|
|
| |||||||||||||||||||||||||||||||||||||||||
Background During a record-setting heat wave in Chicago in July 1995, there were at least 700 excess deaths, most of which were classified as heat-related. We sought to determine who was at greatest risk for heat-related death.
Methods We conducted a case–control study in Chicago to identify risk factors associated with heat-related death and death from cardiovascular causes from July 14 through July 17, 1995. Beginning on July 21, we interviewed 339 relatives, neighbors, or friends of those who died and 339 controls matched to the case subjects according to neighborhood and age.
Results The risk of heat-related death was increased for people with known medical problems who were confined to bed (odds ratio as compared with those who were not confined to bed, 5.5) or who were unable to care for themselves (odds ratio, 4.1). Also at increased risk were those who did not leave home each day (odds ratio, 6.7), who lived alone (odds ratio, 2.3), or who lived on the top floor of a building (odds ratio, 4.7). Having social contacts such as group activities or friends in the area was protective. In a multivariate analysis, the strongest risk factors for heat-related death were being confined to bed (odds ratio, 8.2) and living alone (odds ratio, 2.3); the risk of death was reduced for people with working air conditioners (odds ratio, 0.3) and those with access to transportation (odds ratio, 0.3). Deaths classified as due to cardiovascular causes had risk factors similar to those for heat-related death.
Conclusions In this study of the 1995 Chicago heat wave, those at greatest risk of dying from the heat were people with medical illnesses who were socially isolated and did not have access to air conditioning. In future heat emergencies, interventions directed to such persons should reduce deaths related to the heat.
From July 12 through July 16, 1995, in Chicago, the maximal and minimal temperatures reached unprecedented highs, and the high temperatures were accompanied by extremes of relative humidity (Figure 1).10 Within a few days of the onset of the heat spell, the Cook County Medical Examiner's Office reported a sharp increase in the number of heat-related deaths.11
|
Methods
Study Design and Case Definition
We conducted a case–control study from July 21 through August 18, 1995. Eligible case subjects were persons older than 24 years of age who died in Chicago from July 14 through July 17, for whom the cause of death listed on the death certificate met one of three criteria: heat was listed as the immediate or underlying cause of death, with no reference to cardiovascular disease; cardiovascular disease was listed as the primary cause of death, with no reference to heat; or cardiovascular disease was listed as the primary cause and heat as a contributing cause of death. We included deaths due to cardiovascular causes in the case definition because previous studies had demonstrated an excess of deaths from cardiovascular disease during periods of high heat.1,5
In order to avoid the delay that might be involved in obtaining the death certificates of all eligible subjects, all death certificates issued in July 1995 were collected by the Vital Statistics Division of the Department of Public Health in Chicago and assigned temporary identification numbers. From a review of those death certificates, we identified 680 potential subjects who met our case definition; we then selected a random sample from each cause-of-death category, with stratification according to age, race, and date of death (Table 1). We chose the sample sizes for each category to give the study a statistical power of 80 percent for detecting an odds ratio of 2.0 among subjects with any predictive factor as compared with subjects without it, assuming that 18 percent of controls were exposed to any given risk factor, at a significance level of 0.05. We visited the residences of 420 potential case subjects and interviewed family members, neighbors, or friends of 372 of them (89 percent).
|
Using a uniform questionnaire, we obtained information by interviewing controls and surrogate respondents for the case subjects, such as family members, friends, or neighbors who were encountered when staff members visited the address on the death certificate. Additional information was obtained from the death certificate, police reports, and phone calls to the case subjects' next of kin. Information on the case subjects' living conditions was collected by the inspection of residences. Interviews were conducted by staff members of the Centers for Disease Control and Prevention and by trained volunteers recruited by the Chicago Department of Public Health and the Office of the Mayor.
Statistical Analysis
We used matched-pairs analysis to estimate odds ratios and confidence intervals for each potential risk factor and cause-of-death category.12,13 For each risk factor, we calculated a summary odds ratio that combined the three odds ratios for the cause-of-death categories. This summary odds ratio was the weighted average of the three cause-specific odds ratios, with weights calculated from the corresponding sampling fraction and the relative frequency of each cause of death. We estimated the variance for the summary measure by using the delta method14 and treating the weights as constant. In univariate analysis of data on the total population of study subjects, we used 160 potential risk factors derived from the questionnaire.
