Abstract
Background and Objectives
Previous evidence showed that frailty in older age and precarious housing characteristics, respectively, contribute to poorer heat-related health outcomes. It is not known whether these conditions would interact with each other to produce a larger impact on older people’s health.
Research Design and Methods
A cross-sectional questionnaire survey was conducted in May–July 2024 in a sample of older people aged 60 or over in Hong Kong, a city on the Southern coast of China. Frailty was measured using the Fried phenotype (five items). Housing characteristics were measured by whether it is a small flat (<100 square feet per person), living alone, inadequate housing (subdivided units or other forms), or whether the respondent had an air conditioner at home. Heat-related health outcomes were self-rated health, thermal comfort at home, and any heat-related signs/symptoms during summertime. Multivariable Poisson regression with robust standard error was used. Relative risk due to interaction was used to characterize additive interaction between housing characteristics and frailty.
Results
Among the 1,393 respondents who completed the questionnaire, about 60% reported being frail. Those who were frail were less likely to report thermally comfortable at home (RR: 0.75; 95% CI: 0.66, 0.87) and had a higher chance of reporting any heat-related signs/symptoms (RR: 1.28; 95% CI: 1.18, 1.38). We found evidence of additive interaction between frailty and living alone, in which, only for those who were robust, living alone is related to a lower risk of heat-related signs/symptoms.
Discussion and Implications
Targeted interventions to improve the well-being of community-dwelling older adults during periods of extreme heat should be designed especially for those who are frail.
Keywords: Frailty, Precarious housing, Extreme heat, Heat-health
Innovation and Translational Significance.
This study in Hong Kong addresses how frailty and poor housing together worsen heat-related health risks in older adults. It found that frail older adults face higher chances of heat-related symptoms and discomfort, whereas those living in small and crowded conditions face poorer thermal comfort. Living alone appeared to be protective against heat-related symptoms, but it was only evident for those who were robust. Our findings underscore the urgency of targeted interventions, particularly for frail individuals and those living alone, in the face of escalating climate threats. Social programs must address isolation by fostering community networks and providing targeted support during heatwaves.
Extreme heat events are an escalating public health concern, particularly in places with rapid urbanization such as Hong Kong, where demographic ageing and housing precarity converge to exacerbate vulnerability among older adults. The health risks associated with heat exposure are well-documented, with older individuals experiencing increased rates of morbidity and mortality during heatwaves (Whitman et al., 1997), as well as higher rates of emergency department presentations and hospitalizations on hot days (Lai et al., 2025). Frailty—a multidimensional syndrome marked by decreased physiological reserve and heightened susceptibility to external stressors—further amplifies these risks, rendering older adults particularly prone to adverse outcomes during periods of elevated ambient temperature (Browning et al., 2006). In addition to individual-level vulnerabilities, environmental and social determinants, such as inadequate housing, limited access to air conditioning and social isolation, have been shown to significantly influence heat-related health outcomes (Samuelson et al., 2020).
Frailty could pose individuals at increased risk for heat-related illnesses, compared with the general adult population, through several physiological and functional pathways (Ebi et al., 2021). Ageing is associated with diminished cardiovascular and thermoregulatory responses, including attenuated skin blood flow and reduced nitric oxide-mediated vasodilation, which impairs the body’s ability to dissipate heat. Frail individuals often have further compromised cardiovascular, pulmonary, and musculoskeletal systems, as well as weakened immune responses, all of which limit their ability to cope with thermal stress (Mackenbach et al., 1997). Additionally, frail older adults may have reduced mobility, cognitive impairment, and dependence on others for daily activities, rendering them difficult to seek cooler environments or implement adaptive behaviors during heat events (Dent et al., 2020). These physiological and functional limitations collectively increase susceptibility to heat exhaustion, dehydration, and other adverse outcomes during periods of extreme heat.
Precarious housing conditions independently contribute to adverse heat-health outcomes through both physical and psychosocial mechanisms. Small flats, subdivided units, and crowded households are often characterized by poor ventilation, inadequate insulation, and limited access to cooling appliances, leading to a more uncomfortable thermal environment in the indoor space than the outdoor space (Lomas, 2021). In Hong Kong, subdivided units are formed by dividing a larger flat into smaller compartments. These units often arise in response to the keen demand for housing in Hong Kong. The way that a flat is divided might render some subdivided units lacking windows or proper airflow, resulting in poor ventilation and persistent indoor heat, which significantly increases heat stress and the risk of heat-related illness (Cheung & Jim, 2020). Crowded living conditions can also exacerbate psychological distress, which may compound physical vulnerability during heatwaves (Ruiz-Tagle & Urria, 2022). Social isolation, common among older adults living alone, further heightens risk by limiting access to assistance, information, and social support during heat emergencies (Whitman et al., 1997). These environmental and social factors interact to create a hazardous living environment, particularly for low-income and marginalized groups (Estoque et al., 2020).
Hong Kong’s subtropical climate is characterized by hot, humid summers, and the frequency and intensity of extreme heat events have been rising in recent decades. According to climate projections, the annual mean temperature in Hong Kong is expected to increase by 2.0–3.6°C by the end of the century under intermediate to high greenhouse gas emission scenarios, with more frequent and prolonged periods of very hot weather (Zuo et al., 2023). This trend coincides with rapid population ageing: Hong Kong’s proportion of residents aged 65 and above continues to grow (Census and Statistics Department, 2020), amplifying the number of individuals at risk of heat-related morbidity and mortality. The intersection of an ageing demographic with high-density housing and widespread economic inequality compounds the public health challenge posed by extreme heat in this urban context (Hua et al., 2020).
While previous studies have established the independent effects of frailty and housing precarity on heat-health outcomes, there is limited understanding of how these risk factors may interact to produce synergistic effects among community-dwelling older adults, especially in high-density urban environments such as Hong Kong. Few studies have examined whether the co-occurrence of frailty and precarious housing conditions confers a disproportionately higher risk of adverse health outcomes during heat events. Furthermore, research on this topic in the context of Hong Kong’s unique housing landscape and rapidly ageing population remains scarce. Specifically, this study addresses this gap by investigating the interaction between frailty (a common geriatric syndrome among older adults) (Fried et al., 2021) and precarious housing conditions in the relationship with heat-related health outcomes.
