SECTION 5 CONCLUSIONS
The evaluation examined the impact of the CHP on a wide range of outcome measures. Some of these - for example, those relating to specific symptoms and health conditions, and use of health services - were direct indicators of human health, while others ( e.g. those outcomes which assessed social contacts, or financial strain) bore indirect relationships with health. It is clear from the results presented in Section 4 that the set of outcomes which can reasonably be viewed as direct expressions of health provided only limited evidence of a statistically significant effect of the CHP. Of 17 measures representing specific symptoms and health conditions (Section 4.9), only two (namely, reduced likelihood of first diagnoses with heart disease and with high blood pressure) indicated a significant positive effect of the CHP (that is, an effect indicative of improved health), while a third (namely, increased likelihood of first diagnosis of nasal allergy) showed a significant association with the Programme which cannot unambiguously be characterised as positive. The two measures which exhibited significant links with a positive effect (namely, reduced likelihood of first diagnoses with heart disease and with high blood pressure) are potentially important, and are moreover to some degree consistent with current knowledge of the physiological effects of cold on human cardiovascular functioning  .
However, the finding that provision of heating under the CHP appears to be associated with a reduced likelihood of being diagnosed with heart disease and / or high blood pressure must be treated with caution, for the following reasons. First, the evaluation only examined respondents' experiences across a relatively short time period (two years), and without further research it is impossible to say whether the observed reductions in the incidence of heart disease / high blood pressure would be sustained in the longer term. Second, it is paradoxical that the apparently reduced experience of these two classes of condition among heating recipients was not matched by corresponding reductions in the use of health services, which is what might reasonably have been expected. Rather, utilisation of health services exhibited no significant differences between recipients and comparison respondents. Third, the outcome measures representing diagnosis with heart disease and high blood pressure were based on the respondent's self-report, rather than being drawn from objective clinical sources such as GP or hospital records. This, inevitably, means that a degree of uncertainty or imprecision attaches to the outcomes. For example, one respondent might report a recent diagnosis of angina as 'heart disease', while a second might not. These three factors - the limited time period examined, the apparent lack of effect on respondents' use of health services, and the self-reported nature of the outcomes - suggest that the apparently positive effect of the CHP on cardiovascular health, while interesting and potentially important, must be viewed with some reservations. Broadly speaking, it would be prudent to regard the direct impact of the CHP on health - based on these two findings - as limited.
Additional evidence that the direct influence of the CHP on health is limited comes from two further considerations. First, one outcome measure - the likelihood of a first diagnosis with nasal allergy such as hayfever - exhibited a significant association with the CHP which cannot be characterised as positive: recipients were found to be more likely to report such a diagnosis than comparison respondents. This 'adverse' finding goes some way towards negating the positive impact implied by the heart disease / high blood pressure results, leading to a more neutral view of the overall health impact of the Programme. Second, it must be borne in mind that the remaining 14 symptom- and health-related outcomes in this area (most of which related to respiratory health) failed to provide evidence of any significant influence of the CHP.
The findings relating to self-reported health-related quality of life (Section 4.10) present a similar picture, in that they suggest a limited positive effect of the Programme. Of nine measures drawn from the SF-36, two indicated a 'positive' effect of the CHP (that is, recipients experiencing more favourable outcomes than comparison group respondents). However, it is evident that these positive effects, while statistically significant, are modest in size, and their practical or clinical significance is unlikely to be great.
A further area directly reflecting human health is that of long-standing illness or disability (Section 4.11). The single outcome measure featured in this area exhibited no significant effect of the CHP, strengthening the impression that the impact of the Programme on health over the period examined was limited. Further support for this conclusion is provided by the findings relating to use of medications (Section 4.12). While this area is arguably not a direct expression of health, it would reasonably be expected that any appreciable impact of the CHP on recipients' health status would be reflected in altered levels of behaviour associated with the maintenance of health, including use of medications. The findings of the evaluation suggest that no such change was observed during the period examined.
A limited number of subgroup analyses was performed in order to assess whether the CHP operated differentially on specific subgroups of recipients. Differential effects were investigated for three key outcomes: respiratory health problem, general health and environmental problem in main rooms in the respondent's dwelling. In investigating differential effects, attention focussed on the possible influence of housing tenure, physical house type ( e.g. 'semi-detached'), gender and age. None of the results obtained indicates that the effect of the Programme varied substantially across subgroups for the range of outcomes examined.
