Docstoc

THE REPUBLIC OF MALAWI

Document Sample
THE REPUBLIC OF MALAWI Powered By Docstoc
					               THE REPUBLIC OF MALAWI

  MINISTRY OF HEALTH AND POPULATION




                
HIV SENTINEL SURVEILLANCE
          REPORT
            2003




            National AIDS Commission

                NOVEMBER 2003
ACKNOWLEDGEMENTS

The National AIDS Commission (NAC) Secretariat extends its appreciation to
all members of staff in the sentinel sites for their support in planning field
visits and actual specimen collection, the Community Health Sciences Unit
(CHSU) Mr J. Kandulu, Mr L. Zungu, Mr A. Phiri and Mr Kanamazina; Lilongwe
Central Hospital (LCH) Mr H. Feluzi and Mr G. Gamadzi; Queen Elizabeth
Central Hospital (QECH) and Mzuzu Central Hospital, for assisting with
specimen collection and analysis. The Unit is also greatly indebted to all
women sampled, too numerous for individual mention for providing the data
with which to compile this report. Gratitude is extended to Centres for
Diseases Control and Prevention Global AIDS Program (CDC-GAP) for financial
and technical support, Reproductive Health Unit (MOHP) for providing
vironostika ELISA kits, and World Health Organisation (WHO) for successfully
managing the sentinel surveillance transition from NAC to the Ministry of
Health and Population.

Special thanks go to Mr Henry JC Damisoni (WHO), John Aberle-Grasse
(CDC), John B. Chipeta (NAC), George Bello (MOHP), Mike Zulu (NAC), Simon
Makombe (MOHP) and Doreen Ligomeka (NAC) for taking the lead in
conducting the survey and compiling this report.

The following are acknowledged for their invaluable contributions towards this
report; Dr Biziwick Mwale, Mr Roy Hauya, Mr Cosby Nkwazi, Dr Andrina
Mwansambo, Miss Marelize Gorgens and Miss Len van der Hoeven. NAC
further wishes to acknowledge Mrs Ellen Kumwenda and Mr. Felix Phambala
for entering the HIV sentinel surveillance data.




                                      2
EXECUTIVE SUMMARY

The Primary objective of the Malawi 2003 sentinel surveillance was to provide
data on the occurrence, spread, and distribution of HIV infection among
women attending antenatal clinics. The country monitors HIV prevalence
through antenatal clinic (ANC) sentinel surveillance. HIV sentinel surveillance
data does not represent the general population and prevalence in this report
does not represent national prevalence. However sentinel surveillance data
can be used to estimate HIV prevalence in the general population.

The 2003 ANC sentinel surveillance was carried out in nineteen sites in all the
three regions of the country. A total of 7,977 pregnant women were captured
in the survey from 18 February 2003 through mid April 2003. Over 80% of
the women sampled were less than 30 years of age and about 60% aged less
than 25 years. The age pattern of the women recruited in the survey is very
similar to the previous surveys done in 2001, 1999 and 1998. Over 25% of
the sample presented with the first pregnancy while over 80% presented with
no more than their fourth pregnancy. About 75% of the women had some
level of education. However, the sample predominantly comprised women
who had attended primary school (60%). Marriage was almost universal
(97.6%) in the study population, and 87% of the women were housewives.


Site-specific HIV prevalence ranged from 6.7% at Kamboni Health Centre, a
rural site in the central region to 32.9% at Nsanje District Hospital a semi-
urban site in the southern tip of the country. The overall HIV prevalence for
antenatal attendees in 2003 (19.8%) has remained stable from 2001
(19.5%), but lower than in 1999 (24.1%). HIV prevalence in the urban
(21.7%) and semi-urban (20.8%) areas continued to be high and significantly
different from rural areas (14.5%). Over time the gap in HIV prevalence
between urban and semi-urban has been narrowing. However, while still
much lower than the urban and semi-urban figures, prevalence in the rural
sites increased significantly from 12.1% in 1999 to 14.5% in 2003.


HIV prevalence for those without any formal education (19.2%) and those
with a primary education (19.1%) was similar, as was the case in 1999. HIV
prevalence was significantly higher among women who have gone to
secondary school (23.2%) as compared to women with no education (19.2%)
and those reporting going to primary school (19.1%).

HIV prevalence in the southern region (23.7%) remains higher than in the
northern (20.0%) and central (15.5%) regions. In Nsanje, HIV prevalence
remains elevated (32.9%). There are upwards trends in many rural sites, as
seen in Mbalachanda, Thonje, and Mianga Health Centres. Lilongwe, on the
other hand, has shown a consistent decline since 1999.




                                       3
HIV prevalence remains unacceptably high in all the age groups. The stable
high prevalence could be attributed to high incidence in the younger age
groups, especially adolescents (15-19 years).

Overall syphilis prevalence was 2.7%. Prevalence of syphilis ranged from 0%
in Mbalachanda and Rumphi to 6.9% in Mianga health centre (Thyolo
district). There was an inverse relationship between level of education and
syphilis infection. Syphilis prevalence among women by education levels were
as follows; secondary (1.7%), primary (2.5%) and no education (3.6%). The
observed syphilis prevalence rates across regions are statistically different.
Syphilis prevalence in the southern region (4.3%) remains higher than the
central (2.3%) and northern (1.0%) regions. Overall, there appears to be a
decline in syphilis prevalence over the years.

There was a general association between HIV and syphilis infection in the
study sample. This result shows that syphilis and HIV were not occurring
independently.




                                      4
ACRONMYS

AIDS       Acquired Immunodeficiency Syndrome
ANC        Antenatal clinic
CDC        Centres for Disease Control
ARV        Antiretroviral
CI         Confidence interval
CHSU       Community Health Sciences Unit
ELISA      Enzyme-Linked Immunosorbent Assay
GAP        Global AIDS Programme
HIV        Human Immunodeficiency Virus
KAP        Knowledge, Attitude and Practices
LCH        Lilongwe Central Hospital
MDHS       Malawi Demographic and Health Survey
MOHP       Ministry of Health and Population
NGO        Non governmental Organization
NAC        National AIDS Commission
QECH       Queen Elizabeth Central Hospital
RPR        Rapid Plasma Reagin test
STI        Sexually Transmitted Infections
TPHA       Treponema Pallidum Heamagglutination Assay
UNAIDS     Joint United Nations Programme on HIV/AIDS
VDRL       Venereal Disease Research Laboratory
WHO        World Health Organization




                          5
TABLE OF CONTENTS


Acknowledgement                                                  2

Executive summary                                               3

Acronyms                                                         5

Table of contents                                                6

List of tables                                                   7

List of figures                                                  8




1. Introduction                                                  9

2. Background                                                   10

3. Objectives                                                   10

4. Methodology                                                  12

5. Results                                                      15

6. Discussion and Trends                                        26

7. Conclusion                                                   38

8. Recommendations                                              39

9. References                                                   40




Annex   1 Brief description of second generation surveillance   41
Annex   2 Glossary                                              42
Annex   3 Spouse occupation categorization                      43
Annex   4 Data collection form                                  44




                                          6
LIST OF TABLES


Table                                                                 Page
Table 1: Number of women sampled by sentinel site                       15
Table 2: Number of women sampled by Locality                            16
Table 3: Number of women sampled by Region                              16
Table 4: Age distribution of women enrolled in the survey               17
Table 5:   Age patterns of women involved in sentinel sites in this     17
           and    previous three surveys
Table 6: Gravidity                                                      18
Table 7: Level of education                                             18
Table 8: Prevalence of HIV among women by site                          19
Table 9: HIV Prevalence by locality                                     19
Table 10: HIV Prevalence by Region                                      20
Table 11: Age specific prevalence for women in the sample               20
Table 12: Age specific prevalence for aggregated urban sites            21
Table 13: Age specific prevalence for aggregated semi-urban sites       21
Table 14: Age specific prevalence for aggregated rural sites            22
Table 15: HIV Prevalence by level of education                          22
Table 16: HIV prevalence by partners‟ occupation                        23
Table 17: Site specific syphilis prevalence                             24
Table 18: Syphilis prevalence by education level                        24
Table 19: Syphilis prevalence by region                                 25
Table 20: Syphilis prevalence by locality                               25
Table 21: Syphilis prevalence by age                                    25
Table 22: Trends in Syphilis prevalence                                 37