Theoretical framework
To understand the interplay between frailty, housing characteristics, and heat-health outcomes, this study employed the social ecological model of health, which posited that an individual’s health is determined by the dynamic and cumulative interplay of factors across multiple levels of social and environmental influence, in addition to the intrinsic human biological factors (Bronfenbrenner, 1977). Different levels of social and environmental factors would have an impact on each other (Golden & Earp, 2012). In the context of heat-health outcomes, we can specifically frame how individual-level vulnerabilities are shaped and exacerbated by the surrounding physical and social environments, thereby providing a robust theoretical motivation for examining frailty, housing, and heat-health outcomes simultaneously.
At the core of the socioecological model are individual constitutional factors. The primary individual-level factor in this study is frailty, a geriatric syndrome characterized by a decline in physiological reserve across multiple organ systems. Frailty is a state in which individuals experience increased sensitivity to environmental stress, such as extreme heat. As elaborated in the above section, ageing and frailty-associated physiological changes would pose a greater risk for adverse heat-health outcomes (Ebi et al., 2021). Importantly, frailty serves as a more holistic and powerful predictor of vulnerability than chronological age alone, capturing a multidimensional state of reduced resilience (Fried et al., 2021).
The next level on the social ecological model is the physical and social environment in which the individual is situated. For the physical (built) environment, housing conditions directly shape an individual’s physical exposure to heat. As noted earlier, factors such as living in a small, crowded flat or inadequate housing (e.g., subdivided units) can lead to poor ventilation, heat retention, and a more intense indoor thermal environment, increasing the duration and intensity of heat exposure (Cheung & Jim, 2020). Furthermore, access to protective resources, such as possessing air conditioning at home, is a critical component of an individual’s adaptive capacity. Lack of access, or the inability to use it due to economic constraints, directly limits behavioral responses to mitigate heat stress. For the social environment, older adults who live alone could be more vulnerable through pathways related to social support and isolation (Lai, Ho, & Woo, 2023). It is also often associated with a lack of immediate assistance during a health emergency, reduced social monitoring, and potential barriers to accessing information or resources (Browning et al., 2006).
While not measured directly, the next level of the theoretical framework operates within the broader macrosystem of Hong Kong’s unique context: a subtropical climate facing accelerating climate change, rapid urbanization leading to urban heat island effects, a notoriously expensive and constrained housing market, and significant economic inequality. These large-scale forces create and perpetuate the environmental risks and social inequities experienced at the lower levels of the model.
This social-ecological framework provides the theoretical impetus for our study. We could hypothesize that the impact of the risk factors for adverse heat-health outcomes is not merely additive. Instead, we reckon that there are synergistic interactions across these levels. For example, the negative health impact of living in a subdivided flat is likely amplified for an individual who is also frail. The central contribution of this research, therefore, is to empirically test for these cross-level interactions, providing evidence to support a multi-level approach to public health interventions.
Methods
Subjects
A cross-sectional questionnaire survey was conducted in May–July 2024. The aim of the questionnaire was to assess the older adults’ awareness of the health impact of hot weather, hot weather warning signals, and whether they have the knowledge of adopting appropriate strategies to avoid extreme heat in summertime. Participants were sampled by recruiting individuals aged 60 or above who were walking on the street in two districts in Hong Kong—the Yau Tsim Mong district and the Kwun Tong district. The two districts were chosen on the basis that they are hotspots that are susceptible to more urban heat than other districts in the Hong Kong territory (Hua et al., 2020; Shi et al., 2019). In brief, Kwun Tong District mainly features high-rise public and private housing estates with very high population density (59280 people per square kilometer), reflecting its industrial and urban redevelopment history, while Yau Tsim Mong District also features a high population density (43510 people per square kilometer) is dominated by private permanent housing, mostly old tenement buildings, many subdivided into small units (Census and Statistics Department, 2025).
Questionnaires were administered by trained research assistants. Only community-dwelling older people were invited to participate, whereas those who were institutionalized were excluded. The sample size needed for this survey was computed based on the mid-year population for people who are aged 60 or above in these two districts in 2020, published by the Census and Statistics Department, as well as taking into consideration a confidence level of 95% and a margin of error of 3%. Accordingly, a sample of 1,514 participants was recruited to ensure that we have a buffer to allow for incomplete questionnaires and invalid responses. The questionnaire was designed based on a similar population knowledge survey used in Europe to assess knowledge and awareness of hot weather by Gil Cuesta et al. (2017), but we added items according to the Hong Kong context, e.g., whether the respondents are aware of the very hot weather warning issued by the Hong Kong Observatory. Ethics approval was obtained from the Survey and Behavioural Research Ethics Committee of the Chinese University of Hong Kong (SBRE-23-0563).
Heat-related health outcomes
There were three heat-related health outcomes included in the current study: self-rated health, thermal comfort at home, and signs or symptoms of heat illness. Self-rated health was measured using a Likert scale, asking the respondent to rate his or her overall health conditions on a scale of 1 to 5. The scale of self-rated health was then recoded as a dichotomous variable, in which scores 3 (good) to 5 (excellent) were defined as good, while 1 (poor) and 2 (fair) were defined as poor (Lee et al., 2021). For thermal comfort, respondents were asked to rate how comfortable their home was when they did not use an air conditioner. A seven-point Likert scale was used for assessing thermal sensation, which corresponds to the American Society of Heating, Refrigerating and Air-Conditioning Engineers seven-point thermal sensation scale (ie, −3 [very uncomfortable] to +3 [very comfortable], in which 0 indicates neutral sensation) (ASHRAE, 2010). Accordingly, when a respondent rated an environment in which he or she was situated as neutral or above, it could be considered thermally comfortable (Lai et al., 2020). We dichotomized thermal comfort at home as poor thermal comfort for those who reported having scores ≤−1. For signs or symptoms of heat-related illnesses, we asked the respondents whether they experienced the following in summertime in recent years: tiredness, muscle cramping, nausea/vomiting, dizziness, headache, palpitation, shortness of breath, or feeling agitated. We then dichotomized this variable as whether the respondents experienced any of the above signs or symptoms.
Housing characteristics
We assessed the respondents’ housing characteristics using the following features: small flat, living alone, inadequate housing, and having air conditioning at home. The size of the respondents’ flat was obtained as a self-reported measure in the questionnaire, which was measured in square feet. We then obtained the variable for small flats by analyzing the average flat size per person in the household and its eventual distribution. A small flat was then defined as a flat size per person smaller than the quartile of the distribution (100 square feet ≈ 9.29 square meters). Our definition of a small flat was slightly larger than the definition suggested by a previous local study (Wong & Chan, 2019), who defined overcrowded flats as smaller than 7 square meters. Nonetheless, we retained our data-driven definition to preserve adequate power to allow comparison. Living alone was assessed by asking the number of persons in a household, in which “1” person indicated he or she lives alone. The type of housing in which the respondent lived was assessed with a question with the following options to choose from: public housing, subsidized housing, subdivided units, private housing (owner-occupied or rental), squatter, and others. From the responses for “others” we collected, there were two respondents who indicated that they lived in a unit in an industrial building, while one indicated that he lived in a unit in a commercial building. We defined inadequate housing as a dichotomous variable, indicating those who lived in subdivided units, squatters, and the three aforementioned “others” cases. In addition, the respondent was also asked in the questionnaire whether he or she possessed an air conditioner at home.
