Abstract
Background: Excess winter mortality and morbidity are often greatest in countries with relatively mild climates, probably due to poorer thermal housing standards compared to well-insulated houses in colder climates. Cold houses are also associated with indoor dampness and mould. Cold temperatures have been linked to worsening of cardiovascular disease symptoms, whilst indoor dampness and mould have been consistently linked to increased risk of respiratory symptoms. Most New Zealand homes are cold by international standards and do not meet the World Health Organization (WHO) recommended minimum indoor temperature of 18°C. Prior to a 1977 amendment to the building code, there was no requirement to insulate homes. Since then, insulation of new homes has been mandatory, but retrofitting insulation into existing homes has been at the discretion of the homeowner. In 2009, the Energy Efficiency and Conservation Authority (EECA) established an insulation subsidy programme for existing homes, Warm Up New Zealand Heat Smart (WUNZ). WUNZ ran from June 2009 to July 2014, with the aim of providing “warm, dry and more energy efficient homes” and represented a NZ$347 million government investment.
Aim: To assess whether residents experienced a reduction in severity and frequency of cold-associated ill health following insulation retrofits in their homes. Four measures of health outcome were used: pharmaceuticals dispensed for respiratory illness, respiratory disease incidence (based on pharmaceuticals dispensed), hospital admissions and mortality.
Objectives: To develop a data linkage method for tracking residents of houses over time. To investigate the modifying effect between different groups within the population based on socio-demographic and environmental influences.
Methods: A quasi-experimental retrospective cohort study design linked a dataset of houses insulated through the WUNZ subsidy programme to health and socio-demographic records. A difference-in-difference model compared changes in health outcomes in an intervention group that had received insulation through WUNZ to a control group from the same WUNZ dataset that subsequently received insulation. Health outcomes measured by this method included: pharmaceuticals dispensed for cold-associated respiratory disease; and acute and cold-associated hospital admissions. Direct comparisons between the intervention and control groups determined whether there was any difference in chronic respiratory disease incidence post-insulation, using two prescription medication measures: prevention medication and exacerbation sensitive medication. Prevention medication included prescription medicines taken regularly to manage the illness. Exacerbation sensitive medication included medicines prescribed for use when symptoms became more acute, for example to prevent an asthma attack. Survival analysis determined any differences in cold-associated mortality between the intervention and control groups.
Results: There were statistically significant improvements in residents’ health across the four outcome measures. A statistically significant lower risk of chronic respiratory disease incidence was found amongst the intervention group compared to the control group for both a prevention of symptoms indicator: risk ratio (RR1) 0.92 (0.89-0.98) and for an exacerbation sensitive medication indicator: RR1 0.88 (0.83-0.92). Relative rates of pharmaceuticals dispensed were lower in the intervention group compared to the control group for infectious respiratory disease: relative rate ratio (RRR) 0.98 (95 % confidence interval: 0.98-0.99) and exacerbation sensitive medication used to treat symptoms of chronic respiratory disease: RRR 0.96 (0.96-0.97). Relative rates of acute hospital admissions were also lower in the intervention group RRR 0.89 (0.87-0.91) and fell further when specific cold-associated respiratory and cardiovascular diseases were considered. There was no statistically significant reduction in mortality overall, however, results were statistically significant when heating was installed alongside insulation: hazard ratio (HR) 0.77 (0.61-0.97).
There was not an even distribution of health gains from the WUNZ subsidy across ethnicities. Health gains for Pacific Peoples were more pronounced than other ethnicities whilst those for Māori were less pronounced.
Additional health gains were demonstrated when an energy efficient heater was installed alongside insulation in two of the four health outcomes investigated (pharmaceuticals dispensed and mortality) and were comparable to installing whole house insulation when hospital admissions were examined.
Conclusion: Taken together these results provide strong evidence to support housing interventions that increase thermal efficiency to improve health outcomes for residents. Delivery of such interventions may need further thought however, to ensure equity of health outcomes across socio-demographic groups.
Further research should incorporate intervention datasets into the integrated data infrastructure (IDI) to understand the impact on health outcomes of household tenure, lifestyle factors and comorbidities better. The data generated from the study could update cost-effectiveness analyses of future housing interventions.
Future research should incorporate a holistic approach to improving thermal efficiency that also addresses the more complex nature of the relationship between housing and health. This would help to improve the delivery of structural interventions to increase equity of health outcomes and fully realise health gains.