Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Associations between app usage and behaviour change in a m-health intervention to improve physical activity and sleep health in adults: secondary analyses from two randomised controlled trials

Associations between app usage and behaviour change in a m-health intervention to improve... Background To examine associations between user engagement and activity-sleep patterns in a 12-week m-health behavioural intervention targeting physical activity and sleep. Methods This secondary analysis used data pooled from two Randomised Control Trials (RCT, [Synergy and Refresh]) that aimed to improve physical activity and sleep (PAS) among physically inactive adults with poor sleep. Both RCTs include a PAS intervention group (n = 190 [Synergy n = 80; Refresh n = 110]) and a wait list Control (CON n = 135 [Synergy n = 80; Refresh n = 55]). The PAS groups received a pedometer and accessed a smartphone/tablet “app” with behaviour change strategies, and email/SMS support. Activity-sleep patterns were quantified using the activity-sleep behaviour index (ASI) based on self-report measures. Intervention usage was quantified as a composite score of the frequency, intensity and duration of app usage during intervention (range: 0–30). Assessments were conducted at baseline, 3 and 6 months. Relationships between usage and ASI were examined using generalised linear models. Differences in ASI between the control group and intervention usage groups (Low [0–10.0], Mid [10.1–20.0], High [20.1–30.0]) were examined using generalised linear mixed models adjusted for baseline values of the outcome. Trial Registration: ACTRN12617000376347; ACTRN12617000680369. Results During the 3-month intervention, the mean (± sd) usage score was 18.9 ± 9.5. At 3 months (regression coef- ficient [95%CI]: 0.45 [0.22, 0.68]) and 6 months (0.48 [0.22, 0.74]) there was a weak association between usage score and ASI in the intervention group. At 3 months, ASI scores in the Mid (Mean [95%CI] = 57.51 [53.99, 61.04]) and High (60.09 [57.52, 62.67]) usage groups were significantly higher (better) than the control group (51.91 [49.58, 54.24]), but not the Low usage group (47.49 [41.87, 53.12]). Only differences between the high usage and control group remained at 6 months. *Correspondence: Mitch J. Duncan Mitch.Duncan@newcastle.edu.au Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 2 of 9 Conclusion These findings suggests that while higher intervention usage is associated with improvements in behav- iour, the weak magnitude of this association suggests that other factors are also likely to influence behaviour change in m-health interventions. Trial registration number: ACTRN12617000376347; ACTRN12617000680369. Keywords eHealth, Usage, Engagement, Attrition, Dose-response Introduction composite measures may be more useful in understand- Delivering behaviour change interventions through elec- ing the usage-behaviour change relationship [14–16], tronic health (e-Health) and mobile health (m-Health) although few studies have applied such multidimensional creates the opportunity to deliver cost-effective wide- measures. The overall aim of this study was to examine reaching interventions. Both e- and m-Health interven- how user engagement with a m-health app is associated tions have demonstrated effectiveness for improving with behaviour change during two randomised con- lifestyle behaviours including physical activity [1], sleep trolled trials (RCT) of a m-health intervention targeting [2], alcohol consumption and chronic disease manage- improvements in physical activity and sleep health in ment [3, 4]. Approximately 30% of adults are both physi- physically inactive adults with poor sleep. Two specific cally inactive and have poor sleep health [5, 6] and may objectives were to: (1) examine the relationship between benefit from interventions to improve these behaviours. user engagement and a composite score of overall physi- However, there are few interventions that have targeted cal activity and sleep health (Activity-Sleep Index [ASI]) improvements in both physical activity and sleep [7, 8]. within the intervention group and, (2) compare differ - Consequently, relatively little is known about how par- ences in overall physical activity and sleep health in the ticipants use and engage with digital physical activity and control group and in different levels of app usage (Low, sleep interventions. Characterising engagement with dig- Mid, High) in the intervention group. ital interventions that target multiple lifestyle behaviours is important given the number of adults who engage in Methods multiple higher risk behaviours [9] and the need for Study design interventions to address multiple lifestyle behaviours This study uses data pooled from two separate RCTs of concurrently. the same m-health intervention, which was designed to Understanding how participants use and engage improve physical activity and sleep health behaviours in with digital interventions is important, as while it is physically inactive adults with poor sleep quality [7, 8]. typically assumed that greater usage is associated with Details of the study rationale, methods and main out- greater behaviour change, the magnitude of this rela- comes of each trial, and intervention effects on the ASI tionship appears to be weak [10, 11]. Furthermore, it is are available elsewhere [7, 8, 18–20]. Similarities between consistently reported that usage declines throughout the trials in terms of the behaviours assessed, inter- the intervention period [12, 13]. Additionally, there are vention and control groups, assessment methods and inconsistencies between studies regarding how usage outcomes assessed allowed data from the control and is conceptualised and measured, which limits compari- physical activity and sleep health intervention groups to sons between studies [11, 14–17]. A systematic review be pooled as described previously [20]. Eligible partici- reported that a greater subjective user experience of the pants were those aged 18–55 years (Synergy Study [7, 18]) intervention, greater number of activities completed and or 45–65 years (Refresh Study [8, 19]), who lived in Aus- more frequent logins are consistently associated with tralia, reported < 90 min of moderate to vigorous intensity greater physical activity, but that time on the website physical activity (MVPA) in the last week and rated their was not associated with physical activity [11]. This sug - sleep quality as fairly bad or very bad. Exclusion criteria gests the usage-behaviour change relationship may dif- included being employed in shift-work, diagnosed sleep fer depending on the usage metric examined and that disorder, and current use of a device to track activity or single usage metrics may not adequately characterise sleep (see Additional file 1: Figs. S1, S2). how participants use and engage with the intervention. Both studies primarily recruited participants using To overcome this, Short and colleagues [15] proposed a social media advertising. The Synergy study aimed to composite measure of usage that captures the frequency compare the efficacy of a combined physical activity and (i.e., number of self-monitoring entries or logins), inten- sleep health intervention with a wait-list control. Partici- sity (i.e., number of intervention features used), duration, pants (n = 160; mean age: 41.5 (SD = 9.9); 80% female) and type (i.e., reflective, didactic, or active) of usage. Such were recruited between June–August 2017 and the study M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 3 of 9 was conducted between June 2017 and February 2018. All self-monitoring data entered into the application The Refresh study aimed to compare the efficacy of a were recorded in the application database, including combined physical activity and sleep health interven- the associated timestamp of entry. The waitlist-control tion with a sleep health-only intervention and a wait-list group did not have any access to the application or other control [19, 21]. Participants (n = 275; mean age: 52.0 intervention materials prior to the 6-month assessment. (SD = 6.9); 83% female) were recruited between May– However, they were offered full access to the intervention September 2017 and the study was conducted between including the “Balanced” application after completing the June 2017 and March 2018. The combined physical final 6-month assessment. activity and sleep health intervention in both trials was the same in terms of mode of delivery, theoretical basis, Measures educational content, and behaviour change techniques Sociodemographic variables such as age, gender, educa- used. The sleep health-only intervention arm (n = 110) tion and chronic disease status were assessed at baseline in Refresh was omitted from the current study as partici- and primary and secondary outcomes were measured at pants in that group did not receive any physical activity baseline, three and six months [7, 8]. Primary and sec- intervention content. The active phase of the intervention ondary measures of the original trials included minutes in both studies ceased at the 3-month point. Both studies of MVPA [22], the frequency of resistance training [18, conducted online assessments at baseline, 3 months and 19], sleep quality [23] and insomnia symptoms [24]. 