                                        7
LIST OF FIGURES


Figure                                                       Page
Figure 1: HIV sentinel site locations                          11
Figure 2: Trends of HIV infection by age group and for all     27
          ANC attendees
Figure 3: Trends of HIV infection for all anc attendees        28
Figure 4: Trends of HIV infection by locality                  29
Figure 5: Trends of HIV infection by region                    30
Figure 6: HIV infection among ANC attendees, Blantyre          31
Figure 7: HIV infection among ANC attendees, Lilongwe          31
Figure 8: HIV infection among ANC attendees, Mzuzu             32
Figure 9: HIV infection among ANC attendees, Mbalachanda       32
Figure 10: HIV infection among ANC attendees, Thonje           33
Figure 11: HIV infection among ANC attendees, Mianga           33
Figure 12: HIV infection among ANC attendees, Nsanje           34
Figure 13: HIV infection among ANC attendees, Ntcheu           34
Figure 14: HIV infection among ANC attendees, Nkhata-bay       35
Figure 15: Age specific ANC HIV infection - Blantyre           35
Figure 16: Age specific ANC HIV infection - Lilongwe           36
Figure 17: Age specific ANC HIV infection - Mzuzu              36
Figure 18: Trends in Syphilis prevalence, 1995 to 2003         37




                                        8
1. INTRODUCTION

Globally the HIV/AIDS pandemic continues to spread. UNAIDS and WHO
estimate that over 40 million people were living with HIV/AIDS and more than
20 million had lost their lives because of AIDS by 20011. More than 28.5
million people were infected in Sub-Saharan Africa by the year 20011. Malawi
has one of the highest national prevalence rates in the world, estimated in
2001 at 15% among adults aged 15-49 years of age2. To come up with cost-
effective interventions the country needs to collect information regularly on
trends and patterns of the epidemic. The country monitors HIV prevalence
through antenatal clinic (ANC) sentinel surveillance. Monitoring trends in HIV
infection over time and place requires that sero-surveys be conducted at
repeated intervals using consistent methodology in the same population
group so as to determine any change in HIV prevalence.

HIV sentinel surveillance targets women attending antenatal clinics during
their first visits. Antenatal women are targeted because residual blood meant
for syphilis testing can be used for anonymous and unlinked HIV serology
tests. It is government policy to screen all antenatal mothers for syphilis and
to treat those infected.

The Primary objective of the Malawi 2003 sentinel surveillance was to provide
data on the occurrence, spread, and distribution of HIV infection among ANC
attendees. It should be stressed that ANC sentinel surveillance data does not
represent the general population and prevalence in this report does not
represent national prevalence. These point estimates only reflect the specific
sentinel site samples. However, results from sentinel surveillance can be used
to estimate and project HIV/AIDS prevalence in the general population, when
looked at with data from additional sources.

Collection of blood samples and demographic data for syphilis and HIV tests
began on 18th February 2003, and ended around mid April 2003. HIV and
syphilis prevalence in this report therefore reflect the HIV situation from mid
February to mid April 2003. Syphilis is concurrently monitored with HIV
prevalence because the two diseases are predominantly transmitted through
sexual intercourse. Furthermore, syphilis prevalence reflects unsafe sex
practices in the underlying population. Information on occurrence and
distribution of syphilis infections can give an opportunity for implementing
evidence-based behaviour change interventions.

The 2003 sentinel surveillance activity was initiated under a transition period
where the responsibility of sentinel surveillance was moving from the NAC
(National AIDS Commission) to MOHP (Ministry of Health and Population). In
order to facilitate the transition, the NAC management formed a Technical
Working Committee (TWC). Members of the TWC represented NAC, MOHP,
WHO, and CDC (Centres for Disease Control and Prevention). Main roles of
the TWC were to provide technical assistance in planning and implementation
of the 2003 sentinel surveillance activities and serve as a resource team to


                                       9
build capacity of the MOHP Epidemiology Unit for the responsibility of sentinel
surveillance.


2. BACKGROUND

The history of HIV sero-prevalence surveys in Malawi dates back to 1985
when the first case of AIDS was confirmed and reported. Between 1985 and
1991 a number of cross sectional studies were conducted in different sub-
populations to estimate the magnitude of the problem (HIV prevalence) and
identify risk factors for infection. Due to logistical constraints, these studies
were concentrated in urban areas. It was not until 1990 that data was
routinely collected from women attending antenatal clinics (ANC) in selected
sites across the country. These sentinel sites were originally fewer in number,
but the number steadily increased to nineteen in 1994 as the National AIDS
Secretariat became particularly interested in monitoring prevalence trends.
Since then, sentinel surveillance surveys have been conducted annually in all
the 19 sentinel sites, except for years 2000 and 2002, when surveys were not
conducted. It is only at Queen Elizabeth Central Hospital (University teaching
and national referral facility) where HIV sero-prevalence surveys have been
conducted annually since 1985.

Data for 2003 were collected from all 19 sentinel sites. Figure 1 shows the
distribution of sentinel sites in the country. Of these, five are in the northern
region and seven each in the central and southern regions. Further
stratification of these sites is such that in the northern region two are rural,
two semi-urban and one urban. For the central and southern regions there
are three rural, three semi-urban and one urban site in each region
respectively.


3. OBJECTIVES

1. To assess HIV and syphilis sero-prevalence in women attending antenatal
   clinics
2. To monitor HIV and syphilis prevalence trends over time in women
   attending antenatal clinics;
3. To monitor HIV and syphilis prevalence trends by geographical location in
   women attending antenatal clinics; and
4. To generate data for use in estimation and projection of HIV sero-
   prevalence trends and impacts of AIDS in the general population.

Figure 1: HIV sentinel site locations




                                       10
                                      Kaporo


        Rumphi DH

       Mbalachanda HC
                                        NkhataBay DH
       St. John’s Hosp.



                                       St. Anne’s Hosp.
        Kamboni HC
                                        Thonje HC

       Mchinji DH                         Kasina HC

       Lilongwe CH                              Mangochi DH

                     Ntcheu DH                    Gawanani HC

                      QECH
                                                  Milepa HC

                      Mianga HC
                                                Mulanje M. Hosp.

                          Nsanje DH




Key:

       urban
       semi-urban site
       rural site




                                         11
4. METHODOLOGY

4.1. Sentinel Site Selection

Since 1994, the country has maintained a total of nineteen sentinel facilities
selected for HIV and syphilis sero-surveillance targeting women attending
antenatal clinics. The sentinel sites are classified as urban, semi-urban and
rural. Urban sites constitute QECH, LCH and Mzuzu Central Hospitals. Semi-
urban sites comprise district and mission hospitals. Health centres are
classified as rural sentinel sites. All the three urban sites were purposefully
selected to get a picture of HIV prevalence in the three Malawian cities of
Blantyre, Lilongwe and Mzuzu respectively. On the other hand, the semi-
urban and rural sites were selected through simple random sampling after
stratifying by region and locality (semi-urban and rural).

Sample size problems in some sentinel sites led to selection of alternative
facilities in the same catchment area. In this regard, Kaporo Health Centre,
Mulanje Mission Hospital and St. Anne‟s Mission Hospital, respectively
replaced Kasoba health post, Mulanje and Nkhotakota district hospitals.
Following the introduction of service fees at QECH and Mzuzu Central hospital
free ANC services have been moved to Limbe and Mzuzu health centres. The
2003 round of HIV surveillance in Blantyre and Mzuzu collected specimens
and data from these substitute sites. However, since the catchment
populations are the same for these sites and their replacements, it is believed
that switching of sample points will not significantly bias the characteristics of
the women sampled.


