A Multidisciplinary Evaluation of a Virtually Supervised Home-Based High-Intensity Interval Training Intervention in People With Type 1 Diabetes

Many people with type 1 diabetes lead a sedentary lifestyle (1–3), with lack of time and fear of hypoglycemia identified as key exercise barriers (4,5). High-intensity interval training (HIT) may address these barriers, with studies showing that HIT improves cardiorespiratory fitness and vascular function without the reductions in glycemia associated with moderate-intensity continuous training (MICT) (6). However, during these studies (6,7), HIT was performed under laboratory conditions with strict researcher supervision, meaning the “real-world” potential of HIT is unclear for people with type 1 diabetes. The HIT protocol used a cycle ergometer, introducing additional exercise barriers, such as difficulty accessing equipment or facilities (including distance and cost), and potential embarrassment due to negative body image if performed within a gym (4,5). This study used a multidisciplinary approach to evaluate a novel, virtually monitored home-based HIT (Home-HIT) intervention in people with type 1 diabetes.

Eleven individuals with type 1 diabetes (seven women; age 30 ± 3 years; 2.5 ± 0.2 L/min; duration of diabetes 10 ± 2 years; HbA_1c 8.0 ± 0.6% [64 ± 7 mmol/mol]; BMI 27.3 ± 1.6 kg/m²; daily insulin dose 0.31 ± 0.06 IU/kg/day) completed 6 weeks of Home-HIT. The Home-HIT program was completed in an unsupervised place of the participant’s choosing. Participants performed repeated 1-min bouts of high-intensity exercise interspersed with 1-min of rest. During the intervals they were asked to achieve a heart rate (HR) of ≥80% of their predicted maximum (220 − age). Intervals were composed of two 30-s simple bodyweight exercises (e.g., star jumps then burpees) with no rest in between. Participants were provided with 18 exercises with 9 suggested exercise pairs, detailed in an information pack, and participants were free to choose exercises according to personal preference. Participants were advised to train 3×/week, and complete 6 1-min intervals per session in weeks 1–2, increasing to 8 in weeks 3–4 and 10 in weeks 5–6. Participants were virtually monitored using a HR monitor that connected via Bluetooth to their smartphone (Polar Beat; www.polar.com/beat/uk-en). Although participants were monitored virtually, training was completed without researcher supervision or encouragement. This allowed participants to monitor their HR and provided immediate feedback on exercise intensity. Following each session, HR data were automatically uploaded to a cloud storage site (www.flow.polar.com), which allowed participants to monitor their progression. The website was also available to the research team to monitor whether the program was being completed as advised. The research team used these data to contact participants by text/email every 2 weeks to inquire about training progress and to provide support if required. If participants missed consecutive sessions, messages inquired as to whether there was a specific reason. The monitoring system provided an objective measure of adherence (number of sessions completed) and compliance (whether HR thresholds and correct number of intervals were achieved during each session) (Supplementary Fig. 1).

Throughout the program, participants were asked to only exercise if their blood glucose levels were 7–14 mmol/L, in accordance with Exercising for Type 1 Diabetes (EXTOD) guidelines (8). They were also asked to record their blood glucose pre, post, and 1 h post each session and whether they used additional carbohydrates or insulin during or following each session.

was measured during pre- and posttesting, which took place ∼72 h before the first training session and 72 after the final training session, respectively. During posttesting, participants completed an anonymous online qualitative survey (www.surveymonkey.co.uk) to explore barriers and facilitators to exercise before the intervention and their experiences of Home-HIT (Supplementary Table 1). During the first and final 7 days of the program, participants monitored their insulin dose and blood glucose using an 8-point profile: before and 2 h after each meal, just before bed, and at 2:00 a.m.

The study was approved by the Black Country National Health Service Research Ethics Committee (West Midlands, U.K.), and written informed consent was obtained from all individuals prior to participation.

Owing to the small sample size, data were assessed using the nonparametric Wilcoxon signed rank test, except for change in blood glucose concentration pre, post, and 1 h postexercise, which was assessed with a Friedman test, with the within-group factor exercise (pre vs. post, vs. 1 h post), using IBM SPSS Statistics for Windows. Significance was set at P ≤ 0.05, and data are presented as mean ± SEM. The qualitative survey responses were analyzed using a framework approach (9).

Training adherence was 95 ± 2% (range 83–100%), with participants completing the advised number of intervals at the 80% HR_max target in 99 ± 1% of sessions (range 94–100%). Blood glucose remained stable during and after exercise, with the mean blood glucose concentration immediately postexercise and 1 h postexercise being not different from baseline (P = 0.249) (Supplementary Fig. 2). Carbohydrate was consumed to prevent hypoglycemia in 6 ± 3% sessions (10 of 188 sessions), and insulin was needed for hyperglycemia after 2 ± 1% of sessions (3 of 188 sessions). No severe hypoglycemic episodes requiring third-party intervention were reported.

