Behavior Change Techniques and the Effects Associated With Digital Behavior Change Interventions in Sedentary Behavior in the Clinical Population: A Systematic Review

Background: Sedentary behavior (SB) negatively impact health and is highly prevalent in the population. Digital behavior change interventions (DBCIs) have been developed to modify behaviors such as SB by technologies. However, it is unknown which behavior change techniques (BCTs) are most frequently employed in SB as well as the effect associated with DBCIs in this field. The aim of this systematic review was: (a) to evaluate the BCT most frequently employed in digital health including all technologies available and interventions aimed at increasing physical activity (PA), reducing sedentary time, and improving adherence to exercise in the clinical population, and (b) to review the effect associated with DBCIs in this field. Methods: The database used was Medline, as well as Scopus, Scielo, and Google Scholar. For the search strategy, we considered versions of behavior/behavioral, mHealth/eHealth/telemedicine/serious game/gamification. The terms related to PA and SB were included, the criteria for inclusion were randomized clinical trials (RCTs), adults, intervention based on digital media, and outcome variable lifestyle modification; a last 5 years filter was included. Michie's Taxonomy was used to identify BCTs. The study was registered under the number PROSPERO CRD42019138681. Results: Eighteen RCTs were included in the present systematic review, 5 of them healthy adults, and 13 of them with some illness. Studies included 2298 sedentary individuals who were followed up for 5 weeks−3 years. The most used BCTs were goal setting, problem solving, review outcomes/goals, feedback on behavior and outcomes of behavior, self-monitoring of behavior, social support, information about health consequences, and behavior practice/rehearsal. The effect associated with DBCIs showed improvements, among several related to PA and physiologic self-reported and anthropometric outcomes. Conclusion: The BCTs most used in digital health to change outcomes related to SB were goals and planning, feedback and monitoring, social support, natural consequences, repetition, and substitution. Besides these findings, DBCIs are influenced by several factors like the type of intervention, patients' preferences and values, or the number of BCTs employed. More research is needed to determine with precision which DBCIs or BCTs are the most effective to reduce SB in the clinical population.


INTRODUCTION
Our daily behavior heavily influences our health. According to the County Health Rankings, a 30% variation in health can be accounted for by healthy behaviors, such as smoking, alcohol intake, diet, and physical activity (PA) levels (1). Regarding PA, the World Health Organization (WHO) recommends a minimum of moderate and vigorous-intensity PA per week (2). The WHO defined moderate-intensity PA as any activity with a metabolic equivalent of task (MET) value between 3 and 5, and vigorous intensity PA as ≥6 MET. More recently, the term "sedentary behavior" (SB) was introduced to define any behavior characterized by energy expenditure of 1.5 MET or less, undertaken while in a sitting or reclining posture (3).
The latest research has shown that prolonged SB is positively associated with a range of health concerns including all-cause mortality, cardiovascular disease, type 2 diabetes, metabolic syndrome, and several types of cancers (4,5). Despite this harmful effect to health, modern society provides many opportunities for prolonged sitting (6) and SB is highly prevalent in the majority of people's time (55-69% of the day) (7). Therefore, behavior change in SB is of great interest.
Behavior change techniques (BCTs) have been defined as irreducible, observable, and replicable components of an intervention designed to redirect behavior (8). In this way BCTs refer to a component of an intervention designed to alter or redirect causal processes that regulate behavior. In 2013, Michie et al. classified 93 hierarchically clustered techniques, distributed in 16 groups, called BCT Taxonomy (v1) (9). The BCT Taxonomy helps to identify the effectiveness of BCTs, including those developed by apps for physical activity (10,11). According to Yardley et al. (12), digital behavior change interventions (DBCIs) employ digital technologies to support behavior change and can be used to promote health.
Given the rise of new technologies, DBCIs have been developed using technologies such as mobile applications and websites (13). Technology-enhanced solutions such as mobile applications, activity monitors, prompting software, texts, emails, and websites are being harnessed to reduce SB (14). Previous systematic reviews and meta-analyses have studied the beneficial effect of DBCIs on SB by both increasing PA or reducing sedentary time (14)(15)(16). However, one of these meta-analyses studied technologies limited to computer, mobile, and wearable technology devices in a healthy population (14); and another meta-analysis including all digital devices/technologies, focused on older adults aged ≥50 years (15).
