QUESTION

What are the health and social impacts on residents following improvements to the physical fabric of housing?

RELEVANCE TO NURSING CARE

Housing is a key social determinant of health and poor housing has been found to be associated with poor health. However, much of the research into the links between housing and health has been cross-sectional and what is less clear is how specific housing improvement interventions impact on health outcomes of residents. Evaluations of area-based 'neighbourhood renewal' improvements designed to attract new residents may report improved aggregated health outcomes, but these may reflect the health of new populations and not necessarily improved health outcomes for the original population who may not have benefited from the intervention. It is of interest therefore to determine whether housing improvement interventions have an impact on health outcomes and related socioeconomic outcomes, because the latter may be predictors for future health outcomes.

Nurses espouse a holistic philosophy to care, which includes concern for their patients' social well-being. Those working in community or public health settings may be more aware of the importance of housing conditions to health, although all nurses and health professionals who adopt a biopsychosocial model of care would share such interest. This review1 provides nurses and other health professionals with evidence about the impact of housing improvements on residents' health and social outcomes. This evidence can support their advocacy role on behalf of patients in relation to their need for appropriately heated and sized housing.

STUDY CHARACTERISTICS

Thirty-nine studies that reported quantitative or qualitative data, or both, were included in the review. Before-and-after, retrospective, controlled, uncontrolled, randomized (including clustered samples) and nonrandomized studies of the health and social effects of housing improvements were included. Cross-sectional studies that did not investigate changes in health outcomes following housing improvement were excluded. Studies were included if they evaluated change in any physical or mental health (including self-reported well-being and quality of life) outcome following housing improvements, which involved a physical improvement to the fabric of the house. There was no exclusion to the types of participants; they could be of any family type, socioeconomic status or other equity indicator. Studies were evaluated using the Cochrane Risk of bias tool, including three additional items from the Cochrane Effective Practice and Organisation of Care group. Because of the wide range of study designs included, the Hamilton Assessment Tool was also used for nonrandomized studies, as recommended by the Cochrane Public Health Group, under which this review was conducted.

Thirty-three quantitative studies were identified by the review, of which

1. five were randomized-controlled trials of warmth improvement

2. 10 were nonexperimental studies of warmth improvement

3. 12 were nonexperimental studies of rehousing or retrofitting

4. three were nonexperimental studies of basic improvement in low-income or middle-income countries

5. three were nonexperimental studies of rehousing from slums.

Twelve studies reporting qualitative data were identified and they covered warmth improvement and rehousing/retrofitting. Six of these also included quantitative data and were included in the review. Following assessment of study quality 14 quantitative studies and three qualitative studies were excluded from the data synthesis due to poor quality (considerable risk of bias). Only those studies judged to have a minimal or moderate risk of bias were included in the final synthesis. There was considerable variation in methodology, samples, interventions, context and outcomes, which meant that little quantitative synthesis was possible, even when studies were grouped by interventions and outcomes. Studies were grouped and data synthesised according to the following intervention categories:

1. 17 studies of 'warmth and energy efficiency improvement (post-1985)

2. 14 studies of 'rehousing or retrofitting +/- neighbourhood renewal (post-1995)

3. Three studies of 'provision of basic housing in low or middle income countries (post-1990)'

4. Three studies of rehousing from slums (pre-1970)

Studies of interventions to improve 'warmth or energy efficiency (post-1980)' suggested that improvements to general and respiratory health were possible. The commonest outcome assessed for children and adults was respiratory health. Greater health benefits were found wherein people with existing chronic respiratory disease and inadequate warmth had been targeted in studies. In some studies there were conflicting results across the range of measures used. Changes to mental health were less clear across the better-quality studies. Improvements in warmth were reported to lead to social impacts such as improved social relationships, reduced absences from work or school because of illness, and increased privacy.

For the 'rehousing and retrofitting +/- neighbourhood renewal (post-1995)' studies, the evidence of impact is unclear. Some studies reported health improvements whereas others did not. In these interventions, which are area-based, health impacts were less clear where areas, rather than individuals in most need, were targeted. There was limited evidence from the studies of 'provision of basic housing in low-income and middle-income countries (post-1990)'. Again some studies showed improvement while others did not, and study quality varied. The small number of studies reflects that interventions to improve living conditions in low-income and middle-income countries are often made at the communal/community level, whereas this review only included studies wherein the intervention was delivered at the individual household level. For the 'rehousing from slums (pre-1970)' studies, again evidence was limited and of poor quality.

A logic model mapping the reported health impacts and pathways to health impacts following housing improvement is presented; this model draws on the evidence from groups 1 and 2 above, as they included better quality studies, and qualitative data and are relevant for higher-income country contexts. There was inadequate reporting of differential impacts (by socioeconomic grouping, etc.) to provide evidence about the impact of housing improvements on social and economic inequalities. The delivery of a housing improvement intervention cannot assume actual improvement in housing conditions for residents. The geographical context of studies is particularly important, for example regarding warmth improvements.

IMPLICATIONS FOR NURSING CARE

Nurses act as advocates for people in relation to their social needs, including for housing. A greater understanding of the impact on health and social outcomes of inadequately heated or sized housing can provide a basis for this advocacy. The evidence for warmth and energy efficiency interventions and for those who have existing illness (e.g. respiratory illness) is strongest.

IMPLICATIONS FOR RESEARCH

There has been an increase in quantitative and qualitative research in this area that is welcomed, but there is room for improvement. No single group of studies was sufficiently homogeneous to facilitate robust meta-analysis. There are remaining questions about the timescale of interventions, impacts for specific groups and in different contexts. Poor reporting across the studies also made synthesis difficult. Randomized-controlled trials are possible, although challenging to design and implement for testing public health interventions. Larger studies would offer subgroup analysis (by baseline health status etc.). Qualitative studies included in the review highlighted outcomes not previously identified and the inclusion of qualitative elements to future studies is recommended. It is recommended that future reviews of the broad area of housing improvement be split, with for example, a separate review covering studies of warmth and energy efficiency and separate reviews covering interventions in high- and low-income/middle-income settings.

Acknowledgements

The author reports no conflicts of interest.

Reference