4.2. Sample Collection

Beginning February 18, 2003 every consecutive woman attending antenatal
clinic services for the first time during the current pregnancy in the sentinel
sites were enrolled in the survey. Enrolment continued for 8 to 12 weeks until
the minimum predetermined sample sizes were achieved. The desired
minimum sample sizes were: 200 women in rural sites, 500 women in semi-
urban sites and 800 women in urban sites. Demographic data like age,
gravidity, level of education, marital status, occupation of the mother and her
partner were collected before blood specimens were drawn.

Blood samples were collected using 8 ml tubes without additives. Thereafter,
sera were separated from whole blood and split into two aliquots. One
aliquot was labelled with a unique identification number and the other with
the woman‟s name for onsite syphilis screening.

In all rural sites sample collection was conducted by three mobile teams of
laboratory technicians drawn from the Community Health Sciences Unit
(CHSU) and Lilongwe Central Hospital for the central region, Queen Elizabeth
Central Hospital for the southern region and Mzuzu Central Hospital for the


                                        12
northern region. In semi-urban and urban sites laboratory procedures were
carried out by technicians at the respective sentinel sites. All serum samples
labelled with unique identification numbers were transported in cold chain to
the reference laboratory in Lilongwe for anonymous and unlinked HIV testing
and syphilis confirmatory testing.

Irregularities in specimen collection were noted at St. Anne‟s Mission hospital.
It was observed that HIV and Syphilis seroprevalence was unexpectedly high.
In addition, the pattern (sequence) in which all reactive specimens appeared
was mostly in clusters (groups) of 3 to 13 per cluster. The ELISA reader
absorbencies in these clusters appeared to be almost the same. This raised
suspicion about the way the specimens were collected. As a result the first
500 blood specimens were discarded and new ones collected under intensive
supervision. Due to time constraints only 272 specimens were collected
instead of the targeted sample size of 500. Because of this, results from St.
Anne‟s Mission hospital may not necessarily be comparable with the rest of
the sero-survey and with previous years results.


4.3. Sample Analysis

All serum samples labelled with the woman‟s name were screened for syphilis
on site using the VDRL test. All reactive samples were considered to be
infected with syphilis and the women were offered treatment onsite.

The data collection instruments with unique identification numbers and serum
aliquots with only the corresponding identification numbers were transported
to the Reference Laboratory in Lilongwe for HIV and syphilis confirmatory
testing. In so doing all information linking the samples to the clients was
removed.

In the Reference Laboratory sera were challenged with one Vironostika ELISA
assay in accordance with the recommendations of the World Health
Organisation (WHO) on HIV testing for surveillance purposes. All samples
reactive to this single test were regarded as HIV positive. All specimens were
again tested with RPR card test and the reactive ones were further challenged
with Treponema Pallidum Heamagglutination Assay (TPHA) to confirm syphilis
infection. The reagents used for both syphilis and HIV testing were supplied
by the National AIDS Commission Secretariat and were from the same lot
number.

Potency of reagents, technical irregularities by different technicians and
performance of ELISA reading machines were controlled for by quality control
samples provided with the test kits to standardize the results.

4.4. Case Management




                                       13
All women reactive to the Venereal Disease Research Laboratory (VDRL) test
on site were treated with a single 2.4 MU intramuscular injection of
Benzathine Penicillin G. Women presenting with vaginal sores in addition to
the VDRL reactive test result were also prescribed Erythromycin orally for five
days. In cases of allergy to Penicillin, Erythromycin to be taken orally six-
hourly for fifteen days was provided. VDRL reactive women were also
encouraged to bring their partners and any other contacts for treatment.

4.5. Data Management

All data were entered into an Epi Info computer database at the National
AIDS Commission Secretariat. Verification of data entry was done through
systematic double entry of data from every tenth data collection instrument.
Discrepant entries were examined and data entry errors corrected. Data
analysis was also carried out using the Epi Info 2002 software package.

Confidence intervals were calculated using exact binomial method. Chi-square
statistics were used to assess associations in 2 by 2 tables. Fisher‟s exact test
was used for 2 by 2 tables with small „expected frequencies‟. Chi-square for
trend was used to assess linear trends for HIV and syphilis sero-prevalence.

4.6. Limitations of Sentinel Surveillance Surveys

ANC sentinel surveillance data does not represent the general population but
only reflect the specific sentinel site sample. These are convenience samples
with sequential enrolment. This being the case standard errors cannot be
easily interpreted because subjects have no known non-zero probability of
inclusion. Prevalence estimates are made from dynamic sub-populations
without clearly defined denominators. This is particularly true of sites with
catchment areas containing tobacco and tea estates that have changing
populations due to variation in agricultural activity and migrant worker
movements.       The effect this may have on the sample population is
compounded by the fact that convenience and not random specimens were
taken. Another shortcoming is the over-sampling of antenatal attendees from
semi-urban and urban sentinel sites as compared to rural areas, given
Malawi‟s predominantly rural population.




                                       14
5. RESULTS


5.1.    Number of Women enrolled

A total of 7,977 women attending antenatal clinics were included in analysis.
Of all the sites, Kasina, St. Anne‟s, and Mbalachanda fell short of the desired
sample size. Refer to table 1 showing the distribution of women sampled by
site.


Table 1:         Number of women sampled by sentinel site

 Sentinel Site                  Region   Location     No.       % of
                                                      of        Total
                                                      Women
 Kaporo Health Centre           North    Rural              210       2.6
 Mbalachanda Health Centre      North    Rural              193       2.4
 Rumphi District Hospital       North    Semi-urban         517       6.5
 Nkhata-Bay District Hospital   North    Semi-urban         522       6.5
 Mzuzu Health Centre            North    Urban              846     10.6
 Kamboni Health Centre          Centre   Rural              238       3.0
 Thonje Health Centre           Centre   Rural              219       2.7
 Kasina Health Centre           Centre   Rural              153       1.9
 Mchinji District Hospital      Centre   Semi-Urban         548       6.9
 St. Anne’s Mission Hospital    Centre   Semi-Urban         272       3.4
 Ntcheu District Hospital       Centre   Semi-Urban         500       6.3
 Lilongwe Central Hospital      Centre   Urban              810     10.2
 Gawanani Health Centre         South    Rural              206       2.6
 Milepa Health centre           South    Rural              205       2.6
 Mianga Health centre           South    Rural              203       2.5
 Mangochi District Hospital     South    Semi-Urban         510       6.4
 Mulanje Mission Hospital       South    Semi-Urban         510       6.4
 Nsanje District Hospital       South    Semi-Urban         511       6.4
 Limbe Health Centre            South    Urban              804     10.1
 Total                                                    7,977   100.0


A deliberate effort was made to increase the sample sizes (from 150 to 200
ANC attendees) in rural sites in 2003 to better represent the rural population.
However, the women sampled were predominantly from semi-urban sites
(49%). The distribution of women enrolled by locality (rural – urban) is shown
in table 2.




Table 2:Number of women sampled by Locality




                                          15
 Location            No. of           % of Total
                     Women
 Rural                        1,627                    20.4
 Semi-urban                   3,890                   48.8
 Urban                        2,460                   30.8
 Total                        7,977                  100.0


The smaller sample size in the northern region is attributable to fewer sites
and is reflective of the smaller population size in the northern region
compared to the central and southern regions. The difference between the
central and southern regions is due to the small sample size from St. Anne‟s
Mission Hospital in the Central Region, which was 272, far below the desired
500. The distribution of women enrolled by region is shown in tables 3.



