Home · Level 1 · Session 3
Session 3
The Framework
🧭 Level 1 of 5 · Session 3 of 5

Level 1 · Foundations

The Framework

Framing the Economic Question for AI

Session 2 told you which tools to reach for. This session tells you how to phrase the question so those tools return health economics evidence rather than clinical trial summaries.

Why this session exists

When health economists and financing specialists turn to AI tools for literature support, they often get back clinical trial summaries, drug efficacy tables, or biomedical findings — none of which is what they came for. The reason is almost always the same: the question was framed in clinical language, so the AI searched clinical space.

The framework you use to pose your research question determines what the AI finds. This session introduces a question structure designed specifically for health economics and financing work, and shows you how to translate it into AI prompts that serve WHO India's mandate.

From PICO to PECO-F

The standard clinical framework is PICO: Population, Intervention, Comparator, Outcome. It works well for asking whether Drug A beats Drug B in a randomised trial. It does not work well for asking whether community health insurance schemes reduce catastrophic health expenditure among informal workers — which is closer to the kind of question this team actually works on.

For health financing and priority-setting, the more useful frame is PECO-F:

ElementWhat it meansIndia / LMIC example
P — Population Who bears the burden? Informal sector households in Uttar Pradesh
E — Exposure / Intervention What financing mechanism or policy? Contributory health insurance vs. tax-funded coverage
C — Comparator What is the alternative? No coverage / out-of-pocket only
O — Outcome What do we measure? Catastrophic health expenditure, impoverishment, utilisation rates
F — Feasibility & Equity What constraints apply? Affordability, gender equity, willingness to pay, political acceptability

The F element is not cosmetic. The acceptability and equity dimensions of an intervention can overturn a purely cost-effectiveness ranking — a mass-media health intervention may score poorly on cost-per-DALY but rank highly on acceptability with local stakeholders and on distributional grounds. AI tools will not surface that tension unless you ask for it explicitly.

Translating PECO-F into an AI prompt

Poorly framed

❌ Too broad — will return generic overviews and decade-old WHO summaries "What is the best way to reduce out-of-pocket health spending in India?"

Well framed using PECO-F

✓ Scoped — signals population, mechanism, comparator, outcome, and equity dimension "I am looking for evidence on whether community-based health insurance schemes (E) reduce catastrophic health expenditure (O) among informal sector workers (P) in low- and middle-income countries, compared to out-of-pocket financing with no insurance (C), with attention to equity outcomes and willingness-to-pay thresholds (F). Please prioritise systematic reviews and economic evaluations published after 2015."

The second prompt also signals the methodological register — systematic reviews, economic evaluations — so the tool does not return qualitative ethnographies or clinical guidelines.

The priority-setting variant

When the question is about allocating scarce resources across competing interventions, the frame shifts slightly. The question is no longer just "does this work?" but "should this be funded, given competing claims on the same budget?" For these questions, add two further dimensions:

✓ Priority-setting prompt example "What frameworks have been used to compare the cost-effectiveness and equity impact of expanding community health insurance versus increasing public facility capacity for primary care in LMICs? I need approaches that go beyond cost-per-DALY to include equity weighting and stakeholder acceptability — relevant to a national UHC reform context in a federal system."

That prompt navigates the ACE (Assessing Cost-Effectiveness) and A4R (Accountability for Reasonableness) literatures simultaneously, which is exactly where India UHC reform work sits.

Live Demo — AI Tools in Action
Watch PECO-F framing applied to a real TB screening HTA across Sessions 02 & 03 of the hands-on app

The embedded app below covers the practical workflow for Sessions 02 and 03 — specifically how PECO-F framing shapes what you search for and how you extract from what you find. It uses a real HTA context: CAD4TB chest X-ray AI evaluation across 18 PHC sites in India. Use the Live Demo button inside the app for a guided 10-step walkthrough, or navigate directly to Session 02 or 03.

Loading app…

Try it now

🧪 Live exercise — reframe your question

Type a research question your team is currently working on. The tool below will reframe it using PECO-F and flag the equity and feasibility dimensions you should add to your AI prompt.

PECO-F reframe

🎯 Key takeaway

The framework is not bureaucracy — it is precision. A structured question reduces the chance that AI returns the wrong kind of evidence. For WHO staff working on UHC, financing, and priority-setting, that means explicitly naming populations, mechanisms, comparators, outcomes, and equity constraints every time. Session 4 shows you how to encode this into a reusable search protocol — so the structure becomes automatic rather than effortful.