Fossil fuel energy use in heating and cooling buildings is considered to be a major contributor to observed climate change effects, so there is an environmental imperative to reduce energy use in buildings. We should also improve buildings' energy efficiency on health grounds. Climate change is predicted to produce more frequent extreme weather events, while epidemiological evidence indicates relationships between ambient temperature and ill health. This points to the need for addressing both climate change itself and the way buildings mediate outdoor conditions, for the sake of vulnerable occupants. The UK government requires evidence in support of policy-making concerning energy use in buildings and consequences for health. This paper reviews epidemiological research to illustrate problems associated with measuring the direct health impact of indoor temperatures, for which evidence remains limited. Conventionally, temperature-related health effects are discussed in terms of seasonal excess deaths. The paper goes on to describe a population-based study in London that considers morbidity rather than mortality. A new methodology is developed that links the risk of cold homes with excess winter hospital episodes, demonstrating its potential for identifying small areas for priority action on improving domestic energy efficiency in terms of health as well as the environment. (C) 2007 Elsevier B.V. All rights reserved.