Frailty
We applied a simple 5-item FRAIL instrument, which stands for fatigue, resistance, ambulation, illnesses, and loss of weight (Morley et al., 2012), as per our previous study (Woo et al., 2015). Fatigue was measured by asking respondents whether they felt tired “most of the time” in the past 4 months (1 point). Resistance was assessed by asking the respondents if they had any difficulty walking up 10 steps of stairs without resting and without aids; and Ambulation by asking if they had any difficulty walking for 5 minutes alone and without aids; those who answered “yes” scored as 1 point each. Illness was scored 1 for respondents who reported five or more illnesses that require long-term care and medication. Loss of weight was scored 1 for respondents with a weight decline of 5% or greater within the past 12 months based on self-report. A total possible score of 5 indicates the worst frailty situation. Those who scored 3 to 5 were defined as being in the state of frailty (Morley et al., 2012), while those who were robust (ie, not in the state of frailty) were those who scored 0 to 2.
Covariates
In the adjusted models, we adjusted for the following variables: age, sex, education attainment, residential district, and the date when the survey was collected. Age was measured as the closest rounded integer in the questionnaire. Sex of the participants was identified as either male or female. Education attainment was measured as a categorical variable with the following categories: primary school or below, secondary or tertiary or above. The residential district was identified as either the Yau Tsim Mong district or the Kwun Tong district. The date of the survey was marked in the questionnaire as the research assistant conducted the survey. It was included as a covariate for adjustment given that the surveying period spanned across two months in summer; temperatures and humidity can fluctuate daily, and generally trend higher toward the latter part of the summer, which could influence participants’ perceptions and experiences of heat-related outcomes.
Statistical analysis
Descriptive analyses were performed to summarize the socio-demographic characteristics, housing conditions, frailty, and heat-related health outcomes in the study samples. Chi-square tests were used for categorical variables, and one-way ANOVA was used for continuous variables, comparing the differences between the characteristics of the two sampling districts.
The relationships between housing characteristics and heat-related health outcomes were assessed using multivariable Poisson regression with robust standard error, from which we obtained the relative risks (RR) and the corresponding 95% confidence interval (CI) (Zou, 2004) (Note: robust in this context specifically refers to the statistical technique handling standard error, but not related to frailty status.) We built two models—we built a crude model and an adjusted model, the latter of which adjusted for the aforesaid covariates. Similarly, we also examined the relationship between frailty and heat-related health outcomes.
To evaluate the joint effect, we included an interaction term between the housing characteristics and measure of frailty in the Poisson models and evaluated whether the effect of housing characteristics and frailty is larger than the sum of the individual effects of the two exposures on heat-related health outcomes. We quantified effect modification on both multiplicative (ratio of relative risks) and additive scales (relative excess risks due to interaction, RERI). Multiplicative interactions were represented by the interaction term between the putative housing characteristics and frailty, in which a statistically significant interaction term indicates the presence of multiplicative interaction; whereas positive additive interaction was indicated by an RERI >0 (VanderWeele & Knol, 2014). Confidence intervals of the RERI estimates were computed using the Delta method (Hosmer & Lemeshow, 1992).
We conducted additional analyses by assessing the independent associations of housing characteristics and heat-related health outcomes stratified by frailty status, and associations of frailty and heat-related health outcomes stratified by housing characteristics. Nonetheless, we noted that some models did not converge due to data sparsity in specific strata with zero or very few subjects. Given the fact that usage of air conditioners is an important strategy against extreme heat, we assessed whether the usage of air conditioners among owners was associated with heat-related health outcomes. We assessed in the questionnaire the number of days in a week that the respondent would use the air conditioner. Statistical analyses in this study were conducted on the basis of complete-case analysis, ie, restricted to participants with a complete set of data on all variables of interest. All the analyses were conducted using R programming (version 4.0.3).
Results
Subject characteristics
Among the 1,514 recruited participants in the survey, 1,393 of them provided a complete response on housing characteristics, frailty, heat-related health outcomes, and the relevant covariates. By design, we recruited a similar number of participants from the two districts (687 from the Yau Tsim Mong district and 706 from the Kwun Tong district) (Table 1). The mean age of the respondents was 72.4 (SD = 8.0), and 40% of them were male. There were 60% of the respondents reported that they are in a state of frailty. Notably, 95% of the respondents reported having at least one air conditioner at home. Among the two sampling districts, respondents reported substantially better thermal comfort at home in the Kwun Tong district than in the Yau Ma Tei district. Only one respondent reported living in inadequate housing in the Kwun Tong district.
Table 1.
Characteristics of the respondents of the questionnaire.
| Characteristics | Districts |
ANOVA/χ2 value | p * | |
|---|---|---|---|---|
| Yau Tsim Mong | Kwun Tong | |||
| n | 687 | 706 | ||
| Age in years, Mean (SD) | 72.38 (7.37) | 72.37 (8.52) | 0.001 | .973 |
| Frailty, n (%) | 399 (58.1) | 440 (62.3) | 2.444 | .118 |
| Male, n (%) | 262 (38.1) | 300 (42.5) | 2.567 | .109 |
| Have air conditioner at home, n (%) | 644 (93.7) | 677 (95.9) | 2.864 | .091 |
| Reported thermally comfortable at home, n (%) | 195 (28.4) | 297 (42.1) | 27.943 | <.001 |
| Reported any heat-related symptoms, n (%) | 479 (69.7) | 494 (70.0) | 0.002 | .966 |
| Poor self-rated health, n (%) | 84 (12.2) | 89 (12.6) | 0.018 | .894 |
| Living alone, n (%) | 150 (21.8) | 147 (20.8) | 0.157 | .692 |
| Small flat, n (%) | 234 (34.1) | 263 (37.3) | 1.409 | .235 |
| Inadequate housing, n (%) | 81 (11.8) | 1 (0.1) | 83.192 | <.001 |
| Education attainment, n (%) | 2.415 | .299 | ||
| Primary school or below | 402 (58.5) | 384 (54.4) | ||
| Secondary | 249 (36.2) | 282 (39.9) | ||
| Tertiary or above | 36 (5.2) | 40 (5.7) | ||
ANOVA was performed for the continuous variable (age), and the chi-square (χ2) test was performed for the categorical variables.