Six weeks of Home-HIT increased by 7% (P = 0.017), and there was a 13% decrease in daily short-acting insulin (P = 0.012) (Supplementary Fig. 2). There was no change in mean blood glucose concentration (pre 8.8 ± 0.5 mmol/L; post 8.6 ± 0.4 mmol/L; P = 0.445), measured using a 7-day 8-point diary, and no change in BMI (pre 27.3 ± 1.6 kg/m²; post 27.4 ± 1.6 kg/m²; P = 0.646).

Three key themes and subthemes were developed from the survey responses: 1) flexibility of Home-HIT with the subthemes type 1 and nontype 1 diabetes–related flexibility, 2) motivation with the subthemes Home-HIT and virtual monitoring, and 3) the “HIT” experience. Table 1 shows the frequency of participants’ positive and negative responses relating to each theme. The top three exercise barriers reported were lack of time (91%), fear of hypoglycemia (27%), and lack of motivation (18%). Supplementary Table 2 shows detailed information on participants’ past exercise experiences, current activity level, feelings toward their current activity level, and exercise barriers before the intervention.

We demonstrate that people with type 1 diabetes are able to engage and adhere to a virtually monitored Home-HIT program and that this is safe and effective. Home-HIT increases and reduces insulin dose, while appearing to reduce traditional exercise barriers as well as fear of hypoglycemia, with 95% adherence rates. Training diaries showed that blood glucose remained stable up to 1 h following Home-HIT sessions (Supplementary Fig. 2C), supporting previous laboratory-based research (6). This contrasts with moderate-intensity continuous training, where there is a consistent drop in glycemia in people with type 1 diabetes (6,10–12). The blood glucose data were supported by the survey responses suggesting participants felt comfortable doing Home-HIT because their blood glucose concentrations remained stable.

Home-HIT sessions lasted 12–20 min, meaning the weekly time commitment was at least 90 min less than the recommended 150 min (13). Many participants reported time efficiency of Home-HIT as a major advantage in the survey and appreciated the convenience of not having to travel, which added to the time efficiency. Furthermore, participants liked being able to exercise at home because there was more privacy and the program was free and required no equipment.

Lack of motivation is a common barrier to achieving physical activity targets (4). The survey responses suggest the design of our Home-HIT intervention contributed to improving motivation to exercise. These motivating factors included the range of exercises available and the progression in number of intervals. Participants suggested that the virtual monitoring contributed to their motivation, as it provided instant feedback on exercise intensity and allowed progression to be tracked by exercise professionals who could provide feedback. This feedback probably contributed to the high adherence. HR monitoring is the most accurate way to track the body’s response to activity, providing objective, personalized data that account for age and fitness (14), reflecting exercise intensity regardless of exercise type (14). Such monitoring systems may provide a relatively inexpensive (∼£40 per HR monitor and mobile application) strategy to engage with participants and improve uptake, adherence, compliance, and ultimately, health outcomes.

We decided not to include an untrained control group. Although this would have strengthened the design, it would have reduced the feasibility of completing the study. Our primary aim was to assess safety and acceptability of virtually monitored Home-HIT in people with type 1 diabetes, which would not have benefited from an untrained control group. Secondly, time of day that training was undertaken was not controlled, and recent work has shown that time of day influences glycemic response to exercise in people with type 2 diabetes (15). However, participants were free to complete Home-HIT at any time of day, suggesting a flexible training intervention that can be used in the “real world.” Furthermore, participants stated in the survey that they felt no increased risk of hypoglycemia even when exercising in the evening (Table 1). Future research should use continuous glucose monitoring to investigate how time of day influences the effects of Home-HIT on glycemia and efficacy.

Our study suggests virtually monitored Home-HIT is a safe, effective, and acceptable strategy for supporting people with type 1 diabetes to exercise.

Author Contributions. S.N.S., S.O.S., R.C.A., P.N., T.S.P., F.K., D.J.C., S.A.-G., T.N., A.J.M.W., and M.C. read and approved the final manuscript. S.N.S., S.O.S., R.C.A., P.N., F.K., D.J.C., A.J.M.W., and M.C. drafted and revised the manuscript. S.N.S., F.K., D.J.C., T.N., and M.C. contributed to collection, analysis, and interpretation of the data. S.N.S., S.A.-G., T.N., A.J.M.W., and M.C. contributed to conception and design of the experiments. S.N.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented as oral presentations at the Experimental Biology meeting, Orlando, FL, 6–9 April, 2019, and at the 23rd Annual Congress of the European College of Sport and Exercise Sciences, Dublin, Ireland, 4–7 July 2018.