Our findings showed that most large cities whose residents slept longer had better baseline health and improvement in general health status of their residents over time. However, the maximum baseline value of good health status was 63.9% and maximum change was 13.1% in cities with better sleep health, implying room for improvement in all cities. Except for Walk Score®, baseline values in exposure variables were slightly higher in cities with better sleep health than in cities with poorer sleep health. Change in exposures were not significantly different between the city groups. However, overall values of the environmental indicators reflected the need for improvement. As we hypothesized, sufficient daily sleep showed a moderating effect on the association between environmental indicators and general health status. In the cities with residents who had higher percentages of sufficient daily sleep, the magnitudes of the positive associations were increased, implying synergistic interactions between sufficient daily sleep and better environmental factors on good health status. However, no mediating effect of sufficient daily sleep was observed on the association between environmental indicators and good health status.
We demonstrated the synergistic moderating effect of sufficient daily sleep on the association of each of the following with good health status: parks within a walking distance, Walk Score®, and Bike Score®. Such an interaction effect on general health was also found in recent studies in which unhealthy behaviors were jointly associated with poor-quality built environment, and too little or an excessive amount of sleep were associated with poor health status (36, 37) and the risk of morbidity and mortality (38, 39).
Our findings also show that selected environment factors are positively correlated with intensive physical activity, which implies that sufficient sleep amplifies the beneficial effects of physical activity on general health. In fact, many recent studies have demonstrated the synergistic effect of sleep and physical activity on general health status at the individual level (14, 15). In a recent mid-aged UK Biobank study, the association between low physical activity and risk of all-cause and cause-specific mortality was exacerbated by poor sleep, as classified with a composite sleep score that included duration (14). Another secondary analysis that used the 2017–2018 National Survey of Children’s Health demonstrated the synergistic interaction effect between insufficient sleep duration and low physical activity on mental health disorders (15).
This synergistic interaction between sleep quantity and physical activity level can be explained by their interrelationship (i.e., 2 bi-directional mechanisms). To be specific, physical activity positively affects sleep duration (40) and lack of sleep also may be a key impediment to promoting and maintaining a physically active lifestyle (41). The first mechanism is related to the effects of exercise on central nervous system, body temperature, cardiac/autonomic, endocrine and metabolic functions during sleep and overall mood (42). The beneficial effects of exercise on sleep have been proven in many studies (43–45) in which physical activity is an effective intervention for those having insufficient sleep duration. The second mechanism is more complex and rarely discussed. Insufficient sleep impairs cognitive performance, mood, glucose metabolism, appetite regulation, and immune function (46), which may play a negative role in physical activity. Moreover, the process of sleep may affect the brain at an endocrine level independent of the hormonal regulation of metabolism and waste removal at the cellular level (47), which may lead to changes in body temperature, cytokine concentrations, energy metabolic rate, growth hormone secretion, nervous/cardiovascular system, symptoms and disorders of mood and anxiety, neurotrophic factor secretion, and fitness level (48). Thus, these 2 mechanisms may help to explain the synergistic interaction effect on general health observed in the current study.
Though no mediating effect of sleep quantity on the association between environmental factors and good health status was found here, previous studies have shown sleep’s mediating role in the association between greater physical activity and good health status (16–18). The mediating mechanism can be explained in two steps. The first step is that the activity-related environmental factors affect sleep quantity; in the second step, the change in sleep quantity influences general health status. There is evidence supporting the first mechanism because sleep has been shown to be affected by neighborhood green space, walkability, safety, and built environment (49). In another study, green space exposures were associated with better sleep quality and quantity (50), which may be due to increased physical activity and spending time in a place with better air quality. Additionally, a study reported that step counts and daily active time (which are related to neighborhood walkability) can lead to better sleep quality and longer sleep duration (51). A high-quality environment that enables bicycling safely may result in better sleep as well as stress relief (52). Having a Complete Streets policy can translate into high-quality walkability and a good location for cycling, which in turn can promote and support a resident’s physical activity, because it ensures safe access for all pedestrians, bicyclists, motorists, and transit riders (53). The second mechanism can be partially explained by our finding of a positive correlation between sufficient daily sleep and good health status. In general, sufficient sleep is acknowledged as being essential to optimal general health status (10), and to influence brain performance (11), mental health (12), and physical health (13). Thus, insufficient sleep is more likely to cause significant morbidity and mortality (54). Taken together, the literature suggests a mediating effect of sleep on the association between activity-related environmental factors and good health status. A potential reason that the current study could not replicate the effect is insufficient statistical power, which was likely exacerbated in the stratified analyses. In addition, sleep was collected only for the last 3 years, which could reduce information and reduce power.
We also found that all cities studied still had much room for improvement in both quality of environmental and personal sleep health. City and public health leaders’ collaboration, including setting and sharing best practices of the fittest city, would help to gradually satisfy unmet needs. There have been some federal strategies for integrating active living and health (55). Arlington, VA, the current fittest city in the AFI, shared strategies for advocating cycling paths (56). Smart growth approaches also contributed to development of mixed land use, walkable and active communities, and increased availability of diverse public transportation choices (57). Such approaches could serve as a basis for developing better strategies to improve public health. In addition, updated regulations and continuous public investment are needed for development of walkable districts, greening of abandoned areas, and development of cycling paths and lanes (55).
Our study has several strengths. First, this is the first investigation of the moderating or mediating effect of sleep on the associations between activity-related environmental factors and general health status using US city data. Second, AFI data integrated environmental indicators and personal health indicators. Third, we included only modifiable measures in order to enable city leaders to implement targeted programs and policies to advance population fitness (20). Fourth, the indicators were selected, reviewed, and updated each year by Fitness Index content experts, ACSM staff, and technical consultants to ensure that the data are up-to-date and include the best measures available (20). Fifth, we also considered demographic and socioeconomic factors that allowed us to control for potential confounders in our analyses.
Our study also has several limitations. First, 2022 data on Baton Rouge, Louisiana, were not collected, which necessitated the replacement of the missing values with the city’s 2021 values, based on assuming little change between 2021 and 2022. This might have introduced some bias. However, Baton Rouge indicators changed little between 2018 and 2021. In addition, data imputation of only one city is unlikely to greatly affect overall results, which may help justify our analysis and results. Second, because Bike Score® and Complete Streets policy were not collected in 2018, we used 2 different baseline years (i.e., 2018 or 2019) according to the available baseline year of each exposure. This may disrupt the uniformity of our examination of the interaction effects on general health, but this is a common issue when using secondary data (58), and the current study had no further missing values in the tailored panel data. Moreover, once the AFI data was gathered, the missing data was addressed (20). Third, much of our study used self-reported data gathered via questionnaires and surveys, which might bring about some biases (59). However, the AFI included only indicators from reputable and regularly updated public sources to ensure validity and reliability (20). Fourth, we have only 100 city samples, which might introduce lower statistical power, depending on the unknown population effect size magnitudes. Lower power might explain statistically nonsignificant findings for some associations.
In conclusion, we identified a synergistic interaction effect between sufficient daily sleep among city residents and physical activity-related environmental factors related on good health status of residents. Sleep duration was not found to be a mediator of the association between environmental indicators and good health status. Our findings also suggested that the largest US cities still have much room for improvement in both environment and sleep adequacy, which might be addressed by continuous proactive collaboration between city and public health leaders and by creating a model informed by lessons from the fittest cities as reported in the ACSM AFI. Further research is warranted to better understand the biological mechanisms that modulate the dynamic interplay between physical activity level and sleep quantity, and to examine the mediating role of sleep health on the association between environmental assets and general health status using a larger data set with more than 100 US cities.