   of	
   malaria	
   cases	
   is	
   estimated	
   to	
   be	
   as	
   high	
   as	
   10-­‐15	
   million	
   [3].	
   The	
   major	
   Plasmodium	
  
species	
  causing	
  malaria	
  in	
  Ethiopia	
  are	
  P.	
  falciparum	
  and	
  P.	
  vivax.	
  
Malaria	
  transmission	
  is	
  unstable	
  and	
  seasonal,	
  linked	
  to	
  environmental	
  variables	
  such	
  as	
  altitude	
  
and	
   rainfall.	
   Epidemics	
   due	
   to	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax	
   occur	
   periodically	
   and	
   historically	
  
have	
   led	
   to	
   significant	
   morbidity	
   and	
   mortality	
   in	
   Ethiopia	
   [4].	
   To	
   reduce	
   malaria	
   morbidity	
   and	
  
mortality,	
  prompt	
  and	
  accurate	
  diagnosis	
  must	
  be	
  available	
  to	
  all	
  at-­‐risk	
  populations.	
  Particularly	
  
in	
  low	
  transmission	
  settings,	
  where	
  a	
  large	
  proportion	
  of	
  febrile	
  illness	
  is	
  due	
  to	
  infections	
  other	
  
than	
  malaria,	
  it	
  is	
  essential	
  that	
  all	
  febrile	
  cases	
  have	
  access	
  to	
  parasitological	
  diagnosis.	
  	
  
The	
   Federal	
   Ministry	
   of	
   Health	
   (FMOH)	
   is	
   tackling	
   malaria	
   using	
   several	
   strategies:	
   early	
   case	
  
detection	
  and	
  effective	
  prompt	
  treatment,	
  vector	
  control	
  including	
  indoor	
  residual	
  spraying	
  and	
  
distribution	
   of	
   long-­‐lasting	
   insecticide-­‐treated	
   bed	
   nets	
   (LLINs),	
   education	
   and	
   behavior	
   change	
  
of	
  the	
  at-­‐risk	
  populations,	
  and	
  epidemic	
  prevention	
  and	
  control	
  [3].	
  	
  	
  
The	
  national	
  malaria	
  drug	
  policy	
  recommends	
  the	
  artemisinin-­‐based	
  combination	
  therapy	
  (ACT)	
  
artemether-­‐lumefantrine	
  (CoArtem®)	
  for	
  P.	
  falciparum	
  and	
  chloroquine	
  for	
  treatment	
  of	
  P.	
  vivax	
  
infection.	
   Intravenous	
   quinine	
   therapy	
   is	
   recommended	
   for	
   treatment	
   of	
   severe	
   malaria,	
   and	
   oral	
  
quinine	
  for	
  uncomplicated	
  malaria	
  in	
  pregnant	
  women	
  during	
  the	
  first	
  trimester	
  and	
  infants	
  less	
  
than	
   5	
   kilograms	
   weight.	
   Currently,	
   as	
   per	
   national	
   guidelines,	
   uncomplicated	
   mixed	
   infections	
  
should	
   receive	
   both	
   CoArtem	
   and	
   chloroquine.	
   There	
   is	
   considerable	
   pressure	
   to	
   minimize	
  
wastage	
  of	
  ACT	
  drugs,	
  both	
  to	
  reduce	
  costs	
  incurred	
  by	
  this	
  relatively	
  expensive	
  treatment	
  but	
  
more	
   importantly	
   to	
   prevent	
   unnecessary	
   exposure	
   of	
   Plasmodium	
   parasites	
   that	
   may	
   accelerate	
  
development	
   of	
   drug	
   resistance,	
   as	
   has	
   been	
   shown	
   with	
   antibiotic	
   exposure	
   leading	
   to	
  
development	
   of	
   resistant	
   bacteria	
   [5].	
   Reduction	
   of	
   antimalarial	
   drug	
   consumption	
   and	
  
application	
  of	
  different	
  treatment	
  regimens	
  for	
  P.	
  falciparum	
  and	
  other	
  Plasmodium	
  species	
  can	
  
only	
   be	
   effectively	
   implemented	
   when	
   parasitological	
   diagnosis	
   of	
   malaria	
   is	
   routinely	
   provided	
  
at	
  all	
  levels	
  of	
  the	
  health	
  care	
  delivery	
  system.	
  	
  
Where	
  there	
  are	
  laboratory	
  services	
  at	
  hospitals	
  and	
  health	
  centers,	
  high-­‐quality	
  light	
  microscopy	
  
of	
  Giemsa-­‐stained	
  thick	
  blood	
  films	
  is	
  expected	
  to	
  remain	
  the	
  primary	
  diagnostic	
  tool	
  for	
  malaria	
  
in	
   Ethiopia.	
   Where	
   these	
   services	
   do	
   not	
   exist,	
   RDTs	
   for	
   detection	
   of	
   P.	
   falciparum	
   (ParaCheck-­‐
Pf®,	
  Orchid	
  Biomedical	
  Systems,	
  Goa,	
  India)	
  have	
  been	
  used	
  by	
  HEWs	
  for	
  a	
  number	
  of	
  years	
  [6-­‐8].	
  
Currently,	
   there	
   is	
   no	
   capacity	
   for	
   parasitological	
   diagnosis	
   of	
   non-­‐falciparum	
   malaria	
   at	
   health	
  
facilities	
  without	
  laboratory	
  services.	
  Therefore,	
  febrile	
  patients	
  testing	
  negative	
  with	
  ParaCheck	
  
are	
   often	
   presumptively	
   treated	
   for	
   malaria,	
   leading	
   to	
   drug	
   wastage	
   and	
   under-­‐diagnosis	
   of	
  
other	
  causes	
  of	
  febrile	
  illness	
  (e.g.	
  pneumonia).	
  


2.2. Rapid	
  diagnostic	
  tests	
  
RDTs	
   detect	
   Plasmodium-­‐derived	
   antigens	
   in	
   small	
   volume	
   whole	
   blood	
   samples	
   from	
   infected	
  
individuals.	
   The	
   tests	
   include	
   a	
   nitrocellulose	
   strip	
   containing	
   bound	
   antibodies	
   that	
   bind	
   to	
  


                                                                                8	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                         	
  


various	
  parasite	
  antigens.	
  Binding	
  of	
  antigen	
  to	
  antibody	
  results	
  in	
  coloration	
  of	
  a	
  specific	
  band	
  
on	
  the	
  nitrocellulose	
  strip,	
  indicating	
  presence	
  of	
  the	
  specific	
  antigen	
  in	
  patient	
  blood.	
  RDTs	
  are	
  
available	
   for	
   detection	
   of	
   a	
   range	
   of	
   Plasmodium	
   antigens:	
   i)	
   histidine-­‐rich	
   protein-­‐2	
   (HRP2)	
   is	
  
specific	
  to	
  P.	
  falciparum,	
  ii)	
  Plasmodium	
  lactate	
  dehydrogenase	
  (pLDH)	
  can	
  be	
  species-­‐specific	
  or	
  
pan-­‐specific,	
   and	
   iii)	
   aldolase	
   is	
   specific	
   to	
   all	
   four	
   major	
   Plasmodium	
   species	
   (P.	
   falciparum,	
   P.	
  
vivax,	
  P.	
  ovale	
  and	
  P.	
  malariae).	
  	
  
Persistence	
   of	
   these	
   antigens	
   in	
   the	
   bloodstream	
   after	
   treatment	
   can	
   affect	
   interpretation	
   of	
  
RDT	
   results.	
   HRP2	
   is	
   known	
   to	
   persist	
   after	
   treatment	
   for	
   four	
   weeks	
   or	
   more,	
   while	
   pLDH	
  
remains	
  detectable	
  for	
  approximately	
  one	
  week	
  [9-­‐11].	
  False	
  positive	
  results	
  are	
  therefore	
  likely	
  
with	
   HRP2-­‐detecting	
   RDTs	
   in	
   patients	
   with	
   recently	
   cleared	
   parasitemia.	
   This	
   emphasizes	
   the	
  
importance	
   of	
   considering	
   RDT	
   results	
   in	
   the	
   context	
   of	
   clinical	
   signs	
   and	
   medical	
   history,	
  
particularly	
  in	
  areas	
  of	
  intense	
  transmission.	
  	
  
A	
   large	
   choice	
   of	
   RDTs	
   are	
   available	
   in	
   various	
   formats	
   (e.g.	
   cassette,	
   dipstick	
   and	
   card-­‐based)	
  
and	
  with	
  a	
  range	
  of	
  antibody	
  combinations.	
  A	
  total	
  of	
  41	
  RDTs	
  from	
  21	
  different	
  manufacturers	
  
were	
   evaluated	
   in	
   the	
   first	
   round	
   of	
   WHO	
   product	
   testing	
   [12].	
   From	
   this	
   wide	
   array	
   of	
   RDTs	
  
policy	
   makers	
   must	
   select	
   the	
   most	
   appropriate	
   test	
   for	
   their	
   context.	
   Factors	
   contributing	
   to	
  
this	
   decision	
   include:	
   i)	
   the	
   Plasmodium	
   species	
   endemic	
   in	
   the	
   area	
   of	
   interest,	
   ii)	
   minimum	
  
acceptable	
   parasite	
   load	
   at	
   which	
   infection	
   should	
   be	
   detected	
   and	
   treated,	
   iii)	
   the	
   treatment	
  
policy,	
   i.e.	
   whether	
   species	
   differentiation	
   is	
   required,	
   iv)	
   heat	
   stability	
   of	
   tests,	
   v)	
   training	
   and	
  
education	
  level	
  of	
  RDT	
  users,	
  vi)	
  shelf	
  life,	
  and	
  vii)	
  cost.	
  	
  RDT	
  performance	
  is	
  constantly	
  evolving	
  
as	
   research	
   and	
   development	
   progresses.	
   	
   National	
   policy	
   makers	
   therefore	
   need	
   to	
   regularly	
  
reassess	
  RDT	
  test	
  choices	
  to	
  ensure	
  that	
  the	
  most	
  cost-­‐effective	
  product	
  is	
  procured.	
  


2.3. Rationale	
  for	
  multi-­‐species	
  RDTs	
  in	
  Ethiopia	
  
As	
   the	
   FMOH	
   continues	
   to	
   scale	
   up	
   malaria	
   interventions,	
   the	
   need	
   to	
   accurately	
   diagnose	
   and	
  
differentiate	
  P.	
  falciparum	
  and	
  P.	
  vivax	
  is	
  becoming	
  more	
  important	
  so	
  that	
  all	
  febrile	
  cases	
  can	
  be	
  
correctly	
  managed,	
  further	
  reducing	
  the	
  reservoir	
  of	
  infection,	
  and	
  obtaining	
  reliable	
  information	
  
on	
  endemicity	
  patterns	
  and	
  the	
  impact	
  of	
  control	
  measures	
  over	
  time.	
  	
  
Performance	
   of	
   the	
   P.	
   falciparum-­‐detecting	
   ParaCheck-­‐Pf	
   RDT	
   at	
   health	
   post	
   level	
   remains	
  
variable,	
   primarily	
   due	
   to	
   the	
   possibility	
   of	
   fever	
   due	
   to	
   P.	
   vivax	
   infection	
   which	
   often	
   leads	
   to	
  
presumptive	
   treatment	
   with	
   ACTs	
   regardless	
   of	
   RDT	
   result.	
   To	
   improve	
   management	
   of	
   febrile	
  
illness	
   there	
   is	
   an	
   urgent	
   need	
   to	
   provide	
   HEWs	
   with	
   a	
   reliable	
   and	
   user-­‐friendly	
   tool	
   to	
   diagnose	
  
both	
  falciparum	
  and	
  non-­‐falciparum	
  malaria	
  and	
  to	
  ensure	
  that	
  they	
  have	
  access	
  to	
  chloroquine	
  
for	
  treatment	
  of	
  P.	
  vivax.	
  Furthermore,	
  use	
  of	
  RDTs	
  has	
  been	
  shown	
  to	
  improve	
  case	
  treatment	
  
and	
  health	
  outcome	
  for	
  all	
  patients	
  with	
  fever	
  compared	
  to	
  clinical	
  diagnosis	
  alone	
  [13].	
  However	
  
roll-­‐out	
   of	
   multi-­‐species	
   RDTs	
   should	
   be	
   accompanied	
   by	
   supporting	
   research	
   examining	
   case	
  
management	
   practices	
   at	
   the	
   health	
   post	
   to	
   ensure	
   that	
   results	
   from	
   the	
   RDT	
   are	
   properly	
  
interpreted	
  and	
  acted	
  upon	
  within	
  the	
  treatment	
  algorithm.	
  

Important	
  progress	
  has	
  been	
  seen	
  in	
  the	
  improvement	
  of	
  RDTs	
  to	
  detect	
  non-­‐falciparum	
  malaria,	
  
based	
   on	
   pLDH.	
   While	
   these	
   tests	
   were	
   originally	
   less	
   reliable	
   than	
   HRP2-­‐based	
   RDTs	
   [14-­‐17],	
  
numerous	
   studies	
   have	
   shown	
   sensitivity	
   and	
   specificity	
   results	
   approaching	
   those	
   of	
   falciparum-­‐
only	
  tests	
  [18-­‐21].	
  RDTs	
  able	
  to	
  detect	
  a	
  combination	
  of	
  antigens	
  are	
  now	
  widely	
  available	
  at	
  low	
  
cost	
   and	
   are	
   easy	
   to	
   use.	
   The	
   FMOH	
   is	
   eager	
   to	
   utilize	
   this	
   new	
   technology	
   and	
   replace	
   the	
  



                                                                                  9	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                             	
  


currently	
   used	
   single-­‐species	
   RDT	
   with	
   an	
   appropriate	
   and	
   reliable	
   multi-­‐species	
   product	
   to	
  
improve	
  diagnosis	
  and	
  case	
  management,	
  particularly	
  at	
  community-­‐level	
  health	
  facilities.	
  	
  

The	
  rapid	
  results	
  provided	
  by	
  RDTs	
  and	
  their	
  ease	
  of	
  use	
  and	
  interpretation	
  after	
  minimal	
  training	
  
make	
  them	
   ideal	
  for	
   use	
  in	
  epidemic	
  situations.	
   Since	
   Ethiopia	
   periodically	
   experiences	
   epidemics	
  
of	
  both	
  P.	
  falciparum	
  and	
  P.	
  vivax	
  infection,	
  availability	
  of	
  a	
  reliable	
  multi-­‐species	
  RDT	
  will	
  be	
  of	
  
great	
  benefit	
  in	
  diagnosing	
  and	
  managing	
  cases	
  in	
  an	
  epidemic	
  context.	
  
Rapid	
  diagnostic	
  tests	
  are	
  increasingly	
  used	
  in	
  epidemiological	
  surveys,	
  since	
  they	
  are	
  particularly	
  
suited	
   to	
   field	
   use	
   where	
   resources	
   are	
   limited.	
   A	
   multi-­‐species	
   RDT	
   for	
   Ethiopia	
   could	
   become	
  
part	
  of	
  ongoing	
  monitoring	
  and	
  evaluation	
  activities,	
  including	
  population	
  surveys,	
  to	
  assess	
  the	
  
impact	
  of	
  the	
  various	
  interventions	
  on	
  malaria	
  transmission.	
  Additionally,	
  implementing	
  a	
  simple	
  
reporting	
   system	
   at	
   health	
   post	
   level	
   to	
   record	
   results	
   of	
   RDTs	
   would	
   provide	
   valuable	
  
information	
  on	
  the	
  incidence	
  and	
  species	
  distribution	
  of	
  malaria	
  across	
  the	
  country.	
  	
  

	
  
3.     Objectives	
  
3.1. General	
  objective	
  
To	
   determine	
   the	
   performance	
   of	
   CareStart	
   Pan/Pf,	
   ParaScreen	
   Pan/Pf	
   and	
   ICT	
   Combo	
   Pan/Pf	
  
under	
   both	
   laboratory	
   controlled	
   and	
   operational	
   conditions,	
   including	
   the	
   identification	
   of	
  
factors	
  that	
  affect	
  their	
  use	
  under	
  routine	
  field	
  conditions.	
  


3.2. Specific	
  objectives	
  
          • To	
   measure	
   the	
   sensitivity	
   and	
   specificity	
   of	
   the	
   tested	
   RDTs	
   to	
   detect	
   P.	
   falciparum	
  
            and	
  P.	
  vivax	
  infection	
  as	
  compared	
  to	
  high	
  quality	
  microscopy.	
  
          • To	
   determine	
   the	
   positive	
   and	
   negative	
   predictive	
   values	
   of	
   the	
   tested	
   RDTs	
   in	
  
            detecting	
  P.	
  falciparum	
  and	
  P.	
  vivax	
  infection.	
  
          • To	
   determine	
   the	
   heat	
   stability	
   of	
   the	
   tested	
   RDTs	
   and	
   how	
   it	
   affects	
   RDT	
  
            performance.	
  
          • To	
   evaluate	
   the	
   ease-­‐of-­‐use	
   and	
   acceptability	
   of	
   tested	
   RDTs	
   under	
   routine	
   field	
  
            conditions.	
  




                                                                           10	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                     	
  


4.      Study	
  area	
  	
  
The	
   study	
   was	
   conducted	
   in	
   three	
   health	
   centers	
   in	
   Oromia	
   Regional	
   State,	
   Ethiopia.	
   Health	
  
centers	
  were	
  selected	
  in	
  collaboration	
  with	
  Oromia	
  Regional	
  Health	
  Bureau	
  (ORHB)	
  considering:	
  
i)	
  the	
  number	
  of	
  patients	
  presenting	
  with	
  fever,	
  ii)	
  the	
  proportion	
  of	
  fevers	
  confirmed	
  as	
  malaria,	
  
iii)	
   the	
   endemicity	
   of	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax,	
   and	
   iv)	
   the	
   technical	
   capacity	
   of	
   health	
  
center	
  staff.	
  Three	
  sites	
  in	
  Jimma	
  zone	
  were	
  selected:	
  Asendabo,	
  Serbo	
  and	
  Dimtu	
  health	
  centers	
  
in	
  Omo	
  Nada,	
  Kersa	
  and	
  Tiro	
  Afeta	
  woredas,	
  respectively	
  (see	
  Figure	
  1).	
  	
  
Three	
  functioning	
  and	
  accessible	
  health	
  posts	
  from	
  each	
  of	
  these	
  woredas	
  were	
  chosen	
  for	
  the	
  
ease	
   of	
   use	
   assessment.	
   In	
   Omo	
   Nada	
   woreda	
   Waktola,	
   Burka	
   Asendabo	
   and	
   Gudeta	
   Bula	
   health	
  
posts	
   were	
   included,	
   in	
   Kersa	
   woreda	
   Bulbul	
   clinic	
   (staffed	
   by	
   two	
   nurses	
   rather	
   than	
   HEWs)	
   and	
  
Dogosso	
  and	
  Bala	
  Waggo	
  health	
  posts	
  were	
  included,	
  while	
  from	
  Tiro	
  Afeta	
  woreda	
  Decha	
  Nadhi,	
  
Kejelo	
  and	
  Kotecha	
  Gibe	
  health	
  posts	
  participated.	
  




Figure	
   1:	
   Location	
   of	
   Oromia	
   Regional	
   State	
   (solid	
   shading)	
   and	
   Jimma	
   zone	
   (vertical	
   shading)	
   within	
  
Ethiopia.	
   Detail	
   of	
   Jimma	
   zone	
   indicates	
   location	
   of	
   the	
   three	
   selected	
   health	
   centers	
   (Serbo,	
   Asendabo	
  
and	
  Dimtu)	
  and	
  their	
  administrative	
  woredas	
  (Kersa,	
  Omo	
  Nada	
  and	
  Tiro	
  Afeta,	
  respectively).	
  


Jimma	
   zone	
   is	
   300	
   km	
   southwest	
   of	
   Addis	
   Ababa,	
   and	
   has	
   an	
   elevation	
   of	
   approximately	
   1780	
  
meters	
  above	
  sea	
  level.	
  Malaria	
  transmission	
  in	
  Oromia	
  is	
  perennial,	
  but	
  peaks	
  from	
  April	
  to	
  May	
  
and	
   from	
   October	
   to	
   December,	
   after	
   the	
   seasonal	
   rains.	
   However,	
   unlike	
   most	
   of	
   Oromia	
  
Region,	
  malaria	
  transmission	
  in	
  Jimma	
  zone	
  is	
  not	
  clearly	
  defined	
  as	
  a	
  short	
  and	
  long	
  transmission	
  
season.	
  	
  

	
  
	
  



                                                                               11	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                	
  


5.      Study	
  period	
  
Data	
   collection	
   from	
   health	
   centers	
   and	
   health	
   posts	
   began	
   in	
   April	
   2009	
   after	
   the	
   seasonal	
   rains	
  
began	
   and	
   an	
   increase	
   in	
   malaria	
   cases	
   was	
   seen	
   at	
   health	
   centers	
   in	
   the	
   area.	
   Patient	
   enrolment	
  
was	
  staggered	
  at	
  each	
  health	
  center	
  in	
  order	
  that	
  close	
  supervision	
  could	
  be	
  provided	
  in	
  the	
  early	
  
stages	
   of	
   data	
   collection.	
   Data	
   collection	
   concluded	
   in	
   August	
   2009,	
   with	
   malaria	
   cases	
   identified	
  
consistently	
  throughout	
  this	
  period	
  at	
  all	
  three	
  study	
  sites.	
  
Heat	
   stability	
   assessments	
   were	
   conducted	
   at	
   the	
   Ethiopian	
   Health	
   and	
   Nutrition	
   Research	
  
Institute	
   (EHNRI)	
   located	
   in	
   Addis	
   Ababa,	
   which	
   is	
   a	
   designated	
   regional	
   reference	
   laboratory	
   for	
  
WHO/FIND.	
  Testing	
  began	
  in	
  late	
  August	
  2009	
  and	
  concluded	
  in	
  December	
  2009.	
  



6.      Methods	
  
6.1. Study	
  design	
  
This	
   study	
   has	
   three	
   related	
   components:	
   i)	
   a	
   health	
   center-­‐based	
   longitudinal	
   study	
   where	
   all	
  
consenting	
  febrile	
  patients	
  presenting	
  at	
  the	
  health	
  center	
  were	
  enrolled	
  over	
  a	
  period	
  of	
  several	
  
months;	
   ii)	
   a	
   qualitative	
   peripheral	
   health	
   facility-­‐based	
   study	
   to	
   assess	
   RDT	
   ease	
   of	
   use;	
   iii)	
  
laboratory-­‐based	
  RDT	
  lot	
  testing	
  and	
  heat	
  stability	
  assessment.	
  


6.2. RDT	
  selection	
  for	
  evaluation	
  
The	
   present	
   study	
   investigated	
   several	
   aspects	
   of	
   malaria	
   RDT	
   performance	
   crucial	
   to	
   identifying	
  
the	
  most	
  appropriate	
  product	
  for	
  routine	
  use	
  in	
  Oromia	
  Regional	
  State,	
  Ethiopia.	
  The	
  ideal	
  test	
  
should	
   fulfill	
   a	
   number	
   of	
   criteria:	
   i)	
   high	
   sensitivity	
   and	
   specificity	
   in	
   detecting	
   both	
   P.	
   falciparum	
  
and	
   P.	
   vivax	
   in	
   low	
   transmission	
   settings,	
   ii)	
   easy	
   to	
   use	
   and	
   interpret	
   by	
   HEWs	
   with	
   minimal	
  
formal	
  training,	
  and	
  iii)	
  stable	
  at	
  high	
  temperatures	
  and	
  humidities.	
  	
