The Impact of Weather on Mobility and Participation in Older US Adults

Abstract

Objectives. We examined the impact of weather on the daily lives of US adults to understand which populations are most vulnerable to various weather conditions.

Methods. Data came from a 2013 supplement to the University of Michigan–Thomson Reuters Surveys of Consumers, a nationally representative telephone survey of 502 adults in the contiguous United States. We used logistic regressions to assess the odds of mobility difficulty and participation restriction during different weather conditions, as well as age group differences.

Results. Ice was most likely to change the way respondents got around (reported by 47%). In icy conditions, participants had difficulty leaving home (40%) and driving (35%). Facing ice, older adults (≥ 65 years) had twice the odds of having great difficulty leaving home (odds ratio [OR] = 2.22; 95% confidence interval [CI] = 1.12, 4.42) and curtailing work or volunteer activities (OR = 2.01; 95% CI = 1.01, 4.06), and 3 times the odds of difficulty driving (OR = 3.33; 95% CI = 1.62, 6.86) as younger respondents. We also found significant differences in mobility and participation by gender and region of residence.

Conclusions. Weather can affect social isolation, health, well-being, and mortality among older US adults.

Increasing attention is being paid to the impact of extreme weather (e.g., heat waves, hurricanes) on population health.^1–7 However, the challenges of day-to-day weather have not received as much attention, particularly for older adults, who are more vulnerable to weather-related factors because of their limited mobility and social isolation.^2,8 As a result, we have little knowledge of how weather-related factors (e.g., reduced daylight hours during winter, slippery surfaces caused by rain or snow) affect mobility and independence in day-to-day life and how these effects vary across persons, regions, and seasons.

Mobility, defined as the ability to move about effectively in our surroundings,^9,10 is important for independence, quality of life, and mortality.^11–13 Independent mobility is also important for participation in day-to-day activities, such as work or volunteer activities, grocery shopping, and engaging in physical activity. The International Classification of Functioning, Disability, and Health conceptualizes mobility and participation as part of a complex biopsychosocial model of health and function that includes interactions with environmental factors, such as climate and weather, which can act as either barriers or facilitators to activities and participation.¹⁴

Studies examining the health impact of weather-related factors have tended to focus almost exclusively on mortality or morbidity as an outcome.^3,15–21 But the ability to move and travel and to participate in daily activities is important for health and is also likely to depend on weather-related factors. Some of America’s major cities receive an average of more than 2 meters of snow every year and have snow on the ground for more than 3 months each year. Previous studies showed that icy and slushy surfaces and snow banks are the most frequently named barriers to mobility among pedestrians during winter, particularly for people with underlying limitations in physical functioning.^22,23 Li et al. found that 37% of older adults in Toronto, Canada, reduced their outdoor walking during the winter when expecting slippery road conditions.²⁴ Similarly, qualitative work with older adults in Detroit, Michigan, found that a fear of falling on ice, sidewalks not cleared of snow and ice, and cold and rainy weather discouraged walking outdoors.²⁵ Car use may increase in winter for adults fearful of falling on the way to the bus stop, but can also decrease because of fear of driving on icy roads or in the dark.²⁶

However, a lack of data from a representative sample of adults across the United States limits our understanding of the weather-related factors that are important for mobility and participation across a range of activities and risk groups (e.g., by age, gender, or geographic region). We used nationally representative data to examine the impact of weather-related factors on the daily lives of US adults to understand which populations are most vulnerable to weather conditions and exactly what those conditions are.

METHODS

Data came from the University of Michigan–Thomson Reuters Surveys of Consumers (SCA; formerly the Survey of Consumer Attitudes and Behavior). The SCA is a nationally representative, monthly telephone survey of approximately 500 adults in the contiguous United States selected through a list-assisted random-digit-dialing frame that uses the GENESYS Sampling System.²⁷ The response rate was 27% (according to the American Association for Public Opinion Research response rate method 1²⁸), which is better than or comparable to the response rate from other national telephone surveys.²⁹ The SCA is used to develop the monthly Index of Consumer Sentiment and its subset, the Index of Consumer Expectations, which are included in the Leading Indicator Composite Index published by the US Department of Commerce’s Bureau of Economic Analysis.

In June 2013, at the conclusion of the survey, the SCA administered a supplement to 502 respondents consisting of a 10-minute series of questions on weather and mobility. The question content and wording for the supplement were developed through focus groups with community volunteers and pretesting with community-dwelling residents who represented all age groups.

