Comparison of VCT and Population - PowerPoint

Document Sample
Comparison of VCT and Population - PowerPoint Powered By Docstoc
					    Comparison of VCT and
Population-based HIV Prevalence

      The 2nd Global HIV/AIDS Surveillance Meeting
                   Bangkok, Thailand
                                 March 2-5, 2009


                          Fulgentius Baryarama
                              CDC-Uganda

The views presented are of the author and do not represent views of
                           his employer
              VCT versus Population-based HIV Prevalence   Slide 1
                          Outline
•   Background
    •   ANC
    •   Surveys
    •   PMTCT
    •   VCT
•   Lessons from VCT in Uganda
•   Objectives and Methods
•   Results
•   Conclusions

          VCT versus Population-based HIV Prevalence   Slide 2
                   Background
•   Population based surveys provide more
    representative point prevalence estimates

•   However, ANC data remains the main HIV
    surveillance data source for epidemic trends

•   Yet, ANC data are known to overestimate
    national HIV prevalence



          VCT versus Population-based HIV Prevalence   Slide 3
     Survey : ANC Prevalence Ratio
•   E. Gouws, et al. (2008). Sex Transm
    Infect:84(Suppl) i17-i23

•   National based surveys from 26 countries,
    2001-2007

•   ANC based prevalence overestimates both
    urban and rural HIV prevalence by 20%

•   Recommended an adjustment factor of 0.8

          VCT versus Population-based HIV Prevalence   Slide 4
                     PMTCT data
•   Reviews of PMTCT data have shown
    •   Potential to replace ANC surveillance



•   But PMTCT has biases similar to ANC data
    •   Both share the same target population
    •   Both have high risk young women
    •   In addition: Opting out of HIV testing




             VCT versus Population-based HIV Prevalence   Slide 5
                        VCT Data
•   Data from multiple VCT providers

•   Different VCT provision approaches
        •   Stand alone sites
        •   Health facility based
        •   Home-based VCT
        •   Outreaches

•   Self-selection bias

•   Growing database

•   Utility of VCT data for HIV surveillance
            VCT versus Population-based HIV Prevalence   Slide 6
The AIDS Information Centre (AIC)
•   AIC was formed in 1990 to offer VCT in Uganda

•   AIC has since tested over 2 million clients in
    over 200 sites




            VCT versus Population-based HIV Prevalence   Slide 7
            Annual number of clients tested at AIC


            300,000

            250,000

            200,000
C lien ts




            150,000

            100,000

             50,000

                  0
                      1990   1992     1994    1996     1998     2000      2002   2004    2006     2008


                             VCT versus Population-based HIV Prevalence                 Slide 8
           Lessons - 1992-2005
1: Negative HIV clients tend to take repeat tests

2: Hence the prevalence of first-time testers is
   higher than for the entire VCT population

3: VCT clients who test due to illness have
   extremely high HIV prevalence



            VCT versus Population-based HIV Prevalence   Slide 9
                     Lessons –De-duplication
Lesson 4: Various methods for de-duplicating VCT
  clients in the absence of a unique identifier exist
        Example with 2007 data: N=150,344 not reporting illness
                     (ID= Client ID, UTC=Unique Testing Code from client data)


 100%                                                           12%
                        87.8%         85.6%
                                                                10%     9.6%                                      9.4%
 80%     74.7%                                      73.3%                             9.1%          9.0%

                                                                8%
 60%
                                                                6%
 40%
                                                                4%

 20%
                                                                2%


  0%                                                            0%
        First time    Recent Test   Recent Test   Recent Test         First time   Recent Test   Recent Test   Recent Test
         testers         by ID        by UTC      by ID + UTC          testers        by ID        by UTC      by ID + UTC


                             VCT versus Population-based HIV Prevalence                           Slide 10
                      Lessons -1992-2000
                                                         (a)

                                                     A) 15-19 Years
                                                                                                                                                                           (b)
                                                                                                                                                                                B) 20-24 Years
Lesson 5:
   Consistent                                50

                                             40                                                                                 C
                                                                                                                                VT       AC
                                                                                                                                         N
                                                                                                                                                                  50