6 months and were prospectively registered with the Aus- tralian and New Zealand Clinical Trials Registry as well Activity sleep index as received ethical approval (H-2016-0181, H-2016-0267) The overall pattern of physical activity and sleep was at the University of Newcastle. Participants in both trials quantified using the activity-sleep index (ASI), which provided informed consent. Each study used computer is a 12-item instrument described elsewhere [20]. It is generated permuted block randomisation to develop designed to assess overall healthy patterns of physical the randomisation sequence, with group allocation con- activity and sleep health based on the frequency, dura- cealed in sequentially numbered envelopes. Participants tion, type and intensity of physical activity, the duration were not blinded to group allocation given the nature of of sitting time and the duration, timing, quality and sat- the interventions. isfaction of sleep. The specific items, responses, and scor - ing for the ASI are provided in Additional file  1: Table S2. Study groups The items are briefly summarised here: The physical activity and sleep (PAS) intervention group (n = 190 [Synergy: n = 80 + Refresh: n = 110]) received 1. Frequency–MVPA (Number of sessions of MVPA/ access to a specifically designed mobile application “Bal - wk), anced” that comprised educational resources, personal 2. Frequency–RT (Number of days of resistance goals, self-monitoring logs (manual data entry), and feed- training/wk), back in relation to personal goals, all relative to a range 3. Intensity (Proportion of MVPA that was vigorous of physical activity and sleep health components (i.e., in intensity), activity minutes, step count, resistance training, bedtime, 4. Type (Participation in no MVPA or resistance sleep wake timing and sleep quality). Details of the inter- training, either MVPA or resistance training, or vention are provided in Additional file  1: Tables S1 and both), Fig. S3. Prior to commencement, the intervention group 5. Time (Duration of MVPA/wk), participants were mailed a printed participant handbook 6. Sitting (Duration of sitting time/wk). with guidance on how to use the app, and a pedometer. 7. Daytime alertness (Trouble staying awake during Participants also received weekly reports and short mes- the day), sage service (SMS) prompts to limit disengagement. 8. Sleep Quality (Overall sleep quality rating), Participants used the app for goal setting and action plan- 9. Sleep Timing (Midpoint of sleep between 02:00 am ning to increase their physical activity (i.e., MVPA, step and 04:00 am), counts and resistance training) and to improve their sleep 10. Sleep Regularity (Variability in bed and wake quality and sleep behaviours (stabilising bed/wake times, times), sleep hygiene behaviours and stress management (e.g., 11. Sleep Efficiency (Sleep efficiency ([sleep duration/ progressive muscle relaxation, deep breathing exercises, time in bed] × 100), mindfulness) [20]. All intervention components were 12. Sleep Duration (Meeting age-appropriate sleep delivered either through the application, email or SMS, duration guidelines). and the messaging component ceased at three months. Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 4 of 9 Data analysis To overcome the different metrics used to quantify Descriptive statistics are presented for the sample at each dimension, each dimension was rescaled to a zero to baseline in each intervention group, and also by interven- 10 scale, with higher scores reflecting lower risk behav - tion usage group. To examine the association between iour. Each dimension is summed to create a score ranging engagement and behaviour change, two separate analyses from 0 to 120. The approach used to rescale the individ - were conducted. The first analysis was limited to only the ual dimensions of the ASI was X−X min intervention group as no usage data were available for the Rescaled score = n where X is the observed Range control group. This analysis examined the relationship value, X is the minimum observed value of the original min between the overall usage score and the ASI at follow- variable, X is the difference between the minimum Range up, adjusted for the baseline value of the outcome, using and maximum of the observed values, and n is upper a generalised linear model. The model included fixed limit of the rescaled variable (e.g., n = 10) [20]. effect for the continuous mean centred usage score, study (Refresh, Synergy), assessment (3 months, 6 months) and the interaction between usage score and assessment. The Intervention usage linearity of the relationship between continuous usage An overall usage score was created to capture interven- score and ASI in the intervention group was examined tion group participants’ frequency, intensity, and dura- using residual plots and including a quadratic term for tion of using the “Balanced” intervention platform using usage score in the analysis. The quadratic term was not data recorded in intervention database. There are no statistically significant (p < 0.05) and the residuals plots usage data for the wait-list control group as they did not did not indicate a non-linear relationship. To examine have access to the app during the intervention period. how varying amounts of usage in the intervention group Type of usage (i.e., reflective, gamified, altruistic, didac - were associated with behaviour change relative to the tic, or active) was not examined as these data were not waitlist-control group, overall usage scores in the inter- recorded in the intervention platform. All indicators were vention group were categorised into a three-level group assessed over the initial 84-day (i.e., 3-month) interven- variable: low usage (0–10); mid usage (10.1–20.0); and tion period to align with the ‘active’ component of the high usage (20.1–30.0) and combined with the control intervention. Each day, participants could self-report group to create a four-level variable. This analysis exam - their: (1) minutes of MVPA, (2) resistance training, (3) ined between group differences (Control, Low usage, daily step count, (4) bedtime, (5) wake time and, (6) sleep Mid usage, High usage) in the ASI adjusted for the base- quality. The app was designed to promote daily self-mon - line value of the outcome. The model included a fixed itoring of these metrics, however participants were free effect for study (Refresh, Synergy), assessment (3 months, to self-monitor any number of these metrics on a given 6 months), group (Control, Low usage, Mid usage, High day. These measures were used to create measures of the usage), and the group by assessment interaction. Residual frequency, intensity and duration of usage. Frequency diagnostics were used to inform the choice of model and was measured as the total number of self-monitoring link. Analyses were conducted using Stata MP v17 and entries made during the 3-month (84 days) intervention alpha was set at 0.05. period, with a maximum of six entries per day (one entry per day for each of the self-monitoring entries made). Intensity was measured as the average number of self- Results monitoring entries made each day during the interven- Participant flow throughout each trial is shown in Addi - tion. Duration was measured as the number of days until tional file  1: Figs. S1 and S2. A total of 325 participants a participant succumbed to non-usage attrition, defined completed the baseline survey, 275 (84.6%) completed as the time they first stopped self-monitoring for at least the 3-month assessment, and 215 (66%) completed the 14 consecutive days [7, 21, 25, 26]. Due to the different 6-month assessment. Completers of the 3-month assess- metrics used to characterise each usage dimension (i.e., ment were older (M = 47.41 [SD = 9.73]) and had higher count of self-monitoring entries per day, number of days) levels of intervention usage (M = 20.73 [8.76]) than non- each dimension was rescaled to a zero to 10 scale as fol- completers (Additional file  1: Table S3). The baseline sam - X−X min lows rescaled score = n ; where X is the original ple consisted of 264 female and 61 male participants, and range most were middle-aged and highly educated (Table  1). score, X is the minimum of the observed variable, min At baseline the average BMI and ASI were 28.15 kg/m X is the range of the potential score and n is the range (SD = 4.21) and 47.34 (SD = 10.91), respectively and both upper limit of the rescaled score [20]. The rescaled these variables were similar between intervention and dimensions were summed to create an overall usage control groups. At baseline the low usage group reported score ranging from zero to thirty with higher values indi- lower average ASI scores (M = 44.19 [SD = 14.11]), cating greater usage. M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 5 of 9 Table 1 Baseline descriptive characteristics of participants by study and intervention usage group Group Intervention usage group Control n = 135 Intervention n = 190 Low Usage n = 34 Mid Usage n = 60 High Usage n = 94 Total N = 325 M (SD), n (%) M (SD), n (%) M (SD), n (%) M (SD), n (%) M (SD), n (%) M (SD), n (%) Age (years) 46.19 (10.39) 47.22 (9.71) 47.12 (9.37) 45.47 (10.53) 48.30 (9.29) 46.76 (10.02) Education (years) 16.26 (2.90) 16.11 (2.77) 16.03 (3.05) 15.73 (2.99) 16.38 (2.53) 16.17 (2.82) BMI 27.67 (4.09) 28.50 (4.27) 28.71 (4.48) 29.30 (4.33) 27.89 (4.10) 28.15 (4.21) ASI Score 46.83 (10.03) 47.83 (11.56) 44.19 (14.11) 47.58 (10.54) 49.06 (10.90) 47.34 (10.91) Sex Male 26 (19.26%) 35 (18.42%) 7 (20.59%) 12 (20.00%) 16 (17.02%) 61 (18.89%) Female 109 (80.74%) 155 (81.58%) 27 (79.41%) 48 (80.00%) 78 (82.98%) 262 (81.11%) Income/Yr ≤ $30,000 32 (23.70%) 30 (15.79%) 2 (5.88%) 11 (18.33%) 17 (18.09%) 62 (19.20%) $30,001–$50,000 14 (10.37%) 26 (13.68%) 3 (8.82%) 6 (10.00%) 17 (18.09%) 40 (12.38%) $50,001–$70,000 26 (19.26%) 39 (20.53%) 8 (23.53%) 18 (30.00%) 13 (13.83%) 65 (20.12%) $70,001–$100,000 25 (18.52%) 47 (24.74%) 7 (20.59%) 13 (21.67%) 26 (27.66%) 71 (21.98%) ≥ $100,001 30 (22.22%) 30 (15.79%) 10 (29.41%) 8 (13.33%) 11 (11.70%) 59 (18.27%) Don’t know/no answer 8 (5.93%) 18 (9.47%) 4 (11.76%) 4 (6.67%) 10 (10.64%) 26 (8.05%) Employment group Professional 85 (62.96%) 107 (56.32%) 21 (61.76%) 31 (51.67%) 55 (58.51%) 192 (59.44%) White-collar 18 (13.33%) 37 (19.47%) 6 (17.65%) 16 (26.67%) 14 (14.89%) 54 (16.72%) Blue-collar 5 (3.70%) 4 (2.11%) 0 (0.00%) 1 (1.67%) 3 (3.19%) 9 (2.79%) Not working 27 (20.00%) 42 (22.11%) 7 (20.59%) 12 (20.00%) 22 (23.40%) 68 (21.05%) Employment group not working includes retired, unemployed, home duties, looking for work, student, and other. Low, Mid and High usage defined as usage score of 0–10.0, 10.1–20.0 and 20.1–30.0, respectively with a higher proportion reported higher income lev- differences were only maintained in the high usage group els (≥$100,001/yr) relative to the Mid and High usage (M = 60.82, 95% CI = 57.96, 63.67; difference to control groups. M = 8.86, 95% CI = 5.04, 12.68) at 6 months (Table  2; The average intervention usage score was 18.88 Fig. 2). (SD = 9.54) out of 30, and the average usage scores in the Low, Mid and High usage groups were 3.49 (SD = 4.10), Discussion 14.78 (SD = 2.97), and 27.03 (SD = 3.00), respectively. This study examined the association between usage of In the intervention group, there was a weak association a m-health intervention app and