Variables that were significant in the crude, univariate analysis were entered into a conditional, stepwise logistic-regression model. The final multivariate model contained the variables that were significant predictors of death in one or more cause-of-death categories. Summary odds ratios were calculated with use of the same weights as in the crude analysis. A summary generalized impact fraction was calculated from the summary odds ratio and the weighted numbers of case subjects exposed or control subjects not exposed to a particular factor, as a measure of the impact of these factors and their relevance for public health.13 Assuming that factors that increase risk are causally related to death, the calculation estimates the percentage of deaths attributable to a particular factor. For factors that decrease risk, the generalized impact fraction estimates the number of deaths attributable to the absence of the factor in some proportion of the population.
Results
Subjects
Of 750 persons — both case subjects and controls — for whom we had completed questionnaires, 678 (90 percent), or 339 matched pairs, were included in the analysis (Figure 1 and Table 1). Case subjects whose surrogates were successfully surveyed were similar in terms of age, race, ethnic background, and sex to those whose surrogates were not surveyed. Table 1 shows selected demographic variables for case subjects and controls according to cause-of-death category. For the initial group of 680 eligible case subjects, the median age was 76 years. The proportions of black and white subjects were approximately equal among the case subjects and controls, and case subjects were generally older than controls in each cause-of-death category; the age difference was greatest between persons who died from cardiovascular causes and their controls.
Living Conditions and Social Contacts
There were fewer deaths among people who had a working air conditioner (odds ratio, 0.2; 95 percent confidence interval, 0.2 to 0.4) or had access to an air-conditioned lobby if the subject lived in an apartment building (odds ratio, 0.2; 95 percent confidence interval, 0.1 to 0.5) than among those who did not (Table 2). Visiting an air-conditioned place during the heat wave, other than the subject's residence, was associated with lower mortality (odds ratio, 0.3; 95 percent confidence interval, 0.2 to 0.5). We did not detect any reduction in mortality in association with the use of electric fans (data not shown).
|
The case subjects lived in homes with fewer rooms, on average, than did the controls. In addition, case subjects were less likely to leave home frequently, to have friends in Chicago, to have pets, and to participate in group activities than were controls.
Medical Factors
Preexisting medical conditions were associated with an increased risk of death during the heat wave (Table 3). Confinement to bed was the strongest predictor of death; 16 percent of case subjects were bedridden, as compared with only 4 percent of controls (odds ratio, 5.5). The need for assistance from visiting nurses was associated with a mortality rate approximately six times higher than that among people who did not receive such assistance (odds ratio, 6.2); housekeepers, home health aides, and Meals on Wheels programs were also used by subjects at increased risk (odds ratios, 2.5 to 2.7). The city of Chicago sent social-service workers to visit people at risk and explain the dangers of hot weather; a decreased risk of death was found among the people they contacted (odds ratio, 0.7).
|
|
In multivariate analyses, the strongest protective factor was having a working air conditioner in the home (odds ratio, 0.3; 95 percent confidence interval, 0.2 to 0.6) (Table 5). We estimate that more than 50 percent of the deaths related to the heat wave could have been prevented if each home had had a working air conditioner (generalized impact fraction, 50.2 percent). Also effective in reducing the risk of heat-related death were visiting an air-conditioned place (odds ratio, 0.5; 95 percent confidence interval, 0.3 to 0.9) and having access to transportation (odds ratio, 0.3; 95 percent confidence interval, 0.1 to 0.5). Confinement to bed was correlated with the inability to care for oneself and was a strong risk factor for heat-related death. Living alone, possibly indicating fewer contacts with family and friends and social isolation, was also associated with increased mortality (odds ratio, 2.3; 95 percent confidence interval, 1.2 to 4.4).
|
In an effort to identify practical intervention strategies, we attempted to determine the most effective means of communication with people at risk by considering information provided by the controls in interviews. We analyzed factors associated with a lack of awareness of the potential danger of sustained heat, since the univariate analysis had indicated that increased awareness (indicated by contact with city workers) was protective. The controls who listened to the radio were more likely than others to be aware of the health risks associated with hot weather (odds ratio, 2.1; 95 percent confidence interval, 0.8 to 5.6); the same was true of the controls who read the newspaper (odds ratio, 2.5; 95 percent confidence interval, 1.2 to 5.3). Virtually all controls watched television; thus, its contribution to their awareness of heat-related dangers could not be evaluated.