Currently, there have been different systematic reviews that have studied the beneficial effect of DBCIs on SB by increasing PA, reducing sedentary time, and improving adherence to exercise for digital health in people with different clinical conditions. A systematic review with a meta-analysis (2017) aimed at evaluating the effectiveness of digital technologyenhanced interventions to reduce SB as well as to examine the BCTs used. However, this study was limited to computer, mobile, and wearable technology devices in a healthy population (14). Secondly, there was more recent study (2019) with the design of a systematic review that addressed a wider range of technology devices under the name of Human Computer Interaction. However, it targeted several behaviors. In addition to physical activity and sedentary behavior, it included diet, sleep, smoking, and others. Furthermore, this study also focused on the healthy population (16). In addition to that, another recent systematic review with meta-analysis (2019) was aimed at assessing the efficacy of DBCIs on PA and SB. Although this study included all digital technologies and included people with different clinical conditions, it only focused on older adults aged ≥50 years (15). However, the BCTs used in DBCIs and their usability to reduce SB are not known.
Would it be possible to reduce SB and modify PA habits through exercise proposals with DBCIs in clinical and healthy populations according to published research? The growing research of DBCIs aimed at decreasing SB and the general interest in people with different clinical conditions (also named clinical population) make it necessary to revise this topic. Therefore, this systematic review's objective was to identify more frequently used BCTs and explore the effectiveness of DBCIs used in health interventions to reduce SB, increase PA, and improve adherence to exercise in clinical and healthy populations.

METHODS
A systematic review was conducted following PRISMA 2009 guidelines (17). This study aimed to do a systematic review of the evidence concerning the most frequently used BCTs in digital health and the effect associated with DBCIs for reducing SB by increasing PA levels, reducing sedentary time, and improving adherence to exercise in adults. PROSPERO study registration number: CRD42019138681.
To be included in this review, studies had to meet the criteria outlined below (based on the PICO model).
Population: Adults (healthy or clinical population). Intervention: Exercise proposals through DBCIs with the aim of reducing SB, increasing PA, or exercise adherence.
The database used was Medline through the search engines Pubmed, CINAHL, PEDro, and EMBASE. Results were completed with the databases Scopus, Scielo, and Google Scholar. A filter of the last five years was applied to obtain the best possible evidence. The search was performed between May and June, 2019.
The inclusion criteria for the studies were randomized clinical trials, PEDro scale score more than 5 (18), adult population, intervention based on digital media (mobile phone, tablet, computer, web systems, and recommender systems), BCTs used, and outcome variable lifestyle modification (related to PA, SB, or exercise adherence). The languages of the manuscripts should be English, French, Italian, Portuguese, or Spanish. The defined exclusion criteria were systematic reviews, meta-analyses, no use of BCTs, and absence of goals related to SB or PA.
Three independent reviewers participated in each of the study selection phases. The search results were filtered by title and abstract according to eligibility criteria before full reading. Conflicts of inclusion were resolved by agreement between reviewers or by a fourth reviewer. A 4-phase flow chart of PRISMA (17) was performed to correctly stratify the results of the search (Figure 1). The internal and external validity of the manuscript included in the systematic review was evaluated by means of the PEDro scale (18) ( Table 1). For each of the manuscripts included in the review, the following elements were identified: author, year of publication, population, age of the groups involved, distribution of the groups ( Table 2), outcome Haggerty et al.
Haller et al.
Holmen et al.
Frontiers in Digital Health | www.frontiersin.org related to SB, PA, or exercise adherence, and change relative to BCT ( Table 3). We used Michie's Taxonomy to identify the different BCTs used in the studies (9).
BCTs from Michie's Taxonomy were identified in the 18 RCTs. The most used were goal setting, problem solving, review outcomes/goals, feedback on behavior and outcomes of behavior, self-monitoring of behavior, social support, information about health consequences, and behavior practice/rehearsal ( Table 3). All the interventions used text messages and most of them added other communication tools (calls, emails, among others), mobile apps, and some type of activity tracker. Types of interventions, control comparison, and outcomes measured in the 18 RCTs can be seen in Table 4, where the reader may also find the effects of intervention associated with DBCIs.

DISCUSSION
This systematic review analyzed the BCTs most frequently used and the effectiveness of DBCIs used in health interventions aimed at reducing SB, increasing PA, and improving adherence to exercise in clinical and healthy populations.