Table 3 :Number of women sampled by Region

 Region              No. of           % of Total
                     Women
 North                        2,288                   28.7
 Centre                       2,740                   34.3
 South                        2,949                   37.0
 Total                        7,977                  100.0




5.2.      Characteristics of the Women enrolled in the Survey:

The sample comprised relatively young women in their prime years of sexual
activity. Over 80% of the women were aged less than 30 years and about
60% aged less than 25 years as shown in table 4.




Table 4: Age distribution of women enrolled in the survey

 Age Group (Years)   No. Sampled       % of total
                                       sample

 <15                             15                  0.2
 15-19                         1614                 20.2



                                              16
20-24                         3,203                   40.2
25-29                         1,722                   21.6
30-34                           892                   11.2
35-39                           379                    4.8
40+                             105                    0.6
Unknown                          47                    0.6
Total                         7,977                  100.0




The age pattern of the women recruited in the survey is very similar to the
previous three surveys as shown in table 5. At least 60% of the sample were
women aged between 15 to 24 years.



Table 5: Age patterns of women at sentinel sites in the four most
recent surveys


Age Group (Years)                    % of total sample in each year3,4,5

                                  2003            2001           1999      1998
<15                                 0.2             0.2           0.2        0.2
15-19                              20.2            21.2          22.2       22.2
20-24                              40.2            37.5          39.6       38.0
25-29                              21.6            22.3          20.5       20.6
30-34                              11.2            11.4          10.0       11.6
35-39                               4.8             5.1           4.9        6.0
40+                                 0.6             1.4           1.4        1.4
Unknown                             0.6               1           1.2        0.1




Over 25% of the total women sampled presented with their first pregnancy,
while over 80% presented with at most the fourth pregnancy as shown in
table 6. Fourty-two percent (42%) of women aged 15 to 24 reported a
gravidity of 1, while 80% of women 15 to 19 reported a gravidity of 1.

Table 6: Gravidity

Gravidity           No. Sampled           % of total

1                                 2,095                   26.3
2                                 1,868                   23.4



                                                17
 3                                  1,490                    18.7
 4                                  1,042                    13.1
 5                                    657                     8.2
 6                                    359                     4.5
 7                                    243                     3.0
 8                                     94                     1.2
 9+                                    75                     0.9
 Unknown                               54                     0.7
 Total                              7,977                   100.0


Seventy six percent of the women had some level of formal education as
shown in table 7. However, the sample predominantly comprised women who
had attended only primary school (60%).


Table 7:           Level of education

 Education level                Number Sampled              % of total
 None                                            1,921                   18.0
 Primary                                         4,785                   60.0
 Secondary                                       1,201                   15.1
 Post-Secondary                                     68                    0.9
 Unknown                                             2                    0.0
 Total                                           7,977                   100




Marriage was almost universal (97.6%) in the study population and 87% of
the women were housewives.


5.3.    HIV Sero-Prevalence

Site-specific HIV prevalence for the period beginning mid February to mid
April 2003 ranged from 6.7% at Kamboni Health Centre, a rural site in the
central region to 32.9% at Nsanje District Hospital a semi-urban site in the
southern tip of the country. Overall, 19.8% of the study population tested HIV
positive, similar to the 2001 result of 19.5% as shown in table 8.

Table 8: Prevalence of HIV among women by site

 Sentinel Site                              Locality           Total     HIV+   %      95% CI
                                                               Sampled          HIV+
 Kaporo Health Centre                       Rural              210       43     20.5   15.2-26.6
 Mbalachanda Health Centre                  Rural              193       28     14.5   9.9-20.3
 Rumphi District Hospital                   Semi-urban         517       84     16.2   13.2-19.8
 Nkhata-Bay District Hospital               Semi-urban         522       126    24.1   20.6-28.1
 Mzuzu Health Centre                        Urban              846       176    20.8   18.1-23.7
 Kamboni Health Centre                      Rural              238       16     6.7    3.9-10.7
 Thonje Health Centre                       Rural              219       25     11.4   7.5-16.4


                                                       18
Kasina Health Centre               Rural         153              11          7.2     3.6-12.5
Mchinji District Hospital          Semi-Urban    548              99          18.1    15.0-21.6
St. Anne’s Mission Hospital        Semi-Urban    272              27          9.9     6.6-14.1
Ntcheu District Hospital           Semi-Urban    500              110         22.0    18.5-25.9
Lilongwe Bottom Hospital           Urban         810              137         16.9    14.4-19.7
Gawanani Health Centre             Rural         206              35          17.0    12.1-22.8
Milepa Health centre               Rural         205              31          15.1    10.5-20.8
Mianga Health centre               Rural         203              47          23.2    17.5-29.6
Mangochi District Hospital         Semi-Urban    510              74          14.5    11.6-17.9
Mulanje Mission Hospital           Semi-Urban    510              122         23.9    20.3-27.9
Nsanje District Hospital           Semi-Urban    511              168         32.9    28.9-37.2
Limbe Health Centre                Urban         804              222         27.6    24.6-30.9
Total                                            7,977            1,581       19.8    19.0-20.7




HIV prevalence in the urban (21.7%) and semi-urban (20.8%) areas continue
to be high and significantly different from rural areas (14.5%) as shown in
table 9 below.



Table 9: HIV Prevalence by locality

Location                Total       HIV+    % HIV+            95% CI
                        Sampled
Urban                         2,460     535          21.7                 20.1-23.4
Semi-Urban                    3,890     810          20.8                 19.6-22.1
Rural                         1,627     236          14.5                 12.8-16.3
Total                         7,977   1,581          19.8                 19.0-20.7




HIV prevalence in the southern region (23.7%) remains higher than in the
northern (20.0%) and central (15.5%) regions as shown in table 10.
Including or excluding St. Anne‟s Mission Hospital does not significantly
change the comparison.



Table 10: HIV Prevalence by Region

Region                  Total       HIV+    % HIV+             95% CI
                        Sampled
North                         2,288     457            20.0               18.4-21.7
Centre                        2,740     425            15.5               14.2-16.9
South                         2,949     699            23.7               22.2-25.3
Total                         7,977   1,581            19.8               19.0-20.7




                                          19
HIV prevalence remains unacceptably high in all age groups. Looking at the
95% confidence intervals in table 11, the data suggest similar prevalence
across the 20-24, 25-29 and 30-34 age groups. However, these figures are
statistically different from those observed in women aged 15-19.


Table 11: Age specific prevalence for women in the sample

 Age group          Total             HIV+          % HIV+         95% CI
                    Sampled
 <15                15                         3       20.0                      **
 15-19              1614                     245       15.2               13.5-17.0
 20-24              3,203                    635       19.8               18.5-21.3
 25-29              1,722                    384       22.3               20.4-24.4
 30-34              892                      219       24.6               21.8-27.5
 35-39              379                       65       17.2               13.6-21.4
 40+                105                       19       18.1               11.3-26.8
 Unknown            47                        11       23.4               12.3-38.0
 Total              7,977                  1,581       19.8               19.0-20.7
Note: ** 95% CI cannot be calculated because the sample size is very small.




Age specific HIV prevalence for aggregated urban sites are shown in table 12.
The pattern is the same as that in the overall sample as shown in table 11.




Table 12: Age specific prevalence for aggregated urban sites

 Age group          Total             HIV+          % HIV+          95% CI
                    Sampled
 <15                             2             1            50.0                 **
 15-19                         481            78            16.2          13.1-19.9
 20-24                        1079           225            20.9          18.5-23.4
 25-29                         565           138            24.4          21.0-28.2
 30-34                         251            75            29.9          24.3-36.0
 35-39                          60            12            20.0          10.8-32.3
 40+                            11             3            27.3                 **
 Total                       2,460           535            21.7          20.1-23.4
Note: ** 95% CI cannot be calculated because the sample size is very small. Numbers may not add up exactly
because those whose age is unknown have been excluded from the analysis




Age specific HIV prevalence for aggregated semi-urban sites are shown in
table 13. The pattern is also the same as that in the overall sample as shown
in table 11.




