People with type 2 diabetes (T2D) have higher levels of sedentary behavior (SB) and lower levels of moderate to vigorous intensity physical activity (PA) compared with those without T2D [1]. Device-based measurement of SB has shown that people with T2D spend >64% of their daily waking time sedentary [1], which has subsequently sparked calls for targeted behavior change programs to reduce SB in this population, in addition to the promotion of regular PA [2].

SB has become a distinct target for individuals who are susceptible to or have been diagnosed with T2D [3]. It is suggested that people with T2D break up their sitting time by light PA, which can serve as a means for gradually adopting a more active lifestyle [4]. Reducing sitting time by either standing or light ambulation has been shown to improve glucose homeostasis among people at risk of or diagnosed with T2D [3].

Ubiquitous mobile health (mHealth) technologies such as smartphones and wearable sensors offer considerable potential to improve access to and delivery of behavior change support for reducing SB and promoting PA [5,6]. Progress in mHealth technology has enabled the design of just-in-time adaptive interventions (JITAIs), which aim to deliver behavior change support in real or near time, matched to when users need or want the support [7,8]. JITAIs harness information captured through mHealth technologies (smartphone sensors, GPS, and internet access) to determine which behavior change support component should be delivered and tailor it to an individual’s changing context (where) or status (when) [9]. The delivery of more contextually aware intervention support at more salient times to end users is hypothesized to enhance behavior change compared with other delivery approaches (eg, face-to-face and websites) [9].

It has been argued that conventional experimental designs are not sufficient to support the development of JITAIs because they do not enable researchers to determine empirically when a particular intervention component should be delivered and whether a JITAI that was delivered exerted the intended effect [10]. Traditional randomized controlled trial designs evaluate the overall effect of an intervention on a specific behavior or health outcome, not specific components of that intervention. Moreover, randomized controlled trials do not provide information on the particular time when the intervention had an effect and the moderators that affected the behavior change [11].

Other study designs, such as factorial designs, enable researchers to determine the effects of each intervention component, the interactions between components, and the dosing of the intervention. However, they are unable to delineate the time when the intervention was most receptive and what moderators influenced the intervention [11]. To address these limitations, Klasnja et al [11] proposed the microrandomized trial (MRT), which, they argued, could supplement the use of behavioral theory to guide the development of JITAIs. The MRT design provides data on the effects of specific intervention components, changes in effects over time, and potential moderators, including contextual and psychological factors [11]. Due to the dynamic nature of SB and PA that change throughout the day and in various contexts, adaptive designs such as MRT align well with interventions targeting SB and PA behavior change [12]. The MRT design has been suggested as necessary to leverage the potential of smartphones as a tool to improve PA and reduce SB [13]. MRT enables assessing the immediate effects of antisedentary and walking interventions [14] and therefore contributes to the optimization of SB and PA interventions.

Microrandomization refers to randomly allocating an intervention option at relevant times when a person would be most responsive to an intervention (based on theory, the person’s past behavior, and the current context), thereby enhancing the effectiveness of the intervention [11]. MRTs explain the causal relationship between each randomized intervention component (eg, Sit Less or Move More notifications) and the proximal effects (ie, what happens in a limited time window, eg, within 1 hour, following the randomized intervention component) [11]. The distal outcome results from an aggregation of proximal effects obtained through repeated exposure to intervention components over time [11]. For example, a distal outcome would be people achieving recommended levels of PA and longer-term clinical outcomes such as improved glycemic control or improved fasting or postprandial glucose [15].

The primary aim of this study is to assess the effectiveness of theory-based motivational messages for reducing proximal sitting time and increasing proximal standing or walking time (ie, nonsitting time). The secondary aims involve (1) assessing what type of behavioral components (ie, behavior change techniques) are most effective in reducing sitting time and increasing standing and walking time, measured proximally, and (2) determining relevant contextual moderators for reducing sitting time and increasing standing and walking time.

A 6-week MRT will be conducted (see Figure 1). Using this design, participants will be randomly assigned an intervention option (Sit Less message, Move More message, or no intervention) at each relevant decision point.

Eligible participants will be adults with T2D, aged between 35 and 65 years, who own an Android smartphone (version 7.0 or higher) and currently interact or engage with apps, have no limitations (including physical problems and medical complications) to engage in low- and moderate-intensity PA, are able to communicate in English, and can provide informed consent.

Volunteers who use other activity trackers and mobile apps that target PA, highly active people, those diagnosed with T2D less than 3 months ago, and people with psychiatric problems, attention, or memory issues will be excluded. We assume highly active people as those who meet World Health Organization (WHO) recommendation of 10,000 steps per day or 150 minutes of moderate to vigorous PA per week. The International Physical Activity Questionnaire—Short Form will be used to assess activity levels [17]. A 1-week run-in period will assess nonadherence with the activity tracker. Individuals with more than 50% incomplete data will be excluded.

