Understanding the healthcare costs of temperature-related morbidity under the changing climate

Abstract

Background Exposure to non-optimum temperature may contribute to a substantial morbidity and mortality burden to our society. A growing number of studies have investigated the risks of increasing temperatures on morbidity and health service use. Estimating the economic burden of treating temperature-attributable emergency department (ED) presentations and hospital admissions is paramount to better understand the resource implications of temperature to the healthcare system. As the climate is shifting towards a warmer temperature, government and public health agencies demand evidence of future climate-related morbidity burden and economic cost to design practical guidelines, adaptive strategies, and tailored public health interventions. Aims The overall aim of this thesis was to provide scientific evidence for policymaking and practical guideline development to reduce the public health impacts and economic burden associated with extreme temperature and heatwaves. This thesis encompasses four studies, with specific aims to i) establish the temperaturemorbidity and temperature-cost relationships; ii) quantify the total ED presentations and costs attributed to heatwave intensity and heatwave severity; iii) estimate the current net-, cold-, and heat-attributable ED presentations, hospital admissions, length of hospital stay (LoS), and associated healthcare costs; and v) estimate future health and cost burden attributed to projected temperatures under different climate and population growth scenarios. Methods Based on the exposure and response variables, the study can be broadly divided into three parts, namely: ED presentations and ED costs with mean temperatures; hospital admission, LoS, and associated costs with mean temperatures; and ED presentations and costs with heatwaves. All the studies in this thesis were conducted using daily aggregated patients’ data from all public hospitals in Adelaide metropolitan area, South Australia (SA). Based on the 2016 census, the total population of the city was nearly 1.3 million. International classifications of diseases 10th version (ICD-10) was used to identify specific diagnosis groups including respiratory, renal, mental health disorders, heat-related illnesses (HRI), ischaemic heart diseases (IHD), and diabetes. These diseases were merged to increase the statistical power and termed as temperature related diseases (TRDs). A time-series analysis combined with a distributed lag nonlinear model (DLNM) was used to estimate the exposure-response relationships and delayed effects simultaneously. Using the current relationships, future morbidity and cost burdens of mean temperatures were projected under three representative concentration pathways (RCPs) (RCP2.6, RCP4.5, and RCP8.5) and medium population growth scenarios. All costs were calculated in Australian dollars (AU$). Results In study 1: The study used ED presentations data of all public hospitals in the Adelaide metropolitan area from 2014-2017. Both extreme cold and hot temperatures increased the risks of TRD-ED presentations and costs with much more risks associated with heat exposures. The baseline heat-attributable ED presentations and costs were estimated to be 3,633 (95% empirical confidence interval (eCI): 695, 6,498) and AU$4.7 million (95% eCI: 1.8, 7.5) costs, respectively. Under RCP8.5 and with a constant population, heat-attributable ED presentations and healthcare costs are projected to increase by 1.5% (95% eCI: 0.8, 2.2) and 2.0% (95% eCI: 1.1, 2.8) during 2054-57, respectively. When population change is considered, heat-attributable ED presentations and costs would increase by 1.9% (95% eCI: 0.8, 3.0) and 2.5% (95% eCI: 1.3, 3.7) during 2034-2037 and by 3.7% (95% eCI: 1.7, 5.6) and 5.0% (95% eCI: 2.6, 7.1) during 2054-2057, respectively. There may be no change in cold-attributable ED presentations and costs. In study 2: Using hospital admission data from 2010-2015, a comprehensive impact assessment of non-optimum temperature on TRD hospitalisations, LoS, and healthcare costs showed significant risks associated with heat exposure. During the baseline period, the net temperature-attributable hospital admissions, LoS, and associated costs were estimated to be 3,915 cases (95% empirical confidence interval (eCI): 235, 7,295), 99,766 days (95% eCI: 14,484, 168,457), and AU$159.1 million (95% eCI: 18.8, 269.0), respectively. Hot ambient temperature will have a substantial impact during 2040s and 2060s. Under RCP8.5 and demographic change, heat-attributable hospital admissions, LoS, and costs were projected to increase by 1.1%, 4.7%, and 4.2% during 2040s and further increased by 2.2%, 8.4%, and 7.7% in 2060s, respectively. In study 3: In Adelaide, the risks of ED presentations and cost increase as heatwave intensity and severity increase for all-cause, most diagnosis groups and age categories. During the four warm seasons of 2014-2017, heatwave-attributable allcause ED presentations were estimated to be 1,161 (95% eCI: 342, 1,944) with an associated cost of AU$1,020.3 (95% eCI: 224.9, 1,804.7), costs are in thousands. The HRI was the disease category contributing most to ED presentations and costs. Age groups ≤ 14 and ≥ 65 years were most susceptible to heat. For the elderly population, heatwave attributed to 554 (95% eCI: 228, 834) ED presentations and AU$530.1 (95% eCI: 160.2, 890.4) costs during the study period. Conclusions Heat-related risks and attributable ED presentations, hospital admission, LoS, and associated healthcare costs were dominant during the baseline periods. Young and old populations were the most vulnerable populations to heatwaves contributing a higher cost burden to the healthcare system. The projected estimates indicated that hot ambient temperatures are likely to increase future morbidity and healthcare costs substantially under all climate change scenarios. The broad implication of this thesis is to inform relevant stakeholders and provide evidence for policymaking about the morbidity and economic burden of observed and projected temperature. The results help to track the future burden of climate change and to develop a range of climate change mitigation and public health adaptation responses. Moreover, the results suggest targeted intervention actions such as protecting the vulnerable population and the need to build the capacity of the health system.