                                                                                                                                                                  40                                                               C
                                                                                                                                                                                                                                   VT          N
                                                                                                                                                                                                                                               AC


   disparity in HIV                          30                                                                                                                   30




                              ecn I oi e
                             Pre tHVP stiv




                                                                                                                                                    ecn I oii e
                                                                                                                                                   Pre tHVPstv
                                             20

   prevalence by
                                                                                                                                                                  20

                                             10                                                                                                                   10


   age between                               0
                                                  1992         1993   1994                1995         1996      1997    1998     1999     2000
                                                                                                                                                                   0
                                                                                                                                                                         1992      1993      1994       1995     1996     1997   1998   1999        2000
                                                                                                   Y a oHVT s
                                                                                                    er f I et                                                                                                 er f I et
                                                                                                                                                                                                             YaoHVT s

   VCT and ANC                                                                                                                  (c)

   data                                                                                       45
                                                                                                                        C) 25+ Years
                                                                                                                                                                                 C
                                                                                                                                                                                 VT              N
                                                                                                                                                                                                 AC
                                                                                              40
                                                                                              35
                                                                                              30
                                                                                              25
                                                                              e e t I oi e
                                                                             P rc nHVP stiv




                                                                                              20
                                                                                              15
                                                                                              10
                                                                                              5
  KEY                                                                                         0
                                                                                                     1992       1993    1994      1995    1996                    1997      1998          1999        2000
  Solid Line
                VCT                                                                                                                    er f I et
                                                                                                                                      Y a oHVT s



  Dashed Line
                ANC

                      VCT versus Population-based HIV Prevalence                                                                                                                                       Slide 11
          Objectives of the Study
•   To compare HIV prevalence from routine VCT
    data with a population-based sero-survey

•   To assess the utility of VCT data as a
    supplemental data source for HIV surveillance




             VCT versus Population-based HIV Prevalence   Slide 12
Methods – Population Based Survey
•   Conducted from August 2004 to Jan 2005

•   Based on a national representative sample of
    adults age 15 – 59 years

•   Collected data from 9,529 households in 417
    enumeration areas using a 2-stage cluster
    sample



             VCT versus Population-based HIV Prevalence   Slide 13
             Methods -Survey
• 11,454 women and 9,905 men were eligible

• Questionnaire completed for 95% of women
  and 89% of men

• Blood specimen collected from 90% of women
  and 84% of men




          VCT versus Population-based HIV Prevalence   Slide 14
                Methods - VCT Data
                    (Same period as Survey)

•   8 Stand alone sites                      •     152 Health facilities

•   AIC Staff                                •     MOH Staff

•   Cost sharing                             •     Free VCT Services

•   19, 972 clients                          •     55,668 clients




                VCT versus Population-based HIV Prevalence          Slide 15
                  Methods –GIS
• Location of residence of VCT clients not
  available

• 15km and 30km buffers drawn around each
  VCT site to estimate catchment areas

• 83% of survey clusters within 30 km and 56%
  within 15km of a VCT site


            VCT versus Population-based HIV Prevalence   Slide 16
      Methods –Data Management
• Analysis variables in separate VCT and survey
  data were made compatible

• Exclusions:
  •   VCT clients who reported illness as reason for testing
  •   VCT records with a subsequent visit to avoid
      duplicates

• Data were merged to one database clearly
  labeling records as either VCT or survey

              VCT versus Population-based HIV Prevalence   Slide 17
        Methods –Data Analysis
• Age-standardized prevalence computed with the
  survey as the standard population


• HIV prevalence presented by 5-year age groups
  stratified by rural or urban residence and sex


• Comparisons of VCT and survey prevalence were
  based on 95% CIs


            VCT versus Population-based HIV Prevalence   Slide 18
         Methods – Sub-analyses

• Survey participants within 15km of a VCT site

• VCT clients at stand-alone sites

• VCT clients at Health Facility (HF) sites




             VCT versus Population-based HIV Prevalence   Slide 19
            RESULTS




VCT versus Population-based HIV Prevalence   Slide 20
Map of Uganda showing location of Survey
Clusters, VCT Sites and 15/30 km buffers




