Appraising an Economic Evaluation

A structured check of whether a study comparing the costs and outcomes of interventions is sound and decision-useful. Confirm: a well-defined question and comparator; the appropriate evaluation type (cost-effectiveness / utility / benefit); that all relevant costs and outcomes were identified, measured and valued; an appropriate time horizon with discounting of future costs and effects; an incremental analysis yielding the ICER; a sensitivity analysis for uncertainty; and a clearly stated perspective (e.g. health-system vs societal). Anchor it with CHEERS 2022 reporting or Drummond’s checklist.

The route from a framed question to a usable answer. Boxes 1–3 establish what is being compared and whether all consequences (costs and outcomes) were captured, valued, and discounted over a long-enough horizon. Box 4 is the incremental cost-effectiveness ratio (ICER) — the extra cost per extra unit of effect (commonly per QALY). Box 5 stress-tests that figure and pins down whose costs count. CHEERS 2022 (28 items) governs reporting; Drummond’s 10-question checklist governs the critique.

Trust the conclusion only if each gate holds. A weak comparator flatters the new intervention. The wrong type answers the wrong question (cost-utility uses QALYs; cost-benefit values outcomes in money). A short time horizon or absent discounting distorts long-run value. The ICER is meaningless without an explicit willingness-to-pay threshold (in the UK, NICE’s roughly £20,000–30,000 per QALY) and a sensitivity analysis showing whether the verdict survives plausible changes in assumptions.

Economic evaluation decides whether a clinically effective intervention is actually affordable — the question behind every funding decision and NICE appraisal. A single ICER quoted without its perspective, horizon or uncertainty is the commonest way these studies mislead, and spotting that omission is precisely what the exam rewards.

• ICER = Δcost ÷ Δeffect (cost per extra QALY, etc.)
• Always state: perspective · time horizon · discounting
• ICER needs a sensitivity analysis + a willingness-to-pay threshold
• Tools: CHEERS 2022 (reporting) · Drummond's checklist (critique)

A cost-utility analysis in an appraisal station — a paper reports that a new ED point-of-care test costs an extra £18,000 per QALY gained versus usual care. Work the flow: was the comparator genuinely usual care (box 1); is cost-utility the right type for a quality-of-life outcome (box 2); were all downstream costs and outcomes captured and discounted (box 3); is £18,000/QALY the ICER (box 4); and — the marks — was a sensitivity analysis done and which perspective was taken (box 5)? An ICER just below the NICE threshold means little if a one-way sensitivity analysis pushes it well above. ⚠ An ICER without a sensitivity analysis or a stated threshold is not interpretable — say so.

• Omitted costs or outcomes — downstream or indirect consequences left out.
• The wrong perspective — health-system figures presented as if societal (or vice versa).
• No sensitivity analysis — an ICER accepted without testing its uncertainty.