overall physical activ- between intervention usage score and ASI at 3 months ity and sleep health behaviour. In the intervention group (Β = 0.45, 95% CI = 0.22, 0.68) indicating that for each there was a weak positive relationship between usage and 1 unit increase in intervention usage score there was an behaviour at 3 and 6 months. Consistent with this obser- estimated mean 0.45 increase in ASI. The association vation, when comparing the Low, Mid and High usage between intervention usage score and ASI at 6 months groups to the Control group, only the Mid and High was of a similar magnitude (Β = 0.48, 95% CI = 0.22, usage groups demonstrated significantly higher (better) 0.74). The estimated marginal mean ASI scores at 3 and overall physical activity and sleep health behaviours at 6 months for different levels of intervention usage are 3 months. These differences were only maintained in the shown in Fig.  1. Exploring the effect of Low, Moderate High usage group at 6 months. These results indicate that and High usage scores in the intervention group rela- while there is a weak relationship between intervention tive to the control group, the results indicated that at usage and behaviour change, it appears that only mid-to- 3 months, the mid usage (M = 57.51, 95% CI = 53.99, high levels of usage are associated with small improve- 61.04; difference to control M = 5.60, 95% CI = 1.39, 9.81) ments in behaviour relative to the control. and high usage groups (M = 60.09, 95% CI = 57.52, 62.67; Overall, intervention usage scores were modest and difference to control M = 8.18, 95% CI = 4.68, 11.68) had there was a weak relationship with behaviour at three significantly higher ASI scores relative to the control and six months. Prior analysis of these trials indicates group (M = 51.91, 95% CI = 49.58, 54.24). However, these that overall activity-sleep behaviours did significantly Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 6 of 9 Fig. 1 Baseline adjusted ASI-12 at 3 and 6 months by usage score in intervention group. Model only includes the pooled intervention group. Model adjusted for study (i.e., Synergy, Refresh), baseline ASI-12 score, and includes the mean centered usage score and its interaction with assessment. p-value for interaction between usage score and assessment is = 0.857 The association between usage score and ASI-12 at 3 months is = Β=, 95%CI: b = 0.45,95%CI = 0.22, 0.68 and at 6 months is Β = 0.48, 95%CI = 0.22, 0.74 Table 2 Baseline adjusted ASI-12 by intervention and usage group at 3 and 6 months 3 Months 6 Months M [95%CI] Diff. to control M [95%CI] Diff. to control M [95%CI] M [95%CI] Control 51.91 [49.58, 54.24] 51.95 [49.47, 54.44] Low usage (0–10.0) 47.49 [41.87, 53.12] − 4.42 [− 10.55, 1.71] 48.71 [42.47, 54.94] − 3.25 [− 9.98, 3.49] Mid usage (10.1–20.0) 57.51 [53.99, 61.04] 5.60 [1.39, 9.81] 55.72 [51.18, 60.27] 3.77 [− 1.40, 8.94] High usage (20.1–30.0) 60.09 [57.52, 62.67] 8.18 [4.68, 11.68] 60.82 [57.96, 63.67] 8.86 [5.04, 12.68] Model adjusted for baseline value of the outcome, and study. p-value for group by time interaction = 0.841. There were 34, 60 and 94 participants in the low, mid and high usage groups, respectively improve in the intervention group relative to the con- is considerable variation in these associations in differ - trol group [20]. Collectively this suggests that while the ent studies [11]. Overall these observations are consistent use and engagement with the intervention platform has with conceptual frameworks of engagement-behaviour some influence, it is not a major driver of behaviour change that identify platform usage as one of several fac- change. Previous studies have observed no statistically tors [30, 31], including psycho-social factors related to significant [27] or weak [28] associations between vari - behaviour change [30], personal relevance of informa- ous measures of app usage and behaviour change, while tion provided [32], and inclusion of behaviour change others have observed positive dose-response relation- techniques in the intervention [31], that can influence ships between greater usage and improvements in health behaviour change. This has important implications for outcomes [29]. Similarly, a meta-analysis summarising the design of future m-health interventions. Specifically, the association between app engagement and change in interventions need not only to be designed to promote physical activity behaviour observed that, while there is and foster a certain degree of user engagement with the a weak statistically significant relationship between usage intervention platform, but they also need to incorporate and physical activity (b = 0.08, 95% CI = 0.01–0.14), there other important features related to behaviour change. Of M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 7 of 9 Fig. 2 Baseline adjusted ASI-12 at 3 and 6 months by Intervention and Usage Group. Model adjusted for study (i.e., Synergy, Refresh), baseline ASI-12 score, fixed effects for group (Control, Low Usage, Mid Usage, High Usage), assessment (3 months, 6 months) and the group by assessment interaction relevance to this study which targeted improvements in that the vast majority of m-health, and digital interven- physical activity and sleep, is the bidirectional relation- tions aim to improve participants’ knowledge and skills ship between these behaviours [33–35] which may have to initiate and maintain behaviour change. It is therefore also influenced behaviour change separate to interven - reasonable to assume that at some point, participants tion usage. stop or reduce their usage because they have acquired the The findings from this study suggest that at least a knowledge and skills needed to engage in the behaviour moderate level of usage is required to facilitate greater without further use of the intervention, and this declin- improvements in ASI scores relative to the Control ing usage over time, or non-usage attrition, is very com- group (Table  2; Fig.  1). There was no significant differ - mon in many e- and m-Health interventions [12, 15, 26, ence between the Low usage group and Control group 28, 36]. in ASI scores, and the Low usage group had far lower There are limitations of this study. First, the original average usage than the moderate and high usage groups. trials were powered to detect statistically significant dif - This pattern of results is consistent with suggestions that ferences in their respective primary outcomes and not to some level of usage and engagement with digital health examine the relationship between app usage and behav- interventions is needed to change behaviours [15]. Yet, iour change. Second, the measures of physical activity the optimal amount of intervention usage required to and sleep used to construct the ASI are self-reported promote behaviour change remains unclear [36, 37], and and may be subject to bias. The reporting of sleep qual - is likely to depend on an individual characteristics, the ity using the Pittsburgh Sleep Quality Index has dem- outcome targeted and the inherent requirements of the onstrated good reliability [23] and validity in clinical intervention (i.e., daily self-monitoring vs. module based and non-clinical samples [38]. However, although there intervention) [37]. Related to this is the issue of intended is some evidence that the Active Australia Survey has use relative to actual use of the intervention. While it was acceptable levels of criterion validity, it was designed as intended that participants could self-monitor any of the a population surveillance instrument and may not be six physical activity and sleep metrics daily throughout sensitive to detecting changes over time during the inter- the intervention period, the average usage scores indicate vention [39]. Third, there were some differences between most participants didn’t use the intervention in this way usage groups in terms of baseline behaviour and socio- and could be considered non-adherers to the interven- demographics, which are overcome in part by adjusting tion. This relationship is also compounded by the fact statistical analyses for the baseline value of the outcome. Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 8 of 9 Competing interests Fourth, data on the duration of time spent using the None of the authors have any competing interests. intervention (which has been shown to be associated with behaviour change in other studies [28, 40]) was not Author details School of Medicine and Public Health, Faculty of Health and Medicine, Col- captured in the intervention database so could not be lege of Health, Medicine, and Wellbeing, University of Newcastle, University examined. Drive, Callaghan 2308, NSW, Australia. Active Living Research Program, Hunter Medical Research Institute, New Lambton Heights, NSW 2305, Australia. Applied Sport Science and Exercise Testing Laboratory, School of Environ- Conclusion mental and Life Sciences, University of Newcastle, Ourimbah, NSW 2258, Overall, there was a weak relationship between app usage Australia. School of Education, University of Newcastle, Callaghan, NSW 2308, Australia. School of Human Movement and Nutrition Studies, The University and behaviour change in the intervention. Relative to the of Queensland, Brisbane, QLD 4072, Australia. Faculty of Health Sciences Control group, only the Mid- and High-usage interven- 7 and Medicine, Bond University, Gold Coast, QLD 4226, Australia. School tion groups had improved overall patterns of physical of Health, Medical and Applied Sciences, Appleton Institute, Physical Activity Research Group, Central Queensland University, Rockhampton, QLD 4701, activity and sleep behaviours after 3 months, with only Australia. the High-usage benefits remaining after 6 months. Col- lectively these findings suggest that a multidimensional Received: 2 November 2022 Accepted: 8 December 2022 metric of intervention usage has a small influence on behaviour change and that other factors are likely to be key drivers of behaviour change. References 1. Müller AM, et al. The effectiveness of e-& mHealth interventions to Supplementary Information promote physical activity and healthy diets in developing countries: a The online version contains supplementary material available at https:// doi. systematic review. Int J Behav Nutr Phys Activity. 2016;13(1):109. org/ 10. 1186/ s44167- 022- 00013-1. 