                                                       20
Table 13: Age specific prevalence for aggregated semi-urban sites

 Age group        Total            HIV+         % HIV+         95% CI
                  Sampled
 <15                           5            1           20.0                **
 15-19                       781          127           16.3         13.8-19.1
 20-24                      1544          331           21.4         19.4-23.6
 25-29                       811          190           23.4         20.6-26.5
 30-34                       435           99           22.8         19.0-27.0
 35-39                       229           42           18.3         13.5-24.0
 40+                          59           13           22.0         12.3-34.7
 Total                     3,890          810           20.8         19.6-22.1
Note: ** 95% CI cannot be calculated because the sample size is very small. Numbers may not add up exactly
because those whose age is unknown have been excluded from the analysis




Table 14 presents age specific HIV prevalence for aggregated rural sites. The
pattern is the same as that in the overall sample, only that prevalence is
higher in urban and semi-urban areas compared to the rural.




Table 14: Age specific prevalence for aggregated rural sites

 Age group        Total            HIV+         % HIV+         95% CI
                  Sampled
 <15                           8            1           12.5                **
 15-19                       352           40           11.4          8.3-15.3
 20-24                       580           79           13.6         11.0-16.7
 25-29                       346           56           15.2         12.5-20.6
 30-34                       206           45           21.8         16.4-28.1
 35-39                        90           11           12.2          6.3-20.8
 40+                          35            3            8.6                **
 Total                     1,627          236           14.5         12.8-16.3
Note: ** 95% CI cannot be calculated because the sample size is very small. Numbers may not add up exactly
because those whose age is unknown have been excluded from the analysis




Analyses were carried out to assess the association between level of
education and HIV prevalence. HIV prevalence for those without any formal
education (19.2%) and primary (19.1%) were similar, as was the case in
1999. HIV prevalence was significantly higher among women who had gone
up to secondary school level as compared to women with no education and
those going up to primary school level as shown in table 15.



                                                   21
Table 15: HIV Prevalence by level of education

 Education                Total       HIV+    % HIV+                       95% CI
 Level                    Sampled
 None                           1,921     369                     19.2               17.5-21.1
 Primary                        4,785     913                     19.1               18.0-20.2
 Secondary                      1,201     279                     23.2               20.9-25.7
 Post Secondary                    68      19                     27.9               17.7-40.1
 Total                          7,977   1,581                     19.8               19.0-20.7
Note: ** 95% CI cannot be calculated because the sample size is very small. Numbers may not add up exactly
because those whose education level is unknown have been excluded from the analysis




HIV Prevalence by partner‟s occupation status is shown in table 16.
Occupation was categorized during data entry (see Annex 3 for description).



There was no significant difference in HIV prevalence among women whose
partners were in business, skilled or professionals. However prevalence in
women whose partners are farmers/fishermen was significantly lower than
those partnered to business, skilled and professional men as shown in table
16.




Table 16: HIV prevalence by partners’ occupation


Partner occupation         Total sampled                %HIV+              95% CI

Unskilled                              205                 15.1          10.5-20.8
Skilled                                950                 20.2          17.7-22.9
professional                           930                 25.2          22.4-28.1
Business                              1900                 24.1          22.2-26.1
Military/police                         55                 21.8          11.8-35.0
security agent                         315                 20.6          16.4-25.6
Student                                 65                 23.1          13.5-35.2
Farmer/fisherman                      3446                 16.0          14.8-17.3
None                                   109                 19.3          12.3-27.9
Total                                 7975              100.0%




                                                   22
5.4. Syphilis Sero-prevalence


The syphilis prevalence presented in this report refers to Treponema Pallidum
Heamagglutination Assay (TPHA) confirmatory results. All site VDRL reactive
samples were further challenged by TPHA in the reference laboratory to
confirm current syphilis infection. Site-specific syphilis prevalence rates are
presented in table 17.




Table 17: Site-specific syphilis prevalence

 Sentinel Site           Locality     Total     Syphilis + % Syphilis +    95% CI
                                      Sampled
 Kaporo H. Centre        Rural              210          9           4.3   2.0 – 8.0
 Mbalachanda H. Centre   Rural              193          0           0.0   0.0 – 1.9
 Rumphi D. Hospital      Semi-Urban         517          0           0.0   0.0 – 0.9
 Nkhata-Bay D.Hospital   Semi-Urban         522         11           2.1   1.1 – 3.9
 Mzuzu Health Centre     Urban              846          2           0.2   1.0 – 3.5
 Kamboni H. Centre       Rural              238          3           1.3   0.3 – 3.6
 Thonje H. Centre        Rural              219          6           2.7   1.0 – 5.9
 Kasina H. Centre        Rural              153          2           1.3   0.2 – 4.6
 Mchinji D. Hospital     Semi-Urban         548         12           2.2   1.2 – 3.9
 St Annes M. Hospital    Semi-Urban         272          5           1.8   0.6 – 4.2
 Ntcheu D. Hospital      Semi-Urban         500         11           2.2   1.2 – 4.0
 Bottom Hospital         Urban              810         25           3.1   2.0 – 4.6
 Gawanani H.Centre       Rural              206          2           1.0   0.1 – 3.5
 Milepa H. Centre        Rural              205         10           4.9   2.4 – 8.8
 Mianga H. Centre        Rural              203         14           6.9   3.8 – 11.3
 Mangochi D. Hospital    Semi-Urban         510         27           5.3   3.6 – 7.7
 Mulanje M. Hospital     Semi-Urban         510         25           4.9   3.3 – 7.3
 Nsanje D. Hospital      Semi-Urban         511         14        2.7      1.6 – 4.7
 Limbe Health Centre     Urban              804         35           4.5   3.1 – 6.1
 TOTAL                                   7,977        213            2.7   2.3-3.7




There appears to be an inverse relationship between level of education and
syphilis infection. See table 18. The higher the level of education of the
woman, the lower the likelihood of active syphilis.




                                          23
Table 18: Syphilis prevalence by education level

 Education level             Total             Syphilis      %              95% CI
                             Sampled           +             Syphilis +
 None                        1921              70            3.6            2.9 – 4.6
 Primary                     4631              121           2.5            2.1 – 3.0
 Secondary                   1227              21            1.7    1.3     1.1 – 2.7
 Post Secondary              68                1             1.5            0.0– 7.9
 Total                       7,977             213           2.7            2.3 – 3.7
Numbers may not add up exactly because those whose education is unknown have been excluded from the analysis




The observed syphilis prevalence rates across regions are statistically
different. Syphilis prevalence in the southern region (4.3%) remains higher
than the central (2.3%) and northern (1.0%) regions as shown in table 19.

Table 19: Syphilis prevalence by region

 Region                      Total             Syphilis      %              95% CI
                             sampled           +             Syphilis +
 North                       2288              22            1.0            0.6 – 1.5
 Centre                      2740              64            2.3            1.8 – 3.0
 South                       2949              127           4.3            3.6 – 5.1
 Total                       7,977             213           2.7            2.3 – 3.7

Syphilis prevalence by locality, either urban or rural, is shown in table 20.
There is no significant difference in prevalence between urban and rural.

Table 20: Syphilis prevalence by locality

 Locality           Total            Syphilis +        % Syphilis +               95% CI
                    sampled
 Rural              1627             46                2.8                       2.1 – 3.8
 Semi-Urban         3890             105               2.2                       2.2 – 3.3
 Urban              2460             62                2.5                       2.0 – 3.2
 Total              7,977            213               2.7                       2.3 – 3.7

Syphilis prevalence across age groups is shown in table 21. There is no
significant difference in prevalence across age groups.

