BackgroundThe frequency and severity of heatwaves are projected to rise due to climate change, increasing heat-related illness risks, particularly among older adults. Existing studies have predominantly focused on hazard and exposure factors as a direct cause of heat-related health effects. However, few studies have examined the role of vulnerability factors among susceptible populations from a behavioural insight perspective. Here, we analysed the effects of risk perception biases on heatstroke occurrence among older adults. MethodsWe conducted a survey across 11 welfare centres for older adults (aged 60 years or older) in Kyoto, Japan, in October, 2021. We partnered with the Elderly Welfare Section of the City Government of Kyoto to design our research approach and surveys. The survey covered four primary factors: essential habits (ie, eating, sleeping, and exercising), heat stress (eg, heat stroke, heat exhaustion, heat cramps, and heat rashes), cognitive biases (ie, comparative optimism and loss aversion biases), and social capital. Descriptive statistics were used to summarise the survey results. Findings:3023 respondents participated in the survey (2152 women, 626 men, and 245 respondents who did not report gender; median age 77 years [IQR 73–81]). A substantial proportion of older adults in Kyoto have comparative optimism biases (64 [43%] of 150 male respondents and 154 [41%] 376 of female respondents). The percentage of men recording the highest degrees of comparative optimism bias was 56% higher than that in women. No substantial differences in social capital were found among the respondents with or without comparative optimism biases. In contrast to comparative optimism bases, most of the surveyed population (89 [88%] of 101 male respondents and 403 [95%] of 424 female respondents) did not have loss aversion biases. However, among those who have loss aversion biases, men show 2·5 times higher biases than women. Living alone seemed more conducive to having biases (12 [7%] of 170 respondents) than not living alone (19 [5%] of 355 respondents). InterpretationPublic agencies need to formulate early warning messages or awareness-raising programmes to manage the existence of these biases. There are three approaches to eliminating optimism biases: increasing self-risk estimates; encouraging older adults to provide feedback about the comparison groups through comprehensive information; and restricting comparison groups, given that people with higher comparative optimism biases tend to intentionally or unintentionally compare themselves with other people at higher risk than themselves. FundingResearch Institute for Humanity and Nature (Kyoto, Japan).