2. Shin JC, Kim J, Grigsby-Toussaint D. Mobile phone interventions for sleep disorders and sleep quality: systematic review. J Med Internet Res mHealth uHealth. 2017;5(9):e131. Additional file 1: Table S1. Eligibility criteria, intervention components 3. Jones KR, Lekhak N, Kaewluang N. Using mobile phones and short andoutcome measures used the Synergy and Refresh Studies. Table S1a. message service to deliver self-management interventions for chronic Operationalisation of social cognitive factorsand behaviour change strate- conditions: a meta‐review. Worldviews Evid‐Based Nurs. 2014;11(2):81–8. gies in the Physical Activity and Sleep HealthIntervention as described in 4. Neubeck L, et al. The mobile revolution—using smartphone apps to the original Synergy protocol paper. Table S3 Comparison ofcompleters prevent cardiovascular disease. Nat Reviews Cardiol. 2015;12(6):350. and non-completers in control and intervention groups pooled fromSyn- 5. Oftedal S, et al. Prevalence, temporal trends, and correlates of joint ergy and Refresh trials. Fig. S1. Screenshots ofBalanced app screens for patterns of aerobic activity, muscle-strengthening activity, and sleep self-monitoring and feedback relative to goals. duration: a pooled analysis of 359,019 adults in the NHIS 2004–2018. J Phys Activ Health. 2022;19:246. 6. Rayward AT, et al. A cross-sectional cluster analysis of the combined asso- Author contributions ciation of physical activity and sleep with sociodemographic and health MJD conceptualised the initial study. LLM, MJD BJD drafted the initial manu- characteristics in mid-aged and older adults. Maturitas. 2017;102:56–61. script and refined the study approach. MJD completed statistical analyses 7. Murawski B, et al. Efficacy of an m-Health physical activity and sleep and LLM assisted. ATR compiled the data from the two trials. LLM, BJD, BM, health intervention for adults: a randomized waitlist-controlled trial. Am J ATR, WJB, RCP, CV, EGH, and MJD provided critical review of the manuscript. Prev Med. 2019;57(4):503–14. All authors reviewed the manuscript. All authors read and approved the final 8. Rayward AT, et al. Efficacy of an m-Health physical activity and sleep manuscript. intervention to improve sleep quality in middle-aged adults: the refresh study randomized controlled trial. Ann Behav Med. 2020;54(7):470–83. Funding 9. Ding D, et al. A widening gap? Changes in multiple lifestyle risk behav- This research was in part supported by a Future Leader Fellowship (ID 100029) iours by Socioeconomic Status in New South Wales, Australia, 2002–2012. and Vanguard Grant (ID 100629) from the National Heart Foundation of Aus- PLoS ONE. 2015;10(8):e0135338. tralia, awarded to Mitch J. Duncan. Mitch J. Duncan is supported by a Career 10. Smith N, Liu S. A systematic review of the dose–response relationship Development Fellowship from the National Health and Medical Research between usage and outcomes of online weight-loss interventions. Inter- Council (APP1141606). Corneel Vandelanotte is supported by a Future Fellow- net Intervent. 2020;22: 100344. ship from the Australian Research Council (FT210100234). 11. McLaughlin M, et al. Associations between digital health intervention engagement, physical activity, and sedentary behavior: systematic review Availability of data and materials and meta-analysis. J Med Internet Res. 2021;23(2):e23180. The datasets used and analysed during the current study are available from 12. Donkin L, et al. A systematic review of the impact of adherence on the the corresponding authors on reasonable request. effectiveness of e-therapies. J Med Internet Res. 2011;13(3):e52. 13. Vandelanotte C, et al. Past, Present, and future of eHealth and mHealth Declarations Research to improve physical activity and dietary behaviors. J Nutr Educ Behav. 2016;48(3):219-28.e1. Ethics approval and consent to participate 14. Perski O, et al. Conceptualising engagement with digital behaviour All procedures performed in studies involving human participants were in change interventions: a systematic review using principles from critical accordance with the ethical standards of the institutional and/or national interpretive synthesis. Transl Behav Med. 2017;7(2):254–67. research committee and with the 1964 Helsinki declaration and its later 15. Short CE, et al. Measuring engagement in eHealth and mHealth behavior amendments or comparable ethical standards Informed consent was change interventions: viewpoint of methodologies. J Med Internet Res. obtained from all individual participants included in the study. Approval 2018;20(11):e292. was provided by the University of Newcastle Human Research Ethics Commit- 16. Yardley L, et al. Understanding and promoting effective engage - tee: H-2016-0267; H-2016-0181. ment with digital behavior change interventions. Am J Prev Med. 2016;51(5):833–42. M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 9 of 9 17. Weber R. Evaluating and developing theories in the information systems 40. Edney S, et al. User engagement and attrition in an app-based physical discipline. J Association Inform Syst. 2012;13(1):2. activity intervention: secondary analysis of a randomized controlled trial. 18. Murawski B, et al. Randomised controlled trial using a theory-based J Med Internet Res. 2019;21(11):e14645. m-health intervention to improve physical activity and sleep health in adults: the Synergy Study protocol. BMJ Open. 2018;8(2): e018997. Publisher’s Note 19. Rayward AT, et al. A randomised controlled trial to test the efficacy of an Springer Nature remains neutral with regard to jurisdictional claims in pub- m-health delivered physical activity and sleep intervention to improve lished maps and institutional affiliations. sleep quality in middle-aged adults: the Refresh Study Protocol. Contemp Clin Trials. 2018;73:36–50. 20. Duncan MJ, et al. Eec ff t of a physical activity and sleep m-health interven- tion on a composite activity-sleep behaviour score and mental health: a mediation analysis of two randomised controlled trials. Int J Behav Nutr Phys Activity. 2021;18(1):45. 21. Rayward AT, et al. Efficacy of an m-Health physical activity and sleep intervention to improve sleep quality in middle-aged adults: the refresh study randomized controlled trial. Annal Behav Med. 2020;54:470. 22. Health AIo, Welfare. The active Australia Survey: a guide and manual for implementation, analysis and reporting. AIHW: Canberra; 2003. 23. Buysse DJ, et al. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193–213. 24. Morin CM, et al. The Insomnia Severity Index: psychometric indica- tors to detect insomnia cases and evaluate treatment response. Sleep. 2011;34(5):601–8. 25. Rayward AT, et al. The association between logging steps using a website, app, or fitbit and engaging with the 10,000 steps physical activity pro - gram: observational study. J Med Internet Res. 2021;23(6):e22151. 26. Kolt GS, et al. Using web 2.0 applications to promote health-related physical activity: findings from the WALK 2.0 randomised controlled trial. Br J Sports Med. 2017;51(19):1433–40. 27. Duncan M, et al. Eec ff tiveness of a web- and mobile phone-based intervention to promote physical activity and healthy eating in middle- aged males: randomized controlled trial of the ManUp study. JMIR. 2014;16(6):e136. 28. Vandelanotte C, et al. Are web-based personally tailored physical activity videos more effective than personally tailored text-based interventions? Results from the three-arm randomised controlled TaylorActive trial. Br J Sports Med. 2020;55:336. 29. Li Y, et al. Dose–response effects of patient engagement on health out - comes in an mHealth intervention: secondary analysis of a randomized controlled trial. JMIR Mhealth Uhealth. 2022;10(1):e25586. 30. Short CE, et al. Designing engaging online behaviour change inter- ventions: a proposed model of user engagement. Eur Health Psychol. 2015;17(1):32–8. 31. Perski O, et al. Conceptualising engagement with digital behaviour change interventions: a systematic review using principles from critical interpretive synthesis. Transl Behav Med. 2016;7(2):254–67. 32. Short CE, et al. Exploring the interplay between message format, need for cognition and personal relevance on processing messages about physical activity: a two-arm randomized experimental trial. Int J Behav Med. 2022. 33. Rayward AT, et al. Associations between changes in activity and sleep quality and duration over two years. Med Sci Sports Exerc. 2018;50(12):2425–32. 34. Huang B-H, et al. The bidirectional association between sleep and physical activity: a 6.9 years longitudinal analysis of 38,601 UK Biobank participants. Prev Med. 2020;143:106315. 35. Kline CE. The bidirectional relationship between exercise and sleep: impli- Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : cations for exercise adherence and sleep improvement. Am J Lifestyle Med. 2014;8(6):375–9. fast, convenient online submission 36. Duncan MJ, et al. Efficacy of a multi-component m-Health weight-loss thorough peer review by experienced researchers in your field intervention in overweight and obese adults: a randomised controlled trial. Int J Environ Res Public Health. 2020;17(17):6200. rapid publication on acceptance 37. Ritterband LM, et al. A behavior change model for internet interventions. support for research data, including large and complex data types Ann Behav Med. 2009;38(1):18–27. • gold Open Access which fosters wider collaboration and increased citations 38. Mollayeva T, et al. The Pittsburgh sleep quality index as a screening tool for sleep dysfunction in clinical and non-clinical samples: a systematic maximum visibility for your research: over 100M website views per year review and meta-analysis. Sleep Med Rev. 2016;25:52–73. 39. Vandelanotte C, et al. Validity and responsiveness to change of the active At BMC, research is always in progress. Australia Survey according to gender, age, BMI, education, and physical Learn more biomedcentral.com/submissions activity level and awareness. BMC Public Health. 2019;19(1):407. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Activity Sedentary and Sleep Behaviors Springer Journals

Associations between app usage and behaviour change in a m-health intervention to improve physical activity and sleep health in adults: secondary analyses from two randomised controlled trials

Download PDF
9 pages

Loading next page...
