2. Study design
• Explanatory controlled before and after design
with multiple post-test measurements
• Controlled before and after design with multiple
post-test measurements: collect information
quantitatively on selected indicators before and
after the intervention is launched in both
intervention & control areas
• Explanatory: qualitative element to explain
quantitative results
3. Explanatory controlled before & after
Both in intervention & in control areas
QUAN data QUAN data QUAN data QUAN data
collection collection collection collection
Qual data Qual data Qual data Qual data
collection collection collection collection
Year 0 Year 1 Year 2 Year 3
bHIP bHIP
4. Difference in difference models
Difference-in-Difference with parallel time-trend
4
3,5
3,5
VISTS PER CAPITA
3
2,5
# of hospital visits
2,5
2
2
1,5 1,5
1,5
1
1
0,5
0
t=0 t=1
TIME
5. Design option 1
Enroll in study all 16 Upazila 2013-2018
Year 0
6. Design option 1
Enroll in study all 16 Upazila 2013-2018
Year 0 Year 1
7. Design option 1
Enroll in study all 16 Upazila 2013-2018
Year 0 Year 1 Year 5
8. Design option 2
Enroll in study 4 2013-14 intervention Upazilas & additional 4 controls
Year 0 Year 1 Year 2 Year 3 Year 4 Year 5
9. Choosing the best design
16 Upazilas design: 8 Upazilas design:
• The more clusters, • Less powerful
the better
• Cannot address all
• Can address all
concerns central to
concerns central to
bHIP implementation bHIP implementation
• High-risk design if • Low-risk design even
something goes if implementation is
wrong scaled down
10. Theory of change
Economic
Quality &
development
appropriateness
of care
Providers’
motivation &
performance Preventive Health
care outcomes
utilization
(including
screening)
Loan
repayment
bHIP Curative care Entrepreneurial
utilization investments
Income Poverty
stabilization/ reduction
asset
protection
Household Child labor
OOP
spending
Investments
in education
11. Help needed
• Defining the design
• Selecting indicators to match TOC and relevant
tools for data collection