  
The	
   RDTs	
   chosen	
   for	
   evaluation	
   were	
   selected	
   to	
   allow	
   comparison	
   with	
   results	
   from	
   similar	
  
studies	
   being	
   conducted	
   in	
   other	
   regional	
   states	
   of	
   Ethiopia.	
   Performance	
   of	
   RDTs	
   in	
   other	
  
published	
   studies	
   contributed	
   to	
   selection.	
   Furthermore,	
   the	
   NMCP	
   had	
   stated	
   its	
   intention	
   to	
  
move	
   towards	
   procurement	
   of	
   multi-­‐species	
   RDTs;	
   only	
   multi-­‐species	
   RDTs	
   were	
   therefore	
  
included	
   in	
   the	
   present	
   study.	
   All	
   of	
   the	
   selected	
   products	
   were	
   cassette	
   tests	
   that	
   detect	
   P.	
  
falciparum	
  and	
  other	
  Plasmodium	
  species	
  infection	
  by	
  means	
  of	
  HRP2	
  and	
  pLDH,	
  respectively.	
  	
  
•     CareStart	
  Pan/Pf	
  (Access	
  Bio,	
  USA):	
  A	
  plastic	
  capillary	
  tube	
  is	
  used	
  to	
  collect	
  5	
  µl	
  of	
  blood	
  and	
  
      introduce	
  it	
  into	
  a	
  sample	
  well.	
  Buffer	
  (3	
  drops)	
  is	
  then	
  applied	
  to	
  a	
  buffer	
  well,	
  and	
  results	
  
      are	
   read	
   after	
   20	
   minutes.	
   Each	
   test	
   is	
   individually	
   packaged	
   with	
   desiccant,	
   alcohol	
   swab,	
  
      lancet,	
   plastic	
   capillary	
   tube,	
   buffer	
   and	
   instruction	
   leaflet	
   included	
   (Figure	
   2).	
   The	
  
      recommended	
  storage	
  temperature	
  ranges	
  from	
  4°C	
  to	
  30°C.	
  
•     ParaScreen	
  Pan/Pf	
  (Zephyr	
  Biomedicals,	
  India):	
  5	
  µl	
  of	
  blood	
  is	
  collected	
  by	
  means	
  of	
  a	
  plastic	
  
      sample	
   loop	
   and	
   applied	
   to	
   a	
   sample	
   well.	
   Buffer	
   (4	
   drops)	
   is	
   then	
   added	
   to	
   a	
   buffer	
   well	
   and	
  
      results	
  are	
  read	
  after	
  15	
  minutes.	
  Tests	
  are	
  individually	
  packaged	
  with	
  desiccant	
  and	
  a	
  sample	
  
      loop.	
   Buffer	
   is	
   provided	
   in	
   one	
   dropper	
   bottle	
   per	
   box	
   of	
   25	
   tests	
   (Figure	
   2).	
   Lancets	
   and	
  
      alcohol	
  swabs	
  are	
  packed	
  separately	
  in	
  the	
  box	
  of	
  25,	
  along	
  with	
  an	
  instruction	
  leaflet.	
  Brief	
  
      instructions	
  and	
  guide	
  to	
  interpreting	
  results	
  are	
  printed	
  on	
  the	
  packaging	
  of	
  each	
  test.	
  The	
  
      recommended	
  storage	
  temperature	
  ranges	
  from	
  4°C	
  to	
  30°C.	
  



                                                                                    12	
  
 Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                     	
  


 •      ICT	
  Combo	
  Pan/Pf	
  (R&R,	
  South	
  Africa):	
  A	
  plastic	
  sample	
  pipette	
  is	
  used	
  to	
  collect	
  5	
  µl	
  of	
  blood	
  
        and	
  apply	
  to	
  sample	
  well.	
  Buffer	
  (5	
  drops)	
  is	
  then	
  added	
  to	
  the	
  buffer	
  well	
  and	
  results	
  read	
  
        after	
   15	
   minutes.	
   Tests	
   are	
   individually	
   packaged	
   with	
   desiccant.	
   Lancets,	
   alcohol	
   swabs,	
  
        sample	
   pipettes	
   and	
   instruction	
   leaflet	
   are	
   provided	
   in	
   each	
   box	
   of	
   25	
   tests,	
   together	
   with	
  
        one	
  dropper	
  bottle	
  of	
  buffer	
  (Figure	
  2).	
  The	
  recommended	
  storage	
  temperature	
  ranges	
  from	
  
        4°C	
  to	
  37°C.	
  	
  
         	
  
 	
  




Figure	
  2:	
  Components	
  included	
  in	
  each	
  RDT	
  test.	
  Left	
  to	
  right:	
  CareStart,	
  ParaScreen,	
  ICT	
  Combo	
  
 	
  


 6.3. Inclusion	
  and	
  exclusion	
  criteria	
  
 Patients	
  attending	
  the	
  outpatient	
  department	
  of	
  the	
  three	
  health	
  centers	
  with	
  current	
  fever	
  or	
  
 recent	
   history	
   of	
   fever	
   were	
   considered	
   for	
   inclusion	
   in	
   the	
   study.	
   Specifically,	
   subjects	
   should	
  
 have	
   symptoms	
   indicative	
   of	
   uncomplicated	
   malaria:	
   axillary	
   temperature	
   >37.5°C	
   or	
   report	
   of	
  
 fever	
   in	
   the	
   previous	
   48	
   hours,	
   be	
   over	
   six	
   months	
   of	
   age,	
   and	
   have	
   no	
   other	
   form	
   of	
   life-­‐
 threatening	
  disease.	
  	
  
 Exclusion	
   criteria	
   were	
   refusal	
   to	
   participate,	
   age	
   less	
   than	
   six	
   months,	
   symptoms	
   of	
   severe	
  
 malaria	
  or	
  some	
  other	
  life-­‐threatening	
  disease.	
  Patients	
  without	
  current	
  fever	
  or	
  reported	
  fever	
  
 in	
  the	
  previous	
  48	
  hours	
  were	
  not	
  included	
  in	
  the	
  study.	
  

 At	
   health	
   posts,	
   all	
   patients	
   presenting	
   with	
   suspected	
   malaria	
   were	
   included	
   in	
   the	
   sample.	
  
 Health	
  extension	
  workers	
  also	
  used	
  the	
  RDTs	
  when	
  visiting	
  individuals	
  on	
  home	
  visits	
  as	
  part	
  of	
  
 their	
  routine	
  activities.	
  


 6.4. Sample	
  size	
  
 A	
   sample	
   size	
   sufficient	
   to	
   compare	
   performance	
   of	
   each	
   RDT	
   to	
   microscopy	
   in	
   detecting	
   each	
  
 Plasmodium	
   species,	
   at	
   each	
   site,	
   was	
   calculated	
   [22]	
   assuming	
   RDT	
   sensitivity	
   of	
   ≥	
   95%	
   and	
  
 specificity	
  of	
  ≥	
  80%,	
  and	
  aiming	
  at	
  80%	
  power	
  and	
  a	
  significance	
  level	
  of	
  5%.	
  It	
  was	
  also	
  assumed	
  
 that	
   on	
   average	
   at	
   least	
   15%	
   of	
   patients	
   reporting	
   with	
   febrile	
   illness	
   would	
   have	
   microscopy-­‐
 confirmed	
  malaria	
  infection.	
  We	
  estimated	
  a	
  60:40	
  proportion	
  of	
  P.	
  vivax	
  to	
  P.	
  falciparum.	
  Sample	
  
 size	
  calculations	
  indicate	
  that	
  811	
  individuals	
  would	
  need	
  to	
  be	
  included	
  in	
  the	
  study	
  to	
  determine	
  
 sensitivity	
   within	
   our	
   chosen	
   parameters,	
   while	
   270	
   individuals	
   would	
   be	
   required	
   to	
   determine	
  



                                                                                13	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                              	
  


specificity.	
   Since	
   we	
   wished	
   to	
   compare	
   sensitivity	
   and	
   specificity	
   of	
   each	
   RDT,	
   the	
   greater	
  
sample	
  size	
  was	
  chosen,	
  and	
  800	
  individuals	
  were	
  recruited	
  at	
  each	
  of	
  the	
  three	
  study	
  sites.	
  

At	
  each	
  of	
  the	
  nine	
  health	
  posts	
  included	
  in	
  the	
  qualitative	
  assessment	
  of	
  RDTs,	
  two	
  HEWs	
  were	
  
involved	
  in	
  the	
  study,	
  except	
  Bulbul	
  clinic	
  where	
  two	
  nurses	
  were	
  included.	
  Each	
  HEW	
  carried	
  out	
  
50	
  tests	
  with	
  each	
  type	
  of	
  RDT	
  during	
  their	
  normal	
  duties	
  both	
  at	
  the	
  health	
  post	
  and	
  in	
  house-­‐to-­‐
house	
  visits.	
  This	
  was	
  believed	
  to	
  be	
  sufficient	
  tests	
  for	
  each	
  individual	
  to	
  become	
  familiar	
  with	
  
the	
  RDT	
  and	
  to	
  form	
  opinions	
  on	
  its	
  ease	
  of	
  use	
  and	
  interpretation.	
  Therefore	
  a	
  total	
  of	
  900	
  of	
  
each	
  RDT	
  type	
  were	
  used	
  by	
  HEWs	
  in	
  evaluating	
  ease	
  of	
  use.	
  


6.5. Training	
  
Training	
  of	
  health	
  center	
  staff	
  involved	
  in	
  the	
  study	
  took	
  place	
  over	
  three	
  days.	
  At	
  each	
  site,	
  the	
  
health	
   center	
   head,	
   a	
   nurse	
   or	
   clinical	
   officer	
   and	
   two	
   laboratory	
   technicians	
   were	
   trained.	
   At	
  
Dimtu	
  clinic,	
  an	
  additional	
  nurse	
  was	
  trained	
  since	
  only	
  one	
  laboratory	
  technician	
  is	
  based	
  at	
  the	
  
clinic.	
  On	
  the	
  first	
  day	
  training	
  focused	
  on	
  the	
  general	
  study	
  objectives	
   and	
  background	
  to	
  multi-­‐
species	
  RDTs.	
  The	
  following	
  two	
  days	
  were	
  for	
  refresher	
  training	
  on	
  practical	
  methods:	
  assessing	
  
patients	
   for	
   inclusion	
   in	
   the	
   study,	
   preparation	
   and	
   microscopy	
   of	
   thick	
   and	
   thin	
   blood	
   films	
  
including	
   species	
   identification	
   and	
   parasite	
   counts,	
   and	
   preparation	
   and	
   interpretation	
   of	
   the	
  
three	
   multi-­‐species	
   RDTs.	
   All	
   individuals	
   involved	
   in	
   the	
   study	
   received	
   a	
   field	
   manual	
   outlining	
  
the	
   objectives	
   and	
   standard	
   operating	
   procedures	
   for	
   patient	
   enrolment,	
   RDT	
   preparation	
   and	
  
interpretation,	
   and	
   blood	
   film	
   preparation	
   and	
   microscopy.	
   The	
   study	
   coordinator	
   was	
   in	
  
frequent	
  contact	
  with	
  the	
  study	
  sites	
  throughout	
  the	
  period	
  of	
  data	
  collection	
  and	
  made	
  regular	
  
visits	
  to	
  provide	
  support	
  and	
  refresher	
  training	
  where	
  needed.	
  
The	
  16	
  HEWs	
  and	
  two	
  nurses	
  and	
  three	
  supervisors	
  involved	
  in	
  the	
  RDT	
  ease	
  of	
  use	
  assessment	
  
received	
   one	
   day	
   of	
   training.	
   During	
   the	
   training	
   they	
   were	
   informed	
   of	
   the	
   background	
   to	
   the	
  
study	
   and	
   current	
   malaria	
   situation	
   in	
   Ethiopia,	
   then	
   the	
   majority	
   of	
   the	
   training	
   was	
   spent	
  
practicing	
   preparing	
   and	
   interpreting	
   the	
   RDTs.	
   There	
   was	
   limited	
   opportunity	
   for	
   refresher	
  
training	
   other	
   than	
   direct	
   observation	
   of	
   HEWs	
   performing	
   the	
   RDT	
   as	
   part	
   of	
   the	
   evaluation	
   and	
  
interview	
  after	
  each	
  set	
  of	
  50	
  tests	
  by	
  the	
  health	
  post	
  supervisor.	
  


6.6. Data	
  and	
  sample	
  collection	
  
6.6.1. Health	
  centers	
  
Eligible,	
   consenting	
   patients	
   were	
   enrolled	
   by	
   the	
   health	
   center	
   clinical	
   officer	
   and	
   assigned	
   a	
  
unique	
   study	
   identification	
   number.	
   Demographic	
   information	
   was	
   recorded	
   on	
   a	
   standardized	
  
questionnaire	
   (see	
   enrolment	
   form	
   in	
   appendices),	
   together	
   with	
   clinical	
   details	
   including	
   axillary	
  
temperature,	
  reports	
  of	
  fever,	
  other	
  symptoms,	
  and	
  any	
  malaria	
  medicine	
  taken	
  in	
  the	
  previous	
  
four	
  weeks.	
  Basic	
  information	
  regarding	
  malaria	
  prevention	
  measures	
  (mosquito	
  net	
  ownership	
  
and	
  use,	
  and	
  indoor	
  residual	
  spraying)	
  were	
  also	
  recorded	
  for	
  monitoring	
  and	
  analysis	
  purposes.	
  
After	
  enrolment	
  a	
  single	
  finger	
  prick	
  was	
  performed	
  by	
  a	
  laboratory	
  technician	
  at	
  Asendabo	
  and	
  
Serbo	
   health	
   centers,	
   and	
   by	
   a	
   nurse	
   at	
   Dimtu	
   clinic.	
   Two	
   slides	
   were	
   prepared,	
   each	
   with	
   one	
  
thick	
   and	
   one	
   thin	
   blood	
   film.	
   The	
   same	
   finger	
   prick	
   blood	
   sample	
   was	
   used	
   to	
   carry	
   out	
   all	
   three	
  
RDTs	
   in	
   parallel,	
   following	
   the	
   manufacturer’s	
   instructions.	
   A	
   timer	
   was	
   used	
   to	
   ensure	
   RDT	
  
results	
   were	
   read	
   and	
   recorded	
   at	
   the	
   correct	
   time.	
   Strength	
   of	
   each	
   test	
   band	
   seen	
   on	
   the	
   RDTs	
  




                                                                                   14	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                    	
  


was	
  graded	
  as	
  strong	
  or	
  faint	
  by	
  the	
  reader,	
  no	
  standardized	
  reference	
  was	
  provided.	
  RDTs	
  were	
  
discarded	
  after	
  recording	
  of	
  results.	
  

Blood	
  films	
  were	
  left	
  to	
  air-­‐dry,	
  protected	
  from	
  dust	
  and	
  insects.	
  Thin	
  blood	
  films	
  were	
  fixed	
  in	
  
methanol	
   after	
   air-­‐drying,	
   then	
   slides	
   were	
   stained	
   in	
   10%	
   Giemsa	
   solution	
   for	
   15	
   minutes.	
   One	
  
thick	
   smear	
   from	
   each	
   enrolled	
   patient	
   was	
   examined	
   at	
   x1000	
   magnification,	
   and	
   considered	
  
negative	
  if	
  no	
  parasites	
  were	
  seen	
  after	
  examination	
  of	
  200	
  fields.	
  When	
  positive	
  for	
  parasites,	
  
the	
   number	
   of	
   asexual	
   parasites	
   per	
   200	
   white	
   blood	
   cells	
   (WBCs),	
   or	
   500	
   WBCs	
   for	
   low	
   intensity	
  
infections,	
   were	
   used	
   to	
   calculate	
   the	
   number	
   of	
   asexual	
   parasites	
   per	
   µl	
   of	
   blood,	
   assuming	
   a	
  
standard	
   count	
   of	
   8000	
   WBCs	
   per	
   µl	
   of	
   blood.	
   The	
   thin	
   smear	
   was	
   examined	
   for	
   species	
  
determination.	
   Presence	
   and	
   species	
   of	
   gametocytes	
   was	
   recorded,	
   but	
   a	
   gametocyte	
   count	
   was	
  
not	
  conducted.	
  Both	
  slides	
  from	
  each	
  patient	
  were	
  stored	
  in	
  labeled	
  slide	
  boxes	
  and	
  transferred	
  
to	
  Addis	
  Ababa	
  at	
  the	
  end	
  of	
  the	
  study	
  for	
  quality	
  control	
  purposes.	
  
All	
   data	
   collection	
   forms	
   (clinical	
   questionnaire,	
   RDT	
   results,	
   and	
   microscopy	
   results)	
   were	
  
collated	
  at	
  the	
  end	
  of	
  each	
  day	
  by	
  the	
  assigned	
  site	
  supervisor	
  (head	
  of	
  the	
  health	
  center),	
  and	
  
stored	
  in	
  a	
  secure	
  location	
  at	
  the	
  health	
  center.	
  


6.6.2. Health	
  posts	
  
Two	
  HEWs	
  at	
  each	
  of	
  the	
  nine	
  health	
  posts	
  received	
  basic	
  training	
  on	
  use	
  and	
  interpretation	
  of	
  
the	
  multi-­‐species	
  RDTs.	
  All	
  18	
  HEWs	
  were	
  already	
  familiar	
  with	
  the	
  falciparum-­‐only	
  ParaCheck-­‐pf	
  
test	
  and	
  had	
  been	
  using	
  it	
  for	
  a	
  minimum	
  of	
  one	
  year.	
  	
  
Each	
  HEW	
  used	
  50	
  of	
  each	
  multi-­‐species	
  RDT	
  as	
  part	
  of	
  their	
  normal	
  duties	
  at	
  the	
  health	
  post	
  and	
  
when	
  conducting	
  home	
  visits	
  in	
  the	
  community.	
  Blood	
  slides	
  were	
  not	
  prepared,	
  and	
  RDTs	
  were	
  
discarded	
   after	
   use.	
   The	
   different	
   RDT	
   tests	
   evaluated	
   in	
   the	
   present	
   study	
   were	
   used	
  
consecutively.	
   Each	
   health	
   post	
   was	
   provided	
   with	
   a	
   new	
   RDT	
   test	
   after	
   both	
   HEWs	
   had	
  
completed	
  their	
  allocated	
  50	
  tests.	
  The	
  order	
  in	
  which	
  the	
  HEWs	
  received	
  the	
  three	
  RDTs	
  differed	
  
between	
  health	
  posts,	
  with	
  the	
  intention	
  to	
  minimize	
  bias	
  in	
  the	
  reported	
  ease-­‐of-­‐use	
  that	
  may	
  
have	
  resulted	
  from	
  increasing	
  familiarity	
  with	
  multi-­‐species	
  tests	
  over	
  the	
  study	
  period.	
  
After	
  completion	
  of	
  each	
  set	
  of	
  50	
  RDTs,	
  the	
  HEW	
  was	
  interviewed	
  by	
  a	
  supervisor	
  from	
  the	
  local	
  
health	
  center.	
  Supervisors	
  had	
  received	
  prior	
  training	
  in	
  the	
  interview	
  technique	
  and	
  objectives	
  
from	
  the	
  study	
  coordinator.	
  HEW	
  were	
  asked	
  to	
  grade	
  a	
  number	
  of	
  features	
  of	
  the	
  test,	
  such	
  as	
  
the	
   packaging,	
   blood	
   collection	
   device,	
   buffer	
   and	
   results	
   interpretation,	
   and	
   then	
   asked	
   open-­‐
ended	
   questions	
   to	
   probe	
   for	
   any	
   particular	
   difficulties	
   or	
   preferences	
   compared	
   to	
   previous	
  
RDTs	
  used	
  (see	
  interview	
  form	
  in	
  appendices).	
  For	
  the	
  first	
  evaluation,	
  the	
  HEW	
  was	
  requested	
  to	
  
compare	
  the	
  new	
  RDT	
  product	
  to	
  ParaCheck-­‐Pf,	
  for	
  subsequent	
  evaluations	
  staff	
  were	
  asked	
  to	
  
compare	
   the	
   current	
   test	
   with	
   all	
   previously	
   used	
   RDTs,	
   i.e.	
   ParaCheck-­‐Pf	
   and	
   the	
   multi-­‐species	
  
RDTs.	
  After	
  gaining	
  experience	
  with	
  the	
  use	
  of	
  all	
  three	
  multi-­‐species	
  RDT	
  products,	
  HEWs	
  were	
  
asked	
  to	
  report	
  their	
  final	
  preferences	
  and	
  to	
  compare	
  aspects	
  of	
  all	
  tests	
  in	
  a	
  final	
  interview.	
  


6.6.3. Quality	
  control	
  measures	
  
To	
   ensure	
   laboratory	
   technicians	
   were	
   blinded	
   to	
   the	
   results	
   of	
   the	
   different	
   diagnostic	
  
techniques,	
  RDT	
  and	
  microscopy	
  results	
  were	
  read	
  separately	
  by	
  different	
  staff.	
  Specifically,	
  the	
  
individual	
  (laboratory	
  technician	
  or	
  nurse)	
  who	
  performed	
  the	
  finger	
  prick	
  read	
  the	
  RDT	
  results	
  at	
  
the	
  appropriate	
  time.	
  A	
  different	
  laboratory	
  technician	
  was	
  responsible	
  for	
  reading	
  the	
  blood	
  film	
  


                                                                               15	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                 	
  


after	
  fixing	
  and	
  staining,	
  which	
  took	
  place	
  after	
  the	
  RDT	
  results	
  were	
  read.	
  In	
  addition,	
  results	
  for	
  
RDTs	
   and	
   microscopy	
   were	
   recorded	
   on	
   different	
   sheets	
   (see	
   laboratory	
   results	
   forms	
   in	
  
appendices),	
   and	
   laboratory	
   staff	
   were	
   instructed	
   not	
   to	
   discuss	
   patient	
   results.	
   Microscopy	
  
results	
  were	
  reported	
  to	
  the	
  clinical	
  officer	
  for	
  treatment	
  and	
  case	
  management.	
  
A	
   second	
   reading	
   of	
   all	
   blood	
   films	
   was	
   conducted	
   at	
   central	
   level	
   by	
   an	
   experienced	
  
microscopist	
   at	
   Nazareth	
   malaria	
   training	
   center,	
   a	
   malaria	
   reference	
   laboratory	
   operated	
   by	
  
ORHB.	
   The	
   technician	
   conducting	
   the	
   second	
   reading	
   was	
   blinded	
   to	
   both	
   the	
   RDT	
   and	
   initial	
  
microscopy	
  results.	
  The	
  same	
  slide	
  (A	
  or	
  B)	
  read	
  at	
  the	
  health	
  center	
  was	
  selected	
  to	
  be	
  read	
  by	
  
the	
  second	
  microscopist.	
  Where	
  the	
  original	
  slide	
  was	
  broken	
  or	
  unreadable,	
  the	
  duplicate	
  slide	
  
was	
  read.	
  	
  
A	
   third,	
   blinded,	
   reading	
   was	
   conducted	
   on	
   all	
   slides	
   with	
   discrepant	
   first	
   and	
   second	
   readings:	
  
presence/absence	
   of	
   asexual	
   parasites,	
   difference	
   in	
   species,	
   or	
   greater	
   than	
   50%	
   difference	
   in	
  
parasite	
   count.	
   A	
   total	
   of	
   496	
   slides	
   were	
   included	
   in	
   this	
   third	
   reading	
   by	
   an	
   expert	
   microscopist	
  
at	
   EHNRI,	
   the	
   national	
   reference	
   laboratory.	
   Microscopy	
   results	
   and	
   parasite	
   counts	
   were	
  
corrected	
  according	
  to	
  the	
  third	
  reading.	
  
