Estimating the Magnitude of HIV Epidemic in Ethiopia

1
Estimating the Magnitude of HIV Epidemic in
Ethiopia

• Shabbir Ismail MD, MPH
• Surveillance Medical Epidemiologist
• CDC-Ethiopia
• SI / ME Officers Orientation, Atlanta, GA
• August 16, 2004
• Other Members of the Surveillance Team
• Aseged Woldu MD, (MOH)
• Wolfgang Hladik MD, MSc (CDC-Atlanta)
• Jelaludin Ahmed MA (CDC-Eth)
• Tadesse Wuhib MD, MPH (CDC-Eth)

2
Outline

• Brief Background
• Overview of HIV/AIDS Surveillance
• Estimating HIV Magnitude
• Data processing
• Estimations and projections
• Epidemic Projection Package (EPP)
• Spectrum processes
• Examples of EPP and Spectrum Outputs for
  Ethiopia
  
• Practical Challenges Faced

3
ETHIOPIA

• Ethiopia is located in the horn of Africa
• Population 70 million (end 2003)
• 85 rural
• Predominantly agrarian economy
• Area 1,100,000 km2
• More than 80 ethnic groups
• Federal Democratic Republic
• 9 National Regional States
• 2 City Administrative Councils
• Regions further decentralized in to 68 zones, 550
  districts and 10,000 localities

4
Ethiopia and the Regional States
5
Overview of HIV/AIDS Surveillance in Ethiopia
6
Sources of HIV Related Data

• HIV surveillance
• Antenatal care (ANC) based sentinel surveillance
• AIDS case reporting
• Special population surveys
• Other sources of data related to HIV status
• Blood donors
• Voluntary Counseling and Testing (VCT) clients
• Visa applicants
• Behavior Surveillance Survey (BSS)

7
ANC Based HIV Surveillance

• Started in 1989 with one urban site in the
  capital
  
• Grew to 67 in 2003 (34 rural)
• Uses left-over blood from routine syphilis
  testing
  
• HIV testing is unlinked anonymous

8
(No Transcript)
9
Distribution of Sentinel Surveillance Sites
10
Surveillance Related Activities

• Comprehensive assessment of the system
• Joint planning between partners
• Major areas of CDC support
• Improving rural representation
• Improving quality of data generated
• Instituting proper data reporting and processing
  at MOH
  
• Using appropriate data analysis and estimation
  processes
  
• Improving utilization of surveillance data for
  program planning, monitoring and evaluation

11
Data Collection and Processing

• Data collection duration
• Urban sites 8 weeks
• Rural sites 20 weeks
• Data collection timing synchronized
• Anonymous reporting with test result from site to
  regional health bureaus (RHB)
  
• RHB compiled all site findings and sent to
  Ministry of Health (MOH)
  
• At MOH data double entered on Epi-Info 2002

12
Estimating HIV Prevalence
13
Data Processing / Analysis

• Outputs from Epi Info
• Crude HIV prevalence
• By age
• By site
• By region
• By country
• The required data for estimation prevalence by
  site
  

14
Estimation Software

• Software used Epidemic Projection Package
  (EPP)
  
• Replaced the EpiModel used in previous estimates
• Developed by UNAIDS Reference Group, East-West
  Centre Research Program, Thai Red Cross, DFID
  
• Does curve fitting of HIV prevalence data points
  from multiple ANC sites

15
The Estimation Process Using EPP

• Create Workset
• Define the Epidemic national, regional,
  urban/rural
  
• Define respective populations
• HIV Data Entry Excel data forms
• Projections curve fitting process adjust for
  a median curve
  
• Results the curve or numeric results
• Percent HIV
• Number HIV
• Population size
• Save the output into a comma separated values
  (csv) file (MS Excel file with estimates)
  
• Demonstrate the Ethiopia 03 Weighted Workset

16
Demographic and HIV/AIDS Projections
17
Projection Software

• Spectrum developed by POLICY Project
• Demographic Projection Package (DemProj)
• Projects population by age and sex
• Assumptions base year population, fertility,
  migration, life expectancy
  
• AIDS Impact Model (AIM)
• Basis - Adult HIV prevalence (from EPP)
• Impact of AIDS into demographic projection

18
Demographic Projection Outputs

• Population sizes
• urban/rural, various display intervals
• Fertility
• Mortality
• Vital events
• Ratios
• Age groups
• Summary

19
AIM Outputs

• HIV/AIDS Epidemiology (reported by UNAIDS)
• pf adults living with HIV, by sex
• of children under 15 living with HIV
• of new HIV infections
• of new AIDS cases
• of AIDS deaths
• Impacts
• Deaths, economy, health care burden, TB, other
  childhood diseases, population size,
  
• Orphans
• AIDS orphans maternal, paternal, dual

20
The Ethiopian Example Preparations for Spectrum
Input

• Determination of the base population
• From the start of epidemic in 1982 till 2010
• Base pop 1995 (Census)
• Need for back projection to 1982 using the last
  census assumptions
  
• TFR, mortality rates, proportion urban,
  migration
  
• Comparison between census and Spectrum
  projections
  
• Development of Demographic tables for Spectrum
• National urban, rural
• Regional urban, rural
• Demonstrate the MS Excel Ethiopia Demo Tables

21
The Spectrum Process

• Demonstrate the Ethiopia03_edited.pjn on
  Spectrum

22
Spectrum Outputs Calculated

• Estimated adult HIV prevalence
• Total, Male, female, urban, rural,
• HIV-positive population
• Total, adult, children, male, female
• HIV positive pregnant women
• Annual HIV-pos births
• New HIV infections
• Adult HIV incidence ()
• New AIDS cases - Total, adult, children, male,
  female
  
• Annual AIDS deaths - Total, adult, children,
  male, female
  
• Annual population increase, all ages
• newly needing ART, all ages
• Orphans maternal, paternal, dual, AIDS,
  non-AIDS and total

23
Final Outputs

• AIDS in Ethiopia 5th Report
• Detailed Technical Report
• Eleven Regional Reports
• CD with all the data and results
• Dissemination of findings

24
Examples of EPP and Spectrum Outputs for Ethiopia
25
Estimated Adult HIV Prevalence and Trends
(1982-2008) in Ethiopia
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Practical Challenges Faced

• Definitions urban/rural sites
• Contamination
• No pure sites
• Able to adjust on EPP
• Quality assurance / quality assurance issues
  unable to differentiate
  
• Interpretation of output data by high level
  ranking authorities _at_ MOH

32
Thank you for your attention!!