We also used information supplied by the controls to evaluate factors associated with the lack of air conditioning. Residents of apartment buildings were particularly likely to lack air conditioning (odds ratio, 3.4; 95 percent confidence interval, 2.1 to 5.4), as were subjects who lived in single-room-occupancy or other hotels (odds ratio, 6.2; 95 percent confidence interval, 0.6 to 61.4), as compared with people living in single-family houses. People were more likely to change their daily routines during hot weather — for example, by drinking extra fluids or taking extra baths — if they lived with at least one other person (odds ratio, 2.1; 95 percent confidence interval, 1.3 to 3.4) or if they had a pet (odds ratio, 2.1; 95 percent confidence interval, 1.3 to 3.5).
Discussion
The people who are at greatest risk for death due to heat or to cardiovascular causes include those in frail health, such as elderly persons, and those who are socially isolated. People at risk were often in need of help from visiting nurses, housekeepers, or Meals on Wheels programs — an association that could have implications for preventive public health programs. Living conditions, including the type of building, the floor level, and the number of rooms, were also found to be important determinants of risk. The presence of air conditioning was inversely associated with mortality from both heat and cardiovascular causes.
Limitations
As is true in other studies that rely on information provided by surrogate respondents (in this case, family members, neighbors, or friends of people who had died), our results may underestimate the risk associated with social isolation, since people with few social contacts (and hence no identifiable surrogate) were excluded from the study. We did not find any demographic differences between the case subjects for whom surrogate respondents were available and those for whom they were not; however, we have no details of the lifestyles of the case subjects without surrogates. Nonetheless, even after the exclusion of this most isolated segment of the population, we found a significant association between risk and variables measuring the degree of social isolation. Similarly, subjects interviewed personally and surrogate respondents are not equally credible, particularly in the case of socially isolated persons. Therefore, information concerning the case subjects' personal behavior and awareness and knowledge of health risks was excluded.
A limitation inherent in the use of death certificates is the potential misclassification of causes of death.15,16,17 Almost all death certificates listing heat as a cause of death were signed by personnel from the Medical Examiner's Office rather than by private health care providers. Despite the possibility of misclassification, we found that risk factors overlapped strongly between persons whose deaths were listed as related to heat and those whose deaths were listed as due solely to cardiovascular disease.
Strategies for Prevention
Home health care workers, friends, and the media can be effective in communicating health-protection messages directly to high-risk groups. Some of the people at highest risk of death during a heat wave are those who, as a consequence of preexisting illness, have ongoing connections to community services — through visiting nurses or Meals on Wheels personnel, for example. During periods of hot weather, these and similar public health programs provide an ideal opportunity for intervention. Media coverage of appropriate preventive measures can also be effective in increasing the public's awareness of the dangers of heat. Successful messages should encourage the modification of routine behavior in the direction of prevention.
Certain housing conditions may also entail a substantial risk of heat-related death. Living in an apartment was associated with increased risk, particularly if no air conditioner was available in the lobby. Socioeconomic differences that manifest themselves in housing may confound these findings to some extent; for example, the lack of a working air conditioner was more common among inhabitants of apartment buildings than among inhabitants of private homes. This information points to the importance of directing outreach efforts toward apartment dwellers.
Previous research has yielded conflicting results regarding the effects of electric fans in hot weather1,18,19,20,21 (and Steadman RG: personal communication). In our study, we did not find any evidence that the use of fans was protective, nor did we determine that any level of use of electric fans was associated with increased mortality. Interpretation of the data on the use of fans is complicated by the need to take into account specific environmental factors (for example, whether the fan is used in a room with an open or a closed window) and the health status of individual subjects. The effectiveness of fans in preventing death during periods of high heat and humidity remains a matter of controversy and deserves further attention; nonetheless, the potential preventive value and effectiveness of electric fans seem minimal as compared with those of other clearly beneficial factors, such as increasing use of air conditioning.