RDTs	
  were	
  stored	
  at	
  ambient	
  temperature	
  in	
  health	
  centers	
  and	
  health	
  posts.	
  Temperature	
  and	
  
humidity	
  in	
  the	
  storage	
  room	
  were	
  monitored	
  (Tinytag,	
  Gemini	
  Data	
  Loggers,	
  England)	
  at	
  health	
  
centers,	
   but	
   not	
   controlled.	
   RDTs	
   were	
   transported	
   from	
   Addis	
   Ababa	
   to	
   the	
   health	
   centers	
   by	
  
four-­‐wheel	
   drive	
   vehicle	
   without	
   refrigeration.	
   All	
   RDTs	
   required	
   for	
   ease	
   of	
   use	
   testing	
   were	
  
initially	
   stored	
   at	
   health	
   centers,	
   then	
   provided	
   to	
   health	
   posts	
   one	
   brand	
   at	
   a	
   time	
   by	
   the	
  
supervisor	
  conducting	
  HEW	
  interviews.	
  	
  
Results	
  of	
  RDTs	
  used	
  at	
  health	
  post	
  level	
  were	
  recorded	
  by	
  HEWs,	
  but	
  results	
  were	
  not	
  analyzed.	
  
When	
  HEWs	
  received	
  training,	
  they	
  were	
  observed	
  performing	
  RDTs	
  to	
  ensure	
  compliance	
  with	
  
the	
  procedures.	
  In	
  accordance	
  with	
  routine	
  health	
  post	
  operations,	
  there	
  were	
  no	
  quality	
  control	
  
procedures	
  in	
  place	
  to	
  monitor	
  HEW	
  performance	
  after	
  the	
  training.	
  


6.7. Heat	
  stability	
  assessment	
  
Heat	
   stability	
   and	
   lot	
   testing	
   of	
   RDTs	
   was	
   conducted	
   according	
   to	
   the	
   protocol	
   developed	
   by	
  
WHO/FIND	
  (see	
  appendices),	
  for	
  which	
  EHNRI	
  is	
  a	
  designated	
  regional	
  reference	
  laboratory.	
  Staff	
  
conducting	
   the	
   heat	
   stability	
   testing	
   had	
   received	
   extensive	
   training	
   as	
   part	
   of	
   the	
   WHO	
  
collaboration.	
  
All	
  RDTs	
  were	
  initially	
  subject	
  to	
  WHO/FIND	
  standard	
  lot	
  testing	
  procedures	
  to	
  ensure	
  compliance	
  
with	
  agreed	
  standards.	
  RDTs	
  were	
  then	
  placed	
  into	
  incubators	
  at	
  set	
  temperatures	
  of	
  35°C,	
  45°C	
  
and	
  60°C.	
  At	
  set	
  time	
  intervals	
  (4,	
  12,	
  24,	
  48	
  and	
  72	
  hours	
  for	
  60°C;	
  5,	
  10,	
  20,	
  30,	
  60	
  and	
  90	
  days	
  for	
  
35°C	
   and	
   45°C)	
   seven	
   of	
   each	
   RDT	
   were	
   removed	
   and	
   tested	
   with	
   stored	
   blood	
   samples.	
   The	
  
samples	
  used	
  were:	
  i)	
  a	
  negative	
  control,	
  ii)	
  P.	
  falciparum	
  at	
  200	
  parasites/µl,	
  iii)	
  P.	
  falciparum	
  at	
  
2000	
  parasites/µl,	
  iv)	
  P.	
  vivax	
  at	
  200	
  parasites/µl,	
  and	
  v)	
  P.	
  vivax	
  at	
  2000	
  parasites/µl.	
  At	
  the	
  low	
  
infection	
  intensity,	
  two	
  of	
  each	
  RDT	
  were	
  tested	
  because	
  of	
  the	
  expected	
  lower	
  sensitivity,	
  while	
  
only	
   one	
   of	
   each	
   RDT	
   was	
   tested	
   at	
   each	
   time	
   point	
   with	
   high	
   intensity	
   P.	
   falciparum	
   and	
   P.	
   vivax	
  
samples.	
  Testing	
  was	
  stopped	
  for	
  a	
  particular	
  dilution	
  and	
  species	
  if	
  there	
  were	
  two	
  consecutive	
  
negative	
  results.	
  




                                                                                     16	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                                                                 	
  


6.8. Analysis	
  
6.8.1. Data	
  entry	
  
Data	
   from	
   health	
   centers	
   were	
   entered	
   into	
   a	
   Microsoft®	
   Access	
   database	
   (Microsoft	
  
Corporation),	
   subject	
   to	
   range	
   and	
   type	
   checks.	
   Any	
   errors	
   or	
   inconsistencies	
   were	
   corrected	
  
with	
  reference	
  to	
  the	
  original	
  paper	
  forms.	
  Second	
  and	
  third	
  blood	
  film	
  microscopy	
  results	
  were	
  
entered	
  into	
  a	
  Microsoft®	
  Excel	
  spreadsheet	
  (Microsoft	
  Corporation).	
  Data	
  were	
  exported	
  from	
  
Access	
   and	
   Excel	
   into	
   a	
   combined	
   dataset	
   in	
   STATA	
   8.0	
   (Stata	
   Corporation,	
   Texas	
   USA),	
   which	
  
was	
  used	
  for	
  all	
  data	
  analysis.	
  	
  


6.8.2. RDT	
  validity	
  and	
  reliability	
  compared	
  to	
  microscopy	
  
Since	
  Pf/Pan	
  RDTs	
  do	
  not	
  distinguish	
  between	
  mixed	
  infection	
  and	
  P.	
  falciparum	
  mono-­‐infection	
  
(both	
   can	
   activate	
   the	
   HRP2	
   and	
   pLDH	
   bands),	
   RDT	
   performance	
   was	
   assessed	
   in	
   two	
   categories	
  
only.	
   Firstly,	
   non-­‐falciparum	
   mono-­‐infection,	
   followed	
   by	
   a	
   category	
   of	
   either	
   mixed	
   infection	
  
with	
  P.	
  falciparum	
  and	
  a	
  non-­‐falciparum	
  species,	
  or	
  P.	
  falciparum	
  mono-­‐infection.	
  Therefore	
  from	
  
this	
   point,	
   when	
   summarizing	
   RDT	
   performance	
   according	
   to	
   the	
   following	
   indicators,	
   mixed	
  
infection	
  will	
  be	
  included	
  as	
  P.	
  falciparum	
  infection.	
  
                •          Specificity:	
   the	
   proportion	
   of	
   individuals	
   negative	
   by	
   microscopy	
   correctly	
   identified	
   as	
  
                           negative	
  by	
  the	
  RDT	
  
                •          Sensitivity:	
   the	
   proportion	
   of	
   individuals	
   positive	
   by	
   microscopy	
   correctly	
   identified	
   as	
  
                           positive	
  by	
  the	
  RDT.	
  
                •          Positive	
  predictive	
  value	
  (PPV):	
  the	
  proportion	
  of	
  individuals	
  with	
  a	
  positive	
  microscopy	
  
                           result	
  among	
  all	
  individuals	
  with	
  a	
  positive	
  RDT	
  result.	
  	
  
                •          Negative	
   predictive	
   value	
   (NPV):	
   the	
   proportion	
   of	
   individuals	
   with	
   a	
   negative	
  
                           microscopy	
  result	
  among	
  all	
  individuals	
  with	
  a	
  negative	
  RDT	
  result.	
  	
  
The	
   sensitivity,	
   specificity,	
   NPV	
   and	
   PPV	
   were	
   calculated	
   with	
   95%	
   confidence	
   intervals	
   for	
   each	
  
RDT,	
   and	
   for	
   P.	
   falciparum	
   and	
   non-­‐falciparum	
   infection	
   separately,	
   using	
   standard	
   calculations	
  
for	
  this	
  purpose	
  [23]	
  (Figure	
  3).	
  

 	
                                    	
                                                	
                                                      	
                           	
                               	
  
 	
                                                                       ‘Gold	
  standard’	
                                                                                	
                               	
  
                                       	
                                                                                                        	
  
                                                                            microscopy	
  
 	
                 	
                            	
            Positive	
                      Negative	
                                Total	
                             	
     Specificity	
  =	
  d	
  /	
  (b+d)	
  
 	
             RDT	
                          Positive	
                  a	
                             b	
                            a	
  +	
  b	
                       	
     Sensitivity	
  =	
  a	
  /	
  (a+c)	
  
 	
            result	
                       Negative	
                    c	
                            d	
                             c	
  +	
  d	
                      	
     PPV	
  =	
  a	
  /	
  (a+b)	
  
 	
                 	
                              Total	
            a	
  +	
  c	
                   b	
  +	
  d	
             a	
  +	
  b	
  +	
  c	
  +	
  d	
            	
     NPV	
  =	
  d	
  /	
  (d+c)	
  
 	
     	
                      	
                              	
                              	
                       	
                                            	
                                      	
  
	
      Figure	
  3:	
  2x2	
  table	
  for	
  calculating	
  RDT	
  performance	
  compared	
  to	
  ‘gold	
  standard’	
  microscopy	
  
 	
     	
                      	
                              	
                              	
                       	
                                            	
                                      	
  
	
  
Each	
  of	
  the	
  above	
  indicators	
  was	
  compared	
  between	
  two	
  RDTs	
  using	
  McNemar’s	
  test	
  [24],	
  which	
  
allows	
   for	
   the	
   paired	
   nature	
   of	
   results,	
   according	
   to	
   the	
   defined	
   categories	
   of	
   non-­‐falciparum	
  
infection	
  and	
  P.	
  falciparum	
  infection.	
  Data	
  were	
  analyzed	
  by	
  site	
  (health	
  center)	
  and	
  when	
  pooled	
  
for	
  all	
  sites.	
  	
  




                                                                                                                                17	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                 	
  


The	
  Kappa	
  coefficient,	
  representing	
  the	
  proportion	
  of	
  agreements	
  beyond	
  chance,	
  was	
  used	
  to	
  
quantify	
   the	
   level	
   of	
   agreement	
   between	
   each	
   RDT	
   and	
   ‘gold	
   standard’	
   microscopy	
   [25].	
   The	
  
Kappa	
   coefficient	
   was	
   calculated	
   for	
   each	
   RDT,	
   for	
   both	
   P.	
   falciparum	
   infection	
   and	
   for	
   non-­‐
falciparum	
  infection.	
  A	
  kappa	
  value	
  ≥	
  0.8	
  was	
  considered	
  to	
  indicate	
  very	
  good	
  reliability.	
  


6.8.3. RDT	
  ease-­‐of-­‐use	
  	
  
For	
  each	
  of	
  the	
  eleven	
  specific	
  features	
  of	
  the	
  RDTs	
  graded	
  by	
  HEWs	
  (such	
  as	
  ease	
  of	
  filling	
  and	
  
emptying	
   blood	
   collection	
   device,	
   ease	
   of	
   reading	
   results),	
   the	
   mean	
   score	
   for	
   each	
   feature	
   of	
  
each	
   RDT	
   was	
   calculated.	
   The	
   mean	
   overall	
   rating	
   for	
   each	
   RDT	
   was	
   also	
   calculated	
   and	
   RDTs	
  
ranked	
  according	
  to	
  their	
  overall	
  score.	
  All	
  comments	
  made	
  at	
  interview	
  by	
  HEWs	
  were	
  read,	
  and	
  
common	
  features	
  noted.	
  	
  


6.8.4. RDT	
  heat	
  stability	
  
Failure	
   rate	
   of	
   each	
   RDT	
   at	
   each	
   testing	
   point	
   was	
   calculated.	
   Maximum	
   duration	
   of	
   storage	
   at	
  
the	
  various	
  temperatures	
  was	
  determined.	
  


7.      Ethical	
  considerations	
  
All	
  patients	
  identified	
  at	
  health	
  centers	
  as	
  meeting	
  the	
  inclusion	
  criteria	
  had	
  the	
  study	
  explained	
  
to	
  them	
  in	
  detail,	
  together	
  with	
  potential	
  benefits,	
  risks	
  and	
  their	
  right	
  to	
  withdraw	
  at	
  any	
  time.	
  
Potential	
  study	
  participants	
  were	
  then	
  asked	
  to	
  sign	
  an	
  informed	
  consent	
  form.	
  For	
  children	
  aged	
  
≤	
   16	
   years	
   written	
   consent	
   was	
   requested	
   from	
   a	
   parent	
   or	
   guardian.	
   Consent	
   forms	
   were	
  
available	
  in	
  the	
  local	
  language,	
  Afaan	
  Oromo,	
  if	
  preferred	
  by	
  patients.	
  Individuals	
  that	
  refused	
  to	
  
sign	
  the	
  consent	
  form	
  were	
  not	
  included	
  in	
  the	
  study.	
  	
  
Participants	
   were	
   assigned	
   a	
   unique	
   identification	
   number	
   at	
   enrolment,	
   which	
   was	
   used	
   on	
  
microscope	
   slides	
   and	
   RDTs.	
   To	
   preserve	
   anonymity	
   patient	
   names	
   were	
   not	
   entered	
   into	
   the	
  
electronic	
  database,	
  and	
  paper	
  forms	
  were	
  stored	
  in	
  a	
  secure	
  location	
  to	
  which	
  only	
  the	
  survey	
  
investigators	
  had	
  access.	
  For	
  individuals	
  with	
  microscopy-­‐confirmed	
  malaria	
  infection,	
  treatment	
  
was	
   provided	
   in	
   accordance	
   with	
   national	
   guidelines	
   [26].	
   RDT	
   results	
   were	
   not	
   used	
   to	
   inform	
  
patient	
  treatment.	
  	
  
At	
   health	
   posts,	
   patients	
   were	
   asked	
   to	
   provide	
   verbal	
   consent	
   for	
   the	
   multi-­‐species	
   RDT	
   to	
   be	
  
conducted	
  in	
  addition	
  to	
  ParaCheck.	
  Since	
  no	
  additional	
  procedures	
  were	
  to	
  be	
  carried	
  out	
  on	
  the	
  
patient	
   (a	
   finger-­‐prick	
   blood	
   sample	
   was	
   to	
   be	
   taken	
   for	
   ParaCheck	
   as	
   per	
   routine	
   procedures)	
  
and	
  samples	
  or	
  patient	
  data	
  were	
  not	
  being	
  stored,	
  no	
  further	
  formal	
  consent	
  was	
  required,	
  as	
  
advised	
  by	
  the	
  FMoH	
  Institutional	
  Ethics	
  Review	
  Board.	
  	
  
This	
  study	
  was	
  approved	
  by	
  the	
  London	
  School	
  of	
  Hygiene	
  and	
  Tropical	
  Medicine	
  ethics	
  review	
  
board,	
   and	
   the	
   national	
   health	
   research	
   ethics	
   review	
   committee	
   of	
   the	
   Ethiopian	
   Science	
   and	
  
Technology	
  Agency.	
  




                                                                             18	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                      	
  



8.          Results	
  
       8.1. RDT	
  transportation	
  and	
  storage	
  conditions	
  
During	
   transportation	
   of	
   RDTs	
   and	
   other	
   study	
   materials	
   from	
   the	
   Malaria	
   Consortium	
   office	
   in	
  
Addis	
   Ababa	
   to	
   the	
   study	
   site	
   in	
   Jimma	
   zone,	
   a	
   journey	
   of	
   approximately	
   six	
   hours,	
   temperatures	
  
in	
  the	
  vehicle	
  reached	
  a	
  maximum	
  of	
  36°C.	
  Once	
  stored	
  at	
  health	
  centers,	
  RDTs	
  were	
  out	
  of	
  direct	
  
sunlight	
  and	
  daily	
  temperatures	
  did	
  not	
  exceed	
  30°C	
  during	
  the	
  study	
  period.	
  	
  
	
  
8.2. Study	
  population	
  description	
  
A	
  total	
  of	
  2400	
  febrile	
  patients	
  were	
  enrolled	
  into	
  the	
  study	
  between	
  May	
  1	
  and	
  July	
  31	
  2009,	
  800	
  
at	
  each	
  of	
  the	
  three	
  health	
  centers.	
  Full	
  data	
  are	
  available	
  from	
  2383	
  individuals.	
  	
  
The	
   patients	
   were	
   aged	
   between	
   six	
   months	
   and	
   80	
   years,	
   with	
   49%	
   being	
   male.	
   254	
   (10.6%)	
   of	
  
patients	
   were	
   aged	
   under	
   five	
   years,	
   17.3%	
   were	
   aged	
   between	
   five	
   and	
   14	
   years,	
   with	
   the	
  
remaining	
   72.1%	
   aged	
   fifteen	
   years	
   or	
   over	
   (Table	
   1).	
   170	
   of	
   the	
   women	
   enrolled	
   (14.5%)	
   were	
  
pregnant	
  at	
  the	
  time	
  of	
  the	
  study.	
  The	
  mean	
  ±	
  SD	
  temperature	
  was	
  37.6	
  ±	
  0.8	
  °C	
  (Table	
  1).	
  

Table	
  1:	
  Baseline	
  characteristics	
  of	
  the	
  study	
  population	
  	
  
	
                                                                                           Serbo	
                 Dimtu	
                 Asendabo	
  
                                                          All	
  sites	
  
                                                                                             n	
  (%)	
              n	
  (%)	
                n	
  (%)	
  
Number	
  of	
  patients	
                                   2383	
                            790	
                   800	
                      793	
  
                                                        1169/1214	
                       362/428	
               415/385	
                  392/401	
  
Sex:	
  M/F	
  (%	
  female)	
  
                                                          (51.0)	
                         (54.2)	
                (48.1)	
                   (50.6)	
  
Age*	
  	
                                              22.9	
  ±	
  14.2	
              23.3	
  ±	
  15.1	
     22.5	
  ±	
  14.6	
        23.0	
  ±	
  12.8	
  
(mean	
  ±	
  SD,	
  range)	
                            (0.5-­‐80)	
                     (0.5-­‐80)	
            (0.5-­‐75)	
               (0.5-­‐80)	
  
Pregnant	
  (%	
  of	
  females)	
                       170	
  (14.5)	
                    26	
  (6.2)	
           31	
  (8.2)	
            113	
  (29.8)	
  
Temperature	
  	
                                        37.6	
  ±	
  0.8	
               37.4	
  ±	
  0.7	
      37.2	
  ±	
  0.7	
         38.3	
  ±	
  0.6	
  
(mean	
  ±	
  SD,	
  range)	
                           (32.8-­‐40.2)	
                  (32.8-­‐40.0)	
         (35.5-­‐39.8)	
            (35.0-­‐40.2)	
  
Fever	
  in	
  last	
  48	
  hrs	
                      1706	
  (72.0)	
                 390	
  (49.6)	
         548	
  (68.5)	
             768	
  (98.0)	
  
Slept	
  under	
  mosquito	
  net	
  last	
  
                                                        1533	
  (64.7)	
           403	
  (51.3)	
          538	
  (67.4)	
                  592	
  (75.2)	
  
night	
  
Always	
  sleeps	
  under	
  a	
  mosquito	
  
                                                        1010	
  (42.7)	
           194	
  (24.7)	
          407	
  (51.0)	
                  409	
  (52.1)	
  
net	
  
Household	
  sprayed	
  with	
  
                                                        1550	
  (65.4)	
           351	
  (44.5)	
          588	
  (73.7)	
                  611	
  (77.8)	
  
insecticide	
  in	
  last	
  12	
  months	
  
*Where	
  patient	
  is	
  under	
  one	
  year,	
  decimal	
  year	
  calculated	
  for	
  each	
  completed	
  month.	
  

64.7%	
   of	
   patients	
   enrolled	
   reported	
   sleeping	
   under	
   a	
   mosquito	
   net	
   the	
   previous	
   night,	
   with	
   42.7%	
  
of	
   all	
   individuals	
   reportedly	
   sleeping	
   under	
   a	
   net	
   every	
   night.	
   Amongst	
   pregnant	
   women,	
   75.2%	
  
reported	
   having	
   slept	
   under	
   a	
   net	
   the	
   previous	
   night,	
   and	
   52.7%	
   said	
   they	
   always	
   sleep	
   under	
   a	
  
net.	
  Of	
  the	
  children	
  under	
  five	
  years	
  of	
  age,	
  69.4%	
  reportedly	
  slept	
  under	
  a	
  net	
  the	
  night	
  before	
  
enrolment,	
  with	
  48.6%	
  always	
  sleeping	
  under	
  a	
  net.	
  Indoor	
  residual	
  spraying	
  had	
  been	
  carried	
  out	
  
in	
  65.4%	
  of	
  the	
  enrolled	
  patients’	
  households,	
  87.6%	
  of	
  these	
  were	
  sprayed	
  within	
  the	
  last	
  month.	
  

8.3. Microscopy	
  results	
  
According	
   to	
   microscopy,	
   552	
   (23.2%)	
   of	
   individuals	
   enrolled	
   in	
   the	
   study	
   had	
   malaria.	
   P.	
  
falciparum	
  and	
  P.	
  vivax	
  were	
  detected	
  in	
  53.8%	
  and	
  44.4%	
  of	
  confirmed	
  cases,	
  respectively;	
  1.8%	
  of	
  
confirmed	
  cases	
  had	
  mixed	
  infection	
  of	
  P.	
  falciparum	
  and	
  P.	
  vivax.	
  Eight	
  individuals	
  were	
  found	
  to	
  


                                                                                19	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                   	
  


have	
  gametocytes	
  (seven	
  with	
  P.	
  falciparum,	
  one	
  of	
  P.	
  vivax)	
  but	
  no	
  asexual	
  parasites.	
  No	
  cases	
  
of	
   P.	
   malariae	
   were	
   identified	
   by	
   microscopy.	
   Slide	
   positivity	
   rates	
   and	
   species	
   for	
   each	
   health	
  
center	
  are	
  shown	
  in	
  table	
  2.	
  	
  
The	
  arithmetic	
  mean	
  ±	
  SD	
  parasite	
  count	
  was	
  7961	
  p/µl	
  ±	
  14372	
  for	
  P.	
  falciparum	
  and	
  7211	
  p/µl	
  ±	
  
8489	
  for	
  P.	
  vivax.	
  Distribution	
  of	
  parasite	
  densities	
  is	
  presented	
  (Figure	
  4),	
  with	
  most	
  cases	
  of	
  P.	
  
falciparum	
  found	
  to	
  have	
  parasite	
  densities	
  below	
  5000	
  parasites	
  per	
  µl	
  of	
  blood	
  (61.1%).	
  Very	
  low	
  
intensity	
  infection,	
  below	
  200	
  p/µl,	
  was	
  found	
  in	
  20.6%	
  of	
  P.	
  falciparum	
  cases.	
  For	
  those	
  patients	
  
with	
   P.	
   vivax	
   infection,	
   52.6%	
   had	
   parasite	
   density	
   above	
   5000	
   p/µl,	
   while	
   only	
   3.5%	
   had	
   low	
  
intensity	
  infection	
  of	
  below	
  200	
  p/µl.	
  

Figure	
  4:	
  Distribution	
  of	
  parasite	
  densities	
  for	
  all	
  cases	
  of	
  P.	
  falciparum	
  (left)	
  and	
  all	
  cases	
  of	
  P.	
  vivax	
  (right).	
  
Column	
  width	
  500	
  p/µl.	
  One	
  case	
  with	
  173,655	
  p/µl	
  P.	
  falciparum	
  was	
  excluded	
  from	
  the	
  graph.	
  




