OBJECTIVE: To determine the conditions under which the Cold Weather Plan (CWP) for England is likely to prove cost-effective in order to inform the development of the CWP in the short term before direct data on costs and benefits can be collected.

STUDY DESIGN: Mathematical modelling study undertaken in the absence of direct epidemiological evidence on the effect of the CWP in reducing cold-related mortality and morbidity, and limited data or on its costs.

METHODS: The model comprised a simulated temperature time series based on historical data; epidemiologically-derived relationships between temperature, and mortality and morbidity; and information on baseline unit costs of contacts with health care and community care services. Cost-effectiveness was assessed assuming varying levels of protection against cold-related burdens, coverage of the vulnerable population and willingness-to-pay criteria.

RESULTS: Simulations showed that the CWP is likely to be cost effective under some scenarios at the high end of the willingness to pay threshold used by National Institute for Health and Care Excellence (NICE) in England, but these results are sensitive to assumptions about the extent of implementation of the CWP at local level, and its assumed effectiveness when implemented. The incremental cost-effectiveness ratio varied from £29,754 to £75,875 per Quality Adjusted Life Year (QALY) gained. Conventional cost-effectiveness (<£30,000/QALY) was reached only when effective targeting of at-risk groups was assumed (i.e. need for low coverage (∼5%) of the population for targeted actions) and relatively high assumed effectiveness (>15%) in avoiding deaths and hospital admissions.

CONCLUSIONS: Although the modelling relied on a large number of assumptions, this type of modelling is useful for understanding whether, and in what circumstances, untested plans are likely to be cost-effective before they are implemented and in the early period of implementation before direct data on cost-effectiveness have accrued. Steps can then be taken to optimize the relevant parameters as far as practicable during the early implementation period.