Table 21: Syphilis prevalence by age

 Age group          Total sampled       Syphilis +           % Syphilis +               95% CI

 < 15                             15                    0                      0.0        0.0 – 21.8
 15 – 19                        1614                   34                      2.1         1.5 – 3.0
 20 –24                         3203                   81                      2.5         2.0 – 3.1



                                                     24
    25 – 29                   1722                 53                    3.3       2.5 – 4.3
    30 – 34                    892                 23                    2.6       1.7 – 3.9
    35 – 39                    379                 15                    4.0       2.3 – 6.6
    40+                        105                  1                    1.0       0.0 – 5.2
    Unknown                     47                  2                    4.3      0.5 – 14.5
    Total                    7,977                213                    2.7       2.3 - 3.7




6. DISCUSSION AND TRENDS

In the absence of complementary data on changes in sexual behaviours it is
not easy to explain dynamics of the epidemic among antenatal attendees. It
is evident that several factors can drive any epidemic away from a steady
state. These include: factors affecting duration of disease (e.g. new
treatments), depletion of the pool of uninfected individuals and changes in
frequency or nature of behaviours associated with infection among others.
However, very little is known about these factors in Malawi as of now.


HIV and socio-economic factors
HIV prevalence was significantly higher in urban and semi-urban sentinel sites
as compared to rural sites. However there was no significant difference in
prevalence between urban and semi-urban areas. When the epidemic was
young, HIV prevalence in the urban areas were markedly high and statistically
different from semi-urban and rural areas. Over the years, however, the
prevalence gap between urban and semi-urban areas has consistently
narrowed. It is, therefore, important to critically examine the underlying
factors driving the epidemic in these localities to come up with specifically
tailored intervention programmes.


 In the year 2002, 57% of urban households were categorized as ultra poor.
Only 15% of rural households were classified likewise6. Poverty is one of
major structural factors that have an impact on individual sexual behaviour7.
Poverty could be one of the factors resulting in disparities in prevalence
between urban and rural areas. For instance, women tend to stay working as
commercial sex workers for economic reasons as they cannot afford to
survive despite their awareness of risks associated with commercial sex8.

Unmarried women in urban areas are more likely to be sexually active than
rural women9 while sexually active urban men are more likely than sexually
active rural men to have multiple partners9. Furthermore, high levels of
income provide an opportunity to men to engage in sex with multiple partners
in marital or extra marital relationships.10


 Households are grouped into 5 quintiles (20% categories) according to their rank, based on their
household welfare indicators used in poverty analysis (a world bank module).


                                                  25
HIV and the youth
Since prevalence of any disease is crudely estimated by incidence and
duration of disease, the high HIV prevalence in the 30 and older age groups
could imply high incidence rates among adolescents and young adults. This is
particularly so because currently there is no widespread use of antiretroviral
drugs in the country to prolong the lives of those infected.

This survey indicated that HIV and syphilis prevalence among antenatal
attendees aged 15-19 years were 15.2% and 2.1% respectively. The 1996
KAP in Health Survey revealed that STI prevalence was 0.7% among girls
aged 15-19 and 4.8% among boys of the same age group in the general
population11. Though not comparable to the 1996 KAP in health survey, the
2000 DHS reported that STI prevalence in the same age group was 0.9% for
girls and 1.7% for boys9. The median age for sexual debut falls within the age
group 15-19 years and the 2000 MDHS estimated it at 17 years 9. Gravidity
data from this sentinel survey (showing that 80% of ANC attendees aged 15–
19 presented with their first pregnancy) supports MDHS sexual debut
findings. Though HIV prevalence in the 15-24 age group is regarded as a
proxy for incidence (new HIV infections) as per international guidelines, it
may be important in Malawi to also look at women aged 15-19 in developing
a proxy for HIV incidence (see figure 2).

Figure 2: Trends of HIV infection by age group for all ANC attendees

                 40

                 35

                 30

                 25
   Percent (%)




                                                                                 15-19
                 20                                                              15-24

                 15                                                              all

                 10

                 5
                                             *
                 0
                       1995      1996        1997   1998    1999   2001   2003
                      * Data not available




                                                           26
The occurrence of high numbers of new HIV infections and STIs in the age
group 15-19 years is worrisome. This is a very big challenge for behaviour
change intervention programme implementation.
Education status and prevalence of HIV and syphilis
This survey has revealed that the higher the level of education of a woman,
the lower the likelihood of active syphilis. This finding is in line with results in
the 2001 and 1999 sentinel surveillance surveys3,4. This could be due to early
recognition of syphilis infection and health care seeking behaviour, which
generally improves with increasing level of education. In future studies,
however, it may be informative to stratify primary school attendees to assess
any masking effects of broader categorization. However, it seems the higher
the level of education of a woman the higher the risk of testing HIV positive.
HIV prevalence was significantly higher among those who have gone up to
secondary school level as compared to women going up to primary school.

HIV trends
In fourteen of the 19 sentinel sites, the 2003 sero-prevalence did not differ
from the 2001 sero-prevalence by more than would be expected by random
fluctuation3. This observation continues to support the previous years finding
that sentinel site prevalence is stabilizing. Kaporo Health Centre, Nkhata-Bay
District Hospital, and Thonje Health Centre each increased significantly, while
Mchinji District Hospital and St. Anne‟s Mission Hospital each decreased
significantly from 2001. The need for complementary data essential to
interpreting these trends continues to be critical. Such data include HIV
prevalence among VCT and PMTCT clients, blood donors, as well as data for
populations at particularly high risk, and STI prevalence data.

The overall HIV prevalence for antenatal attendees in 2003 (19.8%) has
remained stable compared with 2001, and lower than 1999 as shown in figure
3.

Figure 3: Trends of HIV infection for all ANC attendees




                                        27
                      40

                      35

                      30
                                                                       24.1
        Percent (%)



                      25
                                                 20.6     21.0
                                         19.1                                         19.5          19.8
                      20   17.4   18.2

                      15

                      10

                       5

                       0
                           1994   1995   1996   1997      1998         1999           2001          2003
                                                        YEAR OF SURVEY



Over time the gap between urban and semi-urban prevalence curves has
been narrowing such that from 1999 there has been no statistical difference
between the two. Urban and semi-urban HIV prevalence was 25.3% and
20.9% respectively in 1998, and 21.7% and 20.8% in 2003. 5 There has been
a significant linear decline in HIV prevalence in the urban and the semi-urban
sites from 1999 to 2003, (chi-square =8.5, p<0.005; chi-square=41.8,
p<0.001 respectively (see figure 4). However, while still much lower than the
urban and semi-urban, prevalence in the rural sites increased significantly
from 12.1% in 1999 to 14.5% in 2003, chi sq=4.6, p<0.05. See figure 4. To
reverse this trend, expanding effective behaviour change interventions such
as VCT in rural areas is critical.



Figure 4: Trends of HIV infection by locality

                      30

                      25

                      20
Percent (%)




                      15                                                                              urban
                                                                                                      semi-urban
                      10
                                                                                                      rural
                      5

                      0
                           1995   1996   1997   1998     1999         2000    2001   2002    2003
                                                 YEAR OF SURVEY




                                                                 28
The Southern Region has consistently registered HIV prevalence higher than
the other two regions. From 1999 however, the trend curves for the south
and centre have been drifting further apart as a result of continued declining
prevalence in the Central Region. It is not clear why prevalence is declining in
the central region. On the other hand, the 2003 prevalence in the north has
surpassed the centre as seen in figure 5. Of particular concern is the future
direction of the curve for the northern region. Population and socio-economic
growth in Mzuzu city and areas along major transportation routes may be
having an effect on the observed prevalence.

This survey indicated that the Northern Region has the lowest syphilis
prevalence. However the 2000 MDHS showed the prevalence of self reported
STIs among women higher in the Northern Region (1.7%) than in the Central
(1.5%) and Southern Regions (1%)9. Review of additional STI data is
needed.