Main analysis
As shown in Table 2, Living alone was associated with a higher risk of poor self-rated health (RR: 1.42; 95% CI: 1.04, 1.93), but the association was attenuated upon adjusting for covariates (RR: 1.29; 95% CI: 0.95, 1.76). Owning air conditioner(s) at home was related to lower risk of reporting poor self-rated health (RR: 0.53; 95% CI: 0.33, 0.83). Living in a small and crowded flat was related to a lower chance of reporting better thermal comfort at home (RR: 0.83; 95% CI: 0.71, 0.97). Interestingly, living alone was related to a higher chance of reporting thermally comfortable at home (RR: 1.23; 95% CI: 1.05, 1.44), and owning air conditioner(s) is related to a lower chance of better thermal comfort at home (when the air-conditioner is not in operation). Respondents who lived alone reported a lower risk of reporting any heat-related signs/symptoms (RR: 0.90; 95% CI: 0.82, 0.98). Frailty is strongly related to poor self-rated health (RR: 7.97; 95% CI: 4.57, 13.90), while it is also related to a lower chance of reporting better thermal comfort at home (RR: 0.76; 95% CI: 0.66, 0.87) and a higher risk of any heat-related signs/symptoms reported (RR: 1.28; 95% CI: 1.19, 1.38).
Table 2.
Associations between housing characteristics and heat-related outcomes (n = 1,393).
| Outcomes | Small flat |
Living alone |
Inadequate housing |
Have air conditioner at home |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RR | 95% CI |
RR | 95% CI |
RR | 95% CI |
RR | 95% CI |
|||||
| LL | UL | LL | UL | LL | UL | LL | UL | |||||
| Poor self-rated health | ||||||||||||
| Crude | 0.98 | 0.73 | 1.31 | 1.42 | 1.04 | 1.93 | 0.88 | 0.47 | 1.65 | 0.50 | 0.32 | 0.78 |
| Adjusted | 1.01 | 0.75 | 1.35 | 1.29 | 0.95 | 1.76 | 0.92 | 0.48 | 1.74 | 0.53 | 0.33 | 0.83 |
| Thermally comfortable (binary) | ||||||||||||
| Crude | 0.84 | 0.72 | 0.98 | 1.26 | 1.07 | 1.47 | 0.78 | 0.55 | 1.12 | 0.57 | 0.46 | 0.70 |
| Adjusted | 0.83 | 0.71 | 0.97 | 1.23 | 1.05 | 1.44 | 0.98 | 0.68 | 1.42 | 0.56 | 0.46 | 0.68 |
| Reported any heat-related symptoms | ||||||||||||
| Crude | 1.07 | 1.00 | 1.15 | 0.90 | 0.82 | 0.99 | 1.01 | 0.88 | 1.17 | 1.03 | 0.87 | 1.21 |
| Adjusted | 1.07 | 1.00 | 1.15 | 0.90 | 0.82 | 0.98 | 1.03 | 0.88 | 1.19 | 1.03 | 0.87 | 1.21 |
Note. RR = relative risks; CI = confidence interval; LL = lower limit; UL = upper limit. Models are adjusted for age, sex, residential district, education attainment, and date of the questionnaire conducted. Bolded numbers indicate statistically significant results.
Table 3 shows that there is a joint effect between the status of living alone and frailty on the relationship with any heat-related signs/symptoms reported on the additive scale (RERI 0.22; 95% CI: 0.02, 0.41) as well as on the multiplicative scale (ratio of RR: 1.31; 95% CI: 1.03, 1.66). Our findings did not seem to suggest that there is interaction, either on an additive or multiplicative scale, for other housing characteristics and frailty on heat-related health outcomes. Table 4 shows that, when stratified by frailty status, only for those who were robust, living alone was related to a lower risk of reporting any heat-related signs/symptoms (RR: 0.74; 95% CI: 0.59, 0.93), but not for those who were frail (RR: 0.96; 95% CI: 0.87, 1.05). Table 5 shows results on the flip side—for respondents who lived alone, frailty was related to a higher risk of reporting any heat-related signs/symptoms (RR: 1.60; 95% CI: 1.28, 2.01) than those who did not live alone (RR: 1.22; 95% CI: 1.13, 1.33). Notably, despite the RERI and the interaction term being statistically non-significant, the relationship between frailty and poor self-rated health appeared to be stronger in those who owned an air-conditioner (RR: 10.02; 95% CI: 5.12, 19.62) than those who did not own one (RR: 3.83; 95% CI: 1.01, 14.53). On the contrary, for respondents who did not own an air conditioner, frailty appeared to be related to a slightly higher risk of reporting any signs/symptoms of heat-related illnesses (RR: 1.72; 95% CI: 1.12, 2.64) than owners (RR: 1.26; 95% CI: 1.16, 1.37). When we focused on owners of air conditioners, a higher number of days in a week using air-conditioners was related to a lower chance of reporting better thermal comfort at home without air-conditioners and a higher risk of any signs/symptoms of heat-related illnesses (Supplementary Table 1).
Table 3.
Assessment of additive and multiplicative interaction between frailty and housing characteristics on heat-related health outcomes (n = 1,393).
| Outcomes | Small flat | Living alone | Inadequate housing | Have air conditioner at home |
|---|---|---|---|---|
| Additive interaction, RERI (95% CI) | ||||
| Poor self-rated health |
|
|
|
|
| Reported thermally comfortable at home |
|
|
|
|
| Any symptom reported |
|
|
|
|
| Multiplicative interaction, Ratio of RR (95% CI) | ||||
| Poor self-rated health |
|
|
NAa |
|
| Reported thermally comfortable at home |
|
|
|
|
| Any symptom reported |
|
|
|
|
Note. RERI = relative excess risk due to interaction; CI = confidence interval; RR = relative risks. Models are adjusted for age, sex, residential district, education attainment, and date of the questionnaire conducted. Bolded numbers indicate statistically significant results.
Not applicable; model did not converge.
Table 4.
