Britain has around 40,000 more deaths in winter than expected from death rates in other months of the year. Looking at England, this study investigated whether this large winter excess of deaths could in part be attributed to poorly insulated and difficult-to-heat housing. The study, by researchers from the London School of Hygiene & Tropical Medicine, found:

• There was a 23 per cent excess of deaths from heart attacks and strokes during winter (December to March) compared with non-winter months. Much of this can be attributed to cold. The winter rise in deaths was greatest in older people, but there was some rise at all ages. The risk of excess winter death showed little variation with socio-economic group.
• The magnitude of the winter excess was greater in people living in dwellings that appear to be poorly heated. The percentage rise in deaths in winter was greater in those dwellings with low energy-efficiency ratings, and those predicted to have low indoor temperatures during cold periods. There was also a gradient of risk with age of the property, the risk being greatest in dwellings built before 1850, and lowest in the more energy-efficient dwellings built after 1980. Absence of central heating and dissatisfaction with the heating system also showed some association with increased risk of excess winter death.
• The indoor temperature in many dwellings appears to fall below 16º Celsius during cold periods. Low indoor temperatures were found to be more likely if the dwelling was old, had no or inadequate central heating, was costly to heat, or was occupied by a household with low income. People in local authority or housing association dwellings appeared to be especially likely to have low indoor temperatures during cold periods if their heating costs were high.
• Although there is some uncertainty about the risk estimates, there was evidence that cold-related mortality was greater in homes predicted to have low indoor temperatures during cold spells.
• The findings suggest that people in poorly heated homes are indeed more vulnerable to winter death than those living in well-heated homes. This suggests that substantial public health benefits can be expected from measures that improve the thermal efficiency of dwellings and the affordability of heating them.

Britain has a large winter excess of deaths - some 40,000 more deaths during December to March than expected from death rates in other months of the year. Although influenza, respiratory infections and other seasonal factors may account for part of the winter excess in mortality, around 60 per cent of it can be attributed to the effects of cold. This winter excess is greater than in most other countries of continental Europe and Scandinavia, despite the fact that Britain has comparatively mild winters. A part explanation for this may lie in the quality of our housing stock, which is less thermally efficient than that in most other north European countries and hence may afford less protection against the cold.

This study investigated the relationship between housing and cold-related death. It was based on linking postcoded mortality statistics for the years 1986 to 1996 to housing data from a large national survey carried out in 1991. By linking these two data sets on postcode, it was possible to analyse the seasonal pattern of mortality in relation to the characteristics of housing at the same locations.

The research focused on how dwelling characteristics were related to the size of the winter excess of deaths, to indoor temperatures and to specific cold-related mortality.

People living in older properties (which are more difficult to heat) had a greater excess of winter deaths than those living in recently built properties: the lowest risk was seen in people living in dwellings built since 1980, and the highest risk in those living in dwellings built before 1850. There was also an association between excess winter mortality and low indoor temperatures. Homes predicted to be in the lowest quarter of indoor temperatures ('cold homes') had around 20 per cent greater risk of excess winter death than homes in the top quarter of indoor temperatures ('warm homes').

• Age of property (older homes were colder);
• Absence of/dissatisfaction with the heating system;
• Cost of heating the dwelling;
• Low household income;
• Household size.

Table 1 shows the mean measured temperatures for these household and housing characteristics, together with an estimate of the percentage of households with a hall temperature below 16ºC when the outside temperature falls to 5ºC.

Housing tenure, being on state benefits, and having a poor thermal efficiency rating were not strongly associated with indoor temperature after these other factors were taken into account. The disadvantage of having a difficult-to-heat home appeared to be greater in households with low income.

The one finding that appears counter-intuitive is the absence of a clear socio-economic gradient in risk of excess winter death. Cardio-vascular and all-cause mortality are known to have a strong association with poverty, but this appears not to be shown by excess winter death. However, a similar proportional rise in winter mortality means that poorer social groups must have a greater absolute rise in winter mortality because of their higher underlying death rate. Moreover, the analyses of indoor temperatures show that, as a whole, residents in the social sector (local authority and housing association dwellings) in fact maintain quite good indoor temperatures, in part because much of the housing stock in this sector is relatively new and easy to heat. However, where a dwelling is expensive to heat and occupied by a household with low income, average internal temperatures are low and the risks of excess winter death high.

In summary, the findings suggest that indoor temperature and factors associated with poor thermal efficiency of dwellings, including property age, are associated with increased vulnerability to winter death from diseases of the heart and circulation. Hence, it is likely that substantial health benefits could be achieved by measures aimed at improving the thermal efficiency of homes and the affordability of heating them.

Evaluations now beginning of the Government's new Home Energy Efficiency Scheme should provide evidence on this. But wider debate is also needed to consider the forms of housing, energy and social policy that are likely to reduce winter deaths. In particular, consideration needs to be given to the types of energy-efficiency programme and the methods of population targeting that will provide the greatest benefits to public health.

• An analysis of variation in the magnitude of excess winter mortality and whether such variation is related to dwelling and household characteristics;
• Tabulation and regression analysis of the factors which determine indoor temperature;
• A time-series analysis of mortality in relation to outdoor temperature - the focus being to examine whether housing and other factors influence cold-related mortality. This analysis was based on a modification of methods more commonly used to investigate the relationship between outdoor air pollution and health.