       VCT versus Population-based HIV Prevalence   Slide 21
      Study Population - Sex
                              Percent women
60%
        55%                                                                              54%
                              53%               52%
50%                                                                  46%


40%

30%

20%

10%

0%
      All Survey         Survey - 15km         All VCT          VCT-Stand alone        VCT-HF




                   VCT versus Population-based HIV Prevalence                     Slide 22
      Study Population - Residence
             Percent residing in urban area
60%

50%

40%                                                                    38%
                                                  33%

30%                             28%                                                        28%


20%       16%


10%

0%
        All Survey         Survey - 15km         All VCT          VCT-Stand alone      VCT-Facility




                     VCT versus Population-based HIV Prevalence                     Slide 23
Study Population – Previous HIV Testing
            Percent previously tested for HIV
 60%

 50%
                                                                      42%
 40%

 30%
                                                 23%

 20%                           16%                                                        16%
         13%

 10%

 0%
       All Survey         Survey - 15km         All VCT          VCT-Stand alone        VCT-HF




                    VCT versus Population-based HIV Prevalence                     Slide 24
       Study Population – Age Groups
                                15-19      20-24         25-29      30+
100%


80%                                             39%                 37%                 39%
           47%                44%


60%
                                                19%                 24%                 18%
           15%                16%
40%
           16%                17%                                                       25%
                                                26%
                                                                    28%
20%
           22%                23%
                                                16%                                     18%
                                                                    11%
 0%
         All Survey      Survey - 15km         All VCT         VCT-Stand alone      VCT-Facility




                  VCT versus Population-based HIV Prevalence                     Slide 25
  Study Population – Marital Status
          N/Married        Monogamous               Polygamous            Div/Sep/Wid
100%                                                                                  11%
         12%                12%               11%                 13%

                                              12%                  7%                 14%
80%      18%                16%

                                                                  28%
60%                                           38%                                     42%
         43%                42%
40%

                                                                  53%
20%                                           39%                                     36%
         27%                29%

 0%
       All Survey      Survey - 15km         All VCT         VCT-Stand alone      VCT-Facility




                VCT versus Population-based HIV Prevalence                     Slide 26
       Study Population – Education
             None       Incomplete Primary              Primary          Secondary +
100%
          23%
                             29%
80%                                            48%                                     40%
          12%
                             12%                                   70%
60%
                                                                                       15%
                                               15%
40%       45%
                             42%
                                                                                       31%
                                               27%                 13%
20%
          20%                17%                                   14%
                                               11%                                     13%
 0%                                                                 4%
        All Survey      Survey - 15km         All VCT         VCT-Stand alone      VCT-Facility




                 VCT versus Population-based HIV Prevalence                     Slide 27
                   HIV Prevalence
12%
                                  10.1            9.7
10%                                                       8.8
      8.2
8%
                   5.2                                                      5.9
6%

4%

2%

0%
      VCT         Survey           VCT           Survey   VCT           Survey


            Rural                        Urban                  Total
             VCT versus Population-based HIV Prevalence          Slide 28
      HIV Prevalence – Urban Men
          Survey        Survey - 15km               VCT-stand alone           VCT-HF
25%


20%


15%


10%


5%


0%
      15-19   20-24    25-29      30-34     35-39     40-44   45-49   50-54      55-59




               VCT versus Population-based HIV Prevalence             Slide 29
  HIV Prevalence – Urban Women
          Survey        Survey - 15km               VCT-stand alone           VCT-HF
30%

25%

20%

15%

10%

5%

0%
      15-19   20-24    25-29      30-34     35-39     40-44   45-49   50-54      55-59




               VCT versus Population-based HIV Prevalence             Slide 30
      HIV Prevalence – Rural Men

          Survey         Survey - 15km               VCT-stand alone              VCT-HF
25%


20%


15%


10%


5%


0%
      15-19   20-24     25-29      30-34     35-39     40-44   45-49     50-54      55-59




              VCT versus Population-based HIV Prevalence               Slide 31
      HIV Prevalence – Rural Women
           Survey        Survey - 15km               VCT-stand alone           VCT-HF
25%


20%


15%


10%


5%


0%
       15-19   20-24    25-29      30-34     35-39     40-44   45-49   50-54      55-59