The intervention comprises a bespoke mobile app (iMove app) [18] and a Bluetooth-enabled wearable sensor (SORD) developed and validated by our research group [19]. iMove is a JITAI that aims to support people to decrease sedentary time (Sit Less) and increase PA (Move More) [20]. It consists of a home tab that illustrates sitting, standing, and walking activities, an activity tab to select activities manually in situations where technical issues exist with the app, a goal setting tab to set sitting and walking goals, a notification history tab to compile previous notification messages, and a profile tab (see Figures 2 and 3). SORD is a bespoke thigh-worn device that connects to the iMove app and measures sitting, standing, lying, and PA (eg, walking) [19].

Eligible participants who provide their consent will be provided with the iMove app installer and invited to a web-based individual orientation session through the Zoom videoconferencing platform. During the web-based meeting, participants will be assisted in installing the app and informed about how to use the iMove app and SORD device. Weight, height, and other demographic measures (age, gender, marital status, job status, and education level) will be assessed at baseline as well as after intervention.

Activity sensor data (sitting, standing, and walking time) and other mobile data (location, weather, and calendar information) will be collected automatically through SORD and iMove and sent to a Deakin University secured server. Participant questionnaire data, including demographics questionnaire and system usability, will be captured digitally using Deakin secured Qualtrics surveys.

Time spent sitting (T_sit), standing (T_s), and walking (T_w) within 1 hour after sending the push notification are the proximal outcomes.

The distal outcomes of interest are the “average weekly time spent sanding and walking altogether, that is, nonsitting time” and the “average time spent being in the sitting state.” Moreover, we will also assess total sitting, standing, walking, and exercising time.

Contextual factors, including weather and location, will be assessed using embedded smartphone sensors (GPS) and weather application programming interfaces (APIs; eg, OpenWeatherMap).

Other outcome measures include feasibility and acceptability. Feasibility will be assessed with descriptive statistics, including the proportion of missing data and engagement with the app (the average time the app was opened every day). Usability is assessed using the System Usability Scale through a web-based survey at the end of the intervention [25,26].

Noncompliance with the intervention platform (ie, SORD and iMove) will be evaluated by sending reminders and phone calls. In case no data have been gathered at one decision point, an automated text message reminder will be sent to the user. This reminder will be triggered once a day. If the data remain uncollected for the second day, the system will notify the research team, and a human investigator will contact the user to encourage them to use the platform and offer assistance in resolving any problems. Each user will be expected to continue the study until they provide data for the corresponding number of days × decision times (42 days × 5 = 210).

The study received ethical approval from the Deakin University Human Research Ethics Committee (2021-258).

We present an MRT to evaluate the development of a Sit Less and Move More JITAI intervention in people with T2D [20]. The proximal and distal main effects of the sedentary and PA interventions on reducing sitting time and increasing standing or walking time (ie, nonsitting time), along with their univariate effects, will be evaluated. Moreover, the most effective behavior change technique will be recognized. Furthermore, the most critical contexts (eg, location: home or workplace) in which users are most responsive to interventions will be identified.

Too much sitting is a distinct behavior consideration from too little exercise for increasing the risk of cardiovascular disease [20]. Practical and integrated approaches involving both Sit Less and Move More have been introduced for risk reduction [20]. It is expected that interventions delivering a combination of Sit Less and Move More would add to the overall daily activity of an individual rather than focusing on at least 150 minutes of weekly exercise [20].

To our knowledge, there are no previous MRT studies targeting SB and PA among people with T2D. Other studies have focused on the general public including healthy sedentary adults [14] and university students [29]. The study by Klasnja et al [14] showed that the 6-week MRT intervention (antisedentary or walking vs no suggestion) among healthy sedentary adults (n=44) was effective to increase the average 30-minute step counts (ie, within 30 minutes after the intervention) by 14%. The study by Figueroa et al [29] conducted a 6-week MRT with university students (n=99) to improve their PA level. They showed that receiving the intervention (motivational messages and feedback on step counts vs no suggestion) increased daily steps by 729 steps, but this effect was diminished over time to –33 steps a day. They recommend that future studies need to take into account both user context and preferences [29].

These theory-based MRT studies were mainly focused on walking [14] rather than standing. Incorporating standing into the intervention package could make the Sit Less and Move More concept more practical. Replacing sitting time with standing could be considered a medium in which the most sedentary populations (eg, T2D) may make the initial transition from an inactive lifestyle to a more active lifestyle [30].

Due to the data-intensive nature of this study and the reliance on mobile apps for data transfer, heavy data loads may complicate data transfer to the backend server and generate missing instances. Even though multiple troubleshooting steps have been taken to develop iMove, such issues could still occur on certain smartphones with restricted manufacturers’ system mandates. A key strength of this study is the incorporation of standing as a measured outcome. Because the focus of this MRT is on supporting people with T2D transition from sedentary to more overall physical activities, assessing transitional states, that is, standing, seems relevant. Assessing standing time, distinct from sitting and moving time, is expected to provide more robust evidence on how such Sit Less and Move More MRT performs.

Overall, the results of this study will inform the optimization of digital behavior change interventions to reduce SB and increase PA among people with T2D. Moreover, the findings will provide practical implications to clinicians concerned about recommending high-intensity exercise to older people with T2D. Finally, policy makers will be informed about simple smartphone-based interventions to reduce the burden of sedentariness among people with chronic diseases (eg, diabetes).