 
/lp/springer-journals/associations-between-app-usage-and-behaviour-change-in-a-m-health-ItjV0th8oy

References (42)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2023
eISSN
2731-4391
DOI
10.1186/s44167-022-00013-1
Publisher site
See Article on Publisher Site

Abstract

Background To examine associations between user engagement and activity-sleep patterns in a 12-week m-health behavioural intervention targeting physical activity and sleep. Methods This secondary analysis used data pooled from two Randomised Control Trials (RCT, [Synergy and Refresh]) that aimed to improve physical activity and sleep (PAS) among physically inactive adults with poor sleep. Both RCTs include a PAS intervention group (n = 190 [Synergy n = 80; Refresh n = 110]) and a wait list Control (CON n = 135 [Synergy n = 80; Refresh n = 55]). The PAS groups received a pedometer and accessed a smartphone/tablet “app” with behaviour change strategies, and email/SMS support. Activity-sleep patterns were quantified using the activity-sleep behaviour index (ASI) based on self-report measures. Intervention usage was quantified as a composite score of the frequency, intensity and duration of app usage during intervention (range: 0–30). Assessments were conducted at baseline, 3 and 6 months. Relationships between usage and ASI were examined using generalised linear models. Differences in ASI between the control group and intervention usage groups (Low [0–10.0], Mid [10.1–20.0], High [20.1–30.0]) were examined using generalised linear mixed models adjusted for baseline values of the outcome. Trial Registration: ACTRN12617000376347; ACTRN12617000680369. Results During the 3-month intervention, the mean (± sd) usage score was 18.9 ± 9.5. At 3 months (regression coef- ficient [95%CI]: 0.45 [0.22, 0.68]) and 6 months (0.48 [0.22, 0.74]) there was a weak association between usage score and ASI in the intervention group. At 3 months, ASI scores in the Mid (Mean [95%CI] = 57.51 [53.99, 61.04]) and High (60.09 [57.52, 62.67]) usage groups were significantly higher (better) than the control group (51.91 [49.58, 54.24]), but not the Low usage group (47.49 [41.87, 53.12]). Only differences between the high usage and control group remained at 6 months. *Correspondence: Mitch J. Duncan Mitch.Duncan@newcastle.edu.au Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 2 of 9 Conclusion These findings suggests that while higher intervention usage is associated with improvements in behav- iour, the weak magnitude of this association suggests that other factors are also likely to influence behaviour change in m-health interventions. Trial registration number: ACTRN12617000376347; ACTRN12617000680369. Keywords eHealth, Usage, Engagement, Attrition, Dose-response Introduction composite measures may be more useful in understand- Delivering behaviour change interventions through elec- ing the usage-behaviour change relationship [14–16], tronic health (e-Health) and mobile health (m-Health) although few studies have applied such multidimensional creates the opportunity to deliver cost-effective wide- measures. The overall aim of this study was to examine reaching interventions. Both e- and m-Health interven- how user engagement with a m-health app is associated tions have demonstrated effectiveness for improving with behaviour change during two randomised con- lifestyle behaviours including physical activity [1], sleep trolled trials (RCT) of a m-health intervention targeting [2], alcohol consumption and chronic disease manage- improvements in physical activity and sleep health in ment [3, 4]. Approximately 30% of adults are both physi- physically inactive adults with poor sleep. Two specific cally inactive and have poor sleep health [5, 6] and may objectives were to: (1) examine the relationship between benefit from interventions to improve these behaviours. user engagement and a composite score of overall physi- However, there are few interventions that have targeted cal activity and sleep health (Activity-Sleep Index [ASI]) improvements in both physical activity and sleep [7, 8]. within the intervention group and, (2) compare differ - Consequently, relatively little is known about how par- ences in overall physical activity and sleep health in the ticipants use and engage with digital physical activity and control group and in different levels of app usage (Low, sleep interventions. Characterising engagement with dig- Mid, High) in the intervention group. ital interventions that target multiple lifestyle behaviours is important given the number of adults who engage in Methods multiple higher risk behaviours [9] and the need for Study design interventions to address multiple lifestyle behaviours This study uses data pooled from two separate RCTs of concurrently. the same m-health intervention, which was designed to Understanding how participants use and engage improve physical activity and sleep health behaviours in with digital interventions is important, as while it is physically inactive adults with poor sleep quality [7, 8]. typically assumed that greater usage is associated with Details of the study rationale, methods and main out- greater behaviour change, the magnitude of this rela- comes of each trial, and intervention effects on the ASI tionship appears to be weak [10, 11]. Furthermore, it is are available elsewhere [7, 8, 18–20]. Similarities between consistently reported that usage declines throughout the trials in terms of the behaviours assessed, inter- the intervention period [12, 13]. Additionally, there are vention and control groups, assessment methods and inconsistencies between studies regarding how usage outcomes assessed allowed data from the control and is conceptualised and measured, which limits compari- physical activity and sleep health intervention groups to sons between studies [11, 14–17]. A systematic review be pooled as described previously [20]. Eligible partici- reported that a greater subjective user experience of the pants were those aged 18–55 years (Synergy Study [7, 18]) intervention, greater number of activities completed and or 45–65 years (Refresh Study [8, 19]), who lived in Aus- more frequent logins are consistently associated with tralia, reported < 90 min of moderate to vigorous intensity greater physical activity, but that time on the website physical activity (MVPA) in the last week and rated their was not associated with physical activity [11]. This sug - sleep quality as fairly bad or very bad. Exclusion criteria gests the usage-behaviour change relationship may dif- included being employed in shift-work, diagnosed sleep fer depending on the usage metric examined and that disorder, and current use of a device to track activity or single usage metrics may not adequately characterise sleep (see Additional file 1: Figs. S1, S2). how participants use and engage with the intervention. Both studies primarily recruited participants using To overcome this, Short and colleagues [15] proposed a social media advertising. The Synergy study aimed to composite measure of usage that captures the frequency compare the efficacy of a combined physical activity and (i.e., number of self-monitoring entries or logins), inten- sleep health intervention with a wait-list control. Partici- sity (i.e., number of intervention features used), duration, pants (n = 160; mean age: 41.5 (SD = 9.9); 80% female) and type (i.e., reflective, didactic, or active) of usage. Such were recruited between June–August 2017 and the study M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 3 of 9 was conducted between June 2017 and February 2018. All self-monitoring data entered into the application The Refresh study aimed to compare the efficacy of a were recorded in the application database, including combined physical activity and sleep health interven- the associated timestamp of entry. The waitlist-control tion with a sleep health-only intervention and a wait-list group did not have any access to the application or other control [19, 21]. Participants (n = 275; mean age: 52.0 intervention materials prior to the 6-month assessment. (SD = 6.9); 83% female) were recruited between May– However, they were offered full access to the intervention September 2017 and the study was conducted between including the “Balanced” application after completing the June 2017 and March 2018. The combined physical final 6-month assessment. activity and sleep health intervention in both trials was the same in terms of mode of delivery, theoretical basis, Measures educational content, and behaviour change techniques Sociodemographic variables such as age, gender, educa- used. The sleep health-only intervention arm (n = 110) tion and chronic disease status were assessed at baseline in Refresh was omitted from the current study as partici- and primary and secondary outcomes were measured at pants in that group did not receive any physical activity baseline, three and six months [7, 8]. Primary and sec- intervention content. The active phase of the intervention ondary measures of the original trials included minutes in both studies ceased at the 3-month point. Both studies of MVPA [22], the frequency of resistance training [18, conducted online assessments at baseline, 3 months and 19], sleep quality [23] and insomnia symptoms [24]. 