For	
   a	
   total	
   of	
   2383	
   individuals,	
   blood	
   slides	
   were	
   read	
   twice;	
   once	
   at	
   the	
   health	
   center	
   on	
   the	
   day	
  
of	
  patient	
  enrolment	
  and	
  then	
  at	
  a	
  central	
  level	
  after	
  completion	
  of	
  the	
  sample.	
  A	
  total	
  of	
  496	
  
slides	
  were	
  read	
  a	
  third	
  time	
  to	
  address	
  discrepancies	
  following	
  the	
  first	
  and	
  second	
  microscopy	
  
readings,	
   either	
   for	
   presence/absence	
   of	
   asexual	
   parasites	
   (n=198),	
   species	
   (including	
   single-­‐
species	
   and	
   mixed	
   infection	
   discrepancies)	
   (n=155)	
   or	
   a	
   greater	
   than	
   50%	
   difference	
   in	
   parasite	
  
count	
  (n=143).	
  	
  

Table	
  2:	
  Summary	
  microscopy	
  results	
  
                          	
                                    Total	
                          Serbo	
  	
                     Dimtu	
                    Asendabo	
  
n	
                                                             2383	
                            790	
                           800	
                        793	
  
Negative	
                                                 1831	
  (76.8)	
                  676	
  (85.6)	
                 608	
  (76.0)	
               547	
  (69.0)	
  
Any	
  Plasmodium	
  spp.	
                                 552	
  (23.2)	
                  114	
  (14.4)	
                 192	
  (24.0)	
               246	
  (31.0)	
  
P.	
  falciparum	
  only	
                                  297	
  (12.5)	
                   61	
  (7.7)	
                   74	
  (9.3)	
                162	
  (20.4)	
  
P.	
  vivax	
  only	
                                       245	
  (10.3)	
                   52	
  (6.6)	
                  112	
  (14.0)	
                81	
  (10.2)	
  
Mixed	
                                                      10	
  (0.4)	
                     1	
  (0.1)	
                    6	
  (0.8)	
                  3	
  (0.4)	
  
P.	
  falciparum	
  p/µl	
  	
                                  7961	
  	
                      8894	
                          6838	
                         8141	
  
	
  	
  arithmetic	
  mean	
  (95%	
  CI)	
               (6341-­‐9580)	
                  (5638-­‐12150)	
                 (4600-­‐9076)	
              (5594-­‐10687)	
  
P.	
  falciparum	
  p/µl	
                                      2400	
                          2973	
                          2140	
                         1800	
  
	
  	
  median	
  (range)	
                               (40-­‐173,655)	
                  (40-­‐52,400)	
                 (40-­‐59,040)	
              (40-­‐173,655)	
  
P.	
  vivax	
  p/µl	
                                           7211	
                          6571	
                          7683	
                         6955	
  
	
  	
  arithmetic	
  mean	
  (95%	
  CI)	
               (6166-­‐8256)	
                   (4677-­‐8464)	
                 (6006-­‐9360)	
               (5137-­‐8774)	
  
P.	
  vivax	
  p/µl	
                                           5280	
                          4640	
                          6160	
                         4750	
  
	
  	
  median	
  (range)	
                                (40-­‐68,640)	
                  (40-­‐30,160)	
                 (40-­‐68,640)	
              (240-­‐51,200)	
  


                                                                                      20	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                           	
  


A	
  negative	
  correlation	
  between	
  patient	
  age	
  and	
  infection	
  with	
  malaria	
  (assessed	
  by	
  microscopy)	
  
was	
  observed	
  (Figure	
  5).	
  Prevalence	
  of	
  malaria	
  amongst	
  febrile	
  patients	
  showed	
  a	
  peak	
  between	
  
ages	
   five	
   to	
   nine	
   years,	
   with	
   a	
   general	
   trend	
   of	
   decreasing	
   slide	
   positivity	
   with	
   increasing	
   age	
  
(score	
   test	
   for	
   trend	
   p<0.001).	
   This	
   trend	
   was	
   stronger	
   for	
   P.	
   vivax	
   (p<0.001)	
   than	
   for	
   P.	
  
falciparum	
  infection	
  (p=0.02).	
  




Figure	
   5:	
   Association	
   between	
   age	
   and	
   slide	
   positivity	
   amongst	
   febrile	
   patients.	
   95%	
   Confidence	
   intervals	
  
are	
  indicated	
  for	
  prevalence	
  of	
  any	
  Plasmodium	
  spp.	
  

Mean	
   parasite	
   density	
   was	
   similar	
   for	
   P.	
   falciparum	
   and	
   P.	
   vivax	
   (T	
   test	
   p=0.46),	
   and	
   there	
   was	
   no	
  
significant	
  association	
  between	
  age	
  and	
  parasite	
  density.	
  

Individuals	
  with	
  an	
  axillary	
  temperature	
  over	
  37.5°C	
  were	
  1.83	
  times	
  more	
  likely	
  to	
  have	
  malaria	
  
than	
   those	
   with	
   temperature	
   equal	
   to	
   or	
   below	
   37.5°C	
   (95%	
   CI	
   1.49-­‐2.25,	
   p<0.001).	
   However	
   when	
  
looking	
   at	
   malaria	
   cases	
   by	
   species,	
   odds	
   of	
   P.	
   falciparum	
   infection	
   were	
   2.20	
   times	
   higher	
   in	
  
individuals	
  with	
  fever	
  than	
  without	
  (95%	
  CI	
  1.68-­‐2.89,	
  p<0.001),	
  while	
  current	
  fever	
  had	
  only	
  very	
  
weak	
  evidence	
  of	
  altering	
  odds	
  of	
  P.	
  vivax	
  infection	
  (OR=1.29,	
  p=0.068).	
  


8.4. Microscopy	
  vs.	
  RDT	
  results	
  
Of	
  the	
  2400	
  individuals	
  enrolled	
  into	
  the	
  study,	
  seventeen	
  had	
  incomplete	
  data	
  recorded	
  for	
  RDT	
  
results,	
  or	
  broken	
  or	
  unreadable	
  blood	
  films,	
  and	
  were	
  subsequently	
  excluded	
  from	
  analysis.	
  	
  

Microscopy	
  results	
  were	
  consistent	
  with	
  RDT	
  results	
  for	
  87.6%,	
  88.6%,	
  and	
  88.4%	
  of	
  cases	
  tested	
  
with	
   ParaScreen,	
   CareStart	
   and	
   ICT	
   Combo	
   respectively	
   (Table	
   3).	
   ParaScreen	
   gave	
   23	
   false	
  
negatives	
   (17	
   P.	
   falciparum	
   and	
   6	
   P.	
   vivax),	
   and	
   208	
   false	
   positives	
   (67	
   P.	
   falciparum,	
   69	
   Pan,	
   72	
  
Pf+Pan),	
  CareStart	
  gave	
  26	
  false	
  negatives	
  (18	
  P.	
  falciparum	
  and	
  8	
  P.	
  vivax),	
  and	
  190	
  false	
  positives	
  
(67	
  P.	
  falciparum,	
  51	
  Pan,	
  72	
  Pf+Pan),	
  and	
  ICT	
  Combo	
  gave	
  24	
  false	
  negatives	
  (19	
  P.	
  falciparum,	
  8	
  P.	
  
vivax)	
  and	
  194	
  false	
  positives	
  (58	
  P.	
  falciparum,	
  60	
  Pan,	
  76	
  Pf+pan).	
  




                                                                                  21	
  
               Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                                                           	
  


               Half	
   of	
   all	
   false	
   negative	
   RDT	
   results	
   for	
   P.	
   falciparum	
   (10/19)	
   were	
   infections	
   below	
   200	
  
               parasites/µl.	
  Of	
  the	
  12	
  false	
  negatives	
  for	
   P.	
   vivax	
   infection,	
  four	
  were	
  below	
  200	
  parasites/µl	
  and	
  
               four	
   between	
   200	
   and	
   500	
  p/µl.	
  False	
  positive	
  RDT	
  results	
  did	
  not	
  appear	
  to	
  be	
  due	
  to	
  residual	
  
               antigen	
   from	
   cleared	
   infection,	
   since	
   only	
   25/222	
   (11.3%)	
   patients	
   with	
   false	
   positive	
   results	
   had	
  
               received	
  anti-­‐malarial	
  treatment	
  in	
  the	
  previous	
  four	
  weeks.	
  

               Table	
  3:	
  Comparison	
  of	
  RDT	
  and	
  microscopy	
  results,	
  all	
  sites.	
  
	
                              	
               	
                     CareStart	
                          	
               ParaScreen	
                             	
                           ICT	
  Combo	
  

          Microscopy	
                   N	
     	
      Neg	
         Pf	
   Pf+Pan	
   Pan	
               	
     Neg	
     Pf	
        Pf+Pan	
   Pan	
             	
     Neg	
                 Pf	
         Pf+Pan	
   Pan	
  
Negative	
                             1831	
   	
   1641	
   67	
                      72	
      51	
       	
   1623	
      67	
           73	
       69	
           	
   1637	
                  58	
                  77	
          60	
  
P.	
  falciparum	
                     297	
     	
       18	
         40	
            213	
      26	
       	
      17	
     30	
          223	
       27	
           	
      19	
                 29	
                 224	
          25	
  
P.	
  vivax	
                          245	
     	
        8	
          3	
             26	
     208	
       	
       6	
      2	
           35	
      202	
           	
       8	
                  2	
                  27	
         208	
  
Mixed	
                                 10	
     	
        0	
          0	
              9	
       1	
       	
       0	
      1	
            8	
              1	
     	
       0	
                  0	
                   9	
                 1	
  
                       Total	
   2383	
   	
   1667	
   110	
                          320	
     286	
       	
   1664	
   100	
            338	
      299	
           	
   1664	
                  89	
                 336	
         294	
  

               For	
   a	
   number	
   of	
   Plasmodium-­‐infected	
   patients	
   the	
   RDT	
   result	
   was	
   positive	
   but	
   incorrectly	
  
               identified	
   the	
   species.	
   Most	
   frequently	
   the	
   RDT	
   showed	
   only	
   the	
   Pan	
   band	
   for	
   P.	
   falciparum	
  
               mono-­‐infections	
  (26	
  CareStart,	
  27	
  ParaScreen	
  and	
  25	
  ICT	
  Combo).	
  There	
  was	
  one	
  case	
  of	
  mixed	
  P.	
  
               falciparum	
   and	
   P.	
   vivax	
   infection	
   that	
   resulted	
   in	
   only	
   Pan	
   band	
   activation	
   in	
   all	
   three	
   RDTs.	
  
               Additionally,	
   RDTs	
   occasionally	
   indicated	
  P.	
   falciparum	
   mono-­‐infection	
   for	
   what	
   was	
   identified	
   by	
  
               microscopy	
  to	
  be	
  P.	
  vivax	
  mono-­‐infection	
  (3	
  CareStart,	
  2	
  ParaScreen	
  and	
  2	
  ICT	
  Combo).	
  

               Table	
  4:	
  Comparative	
  performance	
  indicators	
  of	
  each	
  RDT,	
  all	
  sites	
  combined.	
  Significant	
  relationships	
  
               (p<0.05)	
  in	
  bold	
  type.	
  
       	
  
                                                                   CareStart	
                        ParaScreen	
                         ICT	
  Combo	
                                      McNemar	
  p	
  value	
  


       P.	
  falciparum	
                                           n=2383	
                               n=2383	
                           n=2383	
                        CS-­‐PS	
                      CS-­‐ICT	
            ICT-­‐PS	
  
       Sensitivity	
  (95%	
  CI)	
                     85.3%	
  (80.9-­‐89.1)	
                 85.3%	
  (80.9-­‐89.1)	
              85.3%	
  (80.9-­‐89.1)	
                 1.0	
                          1.0	
                 1.0	
  
       Specificity	
  (95%	
  CI)	
                     91.9%	
  (90.7-­‐93.1)	
                 91.5%	
  (90.2-­‐92.7)	
              92.2%	
  (90.9-­‐93.3)	
                0.09	
                         0.06	
               0.002	
  
       PPV	
  (95%	
  CI)	
                             60.9%	
  (56.1-­‐65.6)	
                 59.8%	
  (55.1-­‐64.4)	
              61.6%	
  (56.8-­‐66.3)	
                0.71	
                         0.83	
                0.48	
  
       NPV	
  (95%	
  CI)	
                             97.7%	
  (96.9-­‐98.3)	
                 97.7%	
  (96.9-­‐98.3)	
              97.7%	
  (96.9-­‐98.3)	
                 1.0	
                          1.0	
                 1.0	
  
       κ	
                                                          0.6599	
                               0.6504	
                           0.6659	
                         0.59	
                         0.75	
                0.37	
  
                                                                                                                                                                                        	
                        	
  
       Non-­‐falciparum	
  only	
                                   n=2383	
                               n=2383	
                           n=2383	
                                                                                  	
  
       Sensitivity	
  (95%	
  CI)	
                     84.9%	
  (79.8-­‐89.1)	
                 82.4%	
  (77.1-­‐87.0)	
              84.9%	
  (79.8-­‐89.1)	
                0.11	
                          1.0	
                0.08	
  
       Specificity	
  (95%	
  CI)	
                     96.4%	
  (95.5-­‐97.1)	
                 95.5%	
  (94.5-­‐96.3)	
              96.0%	
  (95.1-­‐96.8)	
               <0.001	
                        0.09	
                0.06	
  
       PPV	
  (95%	
  CI)	
                             72.7%	
  (67.2-­‐77.8)	
                 67.6%	
  (61.9-­‐72.8)	
              70.7%	
  (65.2-­‐75.9)	
                0.17	
                         0.58	
                0.46	
  
       NPV	
  (95%	
  CI)	
                             98.2%	
  (97.6-­‐98.8)	
                 97.9%	
  (97.2-­‐98.5)	
              98.2%	
  (97.6-­‐98.8)	
                0.58	
                          1.0	
                0.58	
  
                                                                                	
                                     	
                               	
  
       κ	
                                                          0.7565                                 0.7099                             0.7430                          0.002	
                         0.37	
                0.03	
  

               When	
   data	
   were	
   pooled	
   for	
   all	
   three	
   health	
   centers	
   (see	
   table	
   4),	
   the	
   sensitivity	
   of	
   all	
   RDTs	
   in	
  
               detecting	
   P.	
   falciparum	
   infections	
   was	
   85.3%.	
   Specificity	
   of	
   CareStart	
   was	
   highest	
   at	
   91.9%,	
  
               compared	
   to	
   91.5%	
   for	
   ParaScreen	
   (CareStart-­‐ParaScreen	
   p=0.059)	
   and	
   92.2%	
   for	
   ICT	
   Combo	
  
               (CareStart-­‐ICT	
   p=0.002).	
   The	
   PPVs	
   for	
   the	
   RDTs	
   were	
   comparable	
   in	
   detecting	
   P.	
   falciparum	
  



                                                                                                                    22	
  
                       Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                	
  


                       infection:	
   61.6%	
   by	
   ICT	
   Combo,	
   60.9%	
   for	
   CareStart,	
   and	
   59.8%	
   for	
   ParaScreen.	
   NPVs	
   were	
   the	
  
                       same	
  for	
  all	
  RDTs	
  at	
  97.7%.	
  Kappa	
  values	
  were	
  similar,	
  at	
  65-­‐66%	
  for	
  each	
  test,	
  with	
  none	
  reaching	
  
                       the	
  80%	
  value	
  indicative	
  of	
  a	
  very	
  good	
  agreement	
  with	
  microscopy.	
  
                       CareStart	
   was	
   the	
   best	
   performing	
   RDT	
   in	
   detecting	
   non-­‐falciparum	
   infections.	
   Sensitivity	
   of	
  
                       ParaScreen	
   was	
   82.4%,	
   while	
   CareStart	
   and	
   ICT	
   Combo	
   had	
   sensitivity	
   of	
   84.9%	
   (p=0.08).	
  
                       Specificity	
   of	
   CareStart	
   was	
   96.4%,	
   ICT	
   Combo	
   96.0%	
   and	
   ParaScreen	
   95.5%	
   (CareStart-­‐ParaScreen	
  
                       p<0.001).	
   The	
   PPV	
   and	
   NPV	
   of	
   all	
   three	
   RDTs	
   were	
   comparable	
   (p>0.05	
   for	
   all	
   pairs).	
   CareStart	
  
                       had	
   the	
   best	
   agreement	
   with	
   microscopy	
   overall	
   with	
   a	
   kappa	
   value	
   of	
   0.7565,	
   ICT	
   Combo	
   had	
  
                       kappa	
   0.7430	
   and	
   ParaScreen	
   kappa	
   0.7099	
   (CareStart-­‐ParaScreen	
   p=0.002,	
   ICT-­‐ParaScreen	
  
                       p=0.03).	
  
                       RDT	
  performance	
  was	
  also	
  assessed	
  at	
  each	
  health	
  center	
  individually	
  (tables	
  5	
  and	
  6).	
  None	
  of	
  
                       the	
  indicators	
  were	
  significantly	
  different	
  for	
  any	
  of	
  the	
  RDTs	
  at	
  Asendabo.	
  	
  

                       Table	
  5:	
  Performance	
  of	
  RDTs	
  at	
  each	
  health	
  center	
  in	
  detecting	
  P.	
  falciparum	
  infection.	
  Significant	
  
                       relationships	
  (p<0.05)	
  in	
  bold	
  type.	
  
               	
  
                                                           CareStart	
                    ParaScreen	
                     ICT	
  Combo	
                     McNemar	
  p	
  value	
  
	
  




               	
                                             n=791	
                         n=791	
                          n=791	
                 CS-­‐PS	
     CS-­‐ICT	
     ICT-­‐PS	
  
               Sensitivity	
  (95%	
  CI)	
           85.5%	
  (74.2-­‐93.1)	
        85.5%	
  (74.2-­‐93.1)	
        85.5%	
  (74.2-­‐93.1)	
           1.0	
          1.0	
               1.0	
  
SERBO	
  




               Specificity	
  (95%	
  CI)	
           95.5%	
  (93.7-­‐96.9)	
        95.2%	
  (93.4-­‐96.6)	
        96.2%	
  (94.5-­‐97.4)	
          0.41	
         0.03	
       0.008	
  
               PPV	
  (95%	
  CI)	
                   61.6%(50.5-­‐71.9)	
            60.2%	
  (49.2-­‐70.5)	
        65.4%	
  (54.0-­‐75.7)	
          0.90	
         0.61	
          0.45	
  
               NPV	
  (95%	
  CI)	
                   98.7%	
  (97.6-­‐99.4)	
        98.7%	
  (97.6-­‐99.4)	
        98.7%	
  (97.6-­‐99.4)	
           1.0	
          1.0	
               1.0	
  
               κ	
                                           0.6878	
                         0.6770	
                        0.7160	
                  0.75	
         0.33	
          0.19	
  

               	
                                             n=800	
                         n=800	
                          n=800	
                 CS-­‐PS	
     CS-­‐ICT	
     ICT-­‐PS	
  
               Sensitivity	
  (95%	
  CI)	
           82.5%	
  (72.4-­‐90.1)	
        82.5%	
  (72.4-­‐90.1)	
        82.5%	
  (72.4-­‐90.1)	
           1.0	
          1.0	
               1.0	
  
DIMTU	
  




               Specificity	
  (95%	
  CI)	
           94.2%	
  (92.2-­‐95.8)	
        93.2%	
  (91.1-­‐94.9)	
        94.0%	
  (92.0-­‐95.6)	
          0.04	
         0.32	
         0.04	
  
               PPV	
  (95%	
  CI)	
                   61.1%	
  (51.3-­‐70.3)	
        57.4%	
  (47.8-­‐66.6)	
        60.6%	
  (50.7-­‐69.8)	
          0.53	
          1.0	
          0.60	
  
               NPV	
  (95%	
  CI)	
                   98.0%	
  (96.6-­‐98.9)	
        98.0%	
  (96.6-­‐98.9)	
        98.0%	
  (96.6-­‐98.9)	
           1.0	
          1.0	
               1.0	
  
               κ	
                                           0.6635	
                         0.6337	
                        0.6591	
                  0.33	
         0.90	
          0.42	
  
               	
                                             n=794	
                         n=794	
                          n=794	
                 CS-­‐PS	
     CS-­‐ICT	
     ICT-­‐PS	
  
               Sensitivity	
  (95%	
  CI)	
           86.7%	
  (80.5-­‐91.5)	
        86.7%	
  (80.5-­‐91.5)	
        86.7%	
  (80.5-­‐91.5)	
           1.0	
          1.0	
               1.0	
  
ASENDABO	
  




               Specificity	
  (95%	
  CI)	
           85.2%	
  (82.2-­‐87.9)	
        85.4%	
  (82.4-­‐88.0)	
        85.4%	
  (82.4-­‐88.0)	
          0.32	
         0.32	
               1.0	
  
               PPV	
  (95%	
  CI)	
                   60.6%	
  (54.0-­‐66.9)	
        60.9%	
  (54.3-­‐67.1)	
        60.9%	
  (54.3-­‐67.1)	
           1.0	
          1.0	
               1.0	
  
               NPV	
  (95%	
  CI)	
                   96.1%	
  (94.1-­‐97.5)	
        96.1%	
  (94.1-­‐97.5)	
        96.1%	
  (94.1-­‐97.5)	
           1.0	
          1.0	
               1.0	
  
               κ	
                                           0.6204	
                         0.6229	
                        0.6229	
                  0.97	
         0.97	
               1.0	
  

                       At	
   Serbo	
   health	
   center,	
   a	
   difference	
   in	
   RDT	
   performance	
   for	
   P.	
   falciparum	
   and	
   non-­‐falciparum	
  
                       infection	
   was	
   seen.	
   For	
   P.	
   falciparum	
   specificity	
   of	
   ICT	
   Combo	
   was	
   higher	
   than	
   both	
   CareStart	
  
                       (p=0.03)	
   and	
   ParaScreen	
   (p=0.008).	
   However	
   the	
   specificity	
   of	
   CareStart	
   for	
   non-­‐falciparum	
  
                       infection	
  was	
  higher	
  than	
  for	
  ICT	
  Combo	
  (p=0.06)	
  and	
  ParaScreen	
  (p=0.01).	
  All	
  other	
  indicators	
  
                       were	
  similar	
  for	
  both	
  species	
  at	
  Serbo.	
  




                                                                                                    23	
  
                              Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                   	
  


                              At	
   Dimtu,	
   ParaScreen	
   had	
   lower	
   specificity	
   for	
   P.	
   falciparum	
   than	
   ICT	
   Combo	
   (p=0.04)	
   and	
  
                              CareStart	
   (p=0.04).	
   A	
   similar	
   relationship	
   was	
   seen	
   with	
   ParaScreen	
   having	
   lower	
   specificity	
   than	
  
                              ICT	
   Combo	
   (p=0.03)	
   and	
   CareStart	
   (p=0.002)	
   for	
   non-­‐falciparum	
   infections.	
   Sensitivity	
   of	
  
                              ParaScreen	
  was	
  also	
  lower	
  at	
  Dimtu	
  than	
  CareStart	
  (p=0.05)	
  and	
  ICT	
  Combo	
  (p=0.01)	
  in	
  detection	
  
                              of	
   non-­‐falciparum	
   species.	
   Furthermore,	
   ParaScreen	
   had	
   the	
   lowest	
   kappa	
   value	
   for	
   non-­‐
                              falciparum	
  infection	
  compared	
  to	
  ICT	
  Combo	
  (p=0.05)	
  and	
  CareStart	
  (p=0.04).	
  

                              Table	
  6:	
  Performance	
  of	
  RDTs	
  at	
  each	
  health	
  center	
  in	
  detecting	
  non-­‐falciparum	
  malaria.	
  Significant	
  
                              relationships	
  (p<0.05)	
  in	
  bold	
  type.	
  

