A Systematic Approach to Selecting Methods for SIM

1
A Systematic Approach to Selecting Methods for
SI/ME

• Philip Setel, PhDDeputy DirectorMEASURE
  Evaluation

2
Getting to Basics

• This presentation lays out a simple (but
  hopefully not simplistic!) strategy to help in
  selecting the right tools to answer ME or SI
  questions
• Should be viewed as integral to the Public
  Health Questions framework introduced in this
  training.

3
(No Transcript)
4
Getting to Basics

• ME/SI is a quickly evolving field with a rapidly
  expanding alphabet soup of tools and toolkits
• DHS, AIS, H/SPA, GIS, PLACE, SAVVY, BSS, HIS, and
  on
• But these are methods
• In your role, how do you advise country partners
  on the selection from this menu of options? What
  is the right tool?
• Whats right for Reporting?
• Whats right for Program Learning or Operations
  Research?
• We may jump straight to deciding on a method
• E.g. We need a DHS we need a qualitative
  study

5
Step 1 What is the question?

• Formulate the particular SI or ME question that
  needs answering in as simple, precise, and clear
  terms as possible operationalize it.
• Important Public Health Questions about risk
• What is the problem contributing factors?
• Operationalized question about risk
• Where are people engaging in behaviors and
  practices known to be associated with risk of HIV
  transmission?
• Important Public Health Questions about ARV
• Are we doing this intervention right, as planned
  and making a difference?
• Operationalized questions about services
• How many people are adhering to the ARV treatment
  being made available?
• Why are some people discontinuing treatment?

6
Step 2. Next, think about the data, not the
method!

• What are the data required to answer the question
  posed or to generate the indicator required?
• Where are people engaging in behaviors and
  practices known to be associated with risk of HIV
  transmission? eed to know something about
• Geographical locations
• Characteristics of the clientele, patrons,
  visitors, residents of those places
• What they are doing there that equates to risk of
  HIV transmission or infection, something about
  the density of these behaviors.

7
Step 3. Then choose the method that generates the
data needed to answer the question.

• The NAC needs to know about locations of risk
• Need data about places and behavior
• Qualitative data describing where behavior and
  place intersect
• Probably GIS to locate those areas relative to
  other spaces places
• Which method PLACE?
• The Department of Curative Services in the MOH
  needs to know why people stop using ARV and how
  this can be addressed?
• Need data that uncovers and documents processes
  of decision-making about care.
• Probably a qualitative study with in-depth
  interviews
• Perhaps some observations of care delivered at
  clinics
• Special study protocol or purpose-designed
  protocol?
• Need to report on national coverage of PMTCT
• Need representative data on pregnancies,
  prevalence of HIV in pregnant women, proportion
  of pregnant women being tested, delivering at
  health facilities, etc.
• Survey with several modules and careful sampling
  procedures
• Possibly some modeled estimates
• Modules in HPSA or some other facility survey

8
Step 4. Interpret with Care using Appropriate
Methods

• Make sure you feel comfortable with
• What the sample does and doesnt represent
• Inputs into the calculation of specific
  indicators
• How the data relate both to the conclusions drawn
  and the original question that needed answering
• Likely sources of bias in the data or things that
  are probably unmeasurable
• If possible
• Understand something basic about the theoretical
  frameworks that have generated the methods youve
  employed.

9
Summing up
START 1. Question formation operationalization
2. Decide on data required to answer question
3. Select method to collect data required
4. Use analytical methods appropriate to data
5. Interpret results with care to answer original
questions develop policy implications
Illuminate new issues
10
One more thing

• In the 21st century, answering the question,
  producing indicators and using information is NOT
  enough to get a 1st class honors in ME.
• MUST take account of and contribute to
• Long-term investments in systems and institutions
  that will make a permanent contribution to
  national health development
• Existing systems, local priorities, other donor
  initiatives that have govt support must factor
  into your information strategies as vastly more
  resource for AIDS (and other health priorities!)
  comes on stream.
• In Tanzania the cynics say UKIMWI ni biashara
• AIDS is a business
• Be aware that our decisions are watched and
  evaluated by and accountable to not just our
  professional peers and the organizations we work
  in, but also the populations we serve.