6 months and were prospectively registered with the Aus- tralian and New Zealand Clinical Trials Registry as well Activity sleep index as received ethical approval (H-2016-0181, H-2016-0267) The overall pattern of physical activity and sleep was at the University of Newcastle. Participants in both trials quantified using the activity-sleep index (ASI), which provided informed consent. Each study used computer is a 12-item instrument described elsewhere [20]. It is generated permuted block randomisation to develop designed to assess overall healthy patterns of physical the randomisation sequence, with group allocation con- activity and sleep health based on the frequency, dura- cealed in sequentially numbered envelopes. Participants tion, type and intensity of physical activity, the duration were not blinded to group allocation given the nature of of sitting time and the duration, timing, quality and sat- the interventions. isfaction of sleep. The specific items, responses, and scor - ing for the ASI are provided in Additional file  1: Table S2. Study groups The items are briefly summarised here: The physical activity and sleep (PAS) intervention group (n = 190 [Synergy: n = 80 + Refresh: n = 110]) received 1. Frequency–MVPA (Number of sessions of MVPA/ access to a specifically designed mobile application “Bal - wk), anced” that comprised educational resources, personal 2. Frequency–RT (Number of days of resistance goals, self-monitoring logs (manual data entry), and feed- training/wk), back in relation to personal goals, all relative to a range 3. Intensity (Proportion of MVPA that was vigorous of physical activity and sleep health components (i.e., in intensity), activity minutes, step count, resistance training, bedtime, 4. Type (Participation in no MVPA or resistance sleep wake timing and sleep quality). Details of the inter- training, either MVPA or resistance training, or vention are provided in Additional file  1: Tables S1 and both), Fig. S3. Prior to commencement, the intervention group 5. Time (Duration of MVPA/wk), participants were mailed a printed participant handbook 6. Sitting (Duration of sitting time/wk). with guidance on how to use the app, and a pedometer. 7. Daytime alertness (Trouble staying awake during Participants also received weekly reports and short mes- the day), sage service (SMS) prompts to limit disengagement. 8. Sleep Quality (Overall sleep quality rating), Participants used the app for goal setting and action plan- 9. Sleep Timing (Midpoint of sleep between 02:00 am ning to increase their physical activity (i.e., MVPA, step and 04:00 am), counts and resistance training) and to improve their sleep 10. Sleep Regularity (Variability in bed and wake quality and sleep behaviours (stabilising bed/wake times, times), sleep hygiene behaviours and stress management (e.g., 11. Sleep Efficiency (Sleep efficiency ([sleep duration/ progressive muscle relaxation, deep breathing exercises, time in bed] × 100), mindfulness) [20]. All intervention components were 12. Sleep Duration (Meeting age-appropriate sleep delivered either through the application, email or SMS, duration guidelines). and the messaging component ceased at three months. Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 4 of 9 Data analysis To overcome the different metrics used to quantify Descriptive statistics are presented for the sample at each dimension, each dimension was rescaled to a zero to baseline in each intervention group, and also by interven- 10 scale, with higher scores reflecting lower risk behav - tion usage group. To examine the association between iour. Each dimension is summed to create a score ranging engagement and behaviour change, two separate analyses from 0 to 120. The approach used to rescale the individ - were conducted. The first analysis was limited to only the ual dimensions of the ASI was X−X min intervention group as no usage data were available for the Rescaled score = n where X is the observed Range control group. This analysis examined the relationship value, X is the minimum observed value of the original min between the overall usage score and the ASI at follow- variable, X is the difference between the minimum Range up, adjusted for the baseline value of the outcome, using and maximum of the observed values, and n is upper a generalised linear model. The model included fixed limit of the rescaled variable (e.g., n = 10) [20]. effect for the continuous mean centred usage score, study (Refresh, Synergy), assessment (3 months, 6 months) and the interaction between usage score and assessment. The Intervention usage linearity of the relationship between continuous usage An overall usage score was created to capture interven- score and ASI in the intervention group was examined tion group participants’ frequency, intensity, and dura- using residual plots and including a quadratic term for tion of using the “Balanced” intervention platform using usage score in the analysis. The quadratic term was not data recorded in intervention database. There are no statistically significant (p < 0.05) and the residuals plots usage data for the wait-list control group as they did not did not indicate a non-linear relationship. To examine have access to the app during the intervention period. how varying amounts of usage in the intervention group Type of usage (i.e., reflective, gamified, altruistic, didac - were associated with behaviour change relative to the tic, or active) was not examined as these data were not waitlist-control group, overall usage scores in the inter- recorded in the intervention platform. All indicators were vention group were categorised into a three-level group assessed over the initial 84-day (i.e., 3-month) interven- variable: low usage (0–10); mid usage (10.1–20.0); and tion period to align with the ‘active’ component of the high usage (20.1–30.0) and combined with the control intervention. Each day, participants could self-report group to create a four-level variable. This analysis exam - their: (1) minutes of MVPA, (2) resistance training, (3) ined between group differences (Control, Low usage, daily step count, (4) bedtime, (5) wake time and, (6) sleep Mid usage, High usage) in the ASI adjusted for the base- quality. The app was designed to promote daily self-mon - line value of the outcome. The model included a fixed itoring of these metrics, however participants were free effect for study (Refresh, Synergy), assessment (3 months, to self-monitor any number of these metrics on a given 6 months), group (Control, Low usage, Mid usage, High day. These measures were used to create measures of the usage), and the group by assessment interaction. Residual frequency, intensity and duration of usage. Frequency diagnostics were used to inform the choice of model and was measured as the total number of self-monitoring link. Analyses were conducted using Stata MP v17 and entries made during the 3-month (84 days) intervention alpha was set at 0.05. period, with a maximum of six entries per day (one entry per day for each of the self-monitoring entries made). Intensity was measured as the average number of self- Results monitoring entries made each day during the interven- Participant flow throughout each trial is shown in Addi - tion. Duration was measured as the number of days until tional file  1: Figs. S1 and S2. A total of 325 participants a participant succumbed to non-usage attrition, defined completed the baseline survey, 275 (84.6%) completed as the time they first stopped self-monitoring for at least the 3-month assessment, and 215 (66%) completed the 14 consecutive days [7, 21, 25, 26]. Due to the different 6-month assessment. Completers of the 3-month assess- metrics used to characterise each usage dimension (i.e., ment were older (M = 47.41 [SD = 9.73]) and had higher count of self-monitoring entries per day, number of days) levels of intervention usage (M = 20.73 [8.76]) than non- each dimension was rescaled to a zero to 10 scale as fol- completers (Additional file  1: Table S3). The baseline sam - X−X min lows rescaled score = n ; where X is the original ple consisted of 264 female and 61 male participants, and range most were middle-aged and highly educated (Table  1). score, X is the minimum of the observed variable, min At baseline the average BMI and ASI were 28.15 kg/m X is the range of the potential score and n is the range (SD = 4.21) and 47.34 (SD = 10.91), respectively and both upper limit of the rescaled score [20]. The rescaled these variables were similar between intervention and dimensions were summed to create an overall usage control groups. At baseline the low usage group reported score ranging from zero to thirty with higher values indi- lower average ASI scores (M = 44.19 [SD = 14.11]), cating greater usage. M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 5 of 9 Table 1 Baseline descriptive characteristics of participants by study and intervention usage group Group Intervention usage group Control n = 135 Intervention n = 190 Low Usage n = 34 Mid Usage n = 60 High Usage n = 94 Total N = 325 M (SD), n (%) M (SD), n (%) M (SD), n (%) M (SD), n (%) M (SD), n (%) M (SD), n (%) Age (years) 46.19 (10.39) 47.22 (9.71) 47.12 (9.37) 45.47 (10.53) 48.30 (9.29) 46.76 (10.02) Education (years) 16.26 (2.90) 16.11 (2.77) 16.03 (3.05) 15.73 (2.99) 16.38 (2.53) 16.17 (2.82) BMI 27.67 (4.09) 28.50 (4.27) 28.71 (4.48) 29.30 (4.33) 27.89 (4.10) 28.15 (4.21) ASI Score 46.83 (10.03) 47.83 (11.56) 44.19 (14.11) 47.58 (10.54) 49.06 (10.90) 47.34 (10.91) Sex Male 26 (19.26%) 35 (18.42%) 7 (20.59%) 12 (20.00%) 16 (17.02%) 61 (18.89%) Female 109 (80.74%) 155 (81.58%) 27 (79.41%) 48 (80.00%) 78 (82.98%) 262 (81.11%) Income/Yr ≤ $30,000 32 (23.70%) 30 (15.79%) 2 (5.88%) 11 (18.33%) 17 (18.09%) 62 (19.20%) $30,001–$50,000 14 (10.37%) 26 (13.68%) 3 (8.82%) 6 (10.00%) 17 (18.09%) 40 (12.38%) $50,001–$70,000 26 (19.26%) 39 (20.53%) 8 (23.53%) 18 (30.00%) 13 (13.83%) 65 (20.12%) $70,001–$100,000 25 (18.52%) 47 (24.74%) 7 (20.59%) 13 (21.67%) 26 (27.66%) 71 (21.98%) ≥ $100,001 30 (22.22%) 30 (15.79%) 10 (29.41%) 8 (13.33%) 11 (11.70%) 59 (18.27%) Don’t know/no answer 8 (5.93%) 18 (9.47%) 4 (11.76%) 4 (6.67%) 10 (10.64%) 26 (8.05%) Employment group Professional 85 (62.96%) 107 (56.32%) 21 (61.76%) 31 (51.67%) 55 (58.51%) 192 (59.44%) White-collar 18 (13.33%) 37 (19.47%) 6 (17.65%) 16 (26.67%) 14 (14.89%) 54 (16.72%) Blue-collar 5 (3.70%) 4 (2.11%) 0 (0.00%) 1 (1.67%) 3 (3.19%) 9 (2.79%) Not working 27 (20.00%) 42 (22.11%) 7 (20.59%) 12 (20.00%) 22 (23.40%) 68 (21.05%) Employment group not working includes retired, unemployed, home duties, looking for work, student, and other. Low, Mid and High usage defined as usage score of 0–10.0, 10.1–20.0 and 20.1–30.0, respectively with a higher proportion reported higher income lev- differences were only maintained in the high usage group els (≥$100,001/yr) relative to the Mid and High usage (M = 60.82, 95% CI = 57.96, 63.67; difference to control groups. M = 8.86, 95% CI = 5.04, 12.68) at 6 months (Table  2; The average intervention usage score was 18.88 Fig. 2). (SD = 9.54) out of 30, and the average usage scores in the Low, Mid and High usage groups were 3.49 (SD = 4.10), Discussion 14.78 (SD = 2.97), and 27.03 (SD = 3.00), respectively. This study examined the association between usage of In the intervention group, there was a weak association a m-health