Associations between housing characteristics and heat-related health outcomes, stratified by frailty status (n = 1,393).
| Outcomes | Frailty status |
|||||
|---|---|---|---|---|---|---|
| Frail |
Robust |
|||||
| RR | 95% CI |
RR | 95% CI |
|||
| LL | UL | LL | UL | |||
| Small flat | ||||||
| Poor self-rated health | 0.95 | 0.71 | 1.26 | 1.09 | 0.29 | 4.06 |
| Reported thermally comfortable at home | 0.87 | 0.70 | 1.07 | 0.80 | 0.64 | 1.00 |
| Any symptom reported | 1.03 | 0.95 | 1.12 | 1.11 | 0.97 | 1.28 |
| Living alone | ||||||
| Poor self-rated health | 1.21 | 0.90 | 1.64 | 1.54 | 0.32 | 7.47 |
| Reported thermally comfortable at home | 1.11 | 0.89 | 1.38 | 1.41 | 1.13 | 1.75 |
| Any symptom reported | 0.96 | 0.87 | 1.05 | 0.74 | 0.59 | 0.93 |
| Inadequate housing | ||||||
| Poor self-rated health | 0.68 | 0.35 | 1.29 | NAa | NA | NA |
| Reported thermally comfortable at home | 0.93 | 0.58 | 1.47 | 1.29 | 0.68 | 2.45 |
| Any symptom reported | 0.99 | 0.84 | 1.16 | 0.97 | 0.66 | 1.42 |
| Having air conditioning at home | ||||||
| Poor self-rated health | 0.60 | 0.38 | 0.94 | 0.23 | 0.04 | 1.19 |
| Reported thermally comfortable at home | 0.52 | 0.41 | 0.67 | 0.62 | 0.45 | 0.86 |
| Any symptom reported | 0.96 | 0.82 | 1.12 | 1.23 | 0.83 | 1.84 |
Note. RR = relative risks; CI = confidence interval; LL = lower limit; UL = upper limit. Models are adjusted for age, sex, residential district, education attainment, and date of the questionnaire conducted. Bolded numbers indicate statistically significant results.
Not applicable; model did not converge.
Table 5.
Associations between frailty and heat-related health outcomes, stratified by housing characteristics (n = 1,393).
| Outcomes | Housing characteristics |
|||||
|---|---|---|---|---|---|---|
| Yes |
No |
|||||
| RR | 95% CI |
RR | 95% CI |
|||
| LL | UL | LL | UL | |||
| Poor self-rated health | ||||||
| Small flat | 8.31 | 3.02 | 22.86 | 9.47 | 4.44 | 20.18 |
| Living alone | 7.32 | 2.36 | 22.69 | 9.60 | 4.67 | 19.75 |
| Inadequate housing | NAa | NAa | NAa | 9.01 | 4.90 | 16.57 |
| Having air-conditioning at home | 10.02 | 5.12 | 19.62 | 3.83 | 1.01 | 14.53 |
| Reported thermally comfortable at home | ||||||
| Small flat | 0.84 | 0.64 | 1.09 | 0.73 | 0.61 | 0.86 |
| Living alone | 0.61 | 0.47 | 0.80 | 0.80 | 0.68 | 0.95 |
| Inadequate housing | 0.63 | 0.30 | 1.31 | 0.76 | 0.66 | 0.88 |
| Having air-conditioning at home | 0.74 | 0.64 | 0.86 | 0.84 | 0.57 | 1.24 |
| Reported any heat-related symptoms | ||||||
| Small flat | 1.21 | 1.07 | 1.36 | 1.32 | 1.20 | 1.47 |
| Living alone | 1.60 | 1.28 | 2.01 | 1.22 | 1.13 | 1.33 |
| Inadequate housing | 1.22 | 0.83 | 1.80 | 1.29 | 1.19 | 1.39 |
| Having air-conditioning at home | 1.26 | 1.16 | 1.37 | 1.72 | 1.12 | 2.64 |
Note. RR = relative risks; CI = confidence interval; LL = lower limit; UL = upper limit. Models adjusted for age, sex, residential district, education attainment, and date of the questionnaire conducted. Bolded numbers indicate statistically significant results.
Not applicable; model does not converge.
Discussion
In a subtropical urban setting, our study, which consisted of a sample of older people in two major residential districts in Hong Kong, showed that frailty was strongly associated with poor self-rated health, lower thermal comfort at home, and a higher risk of heat-related symptoms. Notably, owning an air-conditioner was related to a lower risk of reporting poor self-rated health and better thermal comfort, particularly for those who were frail. Living in a crowded flat was related to poorer thermal comfort. Older people who lived alone had lower risks of reporting any signs or symptoms of heat-related illnesses, but this was only evident in those who were robust. The present study elucidates the complex interplay between frailty, precarious housing conditions, and heat-related health outcomes among community-dwelling older adults in Hong Kong, a rapidly urbanizing subtropical city facing escalating climate challenges.
Our findings underscore that frailty independently confers a substantial risk for poor self-rated health, diminished thermal comfort at home, and increased heat-related symptoms during extreme heat events. These results are consistent with prior research indicating that frail older adults possess diminished physiological reserves and impaired thermoregulatory capacity, rendering them particularly vulnerable to heat stress (Ndlovu & Chungag, 2024). For instance, chronological ageing is associated with attenuated cardiovascular responses to heat, including reduced skin blood flow and impaired vasodilation, which compromises heat dissipation and increases susceptibility to heat-related illnesses (Balmain et al., 2018). Frailty exacerbates these deficits by further impairing cardiovascular, pulmonary, and musculoskeletal functions, limiting the ability to adapt to thermal stress. Functionally, frail individuals often experience mobility limitations and cognitive impairments, which hinder their capacity to seek cooler environments or implement protective behaviors during heatwaves (Ji et al., 2025). These physiological and functional constraints collectively heighten the risk of heat exhaustion, dehydration, and related adverse outcomes.
Housing conditions emerged as critical environmental determinants of heat vulnerability (Samuelson et al., 2020). Our data suggest that living in small and crowded flats was significantly related to worse thermal comfort at home and a higher risk of reporting heat-related signs or symptoms. The psychosocial stress associated with overcrowding may have exacerbated health risks during heat events (Zhang et al., 2023). Our data did not show that inadequate housing is related to heat-related signs or symptoms or has any synergistic effect with frailty. Previous studies in Hong Kong showed that inadequate housing, especially the subdivided units, was strongly related to worse thermal comfort (Cheung & Jim, 2020). This discrepancy might arise because only 6% of our respondents lived in inadequate housing, which may have limited the statistical power to detect such effects. Nonetheless, this study showed that housing design might play a role in indoor thermal comfort. Our data reported that respondents in the Yau Tsim Mong district reported poorer thermal comfort than those who lived in the Kwun Tong district, the former featuring very high population density and a relatively high proportion of subdivided units in Hong Kong, whereas the latter has a high proportion of older population living in public housing. Our previous study showed that older people living in public housing estates were generally satisfied with the thermal comfort at home, possibly due to better ventilation design in general (Lai, Chau, et al., 2023). Further studies are warranted to observe whether housing design synergistically impacts frailty and health or other subjective outcomes.