Regarding the BCTs most frequently used in DBCIs to change outcomes related to SB, results showed that they were goals and planning, feedback and monitoring, social support, natural consequences, repetition, and substitution (Table 3). Therefore, we recognize them as the most important psychological strategies to improve SB change from the 19 groups of BCTs proposed by Michie et al. (9). These findings concur with previous literature showing that the most widely used BCTs are self-monitoring, goal-setting, credible source, feedback on behavior, and social support (13,15). Among these groups of BCTs, the most frequently used was goal setting (n = 10) as part of the goals and planning cluster ( Table 3). More specifically, it was used in five studies involving healthy adults aimed at promoting PA (19,29,30,34,35) and other five studies involving patients suffering from cardiac-related disease (28,31), cancer (33), diabetes (36), or musculoskeletal impairments. (27). The second most used BCT was self-monitoring of behavior (n = 9) as part of the feedback and monitoring cluster. However, only two out of nine studies included healthy adults (19,35). In light of these findings, goal setting is more frequently used in healthy populations than self-monitoring of behavior.
A previous meta-analysis limited to studies in a healthy population found that the most frequently used BCTs were "prompts and cues, ", "self-monitoring of behavior, " "social support, " and "goal settings, " which were coded between 10 and 5 times (14). However, in this study, only 20/93 available BCTs were used (14). Another meta-analysis that focused on adults aged ≥50 years found that the most common BCTs were "social support, " "goal setting, " "feedback on behavior, " and "self-monitoring" (15). These last three BCTs along with "credible source" comprise the most common BCTs in a meta-analysis targeting PA and diet in cancer survivors, although no studies were found to assess SB in this cancer population (16). These BCTs are among the most used in the present systematic review ( Table 3).
Regarding the effect associated with DBCIs, the interventions included in the 18 RCTs showed improvements among several outcomes such as an increase in daily steps, MET, and International Physical Activity Questionnaire (IPAQ). When the clinical population (non-healthy) was evaluated, other complementary outcomes were addressed as well as improvements in self-reported outcomes such as quality of life, self-efficacy, or barriers to exercise, and improvements in anthropometric outcomes such as weight loss or body circumferences ( Table 4). Smart texts were grouped as positive reinforcement messages, sent when a participant was on track to attain or had already attained his or her daily goal, and booster messages, to motivate individuals when they were not tracking to surpass their step goal (29, 32). Short message service (SMS) based on a health promotion model could be used to change beliefs (self-efficacy, family support, coping with barriers) and physical activity (25).
There is evidence that DBCIs to promote PA and/or reduce SB result in increases in total PA and reductions in SB in older adults aged ≥50 years, at least in the short term (15). Stephenson et al. (14), in their meta-analysis on interventions in the workplace, found that SB showed a mean reduction of 40 min/workday in the intervention group at the endpoint follow-up. Greater reductions in SB were found in studies where self-report/proxy measures (53 min/day) of SB were used compared to objective measures (35 min/day) (14).
Besides findings regarding the effects associated with DBCIs, the interventions employed among studies are not common, so it is not possible to know the implication of BCTs among interventions, which concur with previous literature. A systematic review showed that DBCIs are an effective tool to increase PA and reduce body mass index in cancer survivors. However, authors agreed that outcome measures were not common (13). Likewise, the theoretical models applied are not clearly described by the authors (13)(14)(15). In this way, risk of bias and heterogeneity of this type of study was high. This is due to the fact that there is a wide variability of DBCIs, different types of intervention, and different assessment instruments (13,14). Despite this, studies show an increase in vigorous PA, a reduction in body mass index, and a reduction in fatigue in cancer patients (13). Also, the use of computer, mobile, and wearable technology have been shown to be effective tools to reduce SB. Likewise,  Action planning 4 (22%) (19,25,27,33) Review behavior goal(s) 2 (11%) (25,30) Review outcome goal(s) 5 (28%) (22,25,28,29,34) Feedback and monitoring Monitoring of behavior by others without feedback 2 (11%) (23,26) Feedback on behavior 6 (33%) (20,23,25,27,29,32) Self-monitoring of behavior 9 (50%) (17, 19, 19, 23, 23, 24, 24, 27, 28, 28-30, 32, 33, 33-36) Self-monitoring of outcome(s) of behavior 2 (11%) (31,33) Biofeedback 2 (11%) (23,28) Feedback on outcome(s) of behavior 6 (33%) (21,22,25,29,33,35) Social support Social support (unspecified) 5 (28%) (21,24,(26)(27)(28) Social support (practical) 3 (17%) (31,33,34) Natural consequences Information about health consequences 6 (33%) (26,29,30,(32)(33)(34) Salience of consequences 1 (6%) Comparison of behavior Demonstration of the behavior 1 (6%) Social comparison 1 (6%) Repetition and substitution Behavioral practice/rehearsal 6 (33%) (19,21,23,24,29,33) Behavior substitution 1 (6%) Habit formation 4 (22%) (21,24,28,32) Reward and threat Social reward 1 (6%) Incentive (outcome) 2 (11%) (19,31) Reward (outcome) 1 (6%) Scheduled