	
                    	
  
                                                                       CareStart	
                           ParaScreen	
                         ICT	
  Combo	
                       McNemar	
  p	
  value	
  

                      	
                                                  n=791	
                                n=791	
                              n=791	
                   CS-­‐PS	
       CS-­‐ICT	
            ICT-­‐PS	
  
                      Sensitivity	
  (95%	
  CI)	
               90.4%	
  (79.0-­‐96.8)	
               90.4%	
  (79.0-­‐96.8)	
             86.5%	
  (74.2-­‐94.4)	
             1.0	
           0.16	
               0.16	
  
   SERBO	
  




                      Specificity	
  (95%	
  CI)	
               97.8%	
  (96.5-­‐98.8)	
               97.0%	
  (95.5-­‐98.1)	
             97.2%	
  (95.7-­‐98.2)	
            0.01	
           0.06	
               0.74	
  
                      PPV	
  (95%	
  CI)	
                       74.6%	
  (62.1-­‐84.7)	
               68.1%	
  (55.8-­‐78.8)	
             68.2%	
  (55.6-­‐79.1)	
            0.41	
           0.51	
                1.0	
  
                      NPV	
  (95%	
  CI)	
                       99.3%	
  (98.4-­‐99.8)	
               99.3%	
  (98.4-­‐99.8)	
             99.0%	
  (98.0-­‐99.6)	
             1.0	
           0.77	
               0.77	
  
                      Κ	
                                                 0.8032	
                              0.7588	
                             0.7439	
                    0.33	
           0.19	
               0.58	
  

                      	
                                                  n=800	
                                n=800	
                              n=800	
                   CS-­‐PS	
       CS-­‐ICT	
            ICT-­‐PS	
  
                      Sensitivity	
  (95%	
  CI)	
               88.4%	
  (81.0-­‐93.7)	
               82.1%	
  (73.8-­‐88.7)	
             89.3%	
  (82.0-­‐94.3)	
            0.05	
           0.65	
               0.01	
  
       DIMTU	
  




                      Specificity	
  (95%	
  CI)	
               95.5%	
  (93.7-­‐96.9)	
               93.6%	
  (91.5-­‐95.3)	
             95.1%	
  (93.2-­‐96.6)	
           0.002	
           0.41	
               0.03	
  
                      PPV	
  (95%	
  CI)	
                       76.2%	
  (67.9-­‐83.2)	
               67.6%	
  (59.1-­‐75.4)	
             74.6%	
  (66.4-­‐81.7)	
            0.16	
           0.80	
               0.31	
  
                      NPV	
  (95%	
  CI)	
                       98.1%	
  (96.7-­‐99.0)	
               97.0%	
  (95.4-­‐98.2)	
             98.2%	
  (96.9-­‐99.1)	
            0.30	
            1.0	
               0.22	
  
                      Κ	
                                                 0.7860	
                              0.6951	
                             0.7794	
                    0.04	
           0.91	
               0.05	
  
                      	
                                                  n=794	
                                n=794	
                              n=794	
                   CS-­‐PS	
       CS-­‐ICT	
            ICT-­‐PS	
  
                      Sensitivity	
  (95%	
  CI)	
               76.5%	
  (65.8-­‐85.2)	
               77.8%	
  (67.2-­‐86.3)	
             77.8%	
  (67.2-­‐86.3)	
            0.32	
           0.32	
                1.0	
  
       ASENDABO	
  




                      Specificity	
  (95%	
  CI)	
               95.7%	
  (93.9-­‐97.0)	
               95.7%	
  (93.9-­‐97.0)	
             95.7%	
  (93.9-­‐97.0)	
             1.0	
            1.0	
                1.0	
  
                      PPV	
  (95%	
  CI)	
                       66.7%	
  (56.1-­‐76.1)	
               67.0%	
  (56.6-­‐76.4)	
             67.0%	
  (56.6-­‐76.4)	
            0.17	
           0.17	
                1.0	
  
                      NPV	
  (95%	
  CI)	
                       97.3%	
  (95.8-­‐98.4)	
               97.4%	
  (96.0-­‐98.5)	
             97.4%	
  (96.0-­‐98.5)	
            0.30	
           0.30	
                1.0	
  
                      Κ	
                                                 0.6775	
                              0.6855	
                             0.6855	
                    0.88	
           0.88	
                1.0	
  
                              	
  

                              The	
   sensitivity	
   of	
   each	
   RDT	
   at	
   different	
   parasite	
   densities	
   was	
   assessed.	
   Parasite	
   count	
   of	
   each	
  
                              species	
  was	
  classified	
  into	
  categories:	
  fewer	
  than	
  200	
  p/µl	
  of	
  blood;	
  200-­‐500	
  p/µl;	
  500-­‐5000	
  p/µl;	
  
                              and	
  greater	
  than	
  5000	
  p/µl.	
  Sensitivity	
  was	
  calculated	
  for	
  P.	
  falciparum	
  infection	
  and	
  for	
  P.	
  vivax	
  
                              mono-­‐infection,	
  by	
  parasite	
  density	
  category.	
  Full	
  results	
  can	
  be	
  seen	
  in	
  table	
  7.	
  
                              There	
  was	
  no	
  evidence	
  for	
  any	
  statistically	
  significant	
  differential	
  performance	
  of	
  the	
  RDTs	
  within	
  
                              each	
   of	
   the	
   four	
   parasite	
   density	
   categories	
   for	
   P.	
   falciparum.	
   However	
   all	
   three	
   RDTs	
   showed	
   an	
  
                              increasing	
  sensitivity	
  with	
  increasing	
  P.	
  falciparum	
  density	
  (p<0.001)	
  (Figure	
  6).	
  This	
  relationship	
  is	
  
                              not	
   clearly	
   seen	
   with	
   P.	
   vivax	
   however,	
   likely	
   as	
   a	
   result	
   of	
   the	
   few	
   low	
   density	
   P.	
   vivax	
   infections	
  
                              seen.	
   At	
   high	
   density	
   P.	
   vivax	
   infection,	
   ParaScreen	
   has	
   lower	
   sensitivity	
   than	
   both	
   CareStart	
  
                              (p=0.03)	
  and	
  ICT	
  Combo	
  (p=0.01).	
  
                              	
  


                                                                                                                   24	
  
 Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                      	
  

 Table	
  7:	
  Sensitivities	
  of	
  RDTs	
  at	
  different	
  parasite	
  intensities.	
  Significant	
  relationships	
  (p<0.05)	
  in	
  bold	
  type.	
  

 	
                       	
                   CareStart	
                   ParaScreen	
                      ICT	
  Combo	
                      McNemar	
  p	
  value	
  
                                                                                                                                            	
                 	
             	
  
                          	
                          	
                              	
                              	
  
 P.	
  falciparum	
  
 Parasite	
  
 density	
                       N	
                          Sensitivity:	
  Pf	
  or	
  Pf+pan	
  (95%	
  CI)	
                      CS-­‐PS	
          CS-­‐ICT	
     ICT-­‐PS	
  
<200	
  p/µl	
                   63	
     66.7%	
  (53.7-­‐78.0)	
   66.7%	
  (53.7-­‐78.0)	
             66.7%	
  (53.7-­‐78.0)	
      1.0	
              1.0	
           1.0	
  
<500	
  p/µl	
                   90	
     70.0%	
  (59.4-­‐79.2)	
   70.0%	
  (59.4-­‐79.2)	
             70.0%	
  (59.4-­‐79.2)	
       1.0	
              1.0	
          1.0	
  
<5000	
  p/µl	
             188	
   78.7%	
  (72.2-­‐84.3)	
   78.7%	
  (72.2-­‐84.3)	
                   78.7%	
  (72.2-­‐84.3)	
       1.0	
              1.0	
          1.0	
  
>5000	
  p/µl	
             119	
   95.8%	
  (90.5-­‐98.6)	
   95.8%	
  (90.5-­‐98.6)	
                   95.8%	
  (90.5-­‐98.6)	
       1.0	
              1.0	
          1.0	
  
                                                                                                                                            	
                 	
             	
  
 P.	
  vivax	
                     	
                 	
                              	
                              	
  
 Parasite	
  
 density	
                       N	
                         Sensitivity:	
  pan	
  band	
  only	
  (95%	
  CI)	
                      CS-­‐PS	
          CS-­‐ICT	
     ICT-­‐PS	
  
<200	
  p/µl	
                    9	
     77.8%	
  (40.0-­‐97.2)	
   88.9%	
  (51.8-­‐99.7)	
             66.7%	
  (29.9-­‐92.5)	
     0.32	
              0.56	
         0.16	
  
<500	
  p/µl	
                   28	
     82.1%	
  (63.1-­‐93.9)	
   89.3%	
  (71.8-­‐97.7)	
             85.7%	
  (67.3-­‐96.0)	
      0.32	
             0.65	
         0.56	
  
<5000	
  p/µl	
             115	
   82.6%	
  (74.4-­‐89.0)	
   82.6%	
  (74.4-­‐89.0)	
                   82.6%	
  (74.4-­‐89.0)	
       1.0	
              1.0	
          1.0	
  
>5000	
  p/µl	
             135	
   86.9%	
  (79.9-­‐92.2)	
   82.3%	
  (74.6-­‐88.4)	
                   86.9%	
  (79.9-­‐92.2)	
      0.03	
              1.0	
         0.01	
  


 	
  

 Figure	
  6:	
  Graph	
  showing	
  association	
  between	
  RDT	
  sensitivity	
  and	
  parasite	
  density	
  




 In	
  summary,	
  there	
  was	
  no	
  evidence	
  for	
  a	
  difference	
  in	
  sensitivity,	
  PPV	
  or	
  NPV	
  of	
  each	
  of	
  the	
  three	
  
 RDTs	
  in	
  detecting	
  P.	
  falciparum	
  infections	
  when	
  results	
  were	
  pooled	
  for	
  all	
  sites.	
  Sensitivity	
  for	
  all	
  
 three	
   RDTs	
   was	
   85.3%,	
   NPV	
   was	
   very	
   high	
   at	
   97.7%	
   for	
   all	
   RDTs,	
   the	
   PPV	
   was	
   moderate	
   at	
   59.8-­‐
 61.6%.	
   Specificity	
   of	
   ParaScreen	
   was	
   lowest	
   at	
   91.5%	
   and	
   highest	
   for	
   ICT	
   at	
   92.2%	
   (p=0.002).	
  
 Assessing	
  RDT	
  performance	
  for	
  P.	
  falciparum	
  at	
  each	
  site	
  individually,	
  sensitivity	
  of	
  all	
  RDTs	
  was	
  
 similar,	
   while	
   both	
   CareStart	
   and	
   ICT	
   Combo	
   had	
   higher	
   specificity	
   than	
   ParaScreen	
   at	
   Dimtu	
   and	
  
 Serbo.	
  	
  




                                                                                             25	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                   	
  


For	
   detection	
   of	
   non-­‐falciparum	
   malaria	
   when	
   data	
   from	
   all	
   sites	
   were	
   pooled,	
   sensitivity,	
   PPV	
  
and	
   NPV	
   for	
   all	
   three	
   RDTs	
   was	
   statistically	
   similar.	
   Specificity	
   of	
   CareStart	
   was	
   highest	
   (96.4%)	
  
while	
   ParaScreen	
   had	
   lowest	
   specificity	
   (95.5%,	
   p<0.001).	
   When	
   assessing	
   performance	
   at	
   each	
  
site	
  separately,	
  CareStart	
  had	
  higher	
  specificity	
  than	
  ParaScreen	
  at	
  Serbo	
  and	
  Dimtu.	
  Sensitivity	
  
of	
   both	
   CareStart	
   and	
   ICT	
   Combo	
   was	
   higher	
   than	
   ParaScreen	
   at	
   Dimtu.	
   Furthermore,	
   level	
   of	
  
agreement	
  between	
  microscopy	
  and	
  RDT	
  results	
  were	
  significantly	
  better	
  for	
  CareStart	
  and	
  ICT	
  
Combo	
  than	
  for	
  ParaScreen	
  at	
  Dimtu	
  clinic.	
  	
  


8.5. RDT	
  ease-­‐of-­‐use	
  
All	
  HEWs	
  involved	
  in	
  the	
  ease	
  of	
  use	
  assessment	
  had	
  previously	
  used	
  the	
  ParaCheck	
  (falciparum-­‐
only)	
  test,	
  and	
  commented	
  on	
  the	
  importance	
  of	
  using	
  a	
  multi-­‐species	
  RDT	
  in	
  their	
  work	
  at	
  health	
  
post	
  and	
  in	
  the	
  community.	
  “This	
  kind	
  of	
  multi-­‐species	
  RDT	
  is	
  important	
  for	
  our	
  routine	
  activities	
  
and	
   should	
   be	
   immediately	
   applied	
   in	
   our	
   work”	
   noted	
   Amelwork	
   Getachew	
   from	
   Kejelo	
   health	
  
post	
  in	
  Tiro	
  Afeta	
  Woreda.	
  
HEWs	
   graded	
   various	
   aspects	
   of	
   each	
   RDT	
   on	
   a	
   scale	
   of	
   one	
   (difficult	
   to	
   use)	
   to	
   five	
   (very	
   easy	
   to	
  
use).	
   When	
   all	
   scored	
   components	
   were	
   averaged,	
   CareStart	
   had	
   the	
   highest	
   score	
   (4.4),	
   with	
  
ParaScreen	
  a	
  close	
  second	
  (4.3).	
  The	
  overall	
  rating	
  for	
  ICT	
  Combo	
  was	
  4.1	
  (Table	
  8).	
  	
  

Table	
  8:	
  Mean	
  HEW	
  scores	
  for	
  each	
  RDT	
  on	
  various	
  test	
  components	
  
                                                                                                Mean	
  HEW	
  score	
  (1:difficult	
  to	
  5:easy)	
  
	
                                                                         CareStart	
                           ParaScreen	
                         ICT	
  Combo	
  
Format	
                                                                         4.3	
                                 4.4	
                                4.4	
  
Lancet	
                                                                         4.4	
                                 4.5	
                                4.4	
  
Swab	
                                                                          4.6	
                                  3.7	
                                4.4	
  
Writing	
  on	
  the	
  device	
                                                4.6	
                                  4.6	
                                 4.3	
  
Filling	
  blood	
  collection	
  device	
                                       4.3	
                                 4.3	
                                3.8	
  
Emptying	
  blood	
  collection	
  device	
                                     4.8	
                                  4.2	
                                 4.2	
  
Adding	
  buffer	
  drops	
                                                      4.7	
                                 4.5	
                                 4.3	
  
Number	
  of	
  steps	
                                                          4.5	
                                 4.3	
                                4.0	
  
Waiting	
  time*	
  	
                                                           3.7	
                                 4.3	
                                4.4	
  
Interpretation	
  of	
  results	
                                                4.7	
                                 4.7	
                                3.6	
  
Instruction	
  leaflet	
                                                         4.7	
                                 4.6	
                                4.0	
  
                                             Mean	
  score	
                     4.5	
                                 4.4	
                                 4.2	
  
                                                        Rank	
                     1	
                                  2	
                                   3	
  
*(1=too	
  long,	
  5=good	
  length	
  of	
  time)	
  

A	
   number	
   of	
   HEWs	
   commented	
   on	
   the	
   packaging	
   of	
   CareStart	
   in	
   individual	
   ‘lab	
   in	
   a	
   pack’	
   sets	
  
with	
   all	
   components	
   needed	
   for	
   a	
   single	
   test	
   (cassette,	
   swab,	
   lancet,	
   blood	
   collection	
   pipette,	
  
buffer,	
  and	
  instructions)	
  included	
  in	
  each	
  packet.	
  Derartu	
  Jihad	
  from	
  Decha	
  Nedhi	
  health	
  post	
  in	
  
Tiro	
   Afeta	
   woreda	
   commented	
   “I	
   liked	
   having	
   all	
   materials	
   within	
   a	
   single	
   packet,	
   it	
   makes	
   moving	
  
from	
  house-­‐to-­‐house	
  easier	
  as	
  I	
  don’t	
  have	
  to	
  carry	
  a	
  large	
  box,	
  so	
  is	
  very	
  comfortable	
  for	
  my	
  work	
  in	
  
the	
  community.”	
  	
  



                                                                                           26	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                               	
  


Although	
   the	
   majority	
   liked	
   the	
   single	
   packaging	
  
style	
   of	
   CareStart,	
   there	
   were	
   individual	
   differences	
  
over	
   the	
   preferred	
   buffer	
   format.	
   Several	
   HEWs	
  
found	
   the	
   small,	
   individual	
   buffer	
   dispensers	
  
provided	
  with	
  CareStart	
  awkward	
  and	
  preferred	
  the	
  
large	
  dropper	
  bottles	
  of	
  ParaScreen	
  and	
  ICT	
  Combo.	
  
However	
   at	
   one	
   site	
   a	
   HEW	
   commented	
   that	
   it	
   was	
  
common	
   to	
   add	
   too	
   many	
   drops	
   of	
   buffer	
   from	
   the	
  
large	
   bottles,	
   and	
   thus	
   run	
   out	
   of	
   buffer	
   before	
  
completing	
  the	
   box	
   of	
  25	
   RDTs.	
   The	
  single-­‐use	
   buffer	
  
                                                                                                       Figure	
  6:	
  Blood	
  collection	
  devices.	
  
contains	
  the	
  correct	
  amount	
  of	
  drops	
  and	
  therefore	
                              CareStart	
  (top),	
  ParaScreen	
  (middle),	
  
reduces	
  mistakes	
  in	
  measuring	
  buffer.	
                                                    ICT	
  Combo	
  (bottom)	
  

The	
   most	
   common	
   complaint	
   about	
   RDTs	
   was	
   that	
  
swabs	
   provided	
   with	
   the	
   tests	
   were	
   not	
   moist	
   enough	
   to	
   successfully	
   clean	
   the	
   finger	
   before	
  
pricking.	
  This	
  occurred	
  with	
  all	
  three	
  brands,	
  most	
  frequently	
  with	
  ParaScreen.	
  	
  
The	
   three	
   RDTs	
   have	
   different	
   blood	
   collection	
   devices	
   (Figure	
   6),	
   and	
   this	
   was	
   a	
   source	
   of	
  
differing	
   opinions	
   among	
   the	
   HEWs	
   interviewed.	
   ParaScreen	
   has	
   a	
   sample	
   loop,	
   similar	
   to	
   that	
  
used	
  in	
  the	
  ParaCheck	
  device	
  which	
  all	
  HEWs	
  were	
  familiar	
  with.	
  CareStart	
  has	
  a	
  long,	
  thin,	
  plastic	
  
capillary	
   tube,	
   while	
   ICT	
   combo	
   has	
   a	
   shorter	
   plastic	
   pipette.	
   There	
   was	
   no	
   clear	
   favorite,	
   with	
  
individuals	
   having	
   different	
   preferences.	
   The	
   sample	
   loop	
   of	
   ParaScreen	
   was	
   seen	
   as	
   generally	
  
easy	
   to	
   fill,	
   but	
   emptying	
   could	
   be	
   problematic	
   with	
   splashing	
   of	
   sample	
   onto	
   the	
   cassette	
   if	
   due	
  
care	
   was	
   not	
   exercised.	
   One	
   HEW	
   found	
   the	
   sample	
   loop	
   particularly	
   difficult	
   to	
   use	
   with	
   small	
  
children.	
   The	
   comments	
   on	
   the	
   CareStart	
   capillary	
   tube	
   were	
   that	
   the	
   narrow	
   opening	
   could	
  
make	
   filling	
   and	
   emptying	
   difficult,	
   with	
   some	
   HEWs	
   finding	
   it	
   hard	
   to	
   correctly	
   measure	
   the	
  
amount	
   of	
   blood	
   needed	
   for	
   the	
   test.	
   The	
   ICT	
   Combo	
   pipette	
   has	
   a	
   wider	
   opening	
   than	
   the	
  
capillary	
  tube,	
  and	
  two	
  lines	
  for	
  measuring	
  the	
  amount	
  of	
  blood	
  needed.	
  However	
  a	
  number	
  of	
  
HEWs	
  stated	
  that	
  they	
  did	
  not	
  like	
  this	
  pipette	
  and	
  preferred	
  one	
  of	
  the	
  other	
  formats.	
  
It	
  is	
  important	
  to	
  be	
  able	
  to	
  write	
  on	
  the	
  cassette	
  device	
  to	
  correctly	
  identify	
  the	
  patient.	
  Some	
  
found	
   that	
   both	
   the	
   ParaScreen	
   and	
   CareStart	
   cassettes	
   were	
   difficult	
   to	
   write	
   clearly	
   on	
   using	
  
pen	
  and	
  pencil,	
  while	
  the	
  rounded	
  shape	
  of	
  ICT	
  Combo	
  meant	
  that	
  there	
  was	
  very	
  little	
  space	
  to	
  
write	
   on	
   the	
   cassette.	
   Ensuring	
   that	
   health	
   posts	
   have	
   fine-­‐tip	
   marker	
   pens	
   for	
   labeling	
   RDTs	
  
would	
  address	
  this	
  difficulty.	
  
The	
   most	
   salient	
   factor	
   influencing	
   the	
   ease	
   of	
   use	
   of	
   each	
   RDT	
   was	
   the	
   presence	
   of	
   clear	
  
markings	
  on	
  the	
  cassette	
  for	
  interpretation	
  of	
  results.	
  As	
  shown	
  in	
  figure	
  2,	
  ParaScreen	
  has	
  the	
  
most	
  detailed	
  labeling	
  with	
  the	
  band	
  positions	
  marked	
  with	
  ’C’	
  at	
  the	
  control	
  region,	
  then	
  ‘Pan’	
  
and	
  ‘Pf’	
  in	
  the	
  test	
  region.	
  CareStart	
  has	
  ‘C’	
  at	
  the	
  control	
  band,	
  with	
  ‘1’	
  for	
  P.	
  falciparum	
  and	
  ‘2’	
  
for	
   Pan.	
   ICT	
   Combo	
   has	
   very	
   limited	
   markings	
   on	
   the	
   cassette,	
   with	
   only	
   a	
   ‘C’	
   at	
   the	
   control	
   band	
  
region	
   of	
   the	
   strip	
   and	
   ‘T’	
   at	
   the	
   test	
   end.	
   It	
   has	
   no	
   separate	
   markings	
   for	
   Pf	
   and	
   Pan	
   bands,	
  
therefore	
   the	
   user	
   must	
   remember	
   the	
   band	
   location	
   when	
   reading	
   results.	
   This	
   may	
   lead	
   to	
  
incorrect	
  interpretation	
  of	
  positive	
  tests	
  at	
  a	
  much	
  higher	
  frequency	
  from	
  ICT	
  Combo	
  than	
  with	
  
CareStart	
  or	
  ParaScreen.	
  




                                                                                    27	
  
       Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                  	
  


       8.6. Heat	
  stability	
  
       The	
  stability	
  of	
  CareStart	
  and	
  ParaScreen	
  was	
  shown	
  to	
  be	
  superior	
  to	
  that	
  of	
  ICT	
  Combo	
  during	
  
       the	
  heat	
  stability	
  testing.	
  HRP2	
  and	
  pLDH	
  bands	
  were	
  assessed	
  individually	
  for	
  P.	
  falciparum	
  and	
  
       P.	
  vivax	
  at	
  concentrations	
  of	
  200	
  and	
  2000	
  p/µl.	
  




Figure	
  7:	
  Heat	
  stability	
  results	
  presented	
  by	
  storage	
  temperature,	
  band	
  and	
  parasite	
  intensity	
  



       The	
  HRP2	
  band	
  of	
  ICT	
  Combo	
  was	
  stable	
  for	
  only	
  four	
  hours	
  at	
  60°C.	
  The	
  ICT	
  Combo	
  HRP2	
  band	
  
       showed	
   very	
   weak	
   test	
   bands	
   after	
   30	
   days	
   at	
   35°C	
   and	
   45°C.	
   However	
   both	
   CareStart	
   and	
  
       ParaScreen	
   HRP2	
   bands	
   were	
   stable	
   for	
   72	
   hours	
   at	
   60°C,	
   and	
   90	
   days	
   at	
   35°C	
   and	
   45°C,	
   the	
  
       duration	
  of	
  the	
  testing	
  period.	
  