The preexisting medical conditions of individual Chicagoans greatly affected how well they were able to deal with last summer's heat wave. Although obesity is usually associated with a decreased ability to acclimate to heat,22,23 we did not find an association between mortality and a body-mass index (defined as the weight in kilograms divided by the square of the height in meters) above the normal range. We did find that people with a body-mass index below the normal range had increased risk. For the elderly, a body-mass index lower than 24 may indicate a debilitating illness or poor nutritional status. In all three cause-of-death categories, subjects with diagnosed heart or kidney disease or mental illness had an increased risk of death. Other researchers have found that neuroleptic drugs can interfere with the physiologic response to extremes of temperature and therefore increase the risk of heat-related effects on health.1,24 We found that of the 59 case subjects who were at increased risk because of preexisting mental problems, 2 (3 percent) were taking some form of neuroleptic medication. Because of their medical problems, many of the people at highest risk have an established relationship with a health care provider; such providers may be able to increase patients' awareness of the dangers of heat and convey messages about preventive measures, even during annual visits that do not occur during hot weather.
Although previous studies have consistently confirmed that there is an excess of deaths due to cardiovascular causes during periods of high or prolonged heat and have supported the biologic plausibility of the exacerbation of existing ischemic heart disease by heat, little information has been available about specific risk factors for heat-related death3,4,25,26,27,28,29,30,31 (and Wainright S, et al.: personal communication). From our study we know that people who died of cardiovascular disease were somewhat older than those whose death certificates listed heat as the underlying or secondary cause of death (average, 76 vs. 70 years; P<0.001) and had a greater frequency of preexisting medical conditions (84 percent vs. 73 percent, P<0.05). Although the magnitude of risk varied among the cause-of-death categories, we found similar risk factors in these categories. Measures to prevent heat-related death are particularly important for people with cardiovascular disease.
Access to air-conditioned environments is the factor with the greatest protective effect with respect to heat-related mortality.1,8 We found that people who lived in apartments without air conditioning had a lower risk if they had access to an air-conditioned lobby. These results have important implications for public health. Providing air conditioning to common areas in buildings or in the immediate neighborhood or providing readily accessible transportation to public cooling shelters could help to reduce excess mortality during hot weather.
Excess deaths do not typically occur until the second, third, or fourth day of a heat wave.7,10 Specific information on how to avoid heat-related illness should be included in media coverage as soon as possible along with meteorologic forecasts of heat waves.
Conclusions
The risk factors identified in this study reflect the complexity of the environmental, social, and medical components of heat-related mortality. Since heat-related deaths are preventable, targeted actions by public health professionals, health care providers, and the media can save lives each summer. People, especially elderly people, who live alone and do not have networks of social contacts and those with debilitating conditions are at particularly high risk during heat emergencies. These people need to be made aware of simple ways they can adapt their daily routines to prevent death due to hot weather; air-conditioned environments should be made readily available and accessible. During the summer heat wave of 1995 in Chicago, anything that facilitated social contact, even membership in a social club or owning a pet, was associated with a decreased risk of death.
Supported in part by a grant (S070) from the Association of Schools of Public Health (to Dr. Flanders).
We are indebted to Drs. Tiefu Shen, Cindy Whitney, Steve Whitman, Glen Good, Edmund Donoghue, Mike McGeehin, Josephine Malilay, David Olson, Byron J. Francis, and Zhen Zhao; and to Tracey Badsgard, Stefan Weir, Charlotte Harden, Deron Williams, Michael Boehmer, Chris Morris, Janet Pernarelli, Leslie Fiedler, Herminia Guerrero, Lisa Boulden, Bob Rivkin, Jean Franczyk, and Laura Landrum.
Source Information
From the Epidemic Intelligence Service, Epidemiology Program Office (J.C.S., J.D.S.), and the National Center for Environmental Health (J.C.S., C.H.R., K.H.F., J.D.S., W.D.F.), Centers for Disease Control and Prevention, Atlanta; the Illinois Department of Public Health, Springfield (H.L.H.); and the Department of Public Health, Chicago (J.L.W.).
Address reprint requests to Dr. Semenza at CDC-NCEH, 4770 Buford Hwy. (F46), Atlanta, GA 30341-3724.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Deaths in the Chicago Heat Wave
Jones D. S., Cervantes J., Semenza J. C.
Extract |
Full Text
N Engl J Med 1996;
335:1848-1850, Dec 12, 1996.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