intervention app and overall physical activ- between intervention usage score and ASI at 3 months ity and sleep health behaviour. In the intervention group (Β = 0.45, 95% CI = 0.22, 0.68) indicating that for each there was a weak positive relationship between usage and 1 unit increase in intervention usage score there was an behaviour at 3 and 6 months. Consistent with this obser- estimated mean 0.45 increase in ASI. The association vation, when comparing the Low, Mid and High usage between intervention usage score and ASI at 6 months groups to the Control group, only the Mid and High was of a similar magnitude (Β = 0.48, 95% CI = 0.22, usage groups demonstrated significantly higher (better) 0.74). The estimated marginal mean ASI scores at 3 and overall physical activity and sleep health behaviours at 6 months for different levels of intervention usage are 3 months. These differences were only maintained in the shown in Fig.  1. Exploring the effect of Low, Moderate High usage group at 6 months. These results indicate that and High usage scores in the intervention group rela- while there is a weak relationship between intervention tive to the control group, the results indicated that at usage and behaviour change, it appears that only mid-to- 3 months, the mid usage (M = 57.51, 95% CI = 53.99, high levels of usage are associated with small improve- 61.04; difference to control M = 5.60, 95% CI = 1.39, 9.81) ments in behaviour relative to the control. and high usage groups (M = 60.09, 95% CI = 57.52, 62.67; Overall, intervention usage scores were modest and difference to control M = 8.18, 95% CI = 4.68, 11.68) had there was a weak relationship with behaviour at three significantly higher ASI scores relative to the control and six months. Prior analysis of these trials indicates group (M = 51.91, 95% CI = 49.58, 54.24). However, these that overall activity-sleep behaviours did significantly Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 6 of 9 Fig. 1 Baseline adjusted ASI-12 at 3 and 6 months by usage score in intervention group. Model only includes the pooled intervention group. Model adjusted for study (i.e., Synergy, Refresh), baseline ASI-12 score, and includes the mean centered usage score and its interaction with assessment. p-value for interaction between usage score and assessment is = 0.857 The association between usage score and ASI-12 at 3 months is = Β=, 95%CI: b = 0.45,95%CI = 0.22, 0.68 and at 6 months is Β = 0.48, 95%CI = 0.22, 0.74 Table 2 Baseline adjusted ASI-12 by intervention and usage group at 3 and 6 months 3 Months 6 Months M [95%CI] Diff. to control M [95%CI] Diff. to control M [95%CI] M [95%CI] Control 51.91 [49.58, 54.24] 51.95 [49.47, 54.44] Low usage (0–10.0) 47.49 [41.87, 53.12] − 4.42 [− 10.55, 1.71] 48.71 [42.47, 54.94] − 3.25 [− 9.98, 3.49] Mid usage (10.1–20.0) 57.51 [53.99, 61.04] 5.60 [1.39, 9.81] 55.72 [51.18, 60.27] 3.77 [− 1.40, 8.94] High usage (20.1–30.0) 60.09 [57.52, 62.67] 8.18 [4.68, 11.68] 60.82 [57.96, 63.67] 8.86 [5.04, 12.68] Model adjusted for baseline value of the outcome, and study. p-value for group by time interaction = 0.841. There were 34, 60 and 94 participants in the low, mid and high usage groups, respectively improve in the intervention group relative to the con- is considerable variation in these associations in differ - trol group [20]. Collectively this suggests that while the ent studies [11]. Overall these observations are consistent use and engagement with the intervention platform has with conceptual frameworks of engagement-behaviour some influence, it is not a major driver of behaviour change that identify platform usage as one of several fac- change. Previous studies have observed no statistically tors [30, 31], including psycho-social factors related to significant [27] or weak [28] associations between vari - behaviour change [30], personal relevance of informa- ous measures of app usage and behaviour change, while tion provided [32], and inclusion of behaviour change others have observed positive dose-response relation- techniques in the intervention [31], that can influence ships between greater usage and improvements in health behaviour change. This has important implications for outcomes [29]. Similarly, a meta-analysis summarising the design of future m-health interventions. Specifically, the association between app engagement and change in interventions need not only to be designed to promote physical activity behaviour observed that, while there is and foster a certain degree of user engagement with the a weak statistically significant relationship between usage intervention platform, but they also need to incorporate and physical activity (b = 0.08, 95% CI = 0.01–0.14), there other important features related to behaviour change. Of M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 7 of 9 Fig. 2 Baseline adjusted ASI-12 at 3 and 6 months by Intervention and Usage Group. Model adjusted for study (i.e., Synergy, Refresh), baseline ASI-12 score, fixed effects for group (Control, Low Usage, Mid Usage, High Usage), assessment (3 months, 6 months) and the group by assessment interaction relevance to this study which targeted improvements in that the vast majority of m-health, and digital interven- physical activity and sleep, is the bidirectional relation- tions aim to improve participants’ knowledge and skills ship between these behaviours [33–35] which may have to initiate and maintain behaviour change. It is therefore also influenced behaviour change separate to interven - reasonable to assume that at some point, participants tion usage. stop or reduce their usage because they have acquired the The findings from this study suggest that at least a knowledge and skills needed to engage in the behaviour moderate level of usage is required to facilitate greater without further use of the intervention, and this declin- improvements in ASI scores relative to the Control ing usage over time, or non-usage attrition, is very com- group (Table  2; Fig.  1). There was no significant differ - mon in many e- and m-Health interventions [12, 15, 26, ence between the Low usage group and Control group 28, 36]. in ASI scores, and the Low usage group had far lower There are limitations of this study. First, the original average usage than the moderate and high usage groups. trials were powered to detect statistically significant dif - This pattern of results is consistent with suggestions that ferences in their respective primary outcomes and not to some level of usage and engagement with digital health examine the relationship between app usage and behav- interventions is needed to change behaviours [15]. Yet, iour change. Second, the measures of physical activity the optimal amount of intervention usage required to and sleep used to construct the ASI are self-reported promote behaviour change remains unclear [36, 37], and and may be subject to bias. The reporting of sleep qual - is likely to depend on an individual characteristics, the ity using the Pittsburgh Sleep Quality Index has dem- outcome targeted and the inherent requirements of the onstrated good reliability [23] and validity in clinical intervention (i.e., daily self-monitoring vs. module based and non-clinical samples [38]. However, although there intervention) [37]. Related to this is the issue of intended is some evidence that the Active Australia Survey has use relative to actual use of the intervention. While it was acceptable levels of criterion validity, it was designed as intended that participants could self-monitor any of the a population surveillance instrument and may not be six physical activity and sleep metrics daily throughout sensitive to detecting changes over time during the inter- the intervention period, the average usage scores indicate vention [39]. Third, there were some differences between most participants didn’t use the intervention in this way usage groups in terms of baseline behaviour and socio- and could be considered non-adherers to the interven- demographics, which are overcome in part by adjusting tion. This relationship is also compounded by the fact statistical analyses for the baseline value of the outcome. Murphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 8 of 9 Competing interests Fourth, data on the duration of time spent using the None of the authors have any competing interests. intervention (which has been shown to be associated with behaviour change in other studies [28, 40]) was not Author details School of Medicine and Public Health, Faculty of Health and Medicine, Col- captured in the intervention database so could not be lege of Health, Medicine, and Wellbeing, University of Newcastle, University examined. Drive, Callaghan 2308, NSW, Australia. Active Living Research Program, Hunter Medical Research Institute, New Lambton Heights, NSW 2305, Australia. Applied Sport Science and Exercise Testing Laboratory, School of Environ- Conclusion mental and Life Sciences, University of Newcastle, Ourimbah, NSW 2258, Overall, there was a weak relationship between app usage Australia. School of Education, University of Newcastle, Callaghan, NSW 2308, Australia. School of Human Movement and Nutrition Studies, The University and behaviour change in the intervention. Relative to the of Queensland, Brisbane, QLD 4072, Australia. Faculty of Health Sciences Control group, only the Mid- and High-usage interven- 7 and Medicine, Bond University, Gold Coast, QLD 4226, Australia. School tion groups had improved overall patterns of physical of Health, Medical and Applied Sciences, Appleton Institute, Physical Activity Research Group, Central Queensland University, Rockhampton, QLD 4701, activity and sleep behaviours after 3 months, with only Australia. the High-usage benefits remaining after 6 months. Col- lectively these findings suggest that a multidimensional Received: 2 November 2022 Accepted: 8 December 2022 metric of intervention usage has a small influence on behaviour change and that other factors are likely to be key drivers of behaviour change. References 1. Müller AM, et al. The effectiveness of e-& mHealth interventions to Supplementary Information promote physical activity and healthy diets in developing countries: a The online version contains supplementary material available at https:// doi. systematic review. Int J Behav Nutr Phys Activity. 