Estimating the impact of HIV and AIDS in Zimbabwe

1
Estimating the impact of HIV and AIDS in Zimbabwe

• A.D. McNaghten
• CDC Zimbabwe

2
Overview for Generalised Epidemics
Surveillance data from ANC
EPP
Adult HIV Prevalence
UN Population Divisions Population estimates
Spectrum

• PLWHA
• New Infections
• AIDS Deaths

Epidemiology assumptions
3
Process for Zimbabwe Estimates

• WHO/UNAIDS training in estimation methods and
  software (Harare, April 2003)
  
• Ministry of Health and Child Welfare (MOHCW) 2002
  Antenatal Clinic (ANC) Report and National
  Prevalence Estimate Workshop (Harare, May 2003)
  
• Estimate Working Group
• MOHCW, National Microbiology Reference
  Laboratory, Central Statistical Office,
  University of Zimbabwe, Futures Group, UNAIDS,
  Imperial College, CDC Zimbabwe
• Classification of ANC sites as Urban, Rural and
  Other
  
• EPP curve fits
• based on sentinel surveillance data collected
  from women attending antenatal clinics
  
• Estimates and ranges from Spectrum

4
EPP

• ANC sites classified by location as
• urban
• rural (communal lands, resettlements, small-scale
  commercial farms)
  
• other (large scale commercial farms,
  administrative centres, growth points, other
  urban e.g., mines, state land e.g., national
  parks, special e.g., army camp)
• 6 rural sites were adjusted down 30
• Women lived in or near commercial centre

5
EPP

• Adjustments for unreliability of ANC surveillance
  estimates
  
• Implausibly high results at 2 sites additional
  test results used
  
• Inconsistent data points for 2 sites removed
• Representativeness of ANC data of the prevalence
  among women and men 15-49 in the general
  population
  
• EPP curve fits

6
EPP curve fits for Urban, Rural and Other census
strata
Percent
7
Adult (15-49) HIV prevalence by census strata,
2003

• Urban 28.1
• Rural1 20.9
• Other2 34.9
• Overall 24.6

1 Communal land, small scale commercial farms,
resettlements 2 Large scale commercial farms, adm
inistrative centres, growth points, other urban
(e.g., mines), state land (e.g., national parks),
special (e.g., army camp)
8
Overview for Generalised Epidemics
Surveillance data from ANC
EPP
Adult HIV Prevalence
UN Population Divisions Population estimates
Spectrum

• PLWHA
• New Infections
• AIDS Deaths

Epidemiology assumptions
9
Spectrum

• To create a national HIV estimate Spectrum makes
  a demographic and an epidemiologic projection
  
• Demographic and epidemiologic assumptions can be
  accepted or revised
  
• Local and national data can be used

10
Spectrum Demographic Assumptions

• UN Population Division figures
• Total fertility and age-specific fertility rates
  estimates from 1988, 1994 and 1999 Demographic
  Health Survey
  
• Life expectancy at birth (in the theoretical
  absence of HIV) set using UN Population Division
  estimates
  

11
Spectrum Epidemiologic Assumptions

• Adult HIV prevalence was read from EPP file
• Perinatal transmission rate 32 through 1999 30
  2000-2003
  
• Total fertility rate reduction based on
  Manicaland Study data
  

12
Spectrum Epidemiologic Assumptions

• UNAIDS ratio of female to male HIV prevalence
  (sub-Saharan Africa) through 1999 1.35
  2000-2003
  
• Ratio of prevalence by age group to prevalence in
  25-29 year age group from Young Adult Survey
  (15-29) and Manicaland project 1998-2000

13
Spectrum Outputs

• Estimates of prevalence
• males and females
• children and adults
• Estimates of incidence
• Mortality for adults and children
• AIDS orphans

14
Adult HIV prevalence estimate and range, 2003
Percent
2003 estimate 24.6, range 20-28
15
Estimated number of people living with HIV, 2003

• Estimate Range
• Adult (15-49) prevalence 24.6 (20 - 28)
• Adults and Children 1,820,000 (1,473,000 -
  2,020,000)
  
• Adults (15-49) 1,540,000 (1,250,000
  1,710,000)
  
• Women (15-49) 870,000 (700,000 960,000)
• Children (0-14) 165,000 (131,000 186,000)

16
AIDS Deaths During 2003

• Estimate Range
• Adults (15-49) 135,000 (107,000
  151,000)
  
• Women (15-49) 77,000 (61,000 86,000)
• Children (0-14) 36,000 (29,000 41,000)

17
New AIDS Cases During 2003

• Estimate Range
• Adults (15-49) 138,000 (110,000
  154,000)
  
• Women (15-49) 78,000 (62,000 87,000)
• Children (0-14) 36,000 (29,000 41,000)

18
New HIV Infections During 2003

• Estimate Range
• Adults (15-49) 166,000 (137,000
  184,000)
  
• Women (15-49) 88,000 (73,000 97,000)
• Children (0-14) 40,000 (32,000 45,000)

19
Adult HIV prevalence comparison of Zimbabwe and
UNAIDS estimates
Percent
UNAIDS Report on the Global HIV/AIDS Epidemic,
2002
20
Estimated number of people living with HIV a
comparison of 2001 estimates

• New Estimates UNAIDS Estimates
• Adult (15-49) prevalence 24.9 33.7
• Adults and children 1,800,000 2,300,000
• Adults (15-49) 1,520,000 2,000,000
• Women (15-49) 860,000 1,200,000
• Children (0-14) 160,000 240,000

21
Reasons for Differences in Zimbabwe and UNAIDS
Estimates

• Updated and more accurate ANC data entered into
  EPP
  
• Stratification of areas as urban, rural or
  other (e.g., growth points, mines) more
  accurately reflects prevalence in rural and
  other areas
• Updated United Nations population figures reflect
  population declines since the last estimates

22
Changes for 2004 Estimates

• The process and methods are constantly evolving
  and improving
  
• Recent UNAIDS Working Group changes
• Mother to child transmission
• Orphan definition

23
Concluding Remarks

• Methodological changes were the reason for a
  decrease in prevalence
  
• More data are needed to say the epidemic is
  declining or leveling
  
• Estimates are best represented by a range

24
Role of the ME Officer

• Be familiar with sources of data available
• Identification of individuals and organizations
  who can contribute to the estimate process
  
• Understand how estimates are derived (in general
  and country-specific)
  
• Be able to explain country estimates to others
• Data and methods
• Relationship to other country-level data
• Changes over time