       All	
  RDTs	
  had	
  pLDH	
  bands	
  that	
  were	
  more	
  heat-­‐sensitive	
  than	
  HRP2,	
  appearing	
  faint	
  throughout	
  
       the	
   testing	
   with	
   both	
   high	
   and	
   low	
   density	
   P.	
   vivax	
   infections.	
   Both	
   CareStart	
   and	
   ParaScreen	
  
       pLDH	
  bands	
  completed	
  the	
  testing	
  procedure	
  at	
  35°C,	
  45°C	
  and	
  60°C.	
  The	
  ICT	
  Combo	
  pLDH	
  band	
  



                                                                                     28	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                         	
  


failed	
   to	
   detect	
   P.	
   vivax	
   at	
   200	
   p/µl	
   after	
   60	
   days	
   at	
   35°C	
   and	
   after	
   60	
   days	
   at	
   45°C.	
   However	
   after	
  
60	
  days	
  storage	
  at	
  these	
  temperatures,	
  ICT	
  Combo	
  did	
  successfully	
  detect	
  P.	
  vivax	
  at	
  the	
  higher	
  
density	
  of	
  2000	
  p	
  µl.	
  At	
  60°C	
  ICT	
  Combo	
  pLDH	
  failed	
  after	
  72	
  hours.	
  



9.       Discussion	
  
9.1. RDT	
  performance	
  compared	
  to	
  microscopy	
  
RDTs	
   for	
   malaria	
   are	
   being	
   increasingly	
   adopted	
   across	
   endemic	
   countries	
   to	
   strengthen	
  
parasitological	
   diagnosis	
   and	
   appropriate	
   management	
   of	
   all	
   cases	
   of	
   fever.	
   They	
   are	
   particularly	
  
valuable	
   for	
   use	
   in	
   resource-­‐poor	
   areas	
   where	
   microscopy	
   is	
   not	
   available,	
   in	
   epidemic	
   situations,	
  
and	
   to	
   ensure	
   rational	
   use	
   of	
   ACTs.	
   A	
   large	
   number	
   of	
   products	
   are	
   now	
   available,	
   and	
   while	
  
some	
   assessment	
   of	
   performance	
   has	
   been	
   undertaken	
   by	
   WHO/FIND	
   [12],	
   it	
   is	
   important	
   that	
  
the	
  most	
  appropriate	
  RDT	
  for	
  each	
  transmission	
  situation	
  and	
  operational	
  context	
  is	
  identified.	
  
Malaria	
   transmission	
   in	
   Ethiopia	
   is	
   unstable	
   and	
   seasonal,	
   with	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax	
  
endemic,	
   and	
   contributing	
   to	
   the	
   burden	
   of	
   malaria.	
   An	
   appropriate	
   RDT	
   for	
   implementation	
  
across	
   Ethiopia	
   should	
   be	
   able	
   to	
   detect	
   P.	
   falciparum	
   and	
   P.	
   vivax,	
   be	
   highly	
   sensitive	
   and	
  
specific,	
  and	
  also	
  able	
  to	
  detect	
  low	
  density	
  infections.	
  The	
  RDT	
  tests	
  evaluated	
  in	
  this	
  study	
  are	
  
all	
   combination	
   HRP2	
   and	
   pan-­‐pLDH	
   tests,	
   therefore	
   are	
   able	
   to	
   identify	
   P.	
   falciparum	
   or	
   mixed	
   P.	
  
falciparum	
  and	
  non-­‐falciparum	
  infections,	
  as	
  well	
  as	
  non-­‐falciparum	
  infection.	
  
A	
  significant	
  peak	
  in	
  infection	
  was	
  seen	
  in	
  children	
  aged	
  between	
  six	
  and	
  ten	
  years	
  of	
  age.	
  The	
  
reported	
  mosquito	
  net	
  use	
  was	
  similar	
  for	
  all	
  age	
  groups,	
  however	
  it	
  is	
  possible	
  that	
  responses	
  to	
  
this	
  question	
  were	
  biased	
  since	
  many	
  patients	
  expected	
  that	
  they	
  may	
  receive	
  additional	
  nets	
  if	
  
they	
  reported	
  not	
  using	
  a	
  net.	
  Other	
  studies	
  have	
  shown	
  that	
  individuals	
  aged	
  between	
  five	
  and	
  
19	
  are	
  generally	
  less	
  likely	
  to	
  be	
  sleeping	
  under	
  a	
  mosquito	
  net	
  [27].	
  
The	
  sensitivities	
  of	
  the	
  RDTs	
  in	
  detecting	
  P.	
  falciparum	
  compared	
  to	
  P.	
  vivax	
  were	
  similar	
  (85.3%	
  
for	
   P.	
   falciparum,	
   82.4-­‐84.9%	
   for	
   P.	
   vivax).	
   This	
   indicates	
   that	
   development	
   of	
   pLDH	
   tests	
   has	
  
progressed	
  to	
  the	
  point	
  where	
  it	
  is	
  now	
  of	
  comparable	
  performance	
  to	
  that	
  of	
  HRP2	
  detection.	
  	
  
Overall,	
   there	
   was	
   little	
   difference	
   in	
   the	
   performance	
   of	
   the	
   three	
   RDTs	
   when	
   compared	
   to	
  
microscopy.	
  In	
  general,	
  ICT	
  Combo	
  and	
  CareStart	
  showed	
  better	
  performance	
  than	
  ParaScreen.	
  
ParaScreen	
   had	
   the	
   lowest	
   specificity	
   of	
   all	
   RDTs	
   for	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax.	
   When	
  
assessing	
   performance	
   at	
   each	
   health	
   center	
   individually,	
   CareStart	
   and	
   ICT	
   Combo	
   again	
   had	
  
higher	
   specificity	
   than	
   ParaScreen	
   for	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax	
   infection.	
   Furthermore,	
  
sensitivity	
  of	
  ParaScreen	
  to	
  detect	
  P.	
  vivax	
  was	
  lower	
  than	
  for	
  both	
  ICT	
  Combo	
  and	
  CareStart.	
  	
  
Both	
   the	
   HRP2	
   and	
   pLDH	
   lines	
   of	
   all	
   three	
   RDTs	
   showed	
   reduced	
   sensitivity	
   with	
   low	
   density	
  
infections.	
   The	
   linear	
   relationship	
   between	
   parasite	
   density	
   and	
   sensitivity	
   was	
   strongest	
   with	
  
HRP2	
  bands	
  of	
  all	
  RDTs,	
  however	
  the	
  small	
  number	
  of	
  low	
  density	
  P.	
  vivax	
  infections	
  mean	
  that	
  
there	
  is	
  not	
  sufficient	
  power	
  to	
  compare	
  sensitivity	
  at	
  different	
  P.	
  vivax	
  densities.	
  	
  

In	
  a	
  low	
  transmission	
  setting	
  such	
  as	
  Ethiopia,	
  where	
  individuals	
  are	
  unlikely	
  to	
  have	
  any	
  acquired	
  
immunity	
   to	
   Plasmodium	
   infection,	
   and	
   a	
   minority	
   of	
   fevers	
   will	
   be	
   due	
   to	
   malaria,	
   it	
   is	
   more	
  
acceptable	
  for	
  the	
  RDT	
  to	
  result	
  in	
  false	
  positives	
  than	
  false	
  negatives.	
  The	
  RDTs	
  should	
  therefore	
  
reveal	
  few	
  false	
  negatives;	
  a	
  high	
  sensitivity	
  is	
  more	
  important	
  than	
  a	
  high	
  specificity.	
  




                                                                                         29	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                              	
  


There	
   were	
   a	
   total	
   of	
   31	
   individuals	
   with	
   false	
   negative	
   RDT	
   results.	
   Half	
   of	
   the	
   P.	
   falciparum	
  
infections	
   which	
   resulted	
   in	
   false	
   negative	
   RDT	
   results	
   (10/19)	
   were	
   infections	
   with	
   parasite	
  
density	
  below	
  200	
  parasites/µl.	
  However	
  fewer	
  of	
  the	
  false	
  negative	
  RDTs	
  for	
  P.	
  vivax	
  infection	
  
could	
   be	
   attributed	
   to	
   low	
   density	
   infection:	
   only	
   a	
   third	
   (4/12)	
   were	
   below	
   200	
   parasites/µl.	
  
These	
  false	
  negatives	
  are	
  concerning,	
  and	
  emphasize	
  that	
  RDTs	
  are	
  of	
  value	
  in	
  specific	
  contexts,	
  
but	
  will	
  not	
  replace	
  microscopy	
  as	
  the	
  gold	
  standard	
  diagnostic	
  technique.	
  
All	
  three	
  RDTs	
  showed	
  a	
  greater	
  number	
  of	
  false	
  positive	
  results	
  than	
  false	
  negatives.	
  ParaScreen	
  
had	
  208	
  false	
  positive	
  results,	
  ICT	
  194	
  and	
  CareStart	
  190.	
  False	
  positive	
  results	
  did	
  not	
  appear	
  to	
  
be	
  due	
  to	
  residual	
  HRP2	
  antigen	
  from	
  cleared	
  infection,	
  since	
  only	
  11%	
  of	
  the	
  total	
  false	
  positive	
  
results	
   were	
   from	
   patients	
   who	
   had	
   received	
   any	
   anti-­‐malarial	
   treatment	
   in	
   the	
   previous	
   four	
  
weeks.	
   A	
   potential	
   alternative	
   explanation	
   for	
   this	
   level	
   of	
   false	
   positives	
   is	
   sequestration:	
  
erythrocytes	
  containing	
  mature	
  parasites	
  clump	
  together	
  in	
  the	
  microvasculature,	
  therefore	
  are	
  
not	
   seen	
   in	
   the	
   peripheral	
   circulation	
   and	
   blood	
   films,	
   while	
   antigen	
   continues	
   to	
   be	
   released	
  
[28].	
   It	
   may	
   also	
   be	
   possible	
   that	
   the	
   parasite	
   density	
   was	
   too	
   low	
   to	
   be	
   seen	
   by	
   microscopy,	
   but	
  
there	
  was	
  sufficient	
  parasite	
  antigen	
  to	
  results	
  in	
  a	
  positive	
  RDT	
  [29].	
  Whilst	
  previous	
  work	
  has	
  
shown	
  that	
  some	
  RDT	
  false	
  positives	
  are	
  due	
  to	
  patients	
  with	
  residual	
  gametocytemia	
  [30],	
  this	
  
association	
  was	
  not	
  seen	
  in	
  the	
  current	
  study.	
  


9.2. Findings	
  in	
  similar	
  studies	
  
The	
   first	
   round	
   of	
   WHO/FIND	
   product	
   testing	
   of	
   malaria	
   RDTs	
   [12]	
   is	
   the	
   most	
   comprehensive	
  
assessment	
   of	
   malaria	
   RDTs	
   on	
   the	
   market	
   to	
   date,	
   and	
   CareStart,	
   ParaScreen	
   and	
   ICT	
   Combo	
  
were	
   included	
   in	
   the	
   WHO/FIND	
   round	
   one	
   product	
   testing.	
   CareStart	
   and	
   ParaScreen	
   RDTs	
  
evaluated	
   by	
   WHO/FIND	
   were	
   the	
   exact	
   same	
   antigen-­‐detecting	
   combinations.	
   However	
   ICT	
  
Combo	
  tested	
  by	
  WHO/FIND	
  was	
  HRP2-­‐aldose	
  combination,	
  therefore	
  only	
  the	
  HRP2	
  band	
  results	
  
are	
  directly	
  comparable	
  with	
  our	
  study.	
  
CareStart	
  performed	
  better	
  in	
  the	
  WHO/FIND	
  testing	
  in	
  detecting	
  low	
  density	
  P.	
  falciparum	
  and	
  P.	
  
vivax	
  infection	
  than	
  found	
  in	
  the	
  current	
  study.	
  Similarly,	
  the	
  sensitivity	
  of	
  CareStart	
  was	
  found	
  to	
  
be	
   higher	
   in	
   the	
   WHO/FIND	
   testing	
   for	
   high	
   density	
   infection	
   with	
   either	
   species	
   than	
   in	
   the	
  
current	
  findings.	
  ParaScreen	
  however	
  was	
  found	
  to	
  have	
  a	
  better	
  sensitivity	
  in	
  the	
  current	
  study	
  
than	
  the	
  WHO/FIND	
  evaluation	
  in	
  detecting	
  low	
  density	
  infections	
  with	
  either	
  P.	
  falciparum	
  or	
  P.	
  
vivax.	
  However,	
  when	
  detecting	
  high	
  density	
  parasitemia,	
  ParaScreen	
  had	
  superior	
  detection	
  rate	
  
for	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax	
   in	
   the	
   WHO/FIND	
   work	
   compared	
   to	
   the	
   current	
   study.	
   ICT	
  
Combo	
  showed	
  higher	
  sensitivity	
  to	
  low	
  and	
  high	
  density	
  P.	
  falciparum	
  infection	
  in	
  the	
  WHO/FIND	
  
evaluation	
   than	
   the	
   current	
   study.	
   However	
   the	
   ICT	
   Combo	
   aldose	
   band	
   included	
   in	
   the	
   test	
  
evaluated	
  by	
  WHO/FIND	
  was	
  very	
  poor	
  for	
  low	
  density	
  P.	
  vivax	
  infection,	
  detecting	
  none	
  of	
  the	
  
samples	
   tested.	
   The	
   pLDH	
   band	
   of	
   ICT	
   Combo	
   was	
   comparatively	
   better	
   at	
   detecting	
   low	
   density	
  
P.	
  vivax	
  in	
  our	
  findings,	
  indicating	
  that	
  the	
  HRP2/pLDH	
  combination	
  of	
  ICT	
  Combo	
  is	
  superior	
  to	
  
HRP2/aldose	
  in	
  a	
  low	
  transmission	
  setting.	
  

There	
   are	
   a	
   number	
   of	
   possible	
   reasons	
   for	
   this	
   differential	
   performance	
   of	
   the	
   same	
   RDTs	
   in	
   the	
  
two	
   studies.	
   Firstly,	
   the	
   WHO/FIND	
   product	
   testing	
   was	
   carried	
   out	
   by	
   highly	
   experienced	
  
technicians	
   in	
   a	
   controlled	
   laboratory	
   setting,	
   while	
   this	
   study	
   involved	
   health	
   center	
   staff	
  
engaged	
   in	
   their	
   routine	
   activities.	
   Differences	
   in	
   performance	
   could	
   also	
   be	
   attributed	
   to	
  
variation	
   between	
   product	
   lots,	
   or	
   the	
   impact	
   of	
   transportation	
   and	
   storage	
   conditions	
   in	
   the	
  



                                                                                   30	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                           	
  


current	
   study	
   reducing	
   test	
   sensitivity.	
   While	
   this	
   study	
   used	
   fresh	
   finger-­‐prick	
   blood	
   samples	
  
from	
   patients,	
   the	
   WHO/FIND	
   testing	
   used	
   stored	
   wild	
   type	
   samples	
   which	
   had	
   undergone	
   a	
  
freeze-­‐thaw	
   process,	
   and	
   therefore	
   these	
   samples	
   may	
   have	
   different	
   characteristics	
   to	
   that	
   of	
  
fresh	
   blood	
   samples.	
   It	
   is	
   also	
   possible	
   that	
   genetic	
   variation	
   of	
   target	
   antigens	
   can	
   lead	
   to	
  
variation	
   in	
   performance	
   of	
   RDTs	
   in	
   different	
   regions.	
   Baker	
   et	
   al.	
   found	
   significant	
   genetic	
  
variation	
   in	
   PfHRP2	
   isolates	
   from	
   different	
   countries	
   which	
   was	
   associated	
   with	
   reduced	
  
sensitivity	
  of	
  RDTs	
  at	
  low	
  density	
  infections	
  [31].	
  	
  
In	
  the	
  Ethiopian	
  context,	
  a	
  number	
  of	
  studies	
  have	
  been	
  carried	
  out	
  to	
  assess	
  performance	
  of	
  a	
  
variety	
   of	
   multi-­‐species	
   malaria	
   RDTs,	
   to	
   guide	
   the	
   FMOH	
   decision	
   of	
   a	
   replacement	
   for	
  
ParaCheck-­‐Pf.	
  
ParaScreen	
  HRP2-­‐pLDH	
  combination	
  RDT	
  was	
  assessed	
   by	
   Endeshaw	
   and	
   colleagues	
  [32] as	
  part	
  
of	
   a	
   population-­‐based	
   cross-­‐sectional	
   survey.	
   The	
   sensitivity	
   of	
   ParaScreen	
   was	
   found	
   to	
   be	
   51.4%	
  
for	
   P.	
   falciparum	
   and	
   30.7%	
   for	
   non-­‐falciparum	
   infections,	
   both	
   much	
   lower	
   than	
   the	
   sensitivity	
  
found	
   in	
   the	
   current	
   study,	
   which	
   was	
   found	
   to	
   be	
   85.3%	
   and	
   82.4%,	
   respectively.	
   A	
   further	
  
evaluation	
   of	
   ParaScreen	
   HRP2-­‐pLDH	
   was	
   conducted	
   at	
   health	
   facilities	
   in	
   northwest	
   Ethiopia	
  
during	
   the	
   peak	
   transmission	
   season	
   [33].	
   In	
   that	
   study	
   the	
   sensitivity	
   of	
   ParaScreen	
   was	
   79.6%	
  
for	
   P.	
   falciparum	
   and	
   74.4%	
   for	
   P.	
   vivax,	
   significantly	
   higher	
   than	
   performance	
   in	
   the	
   cross-­‐
sectional	
  study	
  and	
  comparable	
  to	
  findings	
  in	
  the	
  current	
  survey.	
  
However	
  it	
  should	
  be	
  noted	
  that	
  the	
  low	
  sensitivity	
  of	
  ParaScreen	
  found	
  by	
  Endeshaw	
  is	
  similar	
  
to	
  that	
  in	
  the	
  WHO/FIND	
  product	
  testing.	
  

A	
   CareStart	
   pf-­‐HRP2/pv-­‐PLDH	
   combination	
   RDT	
   has	
   also	
   been	
   evaluated	
   at	
   two	
   locations	
   in	
  
Ethiopia	
   [6,	
   7].	
   The	
   sensitivity	
   of	
   the	
   test	
   in	
   detecting	
   P.	
   falciparum	
   by	
   HRP2	
   at	
   Wondo	
   Genet	
   was	
  
99.4%	
  and	
  96.4%	
  in	
  Jimma,	
  both	
  very	
  much	
  higher	
  than	
  for	
  CareStart	
  pf/pan	
  in	
  the	
  current	
  study.	
  
The	
  P.	
  vivax-­‐specific	
  pLDH	
  band	
  of	
  the	
  pf/pv	
  RDT	
  also	
  appeared	
  to	
  have	
  much	
  higher	
  sensitivity	
  
(99.4%	
  in	
  Wondo	
  Genet,	
  95.3%	
  in	
  Jimma)	
  than	
  the	
  Plasmodium-­‐specific	
  pLDH	
  band	
  of	
  pf/pan	
  RDT	
  
tested	
  in	
  this	
  study	
  (82.4%).	
  However	
  since	
  transmission	
  of	
  P.	
  malariae	
  occurs	
  in	
  some	
  malarious	
  
areas	
  of	
  Ethiopia,	
  adoption	
  of	
  an	
  RDT	
  which	
  is	
  specific	
  for	
  P.	
  falciparum	
  and	
  P.	
  vivax,	
  but	
  unable	
  to	
  
detect	
  the	
  rarer	
  Plasmodium	
  species	
  is	
  inadvisable.	
  

CareStart	
  also	
  manufactures	
  a	
  pf-­‐pLDH/pan-­‐pLDH	
  combination	
  test,	
  which	
  has	
  been	
  evaluated	
  in	
  
Madagascar	
  [18]	
  and	
  Myanmar	
  [15].	
  Sensitivity	
  of	
  the	
  falciparum-­‐specific	
  pLDH	
  band	
  appears	
  high	
  
(97%	
   in	
   Madagascar,	
   93.5%	
   in	
   Myanmar)	
   compared	
   to	
   the	
   HRP2	
   band	
   in	
   the	
   CareStart	
   test	
  
evaluated	
   in	
   the	
   current	
   study	
   (85.3%).	
   The	
   sensitivity	
   of	
   pan-­‐pLDH	
   in	
   the	
   current	
   study	
   falls	
  
between	
   the	
   values	
   determined	
   in	
   Madagascar	
   and	
   Myanmar	
   (84.9%,	
   92.3%	
   and	
   78.5%,	
  
respectively).	
  These	
  studies	
  indicate	
  that	
  pLDH	
  is	
  reliable	
  in	
  detecting	
  both	
  P.	
  falciparum	
  and	
  non-­‐
falciparum	
  malaria.	
  


9.3. Suitability	
  of	
  RDT	
  for	
  use	
  at	
  health	
  posts	
  
All	
   HEWs	
   involved	
   in	
   the	
   ease	
   of	
   use	
   assessment	
   were	
   comfortable	
   using	
   the	
   selected	
   multi-­‐
species	
   RDTs.	
   The	
   familiarity	
   of	
   HEWs	
   across	
   Ethiopia	
   with	
   RDT	
   procedures	
   from	
   experience	
  
using	
  ParaCheck	
  means	
  that	
  they	
  will	
  be	
  able	
  to	
  adopt	
  a	
  multi-­‐species	
  RDT	
  without	
  the	
  need	
  for	
  
extensive	
  training	
  and	
  supervision.	
  




                                                                                  31	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                        	
  


The	
  ease	
  of	
  use	
  assessment	
  within	
  the	
  current	
  study	
  identified	
  a	
  number	
  of	
  factors	
  which	
  HEWs	
  
feel	
   are	
   important	
   to	
   be	
   considered	
   when	
   choosing	
   a	
   replacement	
   for	
   ParaCheck.	
   The	
   most	
  
salient	
   points	
   raised	
   by	
   HEWs	
   were	
   the	
   packaging	
   of	
   the	
   RDT	
   and	
   labeling	
   on	
   the	
   device.	
   Most	
  
HEWs	
   felt	
   that	
   the	
   ‘lab	
   in	
   a	
  pack’	
  style	
  of	
  CareStart	
  would	
  be	
  very	
  beneficial	
  for	
  their	
  work	
  in	
  the	
  
community.	
  All	
  HEWs	
  also	
  found	
  ICT	
  Combo	
  the	
  most	
  difficult	
  RDT	
  to	
  interpret,	
  since	
  the	
  cassette	
  
did	
   not	
   individually	
   label	
   the	
   P.	
   falciparum	
   and	
   pan	
   bands.	
   They	
   felt	
   that	
   the	
   labeling	
   on	
   both	
  
CareStart	
  and	
  ParaScreen	
  was	
  sufficient	
  to	
  allow	
  them	
  to	
  easily	
  interpret	
  results.	
  	