2016;13(1):109. org/ 10. 1186/ s44167- 022- 00013-1. 2. Shin JC, Kim J, Grigsby-Toussaint D. Mobile phone interventions for sleep disorders and sleep quality: systematic review. J Med Internet Res mHealth uHealth. 2017;5(9):e131. Additional file 1: Table S1. Eligibility criteria, intervention components 3. Jones KR, Lekhak N, Kaewluang N. Using mobile phones and short andoutcome measures used the Synergy and Refresh Studies. Table S1a. message service to deliver self-management interventions for chronic Operationalisation of social cognitive factorsand behaviour change strate- conditions: a meta‐review. Worldviews Evid‐Based Nurs. 2014;11(2):81–8. gies in the Physical Activity and Sleep HealthIntervention as described in 4. Neubeck L, et al. The mobile revolution—using smartphone apps to the original Synergy protocol paper. Table S3 Comparison ofcompleters prevent cardiovascular disease. Nat Reviews Cardiol. 2015;12(6):350. and non-completers in control and intervention groups pooled fromSyn- 5. Oftedal S, et al. Prevalence, temporal trends, and correlates of joint ergy and Refresh trials. Fig. S1. Screenshots ofBalanced app screens for patterns of aerobic activity, muscle-strengthening activity, and sleep self-monitoring and feedback relative to goals. duration: a pooled analysis of 359,019 adults in the NHIS 2004–2018. J Phys Activ Health. 2022;19:246. 6. Rayward AT, et al. A cross-sectional cluster analysis of the combined asso- Author contributions ciation of physical activity and sleep with sociodemographic and health MJD conceptualised the initial study. LLM, MJD BJD drafted the initial manu- characteristics in mid-aged and older adults. Maturitas. 2017;102:56–61. script and refined the study approach. MJD completed statistical analyses 7. Murawski B, et al. Efficacy of an m-Health physical activity and sleep and LLM assisted. ATR compiled the data from the two trials. LLM, BJD, BM, health intervention for adults: a randomized waitlist-controlled trial. Am J ATR, WJB, RCP, CV, EGH, and MJD provided critical review of the manuscript. Prev Med. 2019;57(4):503–14. All authors reviewed the manuscript. All authors read and approved the final 8. Rayward AT, et al. Efficacy of an m-Health physical activity and sleep manuscript. intervention to improve sleep quality in middle-aged adults: the refresh study randomized controlled trial. Ann Behav Med. 2020;54(7):470–83. Funding 9. Ding D, et al. A widening gap? Changes in multiple lifestyle risk behav- This research was in part supported by a Future Leader Fellowship (ID 100029) iours by Socioeconomic Status in New South Wales, Australia, 2002–2012. and Vanguard Grant (ID 100629) from the National Heart Foundation of Aus- PLoS ONE. 2015;10(8):e0135338. tralia, awarded to Mitch J. Duncan. Mitch J. Duncan is supported by a Career 10. Smith N, Liu S. A systematic review of the dose–response relationship Development Fellowship from the National Health and Medical Research between usage and outcomes of online weight-loss interventions. Inter- Council (APP1141606). Corneel Vandelanotte is supported by a Future Fellow- net Intervent. 2020;22: 100344. ship from the Australian Research Council (FT210100234). 11. McLaughlin M, et al. Associations between digital health intervention engagement, physical activity, and sedentary behavior: systematic review Availability of data and materials and meta-analysis. J Med Internet Res. 2021;23(2):e23180. The datasets used and analysed during the current study are available from 12. Donkin L, et al. A systematic review of the impact of adherence on the the corresponding authors on reasonable request. effectiveness of e-therapies. J Med Internet Res. 2011;13(3):e52. 13. Vandelanotte C, et al. Past, Present, and future of eHealth and mHealth Declarations Research to improve physical activity and dietary behaviors. J Nutr Educ Behav. 2016;48(3):219-28.e1. Ethics approval and consent to participate 14. Perski O, et al. Conceptualising engagement with digital behaviour All procedures performed in studies involving human participants were in change interventions: a systematic review using principles from critical accordance with the ethical standards of the institutional and/or national interpretive synthesis. Transl Behav Med. 2017;7(2):254–67. research committee and with the 1964 Helsinki declaration and its later 15. Short CE, et al. Measuring engagement in eHealth and mHealth behavior amendments or comparable ethical standards Informed consent was change interventions: viewpoint of methodologies. J Med Internet Res. obtained from all individual participants included in the study. Approval 2018;20(11):e292. was provided by the University of Newcastle Human Research Ethics Commit- 16. Yardley L, et al. Understanding and promoting effective engage - tee: H-2016-0267; H-2016-0181. ment with digital behavior change interventions. Am J Prev Med. 2016;51(5):833–42. M urphy et al. Journal of Activity, Sedentary and Sleep Behaviors (2023) 2:4 Page 9 of 9 17. Weber R. Evaluating and developing theories in the information systems 40. Edney S, et al. User engagement and attrition in an app-based physical discipline. J Association Inform Syst. 2012;13(1):2. activity intervention: secondary analysis of a randomized controlled trial. 18. Murawski B, et al. Randomised controlled trial using a theory-based J Med Internet Res. 2019;21(11):e14645. m-health intervention to improve physical activity and sleep health in adults: the Synergy Study protocol. BMJ Open. 2018;8(2): e018997. Publisher’s Note 19. Rayward AT, et al. A randomised controlled trial to test the efficacy of an Springer Nature remains neutral with regard to jurisdictional claims in pub- m-health delivered physical activity and sleep intervention to improve lished maps and institutional affiliations. sleep quality in middle-aged adults: the Refresh Study Protocol. Contemp Clin Trials. 2018;73:36–50. 20. Duncan MJ, et al. Eec ff t of a physical activity and sleep m-health interven- tion on a composite activity-sleep behaviour score and mental health: a mediation analysis of two randomised controlled trials. Int J Behav Nutr Phys Activity. 2021;18(1):45. 21. Rayward AT, et al. Efficacy of an m-Health physical activity and sleep intervention to improve sleep quality in middle-aged adults: the refresh study randomized controlled trial. Annal Behav Med. 2020;54:470. 22. Health AIo, Welfare. The active Australia Survey: a guide and manual for implementation, analysis and reporting. AIHW: Canberra; 2003. 23. Buysse DJ, et al. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193–213. 24. Morin CM, et al. The Insomnia Severity Index: psychometric indica- tors to detect insomnia cases and evaluate treatment response. Sleep. 2011;34(5):601–8. 25. Rayward AT, et al. The association between logging steps using a website, app, or fitbit and engaging with the 10,000 steps physical activity pro - gram: observational study. J Med Internet Res. 2021;23(6):e22151. 26. Kolt GS, et al. Using web 2.0 applications to promote health-related physical activity: findings from the WALK 2.0 randomised controlled trial. Br J Sports Med. 2017;51(19):1433–40. 27. Duncan M, et al. Eec ff tiveness of a web- and mobile phone-based intervention to promote physical activity and healthy eating in middle- aged males: randomized controlled trial of the ManUp study. JMIR. 2014;16(6):e136. 28. Vandelanotte C, et al. Are web-based personally tailored physical activity videos more effective than personally tailored text-based interventions? Results from the three-arm randomised controlled TaylorActive trial. Br J Sports Med. 2020;55:336. 29. Li Y, et al. Dose–response effects of patient engagement on health out - comes in an mHealth intervention: secondary analysis of a randomized controlled trial. JMIR Mhealth Uhealth. 2022;10(1):e25586. 30. Short CE, et al. Designing engaging online behaviour change inter- ventions: a proposed model of user engagement. Eur Health Psychol. 2015;17(1):32–8. 31. Perski O, et al. Conceptualising engagement with digital behaviour change interventions: a systematic review using principles from critical interpretive synthesis. Transl Behav Med. 2016;7(2):254–67. 32. Short CE, et al. Exploring the interplay between message format, need for cognition and personal relevance on processing messages about physical activity: a two-arm randomized experimental trial. Int J Behav Med. 2022. 33. Rayward AT, et al. Associations between changes in activity and sleep quality and duration over two years. Med Sci Sports Exerc. 2018;50(12):2425–32. 34. Huang B-H, et al. The bidirectional association between sleep and physical activity: a 6.9 years longitudinal analysis of 38,601 UK Biobank participants. Prev Med. 2020;143:106315. 35. Kline CE. The bidirectional relationship between exercise and sleep: impli- Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : cations for exercise adherence and sleep improvement. Am J Lifestyle Med. 2014;8(6):375–9. fast, convenient online submission 36. Duncan MJ, et al. Efficacy of a multi-component m-Health weight-loss thorough peer review by experienced researchers in your field intervention in overweight and obese adults: a randomised controlled trial. Int J Environ Res Public Health. 2020;17(17):6200. rapid publication on acceptance 37. Ritterband LM, et al. A behavior change model for internet interventions. support for research data, including large and complex data types Ann Behav Med. 2009;38(1):18–27. • gold Open Access which fosters wider collaboration and increased citations 38. Mollayeva T, et al. The Pittsburgh sleep quality index as a screening tool for sleep dysfunction in clinical and non-clinical samples: a systematic maximum visibility for your research: over 100M website views per year review and meta-analysis. Sleep Med Rev. 2016;25:52–73. 39. Vandelanotte C, et al. Validity and responsiveness to change of the active At BMC, research is always in progress. Australia Survey according to gender, age, BMI, education, and physical Learn more biomedcentral.com/submissions activity level and awareness. BMC Public Health. 2019;19(1):407.

Journal

Journal of Activity Sedentary and Sleep Behaviors – Springer Journals

Published: Feb 2, 2023

Keywords: eHealth; Usage; Engagement; Attrition; Dose-response

There are no references for this article.