Health begins at home? An RCT study to test the effects of a housing provider's health interventions

Recent research has been increasingly showing that it is not the houses or the neighbourhoods in which people live that make them sick or poor: rather it is their personal characteristics. Still this does not mean that interventions aimed at occupants of social housing may not be effective in improving health. The structure of social housing provision in Britain means that there are many large social landlords. Landlords have particular access to their tenants and this paper presents the results of an Randomised Control Trial designed to evaluate the impacts of simple health interventions on the health of poorer residents over 50 living in 5 inner London boroughs. The study was carried out between February 2013- May 2015. Participants were randomly divided into three groups: a control group (n= 186), a group receiving additional signposting to health services (n= 172); and a group receiving intensive support from a dedicated health and wellbeing support worker (n= 174). The intervention period was 18 months. The main outcomes measured were self-reported health and wellbeing ratings and NHS usage. Randomisation was computer-generated and participants and assessors were blinded to the allocation. Perhaps the single most important finding was that the initial base line health evaluation revealed 25 (4.5%) of the total sample as having such severe health problems that significant and immediate intervention was required. This set of people was taken out of the trail and directed to treatment. In purely scientific terms this is unfortunate since it means that all intervention effects are likely underestimated since these were people whose health was critical. On the other hand it shows that the simplest possible intervention to a targeted group (older and poorer people), just an individual health status assessment, can produce significant health improvements. For the 95% staying in the trial significant effects were found for planned hospital usage, with the intensive intervention group's usage reducing by 39% in comparison to control group's increase of 11% (p= 0.004). Significant effects were also found for nights in hospital where the signposting intervention's usage decreased (35%) significantly in comparison to the control group's increased usage (13%) (p= 0.022). The intensive intervention group's usage in fact reduced more (62%) but variance was high (20.198), affecting significance. Substantive effects were identified for emergency GP usage, where group 3 reduced their usage by 15%, a substantive difference (p= 0.055) to group 2's 181% increase. *The study was initiated and funded by Family Mosaic