Objective: To assess the short term health effects of improving housing. Design: Randomised to waiting list. Setting: 119 council owned houses in south Devon, UK. Participants: About 480 residents of these houses. Intervention: Upgrading houses (including central heating, ventilation, rewiring, insulation, and re-roofing) in two phases a year apart. Main outcome measures: All residents completed an annual health questionnaire: SF36 and GHQ12 (adults). Residents reporting respiratory illness or arthritis were interviewed using condition-specific questionnaires, the former also completing peak flow and symptom diaries (children) or spirometry (adults). Data on health service use and time lost from school were collected. Results: Interventions improved energy efficiency. For those living in intervention houses, non-asthma-related chest problems (Mann–Whitney test, p = 0.005) and the combined asthma symptom score for adults (Mann–Whitney test, z = 2.7, p = 0.007) diminished significantly compared with control houses. No difference between intervention and control houses was seen for SF36 or GHQ12. Conclusions: Rigorous study designs for the evaluation of complex public health and community based interventions are possible. Quantitatively measured health benefits are small, but as health benefits were measured over a short time scale, there may have been insufficient time for measurable improvements in general and disease-specific health to become apparent.