Interestingly, living alone was related to better self-reported thermal comfort, but it also showed a lower risk of heat-related symptoms. This paradox may reflect behavioral adaptations among solitary older adults, such as increased time spent in air-conditioned public spaces to avoid indoor heat, as it has been reported that in Hong Kong, older residents often rely on malls for free cooling due to high electricity costs in subdivided flats (Tsang & Li, 2024). However, social isolation remains a recognized risk factor for adverse heat outcomes, as it limits access to assistance and health information during heat emergencies (Browning et al., 2006). The joint effect observed between living alone and frailty on heat-related symptoms highlights the compounded vulnerability of socially isolated frail older adults, emphasizing the need for targeted social support interventions.
The protective role of air conditioning was evident, with ownership associated with a lower risk of poor self-rated health. Air conditioning use is a well-established adaptive strategy against heat exposure, capable of significantly reducing heat-related morbidity and mortality (Lundgren-Kownacki et al., 2018). Nevertheless, the benefits of air conditioning are tempered by economic barriers, especially among low-income older adults residing in subdivided units, where electricity costs could be prohibitively high. We have shown in a previous study that older people tend not to use air conditioning due to the concern of costs (Lai, Chau, et al., 2023). This economic constraint leads to underutilization of air conditioning and reliance on less effective cooling strategies, exacerbating heat vulnerability. Our data in this study showed that 95% of the respondents owned at least one air conditioner at home, yet around 20% of the respondents reported that they used air conditioning for fewer than or equal to 2 days a week. This study was limited by not having the information on the reason for their not using air conditioning; nonetheless, this finding implied that this could potentially present as a health inequity issue. Moreover, the environmental impact of widespread AC use, including anthropogenic heat release contributing to urban heat island effects, necessitates balanced public health policies that promote equitable access while mitigating environmental consequences (O’Neill, 2003).
Our findings can be interpreted through the social-ecological framework outlined in the Introduction, which posits that health outcomes arise from the interplay between factors at the individual, environmental, and social levels. At the individual level, our results strongly affirm the role of frailty as a key determinant of intrinsic vulnerability. The consistent associations between frailty and all three adverse heat-related outcomes—poor self-rated health, lower thermal comfort, and increased symptoms—underscore that individuals with diminished physiological reserves are fundamentally more susceptible to the stress of extreme heat. At the environmental level, our findings illustrate how the built environment acts as a critical modulator of exposure and adaptive capacity. The association of living in a small, crowded flat with poorer thermal comfort demonstrates how housing quality directly shapes physical heat exposure. Conversely, the protective effect of owning an air conditioner on self-rated health highlights its role as a key resource for enhancing adaptive capacity, even if its use is limited by other factors. Most importantly, our results provide empirical evidence for a synergistic interaction between levels, a core tenet of the social-ecological model. The finding that living alone was protective against heat-related symptoms only for robust individuals exemplifies this complex interplay. This suggests an “amplified vulnerability” mechanism for the frail group. While robust individuals living alone may have the agency and mobility to utilize cooling centers or public spaces, frail individuals may be limited by functional or cognitive impairments. For them, living alone may translate to social isolation without the capacity for protective behavioral adaptation, thus negating any potential benefits and compounding their vulnerability. This finding demonstrates that interventions cannot target individual or environmental factors in isolation; rather, they must address the specific ways in which they interact to create disproportionate risk.
The confluence of global warming, rapid population ageing, high-density urban living, and widespread economic inequality creates a perfect storm for heat-related health crises among older adults. We have also previously shown that, in Hong Kong, more recent cohorts of older adults had higher levels of frailty than did earlier cohorts (Yu et al., 2017). The current study contributes to the limited but growing body of evidence highlighting the synergistic effects of frailty and precarious housing on heat-health outcomes, a critical gap in the literature particularly relevant to subtropical megacities. To the best of our knowledge, the current study is the first study to assess such a synergistic effect. We also used a well-validated inventory to identify frailty in a sample of community-dwelling older people.
Nonetheless, this study has several limitations. First, this study is a cross-sectional study and cannot ascertain cause and effect between housing characteristics, frailty, and heat-related health outcomes. It could also be plausible that health conditions could affect the respondents’ living conditions. Further study with follow-up surveys in a longitudinal design may help to detect the causal effect. Second, housing characteristics were only represented by the measures of small and crowded flats, living alone, housing type, and possession of air conditioning. Other factors such as flat orientation, window presence and usage, or other cooling devices were not considered. Moreover, housing characteristics in combination, rather than in isolation, could offer more comprehensive insights, but this was not feasible due to the near-ubiquitous ownership of air conditioning and the small proportion of respondents living in inadequate housing. Our current sample was aligned with a previous study in Hong Kong, which also showed a very high ownership of air conditioning (Gao et al., 2020). Future studies could employ stratified sampling to purposively recruit more respondents living in inadequate housing for comparison. Third, convenience sampling of pedestrians at specific points in the two districts may not represent homebound or socially isolated older adults, who may be more vulnerable to heat stress and inadequate housing. More purposive sampling, potentially aided by local community organizations, is warranted. Fourth, we have employed a complete-case analysis approach to conduct the analyses. Nonetheless, the missingness of variables only ranged from 0% to 7%. We therefore reckoned that the impact of possible selection bias on the estimates was minimal.”
The generalizability of our findings warrants careful consideration. Certain aspects of our results are deeply rooted in the unique context of Hong Kong and may be most directly applicable to other high-density, subtropical Asian megacities with significant wealth inequality. For instance, the specific challenge of subdivided units as a primary form of inadequate housing, and the socio-economic paradox of high air conditioner ownership but low utilization due to electricity costs, are characteristic of Hong Kong’s extreme urban environment. Nonetheless, our core findings—that frailty, a state of diminished physiological reserve, significantly amplifies an older adult’s vulnerability to heat stress—are a biological reality that transcends geography. The synergistic effect between physical vulnerability and environmental adversity illustrates the versatility of the theoretical framework of the social determinants of health as well as the social-ecological model of health (WHO Commission on Social Determinants of Health, 2008). While the specific nature of precarious housing or social risk may differ—for example, poorly insulated rural homes in a temperate climate versus a subdivided urban flat—the overarching conclusion that the built and social environments critically mediate heat-related health risks for frail older adults is broadly applicable.
