LA Shafer2, S Biraro2, A Kamali2, H Grosskurth2,

1

HIV Epidemic Trends in Uganda 1989 - 2005

• LA Shafer2, S Biraro2, A Kamali2, H Grosskurth2,
• W Kirungi1, E Madraa1, A Opio1

1National AIDS Control Programme, Ministry of
Health, Uganda 2MRC/UVRI Uganda Research Unit on
AIDS, Uganda Virus Research Institute
2
Background
Effects of not Accounting for Correlated Residuals

• Much published research describing declining
  prevalence/incidence in Uganda throughout 1990s
• Mbulaiteye et al, Lancet 2002, 360, 41-46
• Whitworth et al, Trop Med Internatl Health,
  2002, 712, 1047-1052
• Kirungi et al, Sex Transm Infect, 2006, 82(suppl
  1), i36-i41

• The estimate of ?1 is not biased.

• The estimate of s.e.( ) is biased.
• The estimate of ?1 is inefficient.

3
Background
Effects of not Accounting for Correlated Residuals

• Recently, researchers have noticed evidence that
  decline may be levelling off
• 2004/5 Uganda HIV/AIDS Sero- Behavioural Survey
  report
• Question Is the HIV epidemic in Uganda
• beginning to change?

• The estimate of ?1 is not biased.

4

Objective

• 1.) Examine trends in HIV prevalence from urban
  and semi-urban anti-natal clinic (ANC)
  surveillance (1989-2005)
• 2.) Examine trends in prevalence and incidence
  from a population cohort in a rural area of
  Uganda (1989-2005)

5
Study Area
ANC Sites

UGANDA




















6

Study Population

• Ante-Natal Clinical Surveillance
• multiple sites throughout the country
• followed since 1989
• 24 sites with annual HIV prevalence data
• General Population Cohort Kyamulibwa / Masaka
• cluster of 25 villages (18,000 pop.)
• 15 villages since 1989
• 10 villages added to cohort in 1999
• annual survey
• door-to-door census
• blood collection for HIV serology
• annual HIV prevalence and incidence data

7

Results

• ANC Prevalence in 24 Sites since 2002
• Prevalence declined 7 sites
• Prevalence increased 10 sites
• Prevalence unchanged 7 sites

8
ANC Surveillance sites with subtle increase in
HIV prevalence between 2002 and 2005
Results
9
ANC Surveillance site with subtle decrease in
HIV prevalence between 2002 and 2005
Results
10

• How do the ANC trends compare to prevalence and
  incidence in the Masaka rural cohort?

11

Results
12

Results
Factors Influencing Prevalence
Migration? Mortality? Aging? Incidents?
13

Results
14

Results
15
Results
Males 1.) Second incidence peak in the 40-49
year age group, years 2000-2005 2.) For older
men, incidence higher in 2000-2005 than 1990-1994
Females 1.) Incidence slightly higher in
2000-2005 than 1995-1999 for most age groups
16

Conclusions

• Prevalence and Incidence
• Falling during 1990s
• Fall has stopped since about 2000
• Increasing?
• An incidence peak emerging among men 40-49 yr old
• Changing prevalence
• may partly be explained by changing migration,
  mortality
• partly explained by changing incidence
• Role of changes in risky sexual behaviour?
• Role of natural epidemiologic cycle?

17

• Thank You

18
Study Area
ANC Sites Prevalence up since 2002
Prevalence down
Prevalence stable








19

Ever Had Sex
20

21

22

23

Results
24

Results