By the end of this session

1. State a decision problem precisely enough for someone else to model it
2. Choose a comparator, and say why a bad one inflates every ICER
3. Explain what changes when you change perspective
4. Choose a time horizon and defend it
5. Identify, measure and value costs, and know which are left out
6. Name where Indian cost data actually comes from

Our spine for the rest of the day

The Indian Reference Case: twelve recommendations on eleven principles, written for India by the group behind most of the papers we have cited.

This is the checklist your examiner will use.

Sharma, Prinja et al., Lancet Reg Health Southeast Asia, 2023.

Why it had to be written

A systematic review of Indian evaluations found analysts making different choices on the same questions:

Differences in method produce differences in ICER that have nothing to do with the interventions.

Reported in Sharma et al., 2023, from Prinja et al., Applied Health Economics and Health Policy, 2015.

1 The decision problem

Most papers say "to assess the cost-effectiveness of X". That is a title, not a decision problem.

The Reference Case asks you to answer six questions:

Plus the policy question, the condition, the setting, the target population, and any subgroup whose costs or effects will differ. Identify those subgroups now, not after the results.

2 The comparator

"Any alternative will look good if it is compared to something that is sufficiently bad."

Indian Reference Case, on why comparator choice matters

The recommendation: compare against current practice in use. What the standard treatment guidelines say, or what is most widely used. Not the weakest option, and not "do nothing" unless doing nothing is genuinely what happens now.

Where a health benefit package is involved, the comparator is whatever the new intervention would replace inside it. Comparators need not be alike: a drug can be compared with surgery if they serve the same purpose.

3 Perspective: whose costs count?

India's default is abridged societal, with the payer view reported alongside.

Why India insists on both

There is no single payer. Out-of-pocket spending is a large share of Indian health expenditure, and higher in the private sector, which treats most patients.

An analysis that ignores what families pay can call an intervention affordable while it is bankrupting the households using it.

Productivity losses stay out of the base case for three reasons: the costing methods disagree, Indian data barely exist, and including them favours diseases of the working and better-off over those of the poor, the young and the old.

That third reason is an equity argument inside a methods decision.

4 Where the effectiveness estimate comes from

State the level of evidence you used. If you departed from the hierarchy, say why.

5 Costing, in three steps

Getting step 1 wrong cannot be repaired by doing steps 2 and 3 beautifully.

Micro-costing is accurate and slow. Gross costing is fast and blunt.

Capital costs, and the usual mistake

A cold-chain refrigerator lasts ten years. The programme you are evaluating uses one year of it.

Charge the programme an annuitised share, not the purchase price.

Spread the capital cost over the item's useful life at the discount rate, so a machine bought this year and one bought five years ago are treated alike. Buildings, vehicles, equipment and one-off training all belong here.

Omitting capital costs is the commonest costing error of all, and it always makes the intervention look cheaper.

Your turn: which costs count?

Ten line items from a real programme costing. For each: is it direct or indirect, and is it counted under the payer view, the abridged societal view, or neither?

On your phone: sph-aiimsbhopal.github.io/intro-economic-evaluation-health

Where Indian cost data comes from

Check the database before you plan primary data collection.

healtheconomics.pgisph.in/costing_web/  ·  see also Prinja et al., PharmacoEconomics Open, 2020.

7 Time horizon

Long enough for every cost and effect that differs.

Too short a horizon decides the result in advance.

Vaccinate a girl at 14 and the whole cost falls that year, while the cancers prevented would have come after 30. A one-year horizon counts every rupee and no benefit, so the programme looks like waste. The horizon produced that answer, not the vaccine.

The same framing, on a drug

Row 4 is two answers to one question. Session 5 shows what that did.

Gupta, Verma, Gupta, Prinja, JCO Global Oncology, 2020.

Recap

• The decision problem is six questions, not a title
• The comparator is current practice. A weak one flatters everything downstream
• India's default is abridged societal, payer view alongside
• Costs: identify, measure, value. Annuitise capital
• Check the National Health System Cost Database first
• The horizon must be long enough for benefits to arrive

All of that is the numerator, on paper. Next: the numerator, on your own ward.

Sources

1. Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th ed. Oxford University Press; 2015. Chapter 4, pp. 77–123.
2. Sharma D, Prinja S, Aggarwal AK, Rajsekar K, Bahuguna P. Development of the Indian Reference Case for undertaking economic evaluation for health technology assessment. Lancet Reg Health Southeast Asia 2023;16:100241. PMID 37694178
3. Prinja S, Chauhan AS, Angell B, et al. A systematic review of the state of economic evaluation for health care in India. Appl Health Econ Health Policy 2015;13(6):595–613. PMID 26449485
4. Prinja S, Chauhan AS, Bahuguna P, et al. Cost of delivering secondary healthcare through the public sector in India. PharmacoEcon Open 2020;4(2):249–61. PMID 31468323
5. Prinja S, Gupta A, Verma R, Bahuguna P, et al. Cost of delivering health care services in public sector primary and community health centres in North India. PLoS One 2016;11(8):e0160986. PMID 27536781
6. National Health System Cost Database for India. PGIMER Chandigarh and Department of Health Research. healtheconomics.pgisph.in/costing_web/