  
The	
  WHO/FIND	
  study	
  assessed	
  ease	
  of	
  use	
  according	
  to	
  a	
  number	
  of	
  criteria:	
  i)	
  blood	
  safety;	
  ii)	
  
time	
   to	
   obtain	
   results;	
   iii)	
   number	
   of	
   timed	
   steps;	
   and	
   iv)	
   quality	
   of	
   instructions.	
   Also	
   included	
  
were	
   the	
   format	
   type,	
   blood	
   transfer	
   method	
   and	
   items	
   included	
   in	
   package.	
   The	
   results	
  
indicated	
  a	
  better	
  score	
  for	
  CareStart	
  than	
  ParaScreen	
  due	
  only	
  to	
  the	
  presence	
  of	
  a	
  diagram	
  of	
  
method	
  with	
  CareStart.	
  In	
  the	
  current	
  study,	
  ease	
  of	
  use	
  was	
  assessed	
  by	
  the	
  HEWs	
  who	
  will	
  be	
  
the	
  end	
  users	
  of	
  the	
  selected	
  RDT,	
  and	
  the	
  RDTs	
  were	
  evaluated	
  in	
  an	
  operational	
  setting.	
  This	
  
yields	
  more	
  information	
  about	
  the	
  ease	
  of	
  use,	
  and	
  the	
  criteria	
  which	
  peripheral	
  health	
  workers	
  
feel	
  are	
  important	
  in	
  selecting	
  an	
  RDT.	
  
Results	
   of	
   heat	
   stability	
   assessment	
   for	
   CareStart	
   were	
   consistent	
   between	
   the	
   WHO/FIND	
  
testing	
  and	
  the	
  current	
  study,	
  with	
  strong	
  detection	
  rates	
  after	
  storage	
  at	
  35°C	
  and	
  45°	
  for	
  low	
  
and	
   high	
   density	
   infections	
   of	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax.	
   However	
   ParaScreen	
   performed	
  
better	
  in	
  heat	
  stability	
  during	
  the	
  current	
  study	
  than	
  in	
  the	
  WHO/FIND	
  product	
  testing.	
  While	
  in	
  
the	
   current	
   study	
   ParaScreen	
   consistently	
   detected	
   low	
   and	
   high	
   density	
   P.	
   falciparum	
   after	
  
storage	
   at	
   35°C	
   and	
   45°C,	
   in	
   the	
   WHO/FIND	
   testing	
   ParaScreen	
   failed	
   to	
   detect	
   20%	
   of	
   low	
   density	
  
P.	
   falciparum	
   samples	
   after	
   storage	
   at	
   35°C,	
   and	
   55%	
   of	
   samples	
   after	
   storage	
   at	
   45°C.	
   ParaScreen	
  
failed	
  to	
  detect	
  all	
  low	
  density	
  P.	
  vivax	
  samples,	
  as	
  well	
  as	
  failing	
  to	
  detect	
  a	
  small	
  number	
  of	
  high	
  
density	
   P.	
   vivax	
   samples.	
   ICT	
   Combo	
   performed	
   well	
   in	
   WHO/FIND	
   testing,	
   with	
   the	
   HRP2	
   band	
  
detecting	
   all	
   P.	
   falciparum	
   bands,	
   however	
   the	
   pan	
   aldose	
   band	
   failed	
   for	
   all	
   low	
   density	
  
infections.	
   Since	
   the	
   procedures	
   for	
   heat	
   stability	
   testing	
   were	
   the	
   same	
   in	
   both	
   studies,	
   the	
  
reason	
   for	
   the	
   discrepancy	
   in	
   heat	
   stability	
   results,	
   particularly	
   for	
   ParaScreen	
   pLDH	
   band	
   with	
  
low	
  density	
  infections,	
  is	
  not	
  clear.	
  	
  


9.4. Additional	
  factors	
  determining	
  FMOH	
  RDT	
  choice	
  	
  
In	
   addition	
   to	
   RDT	
   validity,	
   ease	
   of	
   use	
   and	
   heat	
   stability,	
   a	
   number	
   of	
   other	
   factors	
   must	
   be	
  
considered	
  when	
  selecting	
  a	
  multi-­‐species	
   RDT	
   for	
   Ethiopia.	
   Perhaps	
   most	
   pressing	
   is	
   cost	
   of	
   the	
  
RDTs,	
   since	
   a	
   large	
   quantity	
   will	
   need	
   to	
   be	
   procured	
   in	
   order	
   to	
   cover	
   all	
   health	
   posts	
   in	
  
malarious	
  areas	
  of	
  the	
  country.	
  
The	
  shelf	
  life	
  of	
  RDTs	
  should	
  also	
  be	
  considered,	
  since	
  malaria	
  transmission	
  is	
  seasonal	
  in	
  some	
  
regions,	
  and	
  therefore	
  cases	
  may	
  only	
  be	
  seen	
  during	
  a	
  few	
  months	
  of	
  the	
  year.	
  In	
  some	
  highland	
  
fringes	
  malaria	
  transmission	
  does	
  not	
  occur	
  every	
  year,	
  but	
  the	
  risk	
  of	
  epidemics	
  in	
  these	
  areas	
  
highlight	
  the	
  need	
  to	
  maintain	
  the	
  capacity	
  for	
  accurate	
  diagnosis	
  of	
  suspected	
  cases	
  at	
  all	
  times.	
  

The	
  FMOH	
  may	
  also	
  wish	
  to	
  consider	
  the	
  capacity	
  of	
  the	
  manufacturer	
  to	
  provide	
  a	
  large	
   number	
  
of	
   RDTs,	
   and	
   to	
   a	
   consistent	
   high	
   quality.	
   This	
   may	
   involve	
   further	
   investigation	
   of	
   RDT	
  
production	
  sites	
  and	
  quality	
  control	
  methods.	
  
A	
  number	
  of	
  multi-­‐species	
  RDT	
  studies	
  are	
  underway	
  in	
  different	
  areas	
  of	
  Ethiopia.	
  While	
  here	
  we	
  
present	
   results	
   from	
   one	
   area	
   of	
   Oromia	
   Regional	
   State,	
   it	
   may	
   be	
   advisable	
   to	
   make	
   a	
   final	
  


                                                                                32	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                                	
  


decision	
   based	
   on	
   all	
   recent	
   available	
   studies	
   across	
   Ethiopia	
   to	
   ensure	
   that	
   the	
   chosen	
   RDT	
   is	
  
suitable	
  for	
  all	
  malarious	
  areas	
  of	
  the	
  country.	
  

	
  
10. Conclusions	
  and	
  recommendations	
  
10.1. Which	
  RDT	
  to	
  choose	
  for	
  Ethiopia	
  
Here	
  we	
  assessed	
  the	
  performance	
  of	
  three	
  RDTs	
  in	
  a	
  low	
  transmission	
  setting	
  in	
  Ethiopia,	
  where	
  
both	
   P.	
   falciparum	
   and	
   P.	
   vivax	
   transmission	
   takes	
   place.	
   The	
   objective	
   of	
   the	
   study	
   was	
   to	
  
identify	
  a	
  multi-­‐species	
  RDT	
  which	
  is	
  appropriate	
  for	
  use	
  by	
  HEWs	
  at	
  health	
  post	
  level,	
  and	
  is	
  able	
  
to	
   detect	
   both	
   P.	
   falciparum	
   and	
   P.	
   vivax	
   with	
   high	
   sensitivity	
   and	
   specificity,	
   particularly	
   in	
   low	
  
density	
  infections.	
  

There	
   was	
   a	
   limited	
   difference	
   in	
   validity	
   of	
   the	
   three	
   tests	
   in	
   the	
   health	
   center-­‐based	
   data	
   in	
   this	
  
study.	
  The	
  low	
  sensitivity	
  of	
  ParaScreen	
  in	
  detecting	
  low	
  density	
  P.	
   vivax	
  infections	
  in	
  WHO/FIND	
  
product	
  testing	
  is	
  concerning,	
  although	
  not	
  replicated	
  in	
  the	
  current	
  work.	
  This	
  finding	
  perhaps	
  
indicates	
   a	
   significant	
   variation	
   in	
   ParaScreen	
   RDT	
   quality	
   between	
   lots.	
   A	
   discrepancy	
   in	
  
ParaScreen	
   performance	
   was	
   also	
   seen	
   in	
   the	
   heat	
   stability	
   testing	
   for	
   this	
   study	
   and	
   WHO/FIND.	
  
ICT	
   Combo	
   in	
   our	
   study	
   with	
   was	
   found	
   to	
   have	
   low	
   sensitivity	
   to	
   low	
   density	
   P.	
   vivax	
   after	
  
storage	
  at	
  35°C,	
  and	
  to	
  P	
   falciparum	
   indicating	
  that	
  it	
  would	
  not	
  give	
  reliable	
  results	
  when	
  stored	
  
at	
  ambient	
  temperatures	
  in	
  a	
  health	
  post.	
  
Since	
   the	
   chosen	
   multi-­‐species	
   RDT	
   will	
   be	
   primarily	
   used	
   at	
   health	
   post	
   level	
   by	
   HEWs	
   with	
  
limited	
   training	
   and	
   support,	
   their	
   feedback	
   on	
   which	
   RDTs	
   were	
   the	
   simplest	
   to	
   use	
   and	
  
interpret	
  is	
  as	
  important	
  as	
  the	
  quantitative	
  assessment.	
  HEWs	
  felt	
  that	
  ICT	
  Combo	
  was	
  the	
  most	
  
difficult	
   to	
   interpret	
   results	
   from,	
   and	
   therefore	
   could	
   lead	
   to	
   treatment	
   errors,	
   while	
  
interpretation	
   of	
   CareStart	
   and	
   ParaScreen	
   was	
   unambiguous.	
   The	
   packing	
   format	
   of	
   CareStart	
  
was	
  particularly	
  favored,	
  and	
  is	
  highly	
  appropriate	
  for	
  routine	
  HEW	
  activities	
  moving	
  within	
  the	
  
community.	
  	
  

Therefore	
   on	
   the	
   basis	
   of	
   findings	
   from	
   this	
   study,	
   and	
   the	
   published	
   report	
   from	
   WHO/FIND,	
   we	
  
recommend	
  CareStart	
  pf-­‐HRP2/pan-­‐PLDH	
  for	
  adoption	
  by	
  the	
  FMOH	
  to	
  allow	
  for	
  parasitological	
  
diagnosis	
  at	
  health	
  post	
  level.	
  


10.2. Future	
  activities	
  after	
  selection	
  of	
  a	
  multi-­‐species	
  RDT	
  in	
  Ethiopia	
  
Before	
   the	
   chosen	
   multi-­‐species	
   RDT	
   is	
   distributed	
   at	
   health	
   posts	
   across	
   the	
   country,	
   there	
  
should	
  be	
  a	
  detailed	
  training	
  needs	
  assessment	
  to	
  ensure	
  that	
  HEWs	
  are	
  able	
  to	
  correctly	
  use	
  this	
  
RDT	
  and	
  interpret	
  results.	
  Just	
  as	
  important	
  is	
  to	
  ensure	
  that	
  case	
  management	
  procedures	
  are	
  
assessed	
   at	
   health	
   posts,	
   and	
   that	
   HEWs	
   are	
   using	
   the	
   RDT	
   diagnosis	
   to	
   correctly	
   inform	
   their	
  
treatment	
  decisions.	
  
After	
  roll-­‐out	
  of	
  the	
  multi-­‐species	
  RDT,	
  HEWs	
  will	
  be	
  able	
  to	
  confirm	
  diagnosis	
  of	
  non-­‐falciparum	
  
malaria	
  for	
  the	
  first	
  time,	
  and	
  it	
  is	
  vital	
  that	
  they	
  are	
  able	
  to	
  provide	
  appropriate	
  treatment.	
  It	
  may	
  
be	
   necessary	
   to	
   provide	
   refresher	
   training	
   on	
   the	
   treatment	
   procedure	
   and	
   dosages	
   for	
  
chloroquine,	
   as	
   well	
   as	
   to	
   evaluate	
   the	
   drug	
   supply	
   management	
   system	
   to	
   ensure	
   that	
   health	
  
posts	
  have	
  chloroquine	
  available.	
  




                                                                                    33	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                               	
  


Lot	
   testing	
   of	
   RDTs	
   must	
   take	
   place	
   periodically	
   to	
   ensure	
   that	
   the	
   procured	
   product	
   is	
   of	
  
consistent	
  high	
  quality	
  and	
  producing	
  reliable	
  results.	
  Since	
  EHNRI	
  is	
  a	
  designated	
  WHO/FIND	
  lot	
  
testing	
  center,	
  this	
  should	
  be	
  a	
  very	
  simple	
  procedure	
  to	
  implement.	
  
It	
   may	
   be	
   necessary	
   to	
   review	
   the	
   current	
   malaria	
   treatment	
   policy	
   with	
   regard	
   to	
   mixed	
   P.	
  
falciparum	
   and	
   P.	
   vivax	
   infection.	
   A	
   pf/pan	
   combination	
   RDT	
   will	
   show	
   a	
   P.	
   falciparum	
   plus	
   pan	
  
positive	
   result	
   with	
   both	
   mixed	
   and	
   P.	
   falciparum	
   mono-­‐infections.	
   Since	
   the	
   current	
   policy	
  
advises	
   administering	
   both	
   CoArtem	
   and	
   chloroquine	
   for	
   mixed	
   infections,	
   there	
   is	
   a	
   strong	
  
possibility	
  of	
  chloroquine	
  being	
  given	
  to	
  patients	
  unnecessarily.	
  	
  

Development	
  and	
  improvement	
  of	
  malaria	
  RDTs	
  continues	
  to	
  take	
  place,	
  and	
  there	
  is	
  a	
  need	
  for	
  
regular	
   review	
   of	
   literature	
   and	
   new	
   research	
   findings	
   to	
   ensure	
   that	
   Ethiopia	
   continues	
   to	
  
procure	
   the	
   most	
   appropriate	
   RDT.	
   Of	
   particular	
   note	
   is	
   the	
   ongoing	
   WHO/FIND	
   product	
  
evaluation,	
  with	
  round	
  two	
  results	
  due	
  in	
  April	
  2010,	
  and	
  round	
  three	
  during	
  the	
  following	
  year.	
  
After	
   procurement	
   of	
   a	
   chosen	
   multi-­‐species	
   RDT,	
   there	
   should	
   be	
   monitoring	
   and	
   evaluation	
  
processes	
   in	
   place	
   to	
   assess	
   RDT	
   performance	
   in	
   an	
   operational	
   setting	
   across	
   Ethiopia,	
   health	
  
worker	
  performance	
  and	
  malaria	
  case	
  management	
  with	
  this	
  new	
  tool,	
  as	
  well	
  as	
  the	
  collection	
  
of	
   detailed	
   epidemiological	
   information	
   about	
   the	
   incidence	
   and	
   species	
   distribution	
   of	
   malaria	
  
using	
  RDT	
  results.	
  	
  	
  
	
  
	
  
	
  




                                                                            34	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                        	
  


11. References	
  
1.	
        Federal	
  Democratic	
  Republic	
  of	
  Ethiopia,	
  M.o.H.,	
  Guideline	
  for	
  malaria	
  epidemic	
  prevention	
  
            and	
  control	
  in	
  Ethiopia.	
  2nd	
  edition.	
  2004,	
  Addis	
  Ababa,	
  Ethiopia.	
  
2.	
        Federal	
  Democratic	
  Republic	
  of	
  Ethiopia,	
  M.o.H.,	
  Health	
  and	
  Health	
  Related	
  Indicators.	
  
            2007.	
  
3.	
        Federal	
  Democratic	
  Republic	
  of	
  Ethiopia,	
  M.o.H.,	
  National	
  Five	
  Year	
  Strategic	
  Plan	
  for	
  
            Malaria	
  Prevention	
  and	
  Control	
  in	
  Ethiopia.	
  2006,	
  Addis	
  Ababa,	
  Ethiopia.	
  
4.	
        Abeku,	
  T.,	
  et	
  al.,	
  Spatial	
  and	
  temporal	
  variations	
  of	
  malaria	
  epidemic	
  risk	
  in	
  Ethiopia:	
  factors	
  
            involved	
  and	
  implications.	
  Acta	
  Tropica,	
  2003.	
  87:	
  p.	
  331-­‐340.	
  
5.	
        Tacconelli,	
  E.,	
  et	
  al.,	
  Does	
  antibiotic	
  exposure	
  increase	
  the	
  risk	
  of	
  methicillin-­‐resistant	
  
            Staphylococcus	
  aureus	
  (MRSA)	
  isolation?	
  A	
  systematic	
  review	
  and	
  meta-­‐analysis.	
  The	
  
            Journal	
  of	
  Antimicrobial	
  Chemotherapy,	
  2008.	
  61(1):	
  p.	
  26-­‐38.	
  
6.	
        Mekonnen,	
  Z.,	
  et	
  al.,	
  Evaluation	
  of	
  the	
  performance	
  of	
  CareStart	
  Malaria	
  Pf/Pv	
  Combo	
  Rapid	
  
            Diagnostic	
  Test	
  for	
  the	
  diagnosis	
  of	
  malaria	
  in	
  Jimma,	
  southwestern	
  Ethiopia.	
  Acta	
  Tropica,	
  
            2010.	
  113(3):	
  p.	
  285-­‐288.	
  
7.	
        Sharew,	
  B.,	
  et	
  al.,	
  Evaluation	
  of	
  the	
  performance	
  of	
  CareStart	
  Malaria	
  Pf/Pv	
  Combo	
  and	
  
            Paracheck	
  Pf	
  tests	
  for	
  the	
  diagnosis	
  of	
  malaria	
  in	
  Wondo	
  Genet,	
  southern	
  Ethiopia.	
  Acta	
  
            Tropica,	
  2009.	
  111(3):	
  p.	
  321-­‐324.	
  
8.	
        Nigussie,	
  D.,	
  et	
  al.,	
  Evaluation	
  of	
  Paracheck	
  pf	
  and	
  Parascreen	
  pan/pf	
  tests	
  for	
  the	
  diagnosis	
  
            of	
  malaria	
  in	
  an	
  endemic	
  area,	
  South	
  Ethiopia.	
  Ethiopian	
  Medical	
  Journal,	
  2008.	
  46(4):	
  p.	
  
            375-­‐381.	
  
9.	
        Murray,	
  C.K.,	
  et	
  al.,	
  Update	
  on	
  rapid	
  diagnostic	
  testing	
  for	
  malaria.	
  Clinical	
  Microbiology	
  
            Reviews,	
  2008.	
  21(1):	
  p.	
  97-­‐110.	
  
10.	
       Swarthout,	
  T.D.,	
  et	
  al.,	
  Paracheck-­‐Pf	
  accuracy	
  and	
  recently	
  treated	
  Plasmodium	
  falciparum	
  
            infections:	
  is	
  there	
  a	
  risk	
  of	
  over-­‐diagnosis?	
  Malaria	
  Journal,	
  2007.	
  6:	
  p.	
  58.	
  
11.	
       Houze,	
  S.,	
  et	
  al.,	
  PfHRP2	
  and	
  PfLDH	
  antigen	
  detection	
  for	
  monitoring	
  the	
  efficacy	
  of	
  
            artemisinin-­‐based	
  combination	
  therapy	
  (ACT)	
  in	
  the	
  treatment	
  of	
  uncomplicated	
  falciparum	
  
            malaria.	
  Malaria	
  Journal,	
  2009.	
  8(1):	
  p.	
  211.	
  
12.	
       WHO,	
  Malaria	
  rapid	
  diagnostic	
  test	
  performance:	
  results	
  of	
  WHO	
  product	
  testing	
  of	
  malaria	
  
            RDTs:	
  round	
  1.	
  2009.	
  
13.	
       Msellem,	
  M.I.,	
  et	
  al.,	
  Influence	
  of	
  rapid	
  malaria	
  diagnostic	
  tests	
  on	
  treatment	
  and	
  health	
  
            outcome	
  in	
  fever	
  patients,	
  Zanzibar:	
  a	
  crossover	
  validation	
  study.	
  PLoS	
  Medicine,	
  2009.	
  
            6(4):	
  p.	
  e1000070.	
  
14.	
       Hopkins,	
  H.,	
  et	
  al.,	
  Comparison	
  of	
  HRP2-­‐	
  and	
  pLDH-­‐based	
  rapid	
  diagnostic	
  tests	
  for	
  malaria	
  
            with	
  longitudinal	
  follow-­‐up	
  in	
  Kampala,	
  Uganda.	
  The	
  American	
  Journal	
  of	
  Tropical	
  
            Medicine	
  and	
  Hygiene,	
  2007.	
  76(6):	
  p.	
  1092-­‐1097.	
  
15.	
       Ashley,	
  E.A.,	
  et	
  al.,	
  Evaluation	
  of	
  three	
  parasite	
  lactate	
  dehydrogenase-­‐based	
  rapid	
  
            diagnostic	
  tests	
  for	
  the	
  diagnosis	
  of	
  falciparum	
  and	
  vivax	
  malaria.	
  Malaria	
  Journal,	
  2009.	
  8:	
  
            p.	
  241.	
  
16.	
       van	
  den	
  Broek,	
  I.,	
  et	
  al.,	
  Evaluation	
  of	
  three	
  rapid	
  tests	
  for	
  diagnosis	
  of	
  P.	
  falciparum	
  and	
  P.	
  
            vivax	
  malaria	
  in	
  Colombia.	
  The	
  American	
  Journal	
  of	
  Tropical	
  Medicine	
  and	
  Hygiene,	
  2006.	
  
            75(6):	
  p.	
  1209-­‐15.	
  




                                                                               35	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                                 	
  


17.	
       Pattanasin,	
  S.,	
  et	
  al.,	
  Evaluation	
  of	
  a	
  new	
  Plasmodium	
  lactate	
  dehydrogenase	
  assay	
  
            (OptiMAL-­‐IT)	
  for	
  the	
  detection	
  of	
  malaria.	
  Transactions	
  of	
  the	
  Royal	
  Society	
  of	
  Tropical	
  
            Medicine	
  and	
  Hygiene,	
  2003.	
  97(6):	
  p.	
  672-­‐674.	
  
18.	
       Ratsimbasoa,	
  A.,	
  et	
  al.,	
  Which	
  malaria	
  rapid	
  test	
  for	
  Madagascar?	
  Field	
  and	
  laboratory	
  
            evaluation	
  of	
  three	
  tests	
  and	
  expert	
  microscopy	
  of	
  samples	
  from	
  suspected	
  malaria	
  patients	
  
            in	
  Madagascar.	
  The	
  American	
  Journal	
  of	
  Tropical	
  Medicine	
  and	
  Hygiene,	
  2007.	
  76(3):	
  p.	
  
            481-­‐485.	
  
19.	
       Bharti,	
  P.K.,	
  et	
  al.,	
  The	
  usefulness	
  of	
  a	
  new	
  rapid	
  diagnostic	
  test,	
  the	
  First	
  Response	
  Malaria	
  
            Combo	
  (pLDH/HRP2)	
  card	
  test,	
  for	
  malaria	
  diagnosis	
  in	
  the	
  forested	
  belt	
  of	
  central	
  India.	
  
            Malaria	
  Journal,	
  2008.	
  7:	
  p.	
  126.	
  
20.	
       Fogg,	
  C.,	
  et	
  al.,	
  Assessment	
  of	
  three	
  new	
  parasite	
  lactate	
  dehydrogenase	
  (pan-­‐pLDH)	
  tests	
  
            for	
  diagnosis	
  of	
  uncomplicated	
  malaria.	
  Transactions	
  of	
  the	
  Royal	
  Society	
  of	
  Tropical	
  
            Medicine	
  and	
  Hygiene,	
  2008.	
  102(1):	
  p.	
  25-­‐31.	
  
