Intended for healthcare professionals


Economic evaluation of health interventions

BMJ 2008; 337 doi: (Published 29 September 2008) Cite this as: BMJ 2008;337:a1204
  1. Michael Drummond, professor of health economics1,
  2. Helen Weatherly, research fellow1,
  3. Brian Ferguson, director2
  1. 1Centre for Health Economics, University of York, York YO10 5DD
  2. 2Yorkshire and Humber Public Health Observatory, ARRC, University of York, York YO10 5DD
  1. md18{at}

    A broader perspective is needed that includes potential costs and benefits for all stakeholders

    The choice of perspective is important in the economic evaluation of healthcare interventions. For example, in the National Institute for Health and Clinical Excellence’s (NICE) technology appraisal of drugs for Alzheimer’s disease, a major discussion point was whether the costs falling on caregivers should be included as well as costs to the NHS.1

    The main argument for adopting a restrictive perspective is that the budget for the NHS is meant to be for improving health. Therefore, the relevant consideration in evaluating interventions is the opportunity cost (in other treatments forgone) on the healthcare budget. But alternatively shouldn’t the full social benefits of healthcare interventions be considered? If healthcare interventions have benefits outside the healthcare sector—for example in the criminal justice system, transport sector, or education—shouldn’t these be tracked and any budgetary adjustments sorted out separately? And shouldn’t health care aim to provide benefits to families and carers as well as the patient?

    Textbooks say that all costs and benefits of interventions should be considered, no matter on whom they fall.2 However, most evaluations have a narrower perspective and focus on the relevant costs for the agency commissioning the study. Currently, NICE gives mixed messages on the subject of perspective. Technology appraisals normally consider only costs falling on the NHS and personal social services budgets in the primary analysis.3 Public health appraisals, however—once the responsibility of the Health Development Agency, which operated under different statutes on public health—can also consider effects on other government budgets.4 Therefore, an evaluation of a public health intervention to reduce substance abuse would consider the potential advantage gained from a reduction in criminal justice costs, but an evaluation of a drug to treat heroin addiction would not.

    Intersectoral costs were one of the four methodological challenges we considered in a recent review of economic evaluations of public health interventions.5 The review confirmed that the existing literature on economic evaluation generally adopts a restricted perspective, but it also showed that a broader consideration of costs and benefits was informative.

    For example, Byford and colleagues assessed the cost effectiveness of manual assisted cognitive behavioural therapy compared with treatment as usual.6 The cost analysis included the cost to the NHS and that of spill over effects into other sectors of the economy, such as social services, voluntary services, accommodation and living expenses, criminal justice services, and lost productivity from time off work because of illness. They found that behavioural therapy would cost more per patient than treatment as usual if a community health service perspective was taken, but that it would be cost saving if a criminal justice perspective was taken. This shows how the perspective taken can influence the results. For the main analysis the authors took a broad economic perspective, which included the full cost of all service providing sectors, accommodation and living expenses, and productivity losses resulting from time off work as a result of illness. Manual assisted cognitive behavioural therapy was again found to be cost saving.

    Broadening the perspective of economic evaluations is timely and consistent with recent policy developments in England. The Commissioning Framework for Health and Well-being signalled the need for primary care trusts and local authorities to undertake joint strategic needs assessments.7 These will form the basis of a new duty for primary care trusts and local authorities to cooperate with one another.

    For example, a primary care trust and local authority may decide that childhood obesity is a major priority area for joint strategic needs assessments. To facilitate this, NHS budgets to deal with overweight and clinically obese children would need to be identified, and once children are in the healthcare system optimum care pathways would need to be agreed on. However, the existence of such children is an indication that primary prevention has largely failed; a key component might also be to ensure that children do not need to enter the healthcare system with problems related to obesity. This might be achieved by a range of cost effective public health measures across statutory and non-statutory sectors. Input might come from beyond the NHS; for example, from children, their parents, schools, communities, and local authorities. Many questions could be explored. Is advice on diet and nutrition effective and well understood? Are there safe play areas to encourage physical activity? Do nurseries and other childcare facilities provide regular opportunities for active play? Do local transport plans promote safe cycling and walking routes?

    In the past, refusal to adopt a broader perspective has sometimes been justified because of data limitations, measurement difficulties, or limits in budgetary responsibilities. However, it is now time to overcome these practical difficulties and to think more broadly about the costs and benefits of healthcare and public health interventions.


    Cite this as: BMJ 2008;337:a1204.


    • Competing interests: MD and HW have undertaken technology appraisals for the National Institute for Health and Clinical Excellence (NICE). MD and BF are members of advisory committees at NICE.

    • Provenance and peer review: Not commissioned; externally peer reviewed.


    Log in

    Log in through your institution


    * For online subscription