Figure 5: Trends of HIV infection by region


                 30

                 25

                 20
  Perce nt (%)




                                                                                     north
                 15                                                                  centre

                 10                                                                  south

                 5

                 0
                      1995   1996   1997   1998   1999   2000   2001   2002   2003
                                            YEAR OF SURVEY




In most sentinel sites, prevalence peaked in the mid to late 1990s. See
figures 6, 7, 12 and 13. Figures 6 through 8 present HIV prevalence trends
for the three urban centres. Unlike Blantyre where the prevalence has
stabilized, Lilongwe has shown a consistent decline since 1999. Mzuzu on the
other hand has shown no specific pattern.



                                                   29
Figure 6: HIV infection among ANC attendees,
Blantyre
           40
                                                                                                     34
           35                                                                                 32.8
                                                                            30        30.2                30.3 30.4
                                                                                                                                    28.5
           30                                                       27.1                                              27.9                 27.6
                                                              26
           25
 Percent (%)




                                                      21.9
                                               18.6
           20

           15

           10                        8.2

               5        2.5

               0
                       1985          1987      1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999                               2001   2003



                                                                      YEAR OF STUDY




Overall HIV Prevalence among antenatal attendees is declining in Lilongwe as
shown in figure 7. To be understood, this decline needs to be related to
trends in deaths due to AIDS, and behaviour changes, such as increased
condom use and decreased number of sexual partners.


Figure 7: HIV infection among ANC attendees, Lilongwe

                       40

                       35

                       30                                     27
                                                                     25          26          25.2
         Percent (%)




                       25                          22.5
                                                                                                           20.1
                       20                   16.5                                                                             16.9
                       15

                       10

                        5

                        0
                              1992   1993   1994   1995      1996   1997        1998     1999             2001               2003


                                                                       YEAR OF STUDY



                                                                           30
31
Prevalence trends in Mzuzu are not showing any specific pattern and there is
no sign of decrease in prevalence as shown in figure 8.


Figure 8: HIV infection among ANC attendees, Mzuzu

                40

                35

                30
                                                                              23.3
  Percent (%)




                25                         21.2          21.4                                 20.8
                                                  18.5                 18.5           18.6
                20                                              17
                15

                10

                    5

                    0
                            1992    1993   1994   1995   1996   1997   1998   1999   2001     2003



                                                                 YEAR OF STUDY

Figures 9–11 show HIV prevalence trends in selected rural sites. The trend in
many rural sites is on the increase, as seen in Mbalachanda, Thonje, and
Mianga Health Centres. A possible reason for this trend could be increased
interaction between people in rural and urban areas.

Figure 9: HIV infection among ANC attendees,
Mbalachanda

                    40

                    35

                    30
      Percent (%)




                    25

                    20
                                                                                             14.5
                    15       11.6
                                                                              9.5    10.5
                    10               7.8     7     7.3          8.2
                                                          5.9
                        5                                              3.3

                        0
                             1992   1993   1994   1995   1996   1997   1998   1999   2001    2003


                                                         YEAR OF STUDY



Figure 10: HIV infection among ANC attendees, Thonje


                                                                32
                40

                35

                30

                25
  Percent (%)




                20

                15                                                                                                      11.4
                                                                          8.4               9.4
                10                                                                 6.5
                                                  4.7          4.7                                        4.5
                 5    2.4     2.8        3

                 0
                     1992    1993    1994      1995        1996       1997         1998     1999          2001          2003



                                                        YEAR OF STUDY




Figure 11: HIV infection among ANC attendees, Mianga

                40

                35

                30
  Percent (%)




                25                                                                                               23.2
                                                                            20.9     21.7
                                    19
                20          16.5                                                                   16.9
                                                                 13.8
                15                           12
                                                        10.8
                10   6.2
                5

                0
                     1992   1993    1994     1995       1996     1997       1998    1999           2001          2003


                                                           YEAR OF STUDY



The following figures, 12 through 14 show HIV prevalence trends in selected
semi-urban sites. From 1995 to 2001 Nsanje has shown a consistent increase
in HIV prevalence while there is a drop in 2003. There is a need for more data
points before conclusive statements can be made about further trends.
Additional research on behaviour is needed to better understand driving
forces behind the epidemic in Nsanje district. In other sites, such as Ntcheu
and Nkhata-Bay District Hospitals, the trend appears to be more stable.


Figure 12: HIV infection among ANC attendees, Nsanje


                                                                     33
                40                                                                   35.8
                35                                                                            32.9

                30                                                            26
                                                                     23.6
  Percent (%)


                25                    21                     21.3
                20                                   17.2

                15
                                             9.6
                10
                5
                0
                     1992    1993    1994    1995    1996    1997    1998    1999    2001     2003



                                                     YEAR OF STUDY



Figure 13: HIV infection among ANC attendees,
Ntcheu

                40

                35                                                             33

                30
                                                       24
  Percent (%)




                25                                            22.5                               22
                                                                      20.8
                                                                                       18.6
                20                     18     16.8
                15

                10

                 5

                 0
                      1992    1993    1994    1995    1996    1997    1998    1999    2001      2003



                                                     YEAR OF STUDY




Figure 14: HIV infection among ANC attendees, Nkhata-
bay




                                                             34
                  40

                  35

                  30
                                                                         25                         24.1
                  25
    Percent (%)


                                                   22                          21.6
                                          20.5                   20.5
                                                                                      18.6
                  20                                      16
                  15

                  10

                      5

                      0
                          1992   1993     1994    1995   1996    1997   1998   1999   2001          2003




Site-specific trends in age-specific HIV prevalence were examined, focusing
on the younger age groups. Only observations in urban sites are presented,
because of the need for adequate sample sizes for stratified analyses. See
figures 15 to 17. The observed increases in prevalence for 15-19 year olds
between 2001 and 2003 in Blantyre, Lilongwe, and Mzuzu are not statistically
significant. On the other hand HIV prevalence in the age group 20-24 years
shows a decreasing trend for the Lilongwe sentinel site, as shown in table 16.


Figure 15: Age specific ANC HIV infection - Blantyre


            50
            45
            40
            35
 % HIV +




            30                                                                                         15-19
            25                                                                                         20-24
            20                                                                                         25+
            15
            10
                  5
                  0
                          1996     1997          1998     1999                 2001          2003
                                                         YEAR OF SURVEY




Figure 16: Age specific ANC HIV infection - Lilongwe




                                                                 35
               50
               45
               40
               35
     % HIV +




               30                                                           15-19
               25                                                           20-24
               20                                                           25+
               15
               10
                5
                0
                    1996   1997   1998     1999             2001    2003
                                          YEAR OF SURVEY




Figure 17: Age specific ANC HIV infection - Mzuzu


               50
               45
               40
               35
  % HIV +




               30                                                          15-19
               25                                                          20-24
               20                                                          25+
               15
               10
                5
                0
                    1996   1997   1998    1999             2001    2003
                                         YEAR OF SURVEY




Syphilis trends and patterns
On the overall, there is a decline in syphilis prevalence over the years. The
trend in syphilis prevalence from 1995 to early 2003 is shown in table 22 and
figure 18.




Table 22: Trends in Syphilis prevalence



                                                 36
 Year of survey            % Syphilis   95% CI      Sample size

 2003                      2.7          2.3-3.7     7977
 2001                      3.9          3.5-4.4     7361
 1999                      3.5          3.1-4.0     6885
 1998                      4.2          3.8-4.7     6766
 1996                      7.0          6.0-8.0     4163
 1995                      4.0          3.8-5.1     4040



FIGURE 18: Trends in Syphilis prevalence, 1995 to 2003


                8
                                 7
                7
                6
  Percent (%)




                5                       4.2
                     4                                            3.9
                4                                  3.5
                                                                        2.7
                3
                2
                1
                0
                    1995     1996       1998       1999       2001      2003
                                        YEAR OF STUDY




HIV and syphilis relationship
There was a general association between HIV and syphilis infection in the
study sample, chi-square 28.8 p-value <0.001. This result shows that syphilis
and HIV were not occurring independently. The relationship was most clearly
observed in Blantyre, Lilongwe and Mianga health centre because HIV and
syphilis prevalence was relatively high in these sites. Mianga health centre
had the highest syphilis prevalence (6.9%) among all sentinel sites and the
highest HIV prevalence (23.2%) among rural sites.