Measures

The supplement first asked respondents to identify the 1 weather condition that was most likely to change the way they went about their day-to-day activities (separately for summer and winter). Options for summer weather conditions were heat, humidity, poor air quality (an ozone day), rain or thunderstorms, and tornadoes or hurricanes (if volunteered); winter weather conditions were cold temperatures, snow, ice, rain, fog, and wind. Respondents could also indicate that no weather conditions changed the way they went about their activities.

Respondents were then asked to indicate how much difficulty they had when leaving their home to go about their day-to-day activities when the weather condition they identified was present. Response options were no difficulty, a little, some, and a great deal. Respondents could also indicate that they didn’t leave home in the presence of that condition or that they took extra precautions (if volunteered), and we coded these as having a great deal of difficulty. They were also asked how much difficulty they had driving (themselves) when that weather condition was present and given the same set of response options.

The supplement also asked about participation in 3 different activities when the chosen weather condition was present: work and volunteer activities (or other activities for those not working or volunteering), grocery shopping, and outdoor exercise. For example, the question about summer and grocery shopping asked,

Thinking about the last time you had planned to go grocery shopping and (it was hot/it was humid/there was poor air quality or an ozone day/it was raining or thundering in the summer/there were risks of tornadoes or hurricanes), did you use your usual means of transportation, did you use a different means of transportation, or did you stay home on that day?

We created a binary indicator of participation in grocery shopping (and work–volunteer activities) by contrasting those who engaged in the activity (either by usual or different means of transportation) with those who stayed home. Similarly, we created a binary indicator of participation in outdoor physical activity by contrasting those who exercised as planned in that weather condition with those who did not (either by skipping exercise or exercising indoors instead).

Because the consequences of weather have been shown to be greater among those with underlying limitations in physical functioning,^22,23 we used a binary indicator of mobility impairment derived from the self-reported inability to walk a half mile independently (without assistive technology). Respondents were also asked whether they limited their driving to nearby places (within 50 miles) or did not drive.

Other key covariates were age (categorized as < 65 or ≥ 65 years for analyses), gender, and race/ethnicity (non-Hispanic Black, non-Hispanic White, and other [American Indian, Asian, or Pacific Islander]). For our analyses, we categorized education, measured in years, into 3 categories (< high school, high school diploma or some college, and college degree). Because of the small proportion with less than a high school education, we also created a binary indicator contrasting college degree with less than a college degree. Respondents self-reported annual household income in dollars. For the 31 respondents (6.2%) with missing data on income, we assigned the mean value. We categorized region of residence as Northeast, South, West, or Midwest.

Statistical Analyses

We used logistic regression to examine the odds of mobility difficulty (reporting a great deal of difficulty leaving home vs none–a little–some) and participation restriction (staying home from work or volunteer activities, grocery shopping, and outdoor physical activity) in the face of weather conditions. Models also tested for differences in the effects of weather conditions by age group.

We weighted all analyses by the SCA adult weight, which made the results representative of all US adults living in private households. The weight adjusted for unequal selection probability, survey nonresponse, panel attrition, age, and income. We conducted analyses with SAS version 9.3 (SAS Institute, Inc, Cary, NC).

RESULTS

Table 1 describes the characteristics of the study sample weighted to represent the community-dwelling US adult population. Slightly more than half (54%) were female; mean age was 54 years (±18; range = 18–99 years). Most (58%) were aged 55 years or older (39% were aged 18–49 years; 29%, 50–64 years; 32%, ≥ 65 years), 54% had a high school degree, 39% had a college degree, and 7% had not graduated high school. More than half of respondents lived in the southern (34%) or midwestern (27%) United States; 17% lived in the Northeast and 21% in the West. Non-Hispanic Whites were a large majority (77%); non-Hispanic Blacks (10%) were the second largest racial/ethnic group. Because of small numbers, we dichotomized race/ethnicity as minority (non-Hispanic Black, Hispanic, or other) versus non-Hispanic White for analyses. Mean annual household income (in 2012) was $70 000 (range = $5000–$650 000) and modeled in $1000s for analyses. Walking a half mile independently was difficult for 15% of respondents, and 36% reported self-imposed driving limitations (e.g., avoiding driving long distances).

TABLE 1—

Descriptive Statistics of Adults in Survey on Weather and Mobility: University of Michigan–Thomson Reuters Surveys of Consumers, United States, 2013

Variable	Weighted Mean (SD) or %
Age, y
Total sample	54.25 (18.17)
18–49	39.39
50–64	29.04
≥ 65	31.57
Gender
Female	54.73
Male	45.27
Race/ethnicity
Non-Hispanic White	76.84
Non-Hispanic Black	9.76
Hispanic	7.82
Other	5.58
Education
< high school	6.92
High school diploma/some college	53.64
College degree	39.44
Household income, $1000s	70.08 (59.75)
Region of residence
Northeast	17.23
South	34.26
Midwest	27.34
West	21.17
Driving limited	35.94
Mobility impairment	14.77
Weekly activities	85.12
Paid work	65.63
Volunteer work	15.49
School	2.45
Religious services	14.15
Clubs/social groups	2.28
Shop for groceries	90.34
Outdoor exercise
All types	86.55
Walking	69.46
Running	8.09
Biking	5.38
Playing sports	6.97
Gardening/yard work	6.50
Other	3.60

Note. The sample size was n = 502.

The vast majority of US adults reported participating in activities outside the home: 85% regularly left home for work, volunteering, or other activities at least once a week; 90% reported grocery shopping at least once per month; and 87% reported going outdoors to exercise at least once per week (Table 1). Working for pay (66%) or volunteering (15%) were the most common weekly activities away from home, but attending religious services was also popular (for those not working for pay or volunteering). The most prevalent type of outdoor exercise was walking (including walking the dog), reported by 69% of those who regularly exercised outdoors, followed by running (8%), biking (5%), playing sports (7%), and gardening (7%).

The winter weather condition identified as most likely to change the way adults went about their day-to-day activities was ice, chosen by 47% of participants (Table 2). We observed no significant age differences in identifying ice as an impediment, indicating a universal weather challenge for all ages. In summer, rain was the most common challenging condition for the total sample, reported by 55% of respondents. Heat was the second most problematic condition, chosen by 23% of all respondents. However, older adults were more likely than younger adults to indicate that heat changed the way they went about their daily activities (28% vs 20%) and less likely to designate rain (48% vs 58%; P < .05).

TABLE 2—

Weather Conditions Most Likely to Alter Adults' Daily Activities: University of Michigan–Thomson Reuters Surveys of Consumers, United States, 2013

Weather Condition	Total Sample (Unweighted n = 502), %	Aged < 65 Years (Unweighted n = 320), %	Aged ≥ 65 Years (Unweighted n = 182), %
Winter
Cold	11.76	11.86	11.53
Snow	21.71	22.82	19.32
Ice	46.51	44.37	51.14
Rain	9.96	11.26	7.14
Fog	4.11	5.11	1.95
Wind	2.67	2.63	2.76
None	3.29	1.95	6.17
Summer
Heat	22.60	19.97*	28.18*
Humidity	12.54	11.92	13.85
Air quality (ozone day)	4.54	5.24	3.06
Rain/thunderstorm	55.16	58.47*	48.15*
Tornado/hurricane	0.83	0.84	0.81
None	4.33	3.57	5.96

Note. Percentages are weighted to represent the US population.

*P < .05 for differences between age groups.

Although more older than younger adults reported that heat was the summer weather condition most likely to change the way they went about their day-to-day activities, very few respondents reported difficulty leaving home (7%), driving (1%), going to work or volunteer activities (3%), or shopping for groceries (8%) when it was hot (results not shown). However, the majority (62%) avoided exercising outdoors in the heat. Similarly, rain was the summer weather most likely to change the way respondents went about their daily activities, but we found little association between rain and mobility difficulty (8%) or participation restriction (9% stayed home from work or volunteer activities). However, ice was associated with mobility difficulty and participation restrictions, and we conducted separate analyses for the 231 respondents who indicated that ice was the condition most likely to change their day-to-day activities in winter.

Almost half (40%) of our sample reported a great deal of difficulty leaving home when conditions were icy, and 35% reported a great deal of difficulty driving (Figure 1). Although most respondents (56%) left home to engage in work or volunteer activities when it was icy, they tended to put off more discretionary activities, including grocery shopping (73%) and outdoor exercise (92%; Figure 2).

FIGURE 1—

Proportion of adults reporting difficulty leaving home and driving in icy conditions: University of Michigan–Thomson Reuters Surveys of Consumers, United States, 2013.

Note. Unweighted n = 231.

FIGURE 2—

Proportion of adults leaving home for activities in icy conditions: University of Michigan–Thomson Reuters Surveys of Consumers, United States, 2013.

Note. Unweighted n = 231.

We found notable age differences in the effect of ice on mobility and participation. Table 3 reports the results from weighted logistic regression models showing the odds ratios (ORs) and 95% confidence intervals (CIs) for reporting mobility difficulty and participation restrictions across different risk factors. Compared with respondents younger than 65 years, older adults had twice the odds of experiencing a great deal of difficulty leaving home (OR = 2.22; 95% CI = 1.12, 4.42) and 3 times the odds of having difficulty driving (OR = 3.33; 95% CI = 1.62, 6.86) in icy conditions. Older adults were also more likely to stay home from work or volunteer activities in the presence of ice (OR = 2.01; 95% CI = 1.01, 4.06).

TABLE 3—

Weighted Logistic Regression for Mobility and Participation Restrictions in Icy Conditions: University of Michigan–Thomson Reuters Surveys of Consumers, United States, 2013

Variable	Great Deal of Difficulty Leaving Home,a OR (95% CI)	Great Deal of Difficulty Driving,a OR (95% CI)	Stay Home From Work/Volunteer Activities,b OR (95% CI)	No Grocery Shopping, OR (95% CI)	No Outdoor Physical Activity, OR (95% CI)
Age, y
< 65 (Ref)	1.00	1.00	1.00	1.00	1.00
≥ 65	2.22* (1.12, 4.42)	3.33*** (1.62, 6.86)	2.01* (1.01, 4.06)	1.12 (0.51, 2.44)	1.41 (0.40, 5.03)
Gender
Male (Ref)	1.00	1.00	1.00	1.00	1.00
Female	1.86 (0.96, 3.61)	1.90 (0.93, 3.88)	2.27* (1.17, 4.43)	3.15*** (1.55, 6.41)	5.07** (1.48, 17.44)
Race/ethnicity
White (Ref)	1.00	1.00	1.00	1.00	1.00
Minority	1.45 (0.63, 3.32)	1.96 (0.79, 4.85)	1.78 (0.77, 4.12)	1.05 (0.38, 2.95)	4.13 (0.33, 20.08)
Education
< college (Ref)	1.00	1.00	1.00	1.00	1.00
College	1.10 (0.56, 2.13)	1.05 (0.52, 2.15)	1.22 (0.63, 2.38)	1.87 (0.86, 4.06)	0.46 (0.13, 1.56)
Household income, $1000s	1.00 (0.99, 1.01)	1.00 (0.99, 1.01)	0.99 (0.99, 1.00)	1.00 (0.99, 1.01)	1.01 (0.99, 1.02)
Region of residence
Midwest (Ref)	1.00	1.00	1.00	1.00	1.00
Northeast	2.43* (1.03, 5.75)	0.96 (0.37, 2.51)	1.78 (0.78, 4.11)	2.03 (0.81, 5.13)	1.37 (0.22, 8.42)
South	4.68*** (2.17, 10.11)	3.52*** (1.60, 7.77)	3.71*** (1.72, 7.99)	4.25** (1.67, 10.83)	1.04 (0.26, 4.12)
West	1.39 (0.47, 4.06)	1.20 (0.41, 3.51)	0.81 (0.28, 2.34)	0.95 (0.32, 2.78)	0.34 (0.06, 2.09)
Walking impairment
No (Ref)	1.00	1.00	1.00	1.00	1.00
Yes	3.09* (1.08, 8.86)	1.57 (0.52, 4.76)	1.74 (0.52, 5.91)	1.01 (0.28, 3.71)	2.69 (0.13, 57.32)
Driving limited	2.00* (1.01, 4.06)	1.99 (0.92, 4.34)	1.11 (0.52, 2.34)	2.36 (0.97, 5.81)	1.27 (0.27, 6.03)

Note. CI = confidence interval; OR = odds ratio. The sample size was n = 231.

^aReference group was no/a little/some difficulty.

^bIncluded religious services, school, and social clubs.

*P < .05; **P < .01; ***P < .001 (2-tailed tests).

Respondents with underlying mobility impairment were more likely to report difficulty leaving home (OR = 3.09; 95% CI = 1.08, 8.86), and those who usually restricted their driving to short distances had twice the odds of other respondents of reporting a great deal of difficulty leaving home when it was icy (95% CI = 1.01, 4.06). Individuals with self-imposed driving restrictions tended to report more difficulty driving in icy conditions (OR = 1.99; 95% CI = 0.92, 4.34) and to refrain from grocery shopping (OR = 2.36; 95% CI = 0.97, 5.81), although these 95% CIs included 1.0.

Not surprisingly, we detected significant regional differences in the risk of reporting mobility difficulty in the presence of ice. Relative to Midwesterners, adults living in the South had more than 4 times the odds of reporting a great deal of difficulty leaving home (OR = 4.68; 95% CI = 2.17, 10.11) and more than 3 times the odds of difficulty driving (OR = 3.52; 95% CI = 1.60, 7.77) when they faced ice. Southerners were also about 4 times as likely to stay home from work or volunteer activities (OR = 3.71; 95% CI = 1.72, 7.99) and from grocery shopping (OR = 4.25; 95% CI = 1.67, 10.83) to avoid ice. Respondents in the Northeast were also more likely than Midwesterners to report difficulty leaving home in icy conditions (OR = 2.43; 95% CI = 1.03, 5.75). We found no significant interaction effects between age and region, indicating that the regional differences in reporting mobility difficulty or participation restriction because of ice were not any greater among older adults.

Women exhibited a nonsignificant tendency to report greater difficulty with mobility (OR = 1.86; 95% CI = 0.96, 3.61) and driving (OR = 1.90; 95% CI = 0.93, 3.88) in icy conditions than did men, but the 95% CIs included 1 (Table 3). However, we found significant gender differences across the 3 types of participation activities, with women having more than twice the odds of men of staying home from work or volunteer activities (OR = 2.27; 95% CI = 1.17, 4.43), refraining from grocery shopping (OR = 3.15; 95% CI = 1.55, 6.41), and avoiding outdoor physical activity (OR = 5.07; 95% CI = 1.48, 17.44) when ice was present. We found no significant interaction effects between gender and age, indicating that women were at greater risk for participation restriction in icy conditions, regardless of age.

DISCUSSION

Although research has identified extreme heat as a risk for morbidity and mortality,^2,19,20 little work has examined the impact of day-to-day weather conditions. In our analysis of nationally representative data from US adults, we found that ice was the most commonly reported barrier to getting around, resulting in a great deal of difficulty leaving home and driving and in avoiding or delaying activities such as grocery shopping or outdoor exercise. We also found that ice created much more difficulty for older (≥ 65 years) than younger adults in leaving their house to go about their day-to-day activities and in driving. Older adults are also more likely to avoid ice by staying home from work, volunteer, or other social activities. In addition, adults with mobility impairment and self-imposed driving restrictions (which could reflect vision impairment, especially in winter, which has fewer daylight hours³⁰) reported greater difficulty leaving home in icy conditions.

These results underscore the fact that older adults and those with mobility impairments are at risk for being housebound in winter months, with implications for social isolation, health, well-being, and mortality.^12,31–33 Simply being able to get out of the home has been shown to be associated with both functional and psychosocial changes.³¹ Functionally limited older adults are particularly at risk for health decline,^34–36 but getting out the door and walking, even just 2 blocks per day on average, has been shown to be protective against functional decline.¹² Our findings highlight the importance of everyday weather conditions to the ability to be mobile in the local community, which has previously received little attention in research on mobility.³³

Our results also highlight the barriers created by ice for mobility and participation. It is not surprising that we found notable regional differences in the impact of ice on difficulty leaving home and engaging in daily activities. Southern regions of the United States have less experience dealing with ice and fewer resources (salt, snow plows) to maintain clear walkways and roadways.

We also found consistent gender differences in reported difficulty leaving home and engaging in activities in the presence of ice. Women may be more likely than men to be afraid of slipping on ice,³⁷ perhaps because of their smaller skeletal muscle mass³⁸ or even their choice of footwear. Winter weather conditions reduce traction, increasing the risk of slips and falls^39,40 and making balance recovery strategies difficult.⁴¹ Further research is needed to understand the underlying reasons that women and older adults avoid leaving home when it is icy. More generally, making sure that streets, sidewalks, crosswalks, and bus stops are cleared of snow and ice is a simple step toward facilitating mobility and participation in day-to-day activities.

Although our data were collected from a nationally representative sample of US adults, the administration of the survey limited the questions on mobility and participation restrictions in icy conditions to those who reported that ice was the condition most likely to change the way they got around. Thus, we were unable to describe mobility difficulty in the presence of ice among those who indicated that some other weather condition (e.g., cold, snow, rain) was most likely to change the way they got around in winter. A further limitation was a low response rate to the telephone survey. However, we weighted all analyses to adjust for nonresponse.

Ours is the first study to our knowledge to provide national data on the weather conditions most likely to change the way US adults get around and participate in day-to-day activities. The findings underscore the need to prevent weather-related mobility and participation restrictions in high-risk populations, as an important step in maintaining health and preventing functional decline.