21.	
       Gerstl,	
  S.,	
  et	
  al.,	
  Assessment	
  of	
  two	
  malaria	
  rapid	
  diagnostic	
  tests	
  in	
  children	
  under	
  five	
  
            years	
  of	
  age,	
  with	
  follow-­‐up	
  of	
  false-­‐positive	
  pLDH	
  test	
  results,	
  in	
  a	
  hyperendemic	
  falciparum	
  
            malaria	
  area,	
  Sierra	
  Leone.	
  Malaria	
  Journal,	
  2010.	
  9(1):	
  p.	
  28.	
  
22.	
       Jones,	
  S.R.,	
  S.	
  Carley,	
  and	
  M.	
  Harrison,	
  An	
  introduction	
  to	
  power	
  and	
  sample	
  size	
  
            estimation.	
  Emerging	
  Medicine	
  Journal,	
  2003.	
  20:	
  p.	
  453-­‐458.	
  
23.	
       WHO,	
  Methods	
  for	
  field	
  trials	
  of	
  malaria	
  rapid	
  diagnostic	
  tests.	
  2009,	
  World	
  Health	
  
            Organization,	
  Geneva.	
  
24.	
       Lachenbruch,	
  P.A.	
  and	
  C.J.	
  Lynch,	
  Assessing	
  screening	
  tests:	
  extensions	
  of	
  McNemar's	
  test.	
  
            Stat	
  Med,	
  1998.	
  17(19):	
  p.	
  2207-­‐17.	
  
25.	
       McGinn,	
  T.,	
  Tips	
  for	
  learners	
  of	
  evidence-­‐based	
  medicine:	
  3.	
  Measures	
  of	
  observer	
  variability	
  
            (kappa	
  statistic).	
  Canadian	
  Medical	
  Association	
  Journal,	
  2004.	
  171:	
  p.	
  1369-­‐1373.	
  
26.	
       Federal	
  Democratic	
  Republic	
  of	
  Ethiopia,	
  M.o.H.,	
  Malaria	
  diagnosis	
  and	
  treatment	
  
            guidelines	
  for	
  health	
  workers	
  in	
  Ethiopia.	
  2004,	
  Addis	
  Ababa,	
  Ethiopia.	
  
27.	
       Noor,	
  A.,	
  et	
  al.,	
  The	
  use	
  of	
  insecticide	
  treated	
  nets	
  by	
  age:	
  implications	
  for	
  universal	
  
            coverage	
  in	
  Africa.	
  BMC	
  Public	
  Health,	
  2009.	
  9(1):	
  p.	
  369.	
  
28.	
       Dondorp,	
  A.M.,	
  et	
  al.,	
  Estimation	
  of	
  the	
  total	
  parasite	
  biomass	
  in	
  acute	
  falciparum	
  malaria	
  
            from	
  plasma	
  PfHRP2.	
  PLoS	
  Medicine,	
  2005.	
  2(8):	
  p.	
  e204.	
  
29.	
       Bell,	
  D.R.,	
  D.W.	
  Wilson,	
  and	
  L.B.	
  Martin,	
  False-­‐positive	
  results	
  of	
  a	
  Plasmodium	
  falciparum	
  
            histidine-­‐rich	
  protein	
  2-­‐detecting	
  malaria	
  rapid	
  diagnostic	
  test	
  due	
  to	
  high	
  sensitivity	
  in	
  a	
  
            community	
  with	
  fluctuating	
  low	
  parasite	
  density.	
  The	
  American	
  Journal	
  of	
  Tropical	
  
            Medicine	
  and	
  Hygiene,	
  2005.	
  73(1):	
  p.	
  199-­‐203.	
  
30.	
       Tjitra,	
  E.,	
  et	
  al.,	
  Persistent	
  ICT	
  malaria	
  P.f/P.v	
  panmalarial	
  and	
  HRP2	
  antigen	
  reactivity	
  after	
  
            treatment	
  of	
  Plasmodium	
  falciparum	
  malaria	
  is	
  associated	
  with	
  gametocytemia	
  and	
  results	
  
            in	
  false-­‐positive	
  diagnoses	
  of	
  Plasmodium	
  vivax	
  in	
  convalescence.	
  Journal	
  of	
  Clinical	
  
            Microbiology,	
  2001.	
  39(3):	
  p.	
  1025-­‐1031.	
  
31.	
       Baker,	
  J.,	
  et	
  al.,	
  Genetic	
  diversity	
  of	
  Plasmodium	
  falciparum	
  histidine-­‐rich	
  protein	
  2	
  (PfHRP2)	
  
            and	
  its	
  effect	
  on	
  the	
  performance	
  of	
  PfHRP2-­‐based	
  rapid	
  diagnostic	
  tests.	
  The	
  Journal	
  of	
  
            Infectious	
  Diseases,	
  2005.	
  192(5):	
  p.	
  870-­‐877.	
  
32.	
       Endeshaw,	
  T.,	
  et	
  al.,	
  Evaluation	
  of	
  light	
  microscopy	
  and	
  rapid	
  diagnostic	
  test	
  for	
  the	
  
            detection	
  of	
  malaria	
  under	
  operational	
  field	
  conditions:	
  a	
  household	
  survey	
  in	
  Ethiopia.	
  
            Malaria	
  Journal,	
  2008.	
  7:	
  p.	
  118.	
  



                                                                             36	
  
Evaluation	
  of	
  three	
  RDTs	
  in	
  Oromia	
  Region,	
  Ethiopia	
  	
                                                                     	
  


33.	
       Endeshaw,	
  T.,	
  et	
  al.,	
  Comparison	
  of	
  Parascreen	
  Pan/Pf,	
  Paracheck	
  Pf	
  and	
  light	
  microscopy	
  
            for	
  detection	
  of	
  malaria	
  among	
  febrile	
  patients,	
  Northwest	
  Ethiopia.	
  Trans	
  R	
  Soc	
  Trop	
  Med	
  
            Hyg,	
  2010(in	
  press).	
  
	
  
	
  




                                                                          37	
  
12.       Appendices	
  	
  
	
  
        WHO/EHNRI malaria rapid diagnostic test quality assurance initiative,
       Ethiopian Health and Nutrition Research Institute, Addis Abeba, Ethiopia

                    Temperature Stability Testing of Malaria Rapid Diagnostic Tests

              1.       AIM
                        To provide guidelines for temperature stability testing of RDTs using quality control
                        samples to assess if the sensitivity of the RDT batch is retained after exposure to various
                        testing conditions.


              2.       BACKGROUND
                        Malaria RDTs have particular application                 in remote locations where
                        microscopic diagnosis is not available (WHO             2000). However, use of these
                        diagnostic devices in such areas often                  entails storage outside the
                        manufacturer’s recommended temperatures for             prolonged periods, which may
                        affect performance of the tests.

                        A multi-center heat stability study on various products conducted in 2004
                        showed decreasing intensity of RDT test lines when exposed to temperatures
                        above 35oC over time. The extent of degradation (loss of sensitivity at low
                        parasite density) varied widely between the several products tested.

                        Therefore, there is a need for a systematic analysis of the performance of new
                        RDTs to predict how temperature insults are likely to affect their performance.
                        This RDT heat stability testing will provide some information about RDT
                        survival when used and stored outside the specified temperatures.

              3.       PURPOSE
                        This Standard Operating Procedure (SOP) describes the process of heat
                        stability testing of new malaria RDTs.

              4.       SCOPE
                        This procedure applies to the WHO malaria rapid diagnostic test quality
                        assurance initiative. The SOP is not to be modified except by the Designated
                        WHO Officer. The procedure is not part of a formal WHO product testing
                        program, and is carried out independently by laboratories collaborating with
                        WHO in developing testing methods for RDTs. The results do not constitute
                        WHO endorsement of a product, or of product performance.

              5.       REAGENTS, EQUIPMENT
                        RDTs for HS testing
                                Pf-detecting only                                             125*
                                Pf- and Pv-detecting                                          200*


                                                            38	
  
     (*allows for few spare RDTs)


     1-20 µl Pipette                                                  1
     -70°C freezer                                                    1
     Refrigerator thermometer (range: -20°C to 50 °C)                 1
                                        °            °
     Incubator thermometer (range: 0 C to 100 C)                      3
      °
     4 C refrigerator                                                 1
     Incubators (range: 20°C to 80°C)                                 1
     Pipette tips (1-20 µl capacity)
     Quality Control samples: 0 (clean negative), 200, 2000 parasites/µl


6.   PROCEDURE
      6.1 Immediate QA testing
             a. Follow the procedure for receipt of RDTs (SOP 2.02).
             b. Perform QA testing of malaria RDTs (SOP 2.06).
             c. If RDTs pass immediate QA assessment, proceed to temperature stability
                testing. If RDTs fail, follow procedures for confirmatory QA assessment
                for failed results.

     6.2     Temperature stability testing
     a.      Prior to testing, incubators should be set and stabilized (~3 days) at 35,
             45 and 55/60oC. Appropriate thermometer should be placed inside each
             incubator for daily monitoring of temperature.

     b.      After performing the immediate QA testing, put the appropriate
             number of RDTs to be tested in the incubators. The RDT buffers for
             each kit should be kept at 4oC (except for RDTs with buffers inside the
             pouch/package, i.e., Diamed OptiMal- IT). If immediate QA is done
             simultaneously with Day 0 /Time 0, it will also serve as baseline
             testing.

     c.      Record immediate QA/baseline results in RDT Temperature Stability
             Result Sheet (Form X).

     d.      Note the dates and time of future testing intervals on a chart.

     e.      Thirty minutes to one hour prior to each testing interval, thaw the
             required QC samples to be used for testing the RDTs.
             Note: Do not re-use frozen aliquots of QC samples – ONE USE
             ONLY.

     f.      At the appropriate testing interval, take out from the incubators the
             correct number of RDTs to be tested. RDTs should be brought to room
             temperature before testing.



                                            39	
  
              g.     Perform the tests following manufacturer’s procedures, but using a
                     pipette to measure blood volume. The table below describes the
                     number of RDTs to be tested for each time interval:

Pf Test only (1 neg; 1 Pf200 (x 2 RDT); 1 Pf2000 (x 1 RDT) = 4 RDTs/testing interval)
 Temp. Immediate Hr0           Hr4      Hr12 Hr24 Hr48 Hr72                       Total
         QC
          22                                                                        22
 55/60            4             4         4           4     4     4                 24
                 D0             D5       D10         D20   D30   D60    D90
  35              4             4         4           4     4     4      4          28
  45              4             4         4           4     4     4      4          28
                                                                                   102

            Pf/Pv Test (1 neg; 1 Pf200 (x 2 RDT); 1 Pf2000 (x 1 RDT);
        1 Pv200 (x 2 RDT); 1 Pv2000 (x 1 RDT) = 7 RDTs/testing interval)
 Temp. Immediate Hr0 Hr4 Hr12 Hr24 Hr48 Hr72                            Total
          QC
           34                                                            34
 55/60               7      7      7       7    7       7                42
                    D0     D5     D10 D20 D30 D60 D90
  35                 7      7      7       7    7       7      7         49
  45                 7      7      7       7    7       7      7         49
                                                                         174

             Addition of spare RDTs at 35oC and 45oC to allow repeat or longer testing, if
             needed.

              h.     Read the intensity of test bands by visual comparison with the standard
                     RDT rating chart. RDTs should be rated 0, 1, 2 & 3 according to
                     increasing intensity (0 = negative results).

              i.     Record the results on the RDT Temperature Stability Result Sheet
                     (Form X). If 2 consecutive negative results, testing can be stopped at
                     dilution of consecutive failures.




                                            40	
  
                                   INFORMED	
  CONSENT	
  
            Evaluation of Malaria Rapid Diagnostics Test (RDTs) in Ethiopia

Malaria Consortium Ethiopia is carrying out a study to learn more about alternative methods
of malaria diagnosis. Malaria is a disease caused by parasites that can lead to fever,
headaches and muscle pain. In severe cases that are not treated on time, malaria can lead to
death. In order to treat you better, we need a test that detects malaria accurately and quickly,
and that is easy to use. Such tests are now available and are called rapid diagnostic tests.
Rapid diagnostic tests are designed to provide test results within a few minutes to the patient,
so that patients do not have to wait or return to the health facility for their test results. Rapid
diagnostic tests are simple to use, requiring no extra equipment and only little training. The
general procedure for these tests is that a drop of blood is taken from the patient, which is
then mixed with some liquid (called a buffer) and applied to the end of a piece of paper. The
liquid containing the blood is gradually absorbed by the paper, moving over areas that have
been prepared with chemicals that will react depending on whether the patient has malaria or
not. Whether a patient has malaria or not is indicated by different lines on the paper.
You are being asked to participate in a study in which we will compare the performance of
three different rapid diagnostic tests to microscopy. Microscopy is the presently used method
of malaria diagnosis. The reason for this study is that there are problems and limitations
associated with the microscopic diagnosis of malaria. These include the fact that the
diagnosis normally requires at least 60 minutes to conduct and depends on good techniques,
reagents, a well-functioning microscope and good training of the person performing the test.
In most health facilities it is very difficult to meet these requirements. Also, malaria can lead
to a rapid deterioration of patients’ health, and it is therefore important to find ways of
diagnose the disease in the shortest possible time.
If you agree to be involved in this study, a blood sample will be taken by pricking the finger,
and be tested for the presence of malaria parasites. In addition, we will ask you some
questions about yourself and review your clinic card. We estimate that the entire interview
and blood sample collection will take about 15 minutes of your time. If we find that you have
malaria we will treat you with Coartem or chloroquine as in the standard clinic procedures.
These drugs are recommended for treatment of malaria by the Federal Ministry of Health.
Side effects are rare, but treatment with Coartem may cause dizziness, nausea, vomiting or
headache. Possible side effects of chloroquine include nausea, vomiting and diarrhea.
Participation in this study will not cause a loss of privacy. Your name will not be shared with
individuals outside the study team, and all information will be assigned a code in order to
preserve anonymity after data has been entered into a computer. Specimens will not have you
name on them, and will not be stored long-term.
Information gathered will be used to establish which diagnostic method performs best and
thus help the Federal Ministry of Health to decide which rapid diagnostic test to distribute to
the health workers.




                                                41	
  
Your participation in this research study is completely voluntary. If you decide to participate
you may skip any question that you do not wish to answer. If you do not wish to participate
in this study you will still receive the same level of care at the health center. By signing this
form, you are telling the staff that you are willing to participate in the study described above.
If you agree to participate, you should sign below:


Name of participant (printed):____________________________ Patient #__________


Signature/fingerprint of participant:______________________ Date:_____________


Signature of investigator:________________________________ Date:_____________




                                CONTACT INFORMATION

If you have any questions about your rights as a participant in this study or the manner in
which it was performed, you may contact the following person who is not connected with our
study, but who would be prepared to address your concerns or pass them on to the
appropriate authorities:

Oromia Regional Health Bureau
Name: Mr. Damtew Yadeta
Telephone: +251 911 976 746

Malaria Consortium Ethiopia
If at any time you have comments regarding the conduct of this research or questions about
your rights as a research participant, you should contact Malaria Consortium Ethiopia. If you
have general questions regarding this study or a study-related injury (although this is highly
unlikely) which you wish to address to the study representative of Malaria Consortium in
Ethiopia, then you may contact:

Name: Mr. Gezahegn Tesfaye
Telephone: +251 114 674 022
           +251 911 431 860

	
  




                                               42	
  
                                            ENROLMENT QUESTIONNAIRE
                                                      HEALTH CENTER

Health center code:      |____|                                            Health center name:


Clinical officer / nurse code: |____|____|                                 Clinical officer / nurse name:

                                                 DEMOGRAPHIC DETAILS

 Date of               |____|____| / |____|____| / |____|____|             Patient code: |____|____|____|
consultation (GC)          day          month         year


Patient name:                                                              Gender: |____|       1=Male; 2=Female


Age: |____|____| years     99=Unknown                               If < 1 year old, age in months: |____|____| months


Kebele of residence:                                                       How long have you lived there? |____|____| years

                                                 CLINICAL INFORMATION

Do you have a fever?              |____|     0=No; 1=Yes                   Temperature recorded: |____|____| . |____| °C

Have you had a fever
                                                                           If yes, how long ago? |____|____| hours
In the past 48 hours?             |____|    0=No; 1=Yes

Are you currently pregnant?       |____|     0=No; 1=Yes                   Any chronic illness? (specify):
                                             9=N/A




Do you have any of the following symptoms: 0=No; 1=Yes                     Have you taken any malaria medicine in the past
                                                                           month?
                  Headache                                                               0=No; 1=Yes
                                                                           One week ago?              Medicine:______________
                 Muscle pain
                                                                           Two weeks ago?             Medicine:_____________
                 Chills
                                                                           Three weeks ago?           Medicine:_____________
                 Vomiting
                                                                           Four weeks ago?            Medicine:_____________

                                           MALARIA PREVENTION ACTIVITIES

Does your household have any mosquito nets that can be used while sleeping? If yes, how many?                |____|   0=None


Were any of these nets factory pre-treated to kill or repel mosquitoes when you received them?    |____|
                                                                                    0=No; 1=Yes; 8=Don’t know; 9=No net




                                                                 43	
  
Since you got the net, was it ever soaked or dipped in a liquid to kill or repel mosquitoes?         |____|
                                                                                      0=No; 1=Yes; 8=Don’t know; 9=No net


How often do you sleep under a bednet?                              |____|       0=Never; 1=Sometimes; 2=Always


Did you sleep under a bed net last night?                           |____|        0=No; 1=Yes


At any time in the past 12 months has anyone sprayed the interior walls of your dwelling against mosquitoes? |____|
                                                                                                          0=No; 1=Yes


If yes, how many months ago were your walls sprayed?       |____|____|       00=Less than one month ago; 99=Not sprayed


Who sprayed your home?                 |____|
1=Government worker; 2=NGO program; 3=Private company; 4=Household member; 5=Other __________________________;
9=Not sprayed


        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  
        	
  



                                                           44	
  
                                               LABORATORY RESULTS - RDT
                                                HEALTH CENTER DETAILS
Health center code:       |____|                                            Health center name:


Laboratory technician code: |____|____|                                     Laboratory technician name:

                                                        PATIENT DETAILS
Date of           |____|____| / |____|____| / |____|____|                   Patient name:
consultation (GC)     day           month         year

Check: consent given?                 |____|      0=No; 1=Yes               Patient code: |____|____|____|
                 (If consent denied, do not proceed)                            (Ensure this is the same as for enrolment questionnaire)
                                                        PARASCREEN RDT
Time started:   |______| : |______|       - - - - ->     add 15 minutes - - - - - >          Read time:       |______| : |______|


ParaScreen result:          |____|
1=P. falciparum; 2=P. vivax; 3=Mixed infection; 4=Negative; 5=Invalid


If RDT is positive for P. falciparum, P. vivax or mixed infection, how strong is:
   The P. falciparum test line?                   |____|      1=Strong; 2=Faint
   The P. vivax test line?                        |____|      1=Strong; 2=Faint

                                                        ICT COMBO RDT
Time started:   |______| : |______|       - - - - ->     add 15 minutes - - - - - >          Read time:       |______| : |______|


ICT Combo result:            |____|
1=P. falciparum; 2=P. vivax; 3=Mixed infection; 4=Negative; 5=Invalid

If RDT is positive for P. falciparum, P. vivax or mixed infection, how strong is:
   The P. falciparum test line?                   |____|      1=Strong; 2=Faint, 3=Test line not seen
   The P. vivax test line?                        |____|      1=Strong; 2=Faint, 3=Test line not seen

                                                        CARESTART RDT
Time started:   |______| : |______|       - - - - ->     add 20 minutes - - - - - >          Read time:       |______| : |______|


CareStart result:          |____|
1=P. falciparum; 2=P. vivax; 3=Mixed infection; 4=Negative; 5=Invalid
If RDT is positive for P. falciparum, P. vivax or mixed infection, how strong is:
   The P. falciparum test line?                   |____|      1=Strong; 2=Faint, 3=Test line not seen
The P. vivax test line?                        |____|      1=Strong; 2=Faint, 3=Test line not seen




                                                                   45	
  
                                     LABORATORY RESULTS - MICROSCOPY
                                          HEALTH CENTER DETAILS
Health center code:     |____|                                        Health center name:


Laboratory technician code: |____|____|                               Laboratory technician name:

                                                        PATIENT DETAILS
Date of           |____|____| / |____|____| / |____|____|             Patient name:
consultation (GC)     day           month         year

Check: consent given?              |____|            0=No; 1=Yes      Patient code: |____|____|____|
               (If consent denied, do not proceed)                          (Ensure this is the same as for enrolment questionnaire)


        Label the filter paper, slides and RDTs with patient code and date before performing finger
        prick.

                                               BLOOD SMEAR MICROSCOPY

 P. falciparum:

 Slide read (circle one):   A /    B

 Microscopy result: |_____| (0=Negative; 1=Positive)                        Parasite count: |_____________| p / µl

 Gametocytes seen? |_____| (0=No; 1=Yes)


 P. vivax

 Slide read (circle one):   A /    B

 Microscopy result: |_____| (0=Negative; 1=Positive)                        Parasite count: |_____________| p / µl

 Gametocytes seen? |_____| (0=No; 1=Yes)




        NB: To calculate parasite count in p/µl:
        Number of parasites per 200 WBCs x 40 = asexual parasites / µl
        Number of parasites per 500 WBCs x 16 = asexual parasites / µl (this is done when less than
        10 parasites are seen after counting 200 WBCs)


        If microscopy indicates either P. falciparum or P. vivax infection, inform the clinical officer
        in order that treatment can be administered.




                                                                   46	
  
             RDT ‘ease of use’ questionnaire for interview with health post staff
                                HEALTH POST DETAILS
Health post name:                                             Health post code:       |____|____|

Woreda:                                                       Kebele:

HEW name:                                                     HEW code:               |____|____|____|

Date of interview (G.C.):                                     Name of interviewer:

                                                   RDT USED
RDT in use at health post: (circle one)            ICT Combo                   ParaScreen           CareStart


Duration of use:      |____|____| months,             if less than 1 month: |____|____| days

Has the HEW used any other RDT before? (both during this study and as part of routine health post
activities)

RDT 1 name: ________________________________ Length of time used:__________________________

RDT 2 name: ________________________________ Length of time used:__________________________

RDT 3 name: ________________________________ Length of time used:__________________________

RDT 4 name: ________________________________ Length of time used:__________________________

                                                 RDT RATING:
                                    1 = very difficult to use, 5 = very easy to use

Format                                                                     1          2      3       4          5

Lancet                                                                     1          2      3       4          5

Swab                                                                       1          2      3       4      5

Ease of writing details on device                                          1          2      3       4          5

Ease of filling blood collection device                                    1          2      3       4          5

Ease of emptying blood collection device                                   1          2      3       4          5

Buffer drops                                                               1          2       3      4          5

Number of steps to carry out RDT                                           1          2      3       4          5

Time to wait for results (1=too long, 5=good length of time)               1          2       3      4          5

Ease of interpreting test results                                         1           2      3       4      5

Instruction leaflet                                                        1          2      3       4          5



                                                          47	
  
	
  
                                       FURTHER QUESTIONS
In your opinion, how does this RDT compare with other brands of RDT you have used previously?
(Begin by asking for comparison to ParaCheck, then to any other multi-species test used previously in this
study)




Additional comments (probe for specific difficulties with the RDT, or favourable characteristics)




	
  




                                                      48	
  
	
  
                               OBSERVATION OF RDT PROCEDURE
Observe the HEW carrying out an RDT test, to ensure correct procedure is followed.

Check for:
Correct storage of RDTs

Correct labelling of RDTs

Fingerprick procedure

Blood collection using device in RDT test

Emptying of blood collection device

Addition of correct buffer quantity

Waiting correct time before reading results

Interpretation of results

Recording of results on form




Other comments, or specific difficulties observed:




                                                     49	
  

				
DOCUMENT INFO