                                              37
7. CONCLUSION


  HIV prevalence among antenatal attendees has declined in urban and
  semi-urban areas. However there is evidence of an increasing trend in HIV
  prevalence in rural areas. There is also an increasing trend in HIV
  prevalence in the northern region while prevalence shows a decreasing
  trend in the central region and a slight decrease in the southern region.
  Increasing trends in rural Malawi in general, and Nsanje, Mbalachanda and
  Thonje in particular, need special attention.

  It is important to note that national HIV prevalence estimates and
  projections based on findings from this sentinel surveillance survey will be
  generated and disseminated in a forthcoming report.




                                     38
8. RECOMMENDATIONS

1. Include additional sentinel sites in future and increase sample sizes,
   particularly in the young age groups and in the rural areas.

2. MOHP, NAC and NGOs should play a major role in advocating for and
   implementing Second Generation Surveillance for HIV to better understand
   the dynamics of the HIV epidemic in Malawi (see annexe 1 for details on
   second generation surveillance). The country should at least put in place or
   strengthen the following systems:
               i. Periodic behavioural surveillance especially among the youth
              ii. Surveillance of VCT clients HIV serology
             iii. Surveillance of PMTCT clients HIV serology
             iv. Surveillance of STIs
              v. Surveillance of blood donors HIV serology
             vi. Periodic national population based sero-surveillance

3. MOHP with assistance from NAC and DACCs should critically examine and
   document underlying factors driving the epidemic in all districts with the
   purpose of coming up with information for specifically tailored intervention
   programmes. This could be done through district or regional workshops.


4. Operational research should be conducted on why HIV prevalence is:
       still very high among antenatal attendees in Nsanje
       increasing among antenatal attendees in northern region and rural
          sites with special attention to Mbalachanda, Thonje and Mianga health
          centres.
       decreasing amongst attendees in sentinel sites located in central
          region for example Bottom hospital in Lilongwe.

5. Interventions should be put in place to reverse the increasing trend of HIV
   prevalence in rural areas. VCT should be a priority.

6. The data collection form for sentinel surveillance should be updated as
   follows:
         Education level should refer to completion of classes or grades.
         The variable „education‟ on the form should be updates as follows:
           primary school should further be categorized into completed standard
           1-5, 6-8, form 1-2 and form 3-4.
         The variables „mother‟s‟ and „partner‟s occupation should be coded on
           the form.

7. A fixed season for conducting HIV sentinel surveillance surveys should be
   identified. The best season would be the one during which ANC clinics receive
   substantial numbers of clients.




                                     39
9. REFERENCES

  1. UNAIDS. Report on the Global HIV/AIDS Epidemic 2002. Geneva.
  2. NAC. National HIV prevalence estimates from sentinel surveillance data
     2001.
  3. NAC. Sentinel Surveillance Report 2001:HIV/Syphilis Sero-prevalence in
     antenatal clinic attendees. 2001.
  4. NACP. Sentinel Surveillance Report 1999:HIV/Syphilis Sero-prevalence
     in antenatal clinic attendees. 1999.
  5. NACP. Sentinel Surveillance Report 1998:HIV/Syphilis Sero-prevalence
     in antenatal clinic attendees. 1998.
  6. NSO. 2002 Malawi Core Welfare Indicators Questionnaire Survey:
     Report of Survey Results.
  7. Van der Borne, F. Dynamics of bartering sex for subsistence. An
     exploratory study in urban Malawi. 1998.
  8. Mwangulube, K. The Dilemma of AIDS prevention among commercial
     sex workers in Lilongwe City, Malawi: Socio-economic or health
     survival? 2001.
  9. NSO/ORC MACRO. Malawi: Demographic Health Survey 2000. 2001.
  10. Tsoka M.G. Analysis of the HIV/AIDS epidemic and the high population
     growth in Malawi. University of Malawi CSR. 1999.
  11. NSO. Knowledge Attitudes and Practices in health survey 1996.




                                     40
ANNEX 1


SECOND GENERATION SURVEILLANCE

In July 2000 the World Health Organisation (WHO) and other partners released the guidelines
for conducting second-generation surveillance studies. The concept of second generation
HIV/AIDS/STD surveillance systems sets out to achieve the following objectives: to better
understand trends, to understand the behaviours driving the epidemic, to provide a system
that is more focussed on sub-populations at higher risk of infection, to provide a system that
is flexible and moves with the needs and state of the epidemic and finally to provide a system
that will enable better use of data so as to maximise opportunities to plan for prevention and
care.

Second generation HIV/AIDS/STD surveillance systems are meant to be appropriate to the
epidemic, dynamic, use resources where they will generate the most useful information,
compare biological and behavioural data for maximum explanation power, integrate
information from other sources and use data produced to improve and increase the national
response.

In Malawi second generation surveillance may include strengthening surveillance in the age
group 15-24 years. The epidemic in Malawi is well-established “a generalized epidemic”.
Surveillance efforts in generalized epidemics should focus on new infections because it is
difficult to interpret changes in prevalence. Although it is difficult to measure HIV incidence
resources can be concentrated on the youth since their infections are likely to be relatively
recent hence not biased by reduced fertility. Behaviour surveillance among young people is
very critical. Establishing safer behaviour from the beginning of young people‟s sexual lives
could be more effectual in changing the course of the epidemic than changing behaviour in
older groups.

Malawi should put in place behaviour surveillance through repeat cross-sectional household
surveys preferably annually. The following biological data should also be analysed annually to
validate sentinel surveillance findings: 1. blood donor‟s data 2. VCT data 3. PMTCT data 4.
STI data 5. TB data and 6. AIDS cases data




                                              41
ANNEX 2

GLOSSARY

Chi-square test
The statistical test used to test the null hypothesis that proportions are equal or equivalently,
the factors or characteristic are independent or associate

Confidence interval
The interval computed from sample data that has a given probability that an unknown
parameter, such as the proportion is contained within the interval. Common confidence
intervals are 95%. Meaning that the probability that the unknown parameter lies in the
interval is 0.95.
If confidence intervals for groups which are being compared do not overlap, then there is a
statistical difference.

Incidence
Number of new cases in a defined population and period of time.

Null hypothesis
The hypothesis being tested about a population. Null generally means “no difference” and as
a result refers to a situation where no difference exists. E.g. prevalence rates between rural
and urban.

Prevalence
Total number of individuals who have a disease during a particular period divided by the
population at risk of having the disease during the period.

P value
This is the probability of observing a result, as extreme as or more extreme than the one
actually observed from chance alone. Usually, when p value is less than 0.05, we reject the
null hypothesis i.e. we conclude that there is a statistical difference or an association.

Statistical significance
Generally interpreted as a result that would occur by chance, for example 1 time in 20 with a
p-value of 0.05. There is statistical significance when null hypothesis is rejected.

Variable
Any quantity that varies is a variable. Any attribute, phenomenon or event that can have
different values is a variable.



.




ANNEX 3

Occupation categorization


                                               42
Occupation was categorized from open-ended responses as follows:


Unskilled
Labourers or manual workers as well as other jobs that do not require particular skills

Business
Persons reporting occupation as „business‟, as well as small-scale businessmen e.g. owners of
shops, hawkers, bakers, vendors etc.

Skilled
Occupations that need technical skills e.g. mechanics, builders, carpenters, electricians etc

Professional
Occupations requiring professional and academic qualifications e.g. administrator, manager,
teacher, banker e.t.c.

Military/police

Security agent

Student

Farmer/fisherman

None




                                               43

				
DOCUMENT INFO