Predicting Dropouts From an Electronic Health Platform for Lifestyle Interventions: Analysis of Methods and Predictors

A blended face-to-face and eHealth lifestyle intervention on physical activity, diet, and health outcomes in Hong Kong community-dwelling older adults: a study protocol for a randomized controlled trial

Background

Aging individuals are vulnerable to various Noncommunicable Diseases (NCDs). Different behaviors are closely related to a decreased risk of suffering from NCDs: sufficient Physical Activity (PA) (e.g., at least 150 mins Moderate-to-vigorous Physical Activity (MVPA) per week) and a healthy daily diet (e.g., at least five portions of Fruit and Vegetable Intake (FVI), 5-6 taels (189.0-226.8 g) Meat, Fish, Egg and Alternatives (MFEA)). Traditional face-to-face interventions were effective in behavior change. However, it was revealed to be resource-intensive and limited transfer due to poor self-regulation skills outside of face-to-face sessions. Thus, eHealth could be a supplement for older adults outside traditional face-to-face settings. The blended approach combining these two interventions might optimize the intervention effects on lifestyle behavior initiation and maintenance, but little research can be found among Hong Kong older adults. Therefore, the study aims to test a blended intervention to promote PA, diet, and health outcomes among Hong Kong community-dwelling older adults.

Methods

This study will adopt a 10-week three-arm randomized controlled trial. The blended group will receive weekly (1) two 60-min face-to-face sessions with one for PA and one for diet, and (2) two web-based sessions with one for PA and one for diet. The face-to-face group will receive the same intervention content as the face-to-face sessions in the blended group. The control condition will receive a biweekly telephone call. The outcomes will include MVPA (minutes/week), FVI (portions/day), MFEA consumption (taels/day), social-cognitive factors (self-efficacy, planning, social support, action control), physical health outcomes (clinical indicators, senior physical fitness), mental health outcomes (depression, loneliness) and health-related quality of life. Data collection will be implemