Rural-urban differences in meeting of physical activity recommendations by sun exposure and protection behaviors in the United States

Abstract

Objective:

To examine associations between sun-protection behaviors and physical activity (PA) by rural and urban residence status in the United States.

Methods:

We analyzed data from the National Health and Nutrition Examination Survey (2013–2018), restricting to participants ages 20–59 with sun behavior data. Sunburns, sun exposure, and sun protection measures were dichotomized (yes/no): ≥1 sunburn in the past year, 2+ hours outside during workdays or non-work days, and never/rarely/sometimes using sunscreen, wearing long sleeves, and staying in the shade. Meeting PA recommendations (yes/no) was defined as ≥150 minutes of vigorous/moderate or ≥75 minutes vigorous PA per week. Associations between sun behaviors and PA were analyzed using logistic regression models, which accounted for survey-weights and potential confounders, and stratified by rural-urban status.

Results:

Rural and urban individuals meeting PA recommendations had greater odds of spending 2+ hours outside during workdays (OR: 2.26 [1.88, 2.74] and 3.95 [2.72, 5.73]) and non-workdays (OR: 2.06 [1.78, 2.38] and 3.33 [2.47, 4.46]). Among urban residents, odds of staying in the shade were lower among those who met PA recommendations (OR: 0.78 [0.66, 0.92]). We did not observe differences in sunburns or other sun behaviors by PA status, regardless of rurality.

Conclusions:

Meeting PA recommendations was associated with greater sun exposure in both rural and urban populations. Additional exercise location (indoors/outside) data is needed to inform PA and skin cancer prevention interventions to reduce unintended increases in sun exposure and reductions in PA, respectively, especially among rural populations.

INTRODUCTION

Skin cancer is the most common type of cancer in the United States (US), with incidence consistently increasing over the last several decades.¹ Physical activity (PA) is associated with decreased risk of many types of cancer; however, PA has been associated with greater risk of melanoma in some studies.² Greater skin damage from ultraviolet radiation (UVR) from being physically active outdoors may increase skin cancer risk. Most skin cancers could be prevented if exposure to UVR was minimized and sunburns avoided,³ for example through sun protective behaviors: using sunscreen, staying in the shade, wearing protective clothing, and avoiding the sun at peak hours.

Several studies have observed that physically active individuals more frequently report sunburns,^4–8 and one study found this pattern held regardless of sun protective behaviors.⁸ Additionally, those concerned about sun exposure were observed to be less physically active than those who were less concerned about sun exposure.⁹ However, sunburn prevalence has not been associated with PA in every study.¹⁰ Additionally, some studies have found nuances in associations between PA and sun behavior by type of sun protective behavior. For example, one study found physically active people were more likely to wear sunscreen and a hat but less likely to seek shade.⁹ We need to better understand the relationship between sun protective behaviors and PA to optimize both PA for health and wellness benefits as well as sun protection for decreased skin cancer risk. This is particularly relevant as we learn more about the mental and potential physical benefits of nature-based outdoor activities and exercise.¹¹

Rural versus urban residence is an important consideration when assessing associations between sun protective behaviors and PA. Rural residents may be at greater risk of melanoma^12,13 and less often meet PA recommendations than urban residents.¹⁴ Several US studies have observed suboptimal sun protective behaviors among rural adults,^15,16 although results have been mixed.^17,18

As a first step in investigating the potential association between sun protection behaviors and PA status by rural-urban residence in the US, we analyzed data from the National Health and Nutrition Examination Survey (NHANES) on self-reported skin cancer-related sun behaviors and PA. We hypothesized that individuals who reported meeting PA recommendations would report different sun-risk practices—for example, more time spent outside and less frequently seeking shade, but potentially more frequent sunscreen use—than those who did not report meeting PA recommendations, and these associations would differ by rural-urban status.

METHODS

Data source

We analyzed data from the 2013/14, 2015/16, and 2017/2018 cycles of the NHANES by the National Center for Health Statistics within the Centers for Disease Control and Prevention. The NHANES interview data includes demographic and health-related questionnaires. Since 1999, the NHANES sample design has consisted of multiyear, stratified, clustered samples of noninstitutionalized civilians of all ages from across the US, with public-use data releases in 2year cycles. A total of 10,175 individuals completed the 2013/14 NHANES wave, 9,971 the 2015/16 wave, and 9,254 the 2017/18 wave. Our study sample is restricted to participants aged 20–59 years because the dermatology module was only asked among this age group, and to participants with available rural-urban status (total analytic sample size N=11,031). Most NHANES data are publically available (https://www.cdc.gov/nchs/nhanes/index.htm); however, the rural-urban indicator is restricted, so the data were analyzed within the Minnesota Research Data Center, which is part of the Federal Statistical Research Data Centers network. The University of Minnesota Institutional Review Board deemed this study exempt (STUDY00014202), and the NHANES was compliant with the Declaration of Helsinki.

Measures

Questions on sunburn, sun exposure, and sun protection behaviors, the exposures of interest for this analysis, had multiple-choice answers which we dichotomized, consistent with previous studies.¹⁵ Sunburn was categorized as yes (1 or more sunburns) or no (0) in the past year. For time spent outside in the sun, participants were asked, “[…] about the time you spent outdoors during the past 30 days. By outdoors, I mean outside and not under any shade. How much time did you usually spend outdoors between 9 in the morning and 5 in the afternoon on the days that you worked or went to school?” and “on the days when you were not working or going to school?” For both workdays and non-workdays, sun exposure was dichotomized as <2 hours versus 2+ hours. Sun protective behaviors were dichotomized as “optimal” (“always” or “most of the time”) and “suboptimal” (“sometimes”, “rarely”, or “never”): “When you go outside on a very sunny day, for more than one hour, how often do you” “(1) use sunscreen?”, “(2) wear a long sleeved shirt?”, and “(3) stay in the shade?”.

The primary outcome was meeting American Cancer Society recommendations for PA (yes/no).¹⁹ PA was assessed in NHANES as weekly minutes of vigorous or moderate PA a) at work, b) during recreational activities, and c) biking or walking. Weekly total minutes of rigorous and moderate PA were calculated by summing across types and dichotomized by whether or not they met PA recommendations (at least 75 minutes of vigorous PA or a combination of 150 minutes of moderate and/or vigorous PA each week).

NHANES rurality categories were classified based on the 2013 National Center for Health Statistics Urban-Rural Classification Scheme for Counties / census tracts.²⁰ The classification includes six categories: four metropolitan, coded as urban in NHANES (large central metropolitan, large fringe metropolitan, medium metropolitan, small metropolitan), and two nonmetropolitan categories, coded as rural (micropolitan, noncore).

Statistical analysis

Analyses were weighted using the NHANES survey weights to represent the US population. We described demographic population characteristics and frequencies of meeting PA recommendations and self-reported sun exposure and protection measures by rural-urban status. We conducted logistic regression models to assess the associations between PA and sun exposure and protection behaviors for all respondents, stratifying by rural-urban status after testing for interactions between meeting PA recommendations and sun exposure and protection behaviors by rurality. Multivariate models were adjusted for confounders identified apriori including age, (continuous) sex (female, male), annual household income (<$20,000, $20,000–54,999, $55,000–99,999, ≥$100,000), education (high school degree or less, some college, at least college degree), body mass index (BMI; kg/m², continuous), smoked 100+ cigarettes in one’s lifetime (yes, no), and race and ethnicity (measured in NHANES as non-Hispanic [NH] White, Mexican American/Other Hispanic, NH Asian, NH Black, or Other Race, including multiracial). Analyses were conducted using SAS 9.4. We report survey weighted frequencies, adjusted odds ratios (aOR), and 95% confidence intervals (CI); P-values <0.05 were considered significant.

RESULTS

Of the 11,031 participants, 1,394 were classified as rural (weighted 17.4%, 95% CI: 13.2, 21.5%) and 9,637 as urban residents. More rural (54.9%) than urban (44.3%) individuals reported at least one sunburn in the past year, and rural individuals were more likely to spend time outside during work and non-work days (Table 1). Rural and urban residents were similar for meeting US PA recommendations (70.3% and 69.1%, respectively), though the source and location of the activities differed. Individuals living in rural areas reported more PA during work and those in urban areas reported more PA during recreational activities and transportation.

Table 1.

Demographic characteristics and sun behaviors of the study population. Participants in United States National Health and Nutrition Examination Survey (NHANES) cycles 2013–2018, 20–59 years old who completed the dermatology module.

	Urban	Rural
Demographics	Mean±SD	Mean±SD
Body mass index, kg/m 2	28.6±0.2	28.9±0.3
	% (95% CI)	% (95% CI)
Age group, years
20–29	26.8 (25.0, 28.5)	21.1 (19.2, 23.1)
30–39	25.4 (24.3, 26.6)	19.7 (17.0, 22.4)
40–49	23.9 (22.4, 25.3)	26.3 (23.9, 28.6)
50–59	23.9 (22.6, 25.3)	32.9 (30.1, 35.6)
Sex
Female	50.8 (49.6, 52.0)	50.8 (48.9, 52.7)
Male	49.2 (48.0, 50.4)	49.2 (47.3, 51.1)
Race/Ethnicity
Asian	7.4 (5.8, 8.9)	0.9 (0.6, 1.1)
Black/AA, Non-Hispanic	14.2 (11.9, 16.6)	3.2 (1.3, 5.1)
Hispanic	20.0 (16.5, 23.5)	7.9 (5.4, 10.4)
Other, Non-Hispanic	4.2 (3.4, 4.9)	3.3 (2.2, 4.3)
White, Non-Hispanic	54.2 (49.9, 58.6)	84.8 (80.8, 88.7)
Education
High school or less	34.3 (31.3, 37.2)	41.3 (36.3, 46.4)
Some college	31.9 (30.1, 33.6)	37.3 (34.6, 40.0)
At least college degree	33.8 (30.6, 37.0)	21.4 (17.1, 25.6)
Household annual income
<$20,000	11.8 (10.5, 13.1)	11.0 (7.9, 14.1)
$20,000–54,999	29.1 (27.0, 31.2)	29.8 (25.7, 34.0)
$55,000–$99,999	24.3 (22.6, 26.1)	23.9 (19.8, 28.0)
≥$100,000	28.0 (25.0, 31.0)	30.9 (24.2, 37.6)
Prefer not to say / Unknown	6.7 (5.8, 7.7)	4.4 (2.8, 5.9)
Smoked at least 100 cigarettes in life
No	60.5 (58.6, 62.4)	54.0 (50.3, 57.7)
Yes	39.5 (37.6, 41.4)	46.0 (42.3, 49.7)
Sun Exposure and Protection Behaviors
At least one sunburn in past 12 months	44.3 (42.1, 46.6)	54.9 (49.8, 60.0)
2+ hours outside during work day	32.5 (30.4, 34.6)	43.3 (39.4,47.3)
2+ hours outside during non-work day	60.7 (59.1, 62.3)	70.5 (67.5, 73.9)
Sunscreen use – Always/Most of the time	28.4 (26.6, 30.1)	25.3 (23.2, 27.5)
Wear long sleeves outside – Always/Most of the time	10.7 (9.9, 11.5)	7.7 (5.6, 9.9)
Stay in shade – Always/Most of the time	33.6 (31.7, 35.4)	24.3 (21.3, 27.3)
Physical Activity (PA)
Meet PA Recommendations	69.1 (67.7, 70.4)	70.3 (68.0, 72.6)
	Mean±SD	Mean±SD
Minutes vigorous work activity	205.2±10.9	302.5±29.8
Minutes moderate work activity	336.1±12.0	494.3±20.7
Minutes vigorous recreational activities	85.3±2.6	53.4±4.3
Minutes moderate recreational activities	97.3±3.0	86.7±4.4
Minutes walk/bicycle for transportation	74.8±4.1	39.1±3.8

We observed higher rates of sunburn among individuals who met PA recommendations in unadjusted but not adjusted models, and the estimates did not differ by residence (p-value for interaction p=0.70; Table2). Those meeting PA recommendations in both rural and urban populations more often spent at least two hours outside during workdays (aOR: 2.26 [1.88, 2.74] and 3.95 [2.72, 5.73], respectively) as well as during non-workdays (aOR: 2.06 [1.78, 2.38] and 3.32 [2.47, 4.46], respectively). While this association was observed for both rural and urban residents, the effects were greater among rural than urban residents for both the workday and non-workday associations (p-values for interactions p=0.003 and p=0.04, respectively).

Table 2.

Weighted frequencies and associations between sun exposure and protection measures and meeting physical activity (PA) recommendations by rural-urban status among participants 20–59 years old from the United States National Health and Nutrition Examination Survey (NHANES) cycles 2013–2018,

	Everyone			Urban			Rural			P**
	Met PA recom. % (95% CI)	Did not meet PA recom. % (95% CI)	Unadjusted and adjusted * OR (95% CI)	Met PA recom. % (95% CI)	Did not meet PA recom. % (95% CI)	Unadjusted and adjusted * OR (95% CI)	Met PA recom. % (95% CI)	Did not meet PA recom. % (95% CI)	Unadjusted and adjusted * OR (95% CI)
At least 1 sunburn in past year	48.7 (46.2–51.1)	39.9 (36.5–43.4)	1.43 (1.24–1.65) 1.17 (1.00–1.38)	46.9 (44.6–49.2)	37.9 (34.4–41.4)	1.45 (1.26–1.67) 1.16 (0.98–1.36)	56.8 (51.9–61.6)	49.9 (40.8–59.0)	1.32 (0.93–1.87) 1.22 (0.82–1.82)	0.70
During the past 30 days, how much time did you usually spend outdoors between 9 in the morning and 5 in the afternoon …
At least 2 hours during work day	39.9 (37.1–42.7)	20.5 (18.1–22.9)	2.57 (2.20–3.01) 2.53 (2.10–3.05)	37.4 (34.8–40.0)	20.4 (18.0–22.8)	2.33 (1.99–2.73) 2.26 (1.88–2.74)	51.7 (46.1–57.2)	21.1 (17.5–24.7)	4.00 (2.89–5.54) 3.95 (2.72–5.73)	0.003
At least 2 hours during non-work day	69.0 (67.3–70.7)	47.4 (44.4–50.4)	2.47 (2.17–2.81) 2.22 (1.93–2.56)	67.0 (65.4–68.6)	46.5 (43.5–49.6)	2.33 (2.05–2.66) 2.06 (1.78–2.38)	78.2 (75.5–80.9)	51.9 (44.5–59.2)	3.33 (2.43–4.56) 3.32 (2.47–4.46)	0.04
When you go outside on a very sunny day, for more than one hour, how often do you…
Use sunscreen ***	28.5 (26.6–30.3)	26.7 (24.3–29.0)	1.09 (0.98–1.23) 1.01 (0.99–1.24)	29.2 (27.3–31.2)	26.7 (24.2–29.2)	1.13 (0.99–1.29) 1.11 (0.97–1.26)	24.9 (22.4–27.4)	26.6 (24.2–29.1)	0.91 (0.75–1.11) 1.16 (0.93–1.45)	0.05
Wear long sleeves outside ***	9.6 (8.6–10.6)	11.4 (10.2–12.6)	0.82 (0.71–0.95) 0.92 (0.77–1.09)	10.0 (9.0–10.9)	12.2 (10.8–13.5)	0.80 (0.69–0.93) 0.89 (0.74–1.07)	7.8 (5.3–10.3)	7.7 (5.5–9.9)	1.02 (0.72–1.44) 1.16 (0.80–1.66)	0.33
Stay in shade ***	29.0 (27.0–31.1)	38.7 (36.0–41.3)	0.65 (0.58–0.73) 0.79 (0.69–0.91)	30.5 (28.3–32.7)	40.6 (37.9–43.3)	0.64 (0.56–0.74) 0.78 (0.66–0.92)	22.4 (19.2–25.6)	29.1 (23.8–34.4)	0.71 (0.53–0.93) 0.84 (0.60–1.19)	0.74

^*adjusted for age, sex, race, household income, education, body mass index, and smoking history

^**P-value interaction rurality*PA; unadjusted model

^***Always or most of the time

Among urban residents, odds of always or most of the time staying in the shade were lower among those who met PA recommendations (aOR: 0.78 [0.66, 0.92], Table 2). Among rural residents, the estimated association was similar though not statistically significant (aOR: 0.84 [0.60, 1.19]). We observed no differences in wearing long sleeves or sunscreen use by PA status, and the estimates did not differ by rural-urban status.

DISCUSSION

We assessed the associations of sunburns and sun health behaviors with PA, stratified by rural-urban residential location. Meeting PA recommendations was associated with more time spent outdoors, with a larger effect size among rural than urban participants. Those reporting greater PA less often stayed in the shade in both urban and rural populations, but the adjusted association was only statistically significant for the urban population. No other significant differences were observed in reported sun protection behaviors by PA status in either rural or urban participants.

Sun risk behaviors—spending more time outside, being less likely to stay in the shade when outside—among those who are physically active may contribute to why physically active adults have been found to be at greater risk for melanoma.² Similarly, the larger effect size between time spent outside and PA among rural residents may partially explain higher melanoma incidence in rural areas.^12,13 Our results are consistent with previous findings that more active individuals have more frequent melanoma risk factors.^4–8 However, in this study, associations of sunburns with PA were statistically significant only in the unadjusted analysis, among the sample as a whole and urban residents. The weakened associations after adjustment for rural-urban demographic characteristics may partially explain why the association of PA with melanoma risk factors, such as sunburns, vary between studies.^9,10 Other potential contributors to these discrepancies not investigated here and deserving of future research could be lack of data regarding the type and location of PA and work (indoors versus outdoors) and/or differences in lifestyle, cultural factors, and climate by region and country. Further, intersecting identities, including race and ethnicity, may impact PA and sun behaviors. For example, we recently found that adults living in rural areas in the United States generally reported riskier sun behaviors than those living in urban areas, but with nuances depending on race, ethnicity, and specific behavior.²¹

Given the health benefits of both PA and sun protective behaviors,^3,11 considering potential impacts of PA interventions on sun exposure is important. Weighing behavioral health risks versus benefits is especially relevant among rural residents, as our findings suggest that PA and time spent outside are most strongly linked among rural adults. It is unclear if this is due to recreational or occupational sources of PA. Future interventions should develop strategies that maximize health benefits while accounting for intricate links between outdoor PA and sun behaviors to minimize harm. Room for improvement in these behaviors is considerable yet complicated. For example, interventions focused on shade use and decreased time in the sun during peak UVR hours may be difficult to implement without decreasing PA.

A strength of this analysis is the study population from three biennial NHANES cycles that is representative of the US population. Limitations include the cross-sectional nature of the data and self-reported sun behaviors and PA. The binary rural-urban categorization aggregates large populations across the rural-urban continuum. Our findings do not generalize to all population subgroups; future analyses in specific populations are needed to identify nuances within the general rural-urban patterns we found. Further, we did not assess potential mechanisms of why sun risk behaviors or PA would differ between urban and rural populations, for example whether the observed associations were driven by exposures in outdoor-focused jobs and/or leisure activities that involve PA.

CONCLUSIONS

Physically active adults ages 20–59 reported greater sun exposure, and the associations were generally stronger among rural participants. Our findings suggest that maximizing benefit and minimizing harm in this space is complicated, as we must consider the possibility that promoting PA, if performed outdoors, may increase sun exposure and skin cancer risk; and conversely that promoting certain skin cancer prevention behaviors may reduce PA. Additional data on location of exercise (indoors versus outside) and sun protection behaviors during outdoor PA is needed to inform efforts to intervene on PA and sun protection behaviors, especially among rural populations.