consequences Reduce reward frequency 1 (6%) (30) the permanence of life habits acquired in different studies must be analyzed in the long term through follow-up and without complementary behavior change systems (15). Besides differences in DBCIs employed, there are several factors that may influence their effectiveness. One of them is the time slot. For example, immediate rewards have been shown to be more effective at stimulating change behavior than delayed rewards in overweight/obese adults. In addition, adaptive goal-setting encourages better goal keeping in the long-term than static goals (19). Another factor is the number of BCTs employed. Joining different BCTs and self-monitoring of the behavior has a summative effect, as was seen in Kempf 's study (25). Overweight adults equipped with telemonitoring devices (scales and pedometers) and supported by mental coaches, demonstrated large reductions of body weight after 12 months of lifestyle intervention and the within-group analysis revealed a larger reduction in body weight for the intervention with the combination of telemedical coaching and telemonitoring of PA than only telemonitoring (23). Wayne (36) shows similar data about the summative effect between coaching and monitoring, they found accelerated improvement in the group with mobile phone monitoring added to a health coaching strategy than in the group with coaching alone within a diabetes population. They describe a faster improved adoption and adherence to health behaviors with the combined strategy (36). In this way, delivering tailored modules on a monthly schedule and combining them with weekly action planning is a good example to improve PA in breast cancer survivors (33). The use of self-management techniques in exercise programs could enhance PA adherence among healthy subjects and the clinical population [e.g., knee osteoarthritis (27) or heart disease (29)] , encouraging users to keep their motivation through the combination of message strategies and interventions similar to pedometer data through a Smartphone interface.
Lately, other factors influencing the effects of DBCIs have been difficult to analyze, like patients' preferences and values, namely the patient-centered approach. It is associated with a higher rate of patient satisfaction, better adherence to suggested lifestyle changes, and more cost-effective care (37). However, in most systematic reviews, a person-based approach is limited to a few examples, and there is a lack of defining user characteristics in digital health (12). Another factor is the number of BCTs employed, as evidence suggests that incorporating a multitude of BCTs is more effective than those using few or a single BCT in terms of long-term changes (38). For example, in older patients, a minimum of three groups of BCTs are necessary to obtain significant effects on PA (15). Finally, the social, cultural context and environment are key factors influencing SB and must be taken into consideration when developing DBCIs targeting SB (39).
However, the effectiveness and costs of real-time remote exercise monitoring and coaching to traditional face-to-face -Participants reported by survey their socio-demographic data. -BMI (kg/m2) was determined from researcher-measured height and weight.
-The static goal group trajectory crossed and dropped below the activity of the adaptive groups within four months, even with immediate rewards. -The implication is that adaptive goals will not result in immediate changes as large as static goals, but can contribute to a more gradual and perhaps more sustainable behavior change process over four months.

Frederix et al. (20)
An Internet-based, comprehensive, and patient-tailored telerehabilitation program with short message service (SMS) texting support for cardiac patients.  After clinical rating and answering questionnaires, all patients performed a treadmill walking test until exhaustion to determine peak oxygen uptake (VO2 peak) and lactate threshold.
Web-based exercise intervention using goal-planning led to a significant and clinically relevant improvement of depressive symptoms.

Holmen et al. (23)
-Mobile phone-based self-management system Few Touch Application (FTA). -The FTA consisted of a blood glucose-measuring system with automatic wireless data transfer, diet manual, physical activity registration, and management of personal goals, all recorded and operated using a diabetes diary app on the mobile phone. -In addition, health counseling based on behavior change theory and delivered by a diabetes specialist nurse for the first four months after randomization.
-Mobile phone-based self-management system used for one year. -With or without telephone health counseling by a diabetes specialist nurse for the first four months.
-HbA1c level. -Demographic information were self-reported. -Clinical data were obtained from the General Practitioners or self-reported.
-They used personalized goals, self-monitoring, and natural consequences to obtain improvement in HbA1c level.  -Height, weight, waist circumference, and blood pressure measurements were taken twice for accuracy, with a third measurement also taken in cases where either of the first two values fell outside a predetermined acceptable range. -An online survey assessed physical activity, sitting time, dietary intake, and quality of life.
-They found useful social support to obtain positive improvements in body fat and dietary intake.
-Difference in body weight reduction -Freiburger Questionnaire for Physical Activity (FFkA). -Sport per week hour; physically active per week hour.
Self-reported. Self-monitoring and feedback were useful in weight reduction.