From a public health perspective, our findings underscore the urgency of multifaceted and targeted interventions, particularly for the frail individuals and those living alone. Early identification and continuous monitoring of frailty among older adults can enable timely preventive measures. Housing policies should prioritize improving ventilation and access to affordable cooling strategies in subdivided and inadequate housing units. Social programs must address isolation by fostering community networks and providing targeted support during heatwaves. Furthermore, public education campaigns are needed to raise awareness about heat risks and adaptive behaviors.
In conclusion, this study highlights the intertwined roles of individual frailty, housing precarity, and social factors in shaping heat-related health risks among older adults in Hong Kong. Addressing these multifactorial vulnerabilities requires integrated strategies spanning healthcare, housing, social services, and climate adaptation policies. Such efforts are vital to safeguarding the health and well-being of ageing urban populations in the face of escalating climate threats.
Supplementary Material
Acknowledgments
The authors thank the participants of the survey.
Contributor Information
Eric T C Lai, CUHK Institute of Health Equity, The Chinese University of Hong Kong, Hong Kong SAR, China; Department of Medicine and Therapeutics, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China.
Wui Ling Chu, Jockey Club Institute of Ageing, The Chinese University of Hong Kong, Hong Kong SAR, China.
Jean Woo, CUHK Institute of Health Equity, The Chinese University of Hong Kong, Hong Kong SAR, China; Department of Medicine and Therapeutics, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, China.
Supplementary material
Supplementary data are available at Innovation in Aging online.
Funding
This work was supported by the Hong Kong Jockey Club Charity Trusts (grant number 2024-0058-003).
Conflict of interest
None declared.
Data availability
Data from this study cannot be made available to other researchers for reasons of protecting the anonymity of the participants and because they did not provide consent for their raw data to be shared publicly. Analytic methods or materials are available to other researchers for replication purposes. This study was not preregistered.
Author contributions
Eric T. C. Lai and Jean Woo were responsible for conceptualizing this manuscript. E. T. C. Lai was responsible for reviewing the literature, data curation, and writing of the manuscript. W. L. Chu was responsible for coordinating the administration of the questionnaire and data cleaning. All co-authors participated in critical review and intellectual input to the drafts of the manuscript.
References
- ASHRAE. (2010). Standard 55—thermal environmental conditions for human occupancy. American Society of Heating, Refrigerating and Air Conditioning Engineers, Inc. https://www.ashrae.org/technical-resources/bookstore/standard-55-thermal-environmental-conditions-for-human-occupancy
- Balmain B. N., Sabapathy S., Louis M., Morris N. R. (2018). Aging and thermoregulatory control: The clinical implications of exercising under heat stress in older individuals. Biomed Research International, 2018, 8306154. 10.1155/2018/8306154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bronfenbrenner U. (1977). Toward an experimental ecology of human development. American Psychologist, 32, 513. 10.1037/0003-066X.32.7.513 [DOI] [Google Scholar]
- Browning C. R., Wallace D., Feinberg S. L., Cagney K. A. (2006). Neighborhood social processes, physical conditions, and disaster-related mortality: The case of the 1995 Chicago heat wave. American Sociological Review, 71, 661–678. 10.1177/000312240607100407 [DOI] [Google Scholar]
- Census and Statistics Department. (2020). Hong Kong population projections 2020–2069.
- Census and Statistics Department. (2025, May 7). Table 110-02001: Land area, mid-year population and population density by District Council district. https://www.censtatd.gov.hk/en/web_table.html?id=110-02001
- Cheung P. K., Jim C. Y. (2020). Summer thermal discomfort in substandard housing with openable windows in Hong Kong. Science and Technology for the Built Environment, 27, 462–476. 10.1080/23744731.2020.1855934 [DOI] [Google Scholar]
- Dent E., Ambagtsheer R. C., Beilby J., Stewart S. (2020). Editorial: Frailty and seasonality. Journal of Nutritional Health and Aging, 24, 547–549. 10.1007/s12603-020-1367-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ebi K. L., Capon A., Berry P., Broderick C., de Dear R., Havenith G., Honda Y., Kovats R. S., Ma W., Malik A. (2021). Hot weather and heat extremes: Health risks. The Lancet, 398, 698–708. 10.1016/s0140-6736(21)01208-3 [DOI] [PubMed] [Google Scholar]
- Estoque R. C., Ooba M., Seposo X. T., Togawa T., Hijioka Y., Takahashi K., Nakamura S. (2020). Heat health risk assessment in Philippine cities using remotely sensed data and social-ecological indicators. Nature Communications, 11, 1581. 10.1038/s41467-020-15218-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fried L. P., Cohen A. A., Xue Q.-L., Walston J., Bandeen-Roche K., Varadhan R. (2021). The physical frailty syndrome as a transition from homeostatic symphony to cacophony. Nature Aging, 1, 36–46. 10.1038/s43587-020-00017-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gao Y., Chan E. Y., Lam H. C., Wang A. (2020). Perception of potential health risk of climate change and utilization of fans and air conditioners in a representative population of Hong Kong. International Journal of Disaster Risk Science, 11, 105–118. 10.1007/s13753-020-00256-z [DOI] [Google Scholar]
- Gil Cuesta J., Van Loenhout J. A. F., Colaço M. d. C., Guha-Sapir D. (2017). General population knowledge about extreme heat: A cross-sectional survey in Lisbon and Madrid. International Journal of Environmental Research and Public Health, 14, 122. 10.3390/ijerph14020122 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Golden S. D., Earp J. A. L. (2012). Social ecological approaches to individuals and their contexts: Twenty years of health education and behavior health promotion interventions. Health Education & Behavior, 39, 364–372. 10.1177/1090198111418634 [DOI] [PubMed] [Google Scholar]
- Hosmer D. W., Lemeshow S. (1992). Confidence interval estimation of interaction. Epidemiology, 3, 452–456. https://doi.org/10.1097/00001648-199209000-00012 [DOI] [PubMed] [Google Scholar]
