This study describes how positive and negative health behaviors changed during the COVID-19 pandemic and subsequent “Stay-at-Home” orders time period (March–April 2020), who was more likely to change their behaviors, and for what reasons these changes may have occurred, among a convenience sample of U.S. adults. Statewide movement restriction orders have led to immediate social and economic challenges for individuals [2], all of which have potentially adverse effects to health behaviors. Some groups appear more vulnerable to these disruptions, as this study found that among this sample of U.S. adults, females, those with children, and those with moderate-to-severe depression symptom severity were more likely to report increased engagement in negative health behaviors (alcohol consumption and tobacco use) potentially due to increases in available time or additional stressors due to competing responsibilities between family and work. Further, females and those with moderate to severe depression symptom severity reported worsened sleep quality, which may be due to increased feelings of worry and stress. Conversely, females also reported an increase in physical activity, which was most commonly attributed to increased available free time and/or boredom. These results provide preliminary understanding for the behavioral health consequences of COVID-19 beyond the clear immediate toll to physical health and human life.

Of the reasons provided for changing health behaviors during the pandemic, over one third (36.5%) of the reported reasons for behavior change were related to increased engagement in positive health behaviors (physical activity, sleep), which was most commonly attributed to more time available and stress relief (61%). As this sample consists of relatively higher-educated, employed adults, respondents may have been able to engage in alternative forms of physical activity, such as online fitness classes or purchasing their own exercise equipment [41]. In the current study, females were more likely to increase physical activity, which could be related to stress relief/management as about 18% of reasons for increasing positive health behaviors were related to stress relief. As has been shown elsewhere, certain physical activities (e.g., yoga) have been shown to be an effective activity for changes in stress-related psychological outcomes [42]. Among females living with a male, this may also be related to a more equal division of household labor that allows females more time for leisure activities like physical activity [43]. Whereas in 2019, American women spent more time doing housework and less time engaging in physical activity than American men on any given day [44]. Those with children also reported increased activity. Parents may be finding alternative ways to engage children with physical activity in the outdoors to maintain social distancing, such as family walks or hikes, due to closures of schools and organized sports [45]. In addition, less time spent commuting to the workplace may have led to more available time to pursue physical activities. For example, approximately 40 min per day could be added as workers may no longer be commuting to work [46].

Interestingly, in contrast, almost as many of the overall reasons for changing behavior were related to decreasing engagement in positive health behaviors (31.9%). Feelings of worry/stress (30.9%), less motivation (22.4%), and resource concerns (23.5%) were highly reported as the reasons for this adverse behavior change. Those with moderate to severe depression symptom severity were most likely to report decreasing their physical activity behavior. Previous existing evidence also shows significant reductions in U.S. adult’s physical activity during this time period, particularly among those who are Hispanic, lower income, and those not working [21]. Although not statistically significant, non-white and unemployed participants had higher odds of reporting decreased activity in this sample. These similar findings may indicate changes in depression symptom severity may impact changes in physical activity to a greater degree. Thus, efforts are needed not only to increase opportunities for physical activity, as transport- and occupation-related physical activity are likely thwarted due to the pandemic but to also promote mental health resources, such as mindfulness training [47] and access to telehealth care providers [48].

Similarly, out of the 1597 reasons participants reported for increasing negative health behaviors, the most common reasons were reported as boredom and more time available (50.5%). Boredom has been shown to be associated with substance abuse and depression symptoms [47]. While this study did not explore these associations, those living with moderate-to-severe depression symptom severity scores had over 2.2 times higher odds of increased alcohol consumption and three times higher odds of increased marijuana use than those with none to mild depression symptom scores. Further, those with moderate-to-severe depression symptom severity scores revealed increases in other adverse behaviors such as reporting over five-times higher odds of decreasing their physical activity and worsened sleep quality than those with none to mild depression symptoms scores. Similarly, a study conducted in Australian adults during COVID-19 found higher depression symptom severity was associated with worse health behaviors—less physical activity and sleep, more smoking and alcohol drinking [49]. The current findings expand upon this recent study by additionally including specific reasons for behavior change. Elevated population-level depression prevalence due to the pandemic is a public health concern. In representative U.S. samples prior to COVID-19, about 7-8% of U.S. adults had been found to have moderate-to-severe depressive symptoms using the PHQ-9 [18,19]; however, 18.5% of this study sample was categorized as having moderate-to-severe symptoms during the pandemic using the same measure. This high prevalence of moderate-to-severe depression symptom severity is similar to those found in Australia (19.1%) [49] and China (16.8%) [25].

Increases in engagement with negative health behaviors and worsened sleep quality were also seen across a range of sociodemographic factors. College graduates had 1.5 higher odds of increasing alcohol consumption. Alcohol abuse during the pandemic is a concern [28]; as evidenced by 13% of the current sample drinking four or more alcoholic beverages per day. Females were two and half times more likely to increase their tobacco use and 1.4 times more likely to report feeling like their sleep quality had worsened, compared to males in this sample. Those with children were 1.6 times more likely to report increasing alcohol consumption. With the closures of schools, child cares, and summer programs, many parents have had to take on child care needs in addition to completing work responsibilities from home [50]. Nearly 44% of the sample had children, leading many to have to balance work and home responsibilities, or even quit their jobs to take on caregiving duties [51]. In addition, middle-age and older-adults were less likely than younger adults to change their tobacco use, marijuana use, and alcohol consumption. These results may be partially explained by younger adults being the most likely to be furloughed or lose work and income during COVID-19 [52,53]. Similarly, unemployed individuals in this sample were less likely to report decreasing their tobacco use and less likely to report improved sleep. These behaviors may be a result of the high U.S. unemployment rate (14.7% by April 2020), and the almost 50% of Americans reporting a loss of income [54].

While this study shows a number of reasons for changes to health behaviors, increases in stress and anxiety due to COVID-19 have been of increased concern in the recent literature [25,49] as anxiety and stress are risk factors for substance abuse behaviors and poor sleep quality [55,56,57]. Specifically, females, younger adults, lower income, and lower education have been found to have higher stress scores during COVID-19 [49], which may explain our findings. We found approximately 20% of the reported reasons for changing health behaviors were related to feelings of being more worried/stressed. The current study adds to Stanton et al. (2020) by suggesting behavior change coping mechanisms may further differ across specific subpopulations and should be further explored in future research [49]. As COVID-19 continues to spread and full or partial lockdowns remain, access and promotion of mental health services and interventions are critical to preventing increases in negative health behaviors or decreases in positive health behaviors.

The current study’s limitations should be noted. Generalizability is limited due to a convenience sample that is primarily represented by women, non-Hispanic white, and highly educated individuals. The use of an online-only survey that utilized social media platforms for recruitment may have limited participation. Therefore, these findings should not be extrapolated to any other groups of interest and future research should aim to overcome this limitation by including larger, more representative samples. However, the convenience sampling strategy provided the opportunity to provide initial, hypothesis generating information on an important and rapidly evolving topic. In addition, the self-reporting of behaviors may have led to social desirability bias; however, participants were assured of their anonymity prior to agreeing to participate in this survey. Finally, these data are cross-sectional. Although changes to behavior were compared to life prior to the coronavirus outbreak, recall bias may be an issue as local “Stay-at-Home” order duration was 3.8±1.0 weeks at time of survey completion; however, some participants (n = 143) were unsure of their time under a “Stay-at-Home” order, which was significantly (p < 0.05) associated with age, employment status, disability status, comorbidities, and depression scores. These were treated as missing and imputed as described. Future research should examine these health behaviors in a more representative U.S. sample in addition to measuring changes over time with follow-up surveys.