Lari 2018 (26) Training messages within two weeks (two or three messages daily) in the field of physical activity based on health promotion model constructs.
No intervention.
-Body mass index.
-The researchers administered the physical activity questionnaire through interviews and monitored the questionnaires completion.
-The heights and weights of the participants were measured to determine the METs and BMIs.
-Training messages were useful in changing physical activity behaviors of diabetic patients and their beliefs: perceived self-efficacy, family support and lower perceived barriers. -They used self-monitoring, feedback, goal setting, and social support. Participants were less sedentary at 24 weeks. Participants had smaller waist and hip circumferences at 12 weeks.
-Digital physical activity tracking was performed using the Fitbug Orb (Chicago, IL). -They used sequential randomization to individually evaluate the tracking and texting components of the intervention. -After establishing baseline activity during a blinded run-in (week 1), in phase I (weeks 2 to 3) they randomized 2:1 to unblinded versus blinded tracking. -The activity tracker itself did not show activity information, but continuously transmitted it in all participants. -In phase II (weeks 4 to 5), they randomized unblinded participants 1:1 to "smart texts" versus no texts. -Smart texts were automated, personalized, smartphone-delivered coaching messages informed by real-time activity.
-Participant satisfaction through an online survey.
-They used feedback and encouraging messages.
-An automated tracking-texting intervention increased physical activity with, but not without, the texting component.
-Collected by clinical practitioners.
-A mHealth intervention (messages and natural consequences) plus usual care showed a positive effect on adherence to multiple lifestyle behavior changes at three months. -Adherence to healthy lifestyle behaviors measured using a self-reported composite health behavior score.

Sharma et al. (32)
-Telephone call once a month.
-The intervention package also included weekly text messages via SMS. -The text messages included short but catchy slogans and rhymes on the importance of modification of behavioral risk factors.
No mobile-health intervention.
-Collected by clinical practitioners.
-After Intervention Phase (messages and calls), significant reduction was seen in behavioral risk factors (unhealthy diet and insufficient physical activity) in the intervention group compared to control group. -Body mass index (BMI), systolic blood pressure, and fasting blood sugar level also showed significant difference in the intervention group as compared to controls.  -12]). -Self-reported.
-Collected by clinical practitioners.
-The Connected Wellness Platform enabled self-monitoring and health coach interactions with intervention participants, providing a cloud-based platform for mobile phone-based health management.
-This system provided secure, two-way communication between client and health coach, with mobile phone data entry on relevant behaviors entered manually.
Frontiers in Digital Health | www.frontiersin.org rehabilitation must be considered (28). This kind of intervention is at least as favorable as a face-to-face program. Immediate post intervention feedback and at longer-term follow-up (24 weeks) could be considered in this type of intervention (28). Moreover, remote intervention realized a 70% reduction in programmer delivery costs, and very few adverse events were attributed to participation in the programmer. Specialists provided realtime individualized audio coaching, feedback, and social support throughout real-time exercise monitoring (28).
The limitation of the present study refers to data analysis. This systematic review focused on the last five years due to the existence of similar studies; however, this ensured that the most recent studies would be captured. The studies included in the systematic review did not have an outcome variable measured with a common instrument and sometimes did not offer numerical data on the changes produced. This fact makes statistical treatment difficult through meta-analysis. For future RCTs evaluating DBCIs in SB that could be assessed in a meta-analysis, standardized outcomes would be recommended, at least the most studied, such as daily steps, MET, and levels of PA measured by IPAQ (see Table 4). In addition, current evidence showed a high heterogeneity in the clinical population. A greater number of RCTs would allow further metaanalysis in a given clinical population. Besides the mentioned limitations, this is the first study to analyze in depth and compare the results of applying BCTs in increasing PA or interrupting SB in different populations. The demonstrated effectiveness of BCTs in intervention protocols related to PA could be implemented in other types of treatment to increase adherence.

CONCLUSION
The present systematic review showed that the most common BCTs aimed at reducing SB in DBCIs were goal setting, problem solving, review outcomes/goals, feedback on behavior and outcomes of behavior, self-monitoring of behavior, social support, information about health consequences, and behavior practice/rehearsal. The DBCIs among studies included revealed beneficials effects in outcomes such as an increase in MET, PA levels, or daily steps; improvements in self-reported outcomes such as quality of life, self-efficacy, or barriers to exercise; and improvements in anthropometric outcomes such as weight loss or body circumferences. Besides these findings, DBCIs are influenced by several factors like the type of intervention, patients' preferences, and values or the number of BCTs employed. Therefore, more research is required to determine with precision which DBCIs are the most effective to reduce SB in the clinical population.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.