Indicators for Malaria Impact Evaluation

1
Indicators for Malaria Impact Evaluation

• Impact Evaluation Team

2

• Malaria Working Groups
• Biometrics Working Group
• Cognitive and Educational Working Group
• Socio-Economic Working Group ( KAP)
• Cost Effectiveness

3
Malaria Impact Evaluation Team
4
Developing a Common Approach to Measuring the
Biometric Impact of Malaria Control Interventions
Biometrics Working GroupMalaria Impact
EvaluationJoseph Keating (Tulane University)
Simon Brooker (LSHTM)
5
Malaria Impact Indicators I Parasitaemia and
Disease

• Populations -Children lt 5 years old
• -Pregnant women
• -Population in malarious areas
• Data source population based household survey,
  HMIS high versus low transmission seasons
  stable versus unstable transmission areas
• Diagnostic method finger-prick, thick and thin
  blood smear for microscopy (Gold Standard) or
  Rapid Diagnostic Test (RDT) kit
• Indicators -Prevalence of malaria parasite
  infection (lt 5 years old/all ages)
• -All cause mortality in children lt 5 years old
• -Laboratory confirmed malaria death rate (lt 5
  years old/all ages)
• -Malaria incidence
• Costs -RDT USD 1-3 plus cost of training
  personnel Microscopy (Gold Standard) varies
  as a function of existing equipment, reagents,
  and trained personnel

6
Malaria Impact Indicators II Anemia

• Populations Children 6-59 months
• Pregnant women
• Schoolchildren
• Data source population based household survey,
  clinic based survey, school survey
• Diagnostic method finger-prick blood sample,
  portable Hemocue machine
• Costs 0.5/sample
• Accuracy 0.1 g/L
• Anaemia definition age specific, e.g.
  110g/L (under 5s) 115-120 g/L (school-age
  children)
• Alternative methods Haemoglobin Colour
  Scale
• finger-prick blood sample, special
  chromatography paper (0.05/sample but
  accuracy only to 10 g/L
• therefore unsuitable for impact evaluation)

7
Developing a Common Approach for Cognitive and
Educational Assessments
Cognitive and Educational Working Group Malaria
Impact Evaluation Matthew Jukes (Harvard
University) Don Bundy (World Bank)
8
Impact of Early Childhood Malaria Prevention on
Global Cognitive Function
Jukes et al PLOS clinical trials 2006
9
Can IPT in schoolsreduce parasitaemia and
anaemia and improve school performance? A
randomised controlled trial of IPT using SPAQ in
30 primary schools in western Kenya

Malaria Infection in Semi-immune Schoolchildren
Most common
Less common
Clinical Attack
Asymptomatic Parasitaemia
IPT
Anaemia
Absent from School
Reduced Attention During Lessons
Educational Achievement
10
Clinical Attack
Impact of IPT on sustained attention and
education?
Anaemia
Absent from School
Reduced Attention During Lessons
Educational Achievement
Outcome n Mean difference 95 CI p-value Effect size
Counting sounds (max score20) 481 2.12 (-0.17, 4.42) 0.07 0.65
Code transmission (max score40) 469 7.74 (2.83, 10.65) 0.005 1.01

Exam score 6 286 0.55 (-2.26, 3.36) 0.35 0.15
Exam score 7 266 0.69 (-0.93, 2.15) 0.21 0.30
Clarke et al. forthcoming
11
Language Differences in Cognitive Tests
Performance
Jukes et al. forthcoming
12
(No Transcript)
13
Developing a Common Approach to Measuring the
Socio-Economic Impact of Malaria Control
Interventions
Socio-Economic Working Group Malaria Impact
EvaluationJed Friedman (World Bank) Edit V.
Velenyi (World Bank)
14
From Data .. To Impact
Monitor Change in Indicators and Forecast
(ME)
Organize Integrate Analyze (MIS)
Data
Impact
Information
Pathway for Evidence- based Planning
Package (MIS)
Implement the Plan (System)
Action
Evidence
Package Communicate to Planners
Stakeholders (MIS)
Knowledge
Influence the Plan (Planners)
Savigni and Binka (2004)
15
But what data for IE?

• The data we have are not the data we want.
• The data we want are not the data we need.
• The data we need are not available.
• How do we then measure impact?
• What impact do we measure?
• How precise is what we measure?

Quotes Savigni and Binka (2004)
16
Factors Influencing Malaria Burden
Underlying Health Status
Endemicity
Immunological Status
Observed Disease Burden
Socio-Economic Status
Social Organization
Cultural Roles
Cultural Beliefs
Jone and Williams (2004)
17
Health Links to GDP
Macro Economic Impact Poor health reduces GDP
per capita by reducing both labor productivity
and the relative size of the labor force.
Higher Fertility andChild Mortality
Labor Force Reduced byEarly Mortality
Higher Dependency Ratio
Lower GDP per Capita
Child Illness
Adult Illness Malnutrition
Child Malnutrition
Reduced Labor Productivity
Reduced Access to Resources Economy
Reduced Schooling Impaired Cognitive Capacity
Reduced Investment in Physical Capital
Berman, Alilio, and Mills (2004)
18
Data Sources
Types and Levels of Data for Health Information
Systems Important for Malaria Control Programs
Type Type
Level Cross-Sectional Retrospective Longitudinal Prospective
Individual and HH Population Survey (Census, DHS, MICS) Prospective Surveillance (Vital events and DSS)
Health Facility Routine Reporting (HMIS, IDS, DHS) HF Survey
Modeling Risk Mapping (GIS) Remote Sensing and Early Warning Systems
DHS Demographic and health Survey, MICS Multi-Indicators and Cluster Survey, DSS Demographic Surveillance System, HMIS Health Management Information System, IDS Integrated Disease Surveillance, HF Health Facility, GIS Geographic Information System DHS Demographic and health Survey, MICS Multi-Indicators and Cluster Survey, DSS Demographic Surveillance System, HMIS Health Management Information System, IDS Integrated Disease Surveillance, HF Health Facility, GIS Geographic Information System DHS Demographic and health Survey, MICS Multi-Indicators and Cluster Survey, DSS Demographic Surveillance System, HMIS Health Management Information System, IDS Integrated Disease Surveillance, HF Health Facility, GIS Geographic Information System
Savigni and Binka (2004)
19
Conceptual FrameworkEconomic Burden of Illness
for HHs
Individual Household
Health System
Box 3a Direct Costs
Social Resources
Box 3b Indirect Costs
Box 2Treatment Behavior
Box 4 Coping Strategies(Risky, less risky)
Box 6 Access,fees, quality ofcare, insurance
Box 7 SocialNetworks
Box 5 Impact on Livelihood(Assets, income,
food security)
Box 1Reported Illness
Russell (2004)
20
Weak / Missing Link Biomedical Socio-Economic

• Asset v. Consumption Module
• Health Care Seeking Expenditures
• Copying Mechanisms Poverty
• Labor Market / School Participation
• KAP
• Community Effects
• Social Norms (Gender, Vulnerability)

21
Weak / Missing Link (2)Biomedical
Socio-Economic

• Some Operational / Technical Issues
• Are the questions tailored to capture the
  intervention? Is our approach parsimonious?
• Should the sample be expanded?
• What is our knowledge gain, and the marginal
  cost of the information?
• Are we gaining predictive power and making a
  good Biomedical-SE link?