Comparison of Daily Routines Between Middle-aged and Older Participants With and Those Without Diabetes in the Electronic Framingham Heart Study: Cohort Study (Preprint)

<sec> <title>BACKGROUND</title> Daily routines (eg, physical activity and sleep patterns) are important for diabetes self-management. Traditional research methods are not optimal for documenting long-term daily routine patterns in participants with glycemic conditions. Mobile health offers an effective approach for collecting users’ long-term daily activities and analyzing their daily routine patterns in relation to diabetes status. </sec> <sec> <title>OBJECTIVE</title> This study aims to understand how routines function in diabetes self-management. We evaluate the associations of daily routine variables derived from a smartwatch with diabetes status in the electronic Framingham Heart Study (eFHS). </sec> <sec> <title>METHODS</title> The eFHS enrolled the Framingham Heart Study participants at health examination 3 between 2016 and 2019. At baseline, diabetes was defined as fasting blood glucose level ≥126 mg/dL or as a self-report of taking a glucose-lowering medication; prediabetes was defined as fasting blood glucose level of 100-125 mg/dL. Using smartwatch data, we calculated the average daily step counts and estimated the wake-up times and bedtimes for the eFHS participants on a given day. We compared the average daily step counts and the intraindividual variability of the wake-up times and bedtimes of the participants with diabetes and prediabetes with those of the referents who were neither diabetic nor prediabetic, adjusting for age, sex, and race or ethnicity. </sec> <sec> <title>RESULTS</title> We included 796 participants (494/796, 62.1% women; mean age 52.8, SD 8.7 years) who wore a smartwatch for at least 10 hours/day and remained in the study for at least 30 days after enrollment. On average, participants with diabetes (41/796, 5.2%) took 1611 fewer daily steps (95% CI 863-2360; &lt;i&gt;P&lt;/i&gt;&amp;lt;.001) and had 12 more minutes (95% CI 6-18; &lt;i&gt;P&lt;/i&gt;&amp;lt;.001) in the variation of their estimated wake-up times, 6 more minutes (95% CI 2-9; &lt;i&gt;P&lt;/i&gt;=.005) in the variation of their estimated bedtimes compared with the referents (546/796, 68.6%) without diabetes or prediabetes. Participants with prediabetes (209/796, 26.2%) also walked fewer daily steps (&lt;i&gt;P&lt;/i&gt;=.04) and had a larger variation in their estimated wake-up times (&lt;i&gt;P&lt;/i&gt;=.04) compared with the referents. </sec> <sec> <title>CONCLUSIONS</title> On average, participants with diabetes at baseline walked significantly fewer daily steps and had larger variations in their wake-up times and bedtimes than the referent group. These findings suggest that modifying the routines of participants with poor glycemic health may be an important approach to the self-management of diabetes. Future studies should be designed to improve the remote monitoring and self-management of diabetes. </sec> <sec> <title>CLINICALTRIAL</title> <p/> </sec>