- Hua J., Zhang X., Ren C., Shi Y., Lee T.-C. (2020). Spatiotemporal assessment of extreme heat risk for high-density cities: A case study of Hong Kong from 2006 to 2016. Sustainable Cities and Society, 64, 102507. 10.1016/j.scs.2020.102507 [DOI] [Google Scholar]
- Ji H., Shin S. H., Coronado A., Lee H. Y. (2025). Extreme heat, functional disability, and social isolation: Risk disparity among older adults. Journal of Applied Gerontology, 44, 561–570. 10.1177/07334648241282904 [DOI] [PubMed] [Google Scholar]
- Lai D., Lian Z., Liu W., Guo C., Liu W., Liu K., Chen Q. (2020). A comprehensive review of thermal comfort studies in urban open spaces. Science of the Total Environment, 742, 140092. 10.1016/j.scitotenv.2020.140092 [DOI] [PubMed] [Google Scholar]
- Lai E., Ho S. C., Woo J. (2023). Social isolation, socioeconomic status, and development of functional impairments in Chinese older adults aged 70 years and over: A cohort study. Aging Clinical and Experimental Research, 35, 155–165. 10.1007/s40520-022-02259-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lai E. T., Chau P. H., Cheung K., Kwan M., Lau K., Woo J. (2023). Perception of extreme hot weather and the corresponding adaptations among older adults and service providers—A qualitative study in Hong Kong. Frontiers in Public Health, 11, 1–11. 10.3389/fpubh.2023.1056800 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lai E. T.-C., Ho I. Y.-Y., Ho H. C., Chau P.-H., Yip T. C.-F., Wong G. L.-H., Woo J. (2025). Extreme hot weather events and risk of hospitalisation for cardiovascular and respiratory diseases in older people in Hong Kong in 2012 to 2018. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 80, glaf002. 10.1093/gerona/glaf002 [DOI] [PubMed] [Google Scholar]
- Lee J., Abdel-Kader K., Yabes J. G., Cai M., Chang H. H., Jhamb M. (2021). Association of self-rated health with functional limitations in patients with CKD. Kidney and Medicine, 3, 745–752.e741. 10.1016/j.xkme.2021.04.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lomas K. J. (2021). Summertime overheating in dwellings in temperate climates. Buildings & Cities, 2. 10.5334/bc.128 [DOI] [Google Scholar]
- Lundgren-Kownacki K., Hornyanszky E. D., Chu T. A., Olsson J. A., Becker P. (2018). Challenges of using air conditioning in an increasingly hot climate. International Journal of Biometeorology, 62, 401–412. 10.1007/s00484-017-1493-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mackenbach J. P., Borst V., Schols J. M. (1997). Heat-related mortality among nursing-home patients. The Lancet, 349, 1297–1298. 10.1016/s0140-6736(05)62510-x [DOI] [PubMed] [Google Scholar]
- Morley J. E., Malmstrom T., Miller D. (2012). A simple frailty questionnaire (FRAIL) predicts outcomes in middle aged African Americans. The Journal of Nutrition, Health & Aging, 16, 601–608. 10.1007/s12603-012-0084-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ndlovu N., Chungag B. N. (2024). Impact of heat stress on cardiovascular health outcomes of older adults: A mini review. Aging and Health Research, 100189. 10.1016/j.ahr.2024.100189 [DOI] [Google Scholar]
- O’Neill M. S. (2003). Air conditioning and heat-related health effects. Applied Environmental Science and Public Health, 1, 9–12. [Google Scholar]
- Ruiz-Tagle J., Urria I. (2022). Household overcrowding trajectories and mental well-being. Social Science & Medicine, 296, 114051. 10.1016/j.socscimed.2021.114051 [DOI] [PubMed] [Google Scholar]
- Samuelson H., Baniassadi A., Lin A., González P. I., Brawley T., Narula T. (2020). Housing as a critical determinant of heat vulnerability and health. Science of the Total Environment, 720, 137296. 10.1016/j.scitotenv.2020.137296 [DOI] [PubMed] [Google Scholar]
- Shi Y., Ren C., Cai M., Lau K. K.-L., Lee T.-C., Wong W.-K. (2019). Assessing spatial variability of extreme hot weather conditions in Hong Kong: A land use regression approach. Environmental Research, 171, 403–415. 10.1016/j.envres.2019.01.041 [DOI] [PubMed] [Google Scholar]
- Tsang Y., Li S. (2024). Hong Kong’s sweltering summer: Poor elderly struggle to stay cool in subdivided flats. South China Morning Post. https://www.scmp.com/news/hong-kong/society/article/3274714/hong-kongs-sweltering-summer-poor-elderly-struggle-stay-cool-subdivided-flats [Google Scholar]
- VanderWeele T. J., Knol M. J. (2014). A tutorial on interaction. Epidemiologic Methods, 3, 33–72. 10.1515/em-2013-0005 [DOI] [Google Scholar]
- Whitman S., Good G., Donoghue E. R., Benbow N., Shou W., Mou S. (1997). Mortality in Chicago attributed to the July 1995 heat wave. American Journal of Public Health, 87, 1515–1518. 10.2105/AJPH.87.9.1515 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wong H., Chan S. M. (2019). The impacts of housing factors on deprivation in a world city: The case of Hong Kong. Social Policy & Administration, 53, 872–888. 10.1111/spol.12535 [DOI] [Google Scholar]
- Woo J., Yu R., Wong M., Yeung F., Wong M., Lum C. (2015). Frailty screening in the community using the FRAIL scale. Journal of the American Medical Directors Association, 16, 412–419. 10.1016/j.jamda.2015.01.087 [DOI] [PubMed] [Google Scholar]
- Yu R., Wong M., Chong K. C., Chang B., Lum C. M., Auyeung T. W., Lee J., Lee R., Woo J. (2017). Trajectories of frailty among Chinese older people in Hong Kong between 2001 and 2012: An age-period-cohort analysis. Age and Ageing, 47, 254–261. 10.1093/ageing/afx170 [DOI] [PubMed] [Google Scholar]
- Zhang Z., Měchurová K., Resch B., Amegbor P., Sabel C. E. (2023, Jun 21). Assessing the association between overcrowding and human physiological stress response in different urban contexts: A case study in Salzburg, Austria. Int J Health Geogr, 22, 15. 10.1186/s12942-023-00334-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zou G. (2004). A modified Poisson regression approach to prospective studies with binary data. American Journal of Epidemiology, 159, 702–706. 10.1093/aje/kwh090 [DOI] [PubMed] [Google Scholar]
- Zuo Z., Fung J. C., Li Z., Huang Y., Wong M. F., Lau A. K., Lu X. (2023). Projection of future heatwaves in the Pearl River Delta through CMIP6-WRF dynamical downscaling. Journal of Applied Meteorology and Climatology, 62, 1297–1314. 10.1175/JAMC-D-22-0201.1 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data from this study cannot be made available to other researchers for reasons of protecting the anonymity of the participants and because they did not provide consent for their raw data to be shared publicly. Analytic methods or materials are available to other researchers for replication purposes. This study was not preregistered.