                VCT versus Population-based HIV Prevalence             Slide 32
                                  Summary
Residence Data Source       Population            Prevalence 95% CI            Survey:VCT

Urban     Survey            All                         9.7     8.6-10.7              Ref
                            Within 15km                 9.7     8.6-10.8              NA
          VCT               All                        10.1     9.8-10.5              1.04
                            Health Facility            10.0     9.5-10.4              1.03
                            Stand alone                10.3     9.8-10.8              1.06
Rural     Survey            All                         5.2      4.8-5.5              Ref
                            Within 15km                 4.3      3.8-4.7              NA
          VCT               All                         8.2      7.9-8.4              0.63
                            Facility                    7.7      7.5-8.0              0.68
                            Stand alone                11.1     10.3-11.9             0.47
Total     Survey            All clients                 5.9      5.6-6.2               -
          VCT               All Participants            8.8      8.7-9.1              0.67


                   VCT versus Population-based HIV Prevalence              Slide 33
                  Conclusions-1
•   Overall, HIV prevalence higher among VCT
    clients than survey population, 8.8% vs. 5.9%


•   Difference attributed to rural areas where VCT
    prevalence was higher than survey prevalence,
    8.2% versus 5.2%


•   However, for urban areas, survey and VCT HIV
    prevalence were similar, at about 10%


            VCT versus Population-based HIV Prevalence   Slide 34
                 Conclusions-2
•   Survey prevalence within 15km of a VCT site
    was significantly lower than general survey
    prevalence

•   HIV prevalence at Health Facility VCT sites
    showed more stable patterns by age

•   Differences in clientele between VCT and
    survey populations more marked in stand alone
    than health facility sites


           VCT versus Population-based HIV Prevalence   Slide 35
                   Challenges
• Limited access of VCT services to rural
  areas

• Lack of national VCT data sets with
  individual client data from all key VCT
  providers – only aggregated data are
  reported

• Lack of multi-country comparisons of VCT
  and survey data including trends from
  multiple surveys

         VCT versus Population-based HIV Prevalence   Slide 36
                      Questions
•   Does stand alone VCT data have potential for
    supplemental HIV surveillance?
              …Yes, for early warning purposes only


•   Does health facility VCT data have potential for
    HIV prevalence trends?                    …Yes


•   Does health facility VCT data have potential for
    point HIV prevalence estimates
                    …Not until challenges are addressed


          VCT versus Population-based HIV Prevalence   Slide 37
Acknowledgements
•     Rebecca E. Bunnella
•     Jonathan H. Mermina
•     Wolfgang Hladika
•     Livia Montanab
•     Alex Opioc
•     Joshua Musinguzic
•     Wilford Kirungic
•     Laban Waswa-Brightd
•     VCT clients and survey participants
            aCenters   for Disease Control and Prevention – Uganda
            bMacro   International
            cUganda    Ministry of Heath
            dAIDS   Information Centre

VCT versus Population-based HIV Prevalence                  Slide 38
Thank You

 VCT versus Population-based HIV Prevalence   Slide 39
           Survey : ANC Prevalence Ratio*

Country    Year      Survey                 Urban                             Rural
                      Type          Survey         Ratio            Survey             Ratio
                                  prevalence    Survey : ANC      prevalence        Survey : ANC
Lesotho    2004       DHS            29.1          0.901             21.9              0.901
Malawi     2004       DHS            18.3          0.995             11.3              0.883
Zambia     2002       DHS            23.1          0.878             10.8              0.850
Zimbabwe   2005       DHS            18.9          0.863             17.6              0.936
Burundi    2002       HHS             9.4          0.797              2.5              1.000
Ethiopia   2005       DHS             5.6          0.533              0.7              0.368
Kenya      2003       DHS            10.0          0.909              5.6              0.651
Rwanda     2004       DHS             7.3          1.237              2.2              0.815
Tanzania   2004        AIS           10.9          0.973              5.3              0.964
Uganda     2004        SBS           10.1          1.098              5.7              1.326
                                                               *Extracted from: E. Gouws, et al., 2008
                  VCT versus Population-based HIV Prevalence                   Slide 40

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:5
posted:4/23/2010
language:English
pages:40