key: cord-1001067-xv9ga0pv authors: Capodilupo, E. R.; Miller, D. J. title: Changes in health promoting behavior during COVID-19 physical distancing: Utilizing WHOOP data to Examine Trends in Sleep, Activity, and Cardiovascular Health. date: 2020-06-09 journal: nan DOI: 10.1101/2020.06.07.20124685 sha: 29e8f70655bc224e8a1d8b0ef980558ceaf87544 doc_id: 1001067 cord_uid: xv9ga0pv The COVID-19 pandemic incited global and unprecedented restrictions on the behavior of society. The aims of this study were to quantify changes to sleep/wake behavior and exercise patterns (e.g., exercise frequency, modality, and intensity), and the subsequent impact on physiological markers of health (e.g., total sleep duration, social jet lag, resting heart rate, and heart rate variability) with the introduction of physical distancing mandates and recommendations. A retrospective analysis of 50,000 subscribers to the WHOOP platform (mean age = 36.6 {+/-} 10.5; 11,956 females, 38,044 males) was conducted covering the period from January 1st, 2020 through May 15th, 2020. In order to make robust comparisons, this time period was separated into a 68 day baseline period and a 67 day physical distancing period - with a total of 6.3 million sleeps and 4.9 million exercise sessions analyzed. As compared to baseline, during physical distancing, all subjects analyzed in this study dedicated more time to sleep (+0.21 hours), fell asleep earlier (-0.43 hours), woke up earlier (-0.29 hours), obtained more sleep (+0.19 hours) and reduced social jet lag (-0.23 hours). Subjects also increased exercise frequency by an average of 1.1% and increased exercise intensity by spending an average of 1.8% more time in the three highest heart rate zones. These changes to sleep and exercise behavior may have contributed to the observed lowered resting heart rate (-0.9 beats per minute) and increased heart rate variability (+1.3 milliseconds) during physical distancing. A potential explanation for these results is that decreases in business hours-based commitments during physical distancing may have resulted in increased opportunity to engage in exercise and prioritize sleep. Therefore, as the COVID-19 pandemic eases, maintenance of certain aspects of physical distancing (e.g., flexibility to work from home) may result in a healthier population. • Social jet lag: the difference between sleep opportunity onset on weekends (Saturday and 126 Sunday) and weekdays (Monday through Friday) resulting from misalignment between 127 social and internal clocks [23] . 128 • Sleep opportunity onset: the time that each sleep opportunity was initiated relative to local 129 time zone. 130 • Sleep opportunity offset: the time that each sleep opportunity ended relative to local time 131 zone 132 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. • Heart rate variability (milliseconds; ms) -the root-mean-square difference of successive 147 heartbeat intervals sampled during slow wave sleep. Automatically measured during slow 148 wave sleep each night. 149 Throughout, dates are assigned to sleeps and exercise based on the local time zone's date in 150 which they end, for example, a sleep beginning in the final hours of January 1st and ending on the 151 morning of January 2nd is attributed to January 2nd. All analyses were conducted at the cohort-level with data grouped by age (i.e.,18-19, 20-29, 30-154 39, 40-49, 50-59, 60-69, 70-79 years old) and sex (i.e., male and female). Independent non-155 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint P R E -P R I N T parametric significance tests (Mann-Whitney U tests) were performed to examine differences in 156 sleep and exercise variables between the baseline period and the physical distancing period. To 157 assess the magnitude of the differences between these periods, effect sizes and 95% confidence 158 limits were also calculated. Effect sizes were interpreted as: <0.20 (trivial), 0.2 to 0.59 (small), 159 0.60 to 1.19 (moderate), 1.20 to 1.99 (large), and >2.0 (very large) [24] . All analyses were 160 conducted using Python Language Software (version 3.6.2). Sleep 163 A summary of sleep/wake behavior for age and gender cohorts are presented in Table 1 CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 9, 2020. -0.95, moderate) , and males (P<10 -9 , Cohen's d = -1.14, moderate) (Table 1) . . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 9, 2020. Notes: BL = baseline; PD = physical distancing; * = p-value <0.05; ** = p-value <0.001. 195 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint P R E -P R I N T CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. Table 2) . 217 218 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint Notes: BL = baseline; PD = physical distancing; * = p-value <0.05; ** = p-value <0.001. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. Summaries of exercise behavior for age and gender cohorts are presented in Table 3 CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. (Table 4) . . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. Sundays; the vertical gray line delineates between baseline and physical distancing. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. A summary of cardiovascular outcomes for age and gender cohorts are presented in Table 259 5. Average RHR was lower for all subjects during physical distancing when compared to baseline 260 (P<10 -9 , Cohen's d = -1.28, large). When analyzed by cohort, average RHR was lower during (Table 5) . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint P R E -P R I N T . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint The aim of this study was to detail changes in health related behavior and outcomes associated 288 with the introduction of global physical distancing policies in response to the COVID-19 289 pandemic. A retrospective analysis of sleep and activity patterns covering January 1, 2020 through 290 May 15, 2020 including 50,000 randomly selected WHOOP members ranging from age 18 to 80. Fig 1) . This finding reflects the established differences in chronotype 307 across age cohorts -younger individuals tend to have a late chronotype while older individuals 308 skew towards early chronotypes [25] . . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint In addition to exhibiting differences in the magnitude of average sleep onset shifts, Fig 1 310 highlights a change in the differences between weekend and weekday sleep opportunity onset (i.e., 311 social jet lag) across all age groups. All age cohorts reduced social jet lag, with more extreme 312 reductions seen in younger cohorts than in older ones ( Table 2 ). The age group with the largest 313 reduction was 18-19 year-olds. During baseline, this cohort averaged a 0.74 hour (i.e., 44 minutes) 314 later sleep onset on weekends than on weekdays which was reduced during physical distancing to 315 only 0.36 hours (i.e., 22 minutes). The smallest change, -0.08 hours (i.e., -5 minutes), was seen in 316 the oldest cohort, 70-79 year-olds, who also had the least social jet lag in baseline -0.19 hours 317 (i.e.,11 minutes) (Fig 1) . Average nightly sleep duration was significantly higher during physical distancing than 319 during baseline for all age cohorts (Fig 1) . Younger cohorts obtained more sleep than older cohorts 320 during baseline and increased their sleep by a greater percentage during physical distancing (Table 321 2; Fig 2) . The 18-19 year-old cohort averaged 0.56 hours (i.e., 34 minutes) more sleep per night 322 than the 70-79 year-olds during baseline but averaged 0.84 hours (i.e., 50 minutes) more sleep 323 during physical distancing. Among gender cohorts, females slept an average of 0.25 hours (i.e., 324 15.0 minutes) more per night than males during baseline and increased their sleep during physical 325 distancing by more than males (+13 minutes and +11 minutes, respectively; Table 2 ). With larger shifts observed in sleep opportunity onset compared to sleep opportunity offset, 327 it is reasonable to suggest that the extension of sleep opportunities and sleep were primarily 328 achieved by going to bed earlier and not by sleeping in later. These findings suggest that physical 329 distancing may have alleviated societal factors (e.g., work or academic commitments, commuting) 330 that restrict sleep opportunity, allowing individuals to revert towards biological sleep/wake 331 preferences. Practical applications can be made across all cohorts; however, the data highlight 332 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint P R E -P R I N T important changes in behavior amongst the 18-19 year-old cohort. As mentioned above, 333 individuals from this cohort typically have late chronotypes (i.e., their circadian drive for sleep 334 initiates later in the night). However, they are often expected to fulfil academic or sport related that are compliant to physical distancing appear to be those that also require high cardiovascular 351 load (i.e., running, cycling). This is supported by our data highlighting a decrease in weightlifting, 352 and increases in activities like running (Fig 6) . This increase is likely due to the minimal equipment 353 needs (e.g., pair of sneakers for running), and that individuals are likely to prioritize outdoor 354 activities whilst confined to their homes for the majority of each day. Previous studies suggest 355 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint P R E -P R I N T that engagement in new exercise modalities may reduce injury rates [28] , improve athletic 356 performance and improve cardiovascular health [29, 30] . Therefore, alterations to training stimulus 357 (i.e., exercise modality), and the associated physiological adaptations, may be an unexpected side 358 effect of physical distancing restrictions. In addition to changes in exercise modalities performed, our data show that individuals 365 exercised more frequently during physical distancing (Fig 3) . All age cohorts except for the 18 and 366 19 year-olds increased their frequency of exercise during physical distancing as compared to 367 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint baseline. Again, this is likely due to increased flexibility during physical distancing to perform 368 physical activity. In isolation of the impact of physical distancing on exercise frequency, a novel 369 cyclical pattern with a 7-day period in exercise frequency is observed in most age cohorts in Fig 370 3 . The cycle with highest amplitude is in 18 and 19 year-olds, the youngest cohort, with both the 371 highest rates of exercise during the weekdays and the lowest rates of exercise on weekends. Again, 372 this finding emphasizes the impact of professional and academic weekday commitments on 373 exercise behavior. In regard to the intensity of exercise, the highest three heart rate zones (collectively, 70-375 100% HRR) occupy a larger proportion of total exercise time during physical distancing than they 376 did during baseline (Table 4 ). Distributions in relative time spent in heart rate zones 2 through 5 377 show a cyclical pattern with a 7-day period that is more pronounced during baseline than physical 378 distancing; relatively more time is spent in zones 4, 5, and 6 on weekends, while more time is spent 379 in zones 2 and 3 on weekdays. The differentiation between weekend and weekday heart rate zone 380 distributions is less pronounced in all heart rate zones during physical distancing than it was during 381 baseline. An increase in time spent in these higher heart rate zones may mean an increase in 382 anaerobic training, which has been previously demonstrated to reduce RHR and improve 383 endurance [31] . The present study highlights changes in exercise modality and increases in exercise 385 frequency during physical distancing restrictions. Irrespective of societal influences, it is 386 physiologically accurate to suggest that such changes to exercise behavior may improve health 387 outcomes [32]. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint distancing when compared to baseline for all age and gender cohorts (Fig 3) . Both of these 391 outcomes represent an improvement in cardiovascular health, which suggests that improved sleep 392 and exercise behaviors may be conferring positive benefits on the WHOOP members analyzed. Recent research has also demonstrated a relationship between social jet lag and HRV in 394 which higher levels of social jet lag were associated with lower HRV [33] . These findings are 395 consistent with the present study in which all cohorts experienced a reduction in social jet lag and 396 increase in HRV during physical distancing. Analysis of the baseline period shows expected moderate differences across age cohorts in 398 which younger cohorts show signs of greater cardiovascular fitness than older cohorts. While it is 399 well documented that HRV declines with age [34, 35] , previous studies have demonstrated no age-400 related increase in RHR. For example, Kostis et al [36] found no relationship between age and 401 RHR; this apparent discrepancy may be related to the small effect size and the relatively small 402 datasets of previous studies. The demonstration here by analysis of 50,000 individuals that a 403 statistically and practically significant increase in RHR with age occurs is a novel finding of this 404 paper. 405 406 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint This is the first study to report on the sleep and exercise behavioral changes associated with 408 COVID-19 related physical distancing mandates. By leveraging wearable technology, a unique 409 analysis of large population cohorts both prior to and throughout the early stages of the COVID-410 19 pandemic were conducted. The findings suggest that meaningful changes have occurred to sleep 411 and exercise patterns, which may have long-term consequences on the health and wellbeing of the 412 population. While improved sleep and exercise patterns appear to be the mechanism for improved 413 health outcomes during physical distancing, it is unclear which specific barriers were limiting these 414 activities prior to physical distancing. It is reasonable to assume that decreases in business hours-415 based commitments (e.g., commuting) has previously limited the time devoted to exercise and 416 sleep. Therefore, in the context of a post-pandemic society, increased flexibility in how business, 417 academia and other professional endeavors are conducted (i.e., ability to work from home) may 418 result in a healthier general population. The findings of this study should be interpreted with the 419 sample demographic in mind -WHOOP subscribers may not demographically match the general 420 population. It is reasonable to suggest that members of the WHOOP platform are more likely to 421 be health-conscious and fitness-oriented than average. However, these data can be applied to a 422 large population cohort that are seeking to engage more health-enhancing behavior despite 423 professional commitments. Future research will be required as countries begin to reopen to 424 investigate which behaviors are maintained, how these behaviors impact mental health, and if 425 cohorts return to their previous behaviors. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint P R E -P R I N T running distance and risk of running-related injuries: an association which varies according to type 473 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2020. PMID: 19092709. 497 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 9, 2020. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 9, 2020. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 9, 2020. . https://doi.org/10.1101/2020.06.07.20124685 doi: medRxiv preprint Coronaviridae Study Group of the International Committee on Taxonomy of Viruses. The species 430 Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV and naming it 431 Behavioral and physiological consequences of sleep restriction Sleep restriction for 1 week 435 reduces insulin sensitivity in healthy men Association of short sleep duration with obesity, 438 diabetes, fatty liver and behavioral factors in Japanese men Sleep duration predicts cardiovascular 441 outcomes: a systematic review and meta-analysis of prospective studies Correlates in US Adolescents The Impact of Sleep Duration on Performance Among Competitive 447 Athletes: A Systematic Literature Review Sustained Attention and Sleep Pressure Before and During Total Sleep Deprivation and Recovery The Effects of Sleep Extension on Sleep Irregular sleep/wake 456 patterns are associated with poorer academic performance and delayed circadian and sleep/wake 457 timing Duration, timing and quality of sleep are each vital for health, performance and safety Sleep and wakefulness out of phase with 461 internal biological time impairs learning in humans Rotating Night Shift Work and Risk of Coronary Heart Disease Among Women Rotating night 467 shifts and risk of breast cancer in women participating in the nurses' health study The effects of cross-training 470 on fitness and injury in women Excessive progression in weekly 25155475 Aerobic vs anaerobic exercise 476 training effects on the cardiovascular system Effect of Wearables 479 on Sleep in Healthy Individuals: A Randomized Cross-Over Trial and Validation Study Resting heart rate and 482 the risk of cardiovascular disease, total cancer, and all-cause mortality -A systematic review and 483 dose-response meta-analysis of prospective studies The relationship of autonomic imbalance, heart rate 486 variability and cardiovascular disease risk factors Executive Office of the President. Proclamation 9994: Declaring a National Emergency Concerning 491 the Novel Coronavirus Disease (COVID-19) Outbreak Life between clocks: daily temporal patterns of human 493 chronotypes Progressive statistics for studies in sports 495 medicine and exercise science