The findings in relation to the impact on physical health are broadly consistent with previous findings. Earlier studies have suggested that energy efficiency measures may be associated with improvements in respiratory symptoms (not found in the current study). However a systematic review carried out in 2001  (suggested that only one of four studies had adjusted for potential confounding variables and that high rates of attrition (sample loss) in these studies may limit the generalisability of these findings. More recent research points to limited health impacts in the short term, with an indication that improvements in mental health, and perhaps physical functioning ( e.g., mobility) are also plausible health outcomes of energy efficiency measures.
For example, an unpublished evaluation of a fuel poverty programme in the Armagh and Dungannon Health Action Zone in Northern Ireland  found that the installation of energy efficiency measures, including central heating systems, were associated with significant reductions in the numbers of householders reporting arthritis/rheumatism and the numbers reporting the mean number of total illnesses suffered per head in each household compared to the non-intervention group. This finding is consistent with the current study, where significant increases in scores on the physical functioning scale (which assesses fitness and mobility) were recorded. The Northern Ireland study also found no impact on symptoms of angina, but improvements in stress/mental health (compared to the no-intervention group). The number of rooms which had problems with cold, mould and damp also decreased; and though average temperatures reduced little, there was a reduction in the range of temperatures of the monitored rooms, indicating improvements in levels of heating control.
The findings of the Lambeth study which evaluated the impacts of central heating installation in older people  are somewhat similar; with only limited improvements in health in terms of SF-36 scores, which the authors put down to the insensitivity of the SF-36, the small sample size and the short length of follow-up (<1 year). Improvements in controllability and reductions in swings in temperature were again observed. However (contrary to the current study), half of those with central heating felt that their gas bill was 'a bit more' or 'much more' than last winter, compared to a fifth of those without central heating.
The Glasgow Warm homes study , which evaluated the effects of improvements to physical conditions and energy efficiency of homes among Glasgow City Council's housing stock, further illustrates the point that significant changes in health may result from improvements in environmental conditions; here there were improvements in symptoms that tend to be associated with dampness, in particular in relation to respiratory, skin, alimentary, and musculo-skeletal symptoms. The dwellings themselves were warmer, less humid and with reduced fungal spore concentrations, and less damp and easier to heat. This improvement was achieved with no increase in energy costs ( i.e., more rooms could be heated after the intervention, with no significant increase in electricity bills).
The improvements in environmental conditions associated with health (in particular mental health) have been found in other studies   . Somerville et al   also report that installing heating improves child respiratory health.
Overall the more recent research suggests that SF-36 may be relatively insensitive to changes in health associated with non-clinical interventions, but that (based on our own experience) it may pick up changes in physical functioning, in particular mobility, that result from increases in warmth. Changes in health associated with improvements in cold, damp and mould are plausible, on the basis of previous research , but these may emerge over longer periods of follow-up.
In contrast with the limited effects of the Programme on direct measures of health, the effects of the CHP on what may be termed 'thermal comfort' are clear and substantial. Of those measures which relate to perceptions of warmth in the home (Section 4.2), to patterns of heating in the home (that is, the extent and duration of domestic heating in cold weather - Section 4.3) and to internal environmental conditions (Section 4.4), an overwhelming majority show significant effects of the Programme. Recipients of heating under the CHP feel warmer, heat more of their homes for longer, and are less likely to experience undesirable environmental manifestations such as dampness, mould and condensation.
While these effects are clearly desirable in themselves, it is reasonable to assume that, over an extended period of time, the improved residential conditions associated with these effects ( i.e. better-heated homes, with fewer problems of mould etc.) may be conducive to the development and maintenance of improved health. Under this assumption, one possible explanation for the limited direct impact of the CHP on health lies in a factor which has already been highlighted: the relatively short follow-up period (two years). The effects of the Programme on health over a more extended timescale cannot be extrapolated from the evaluation reported here, and further research would be required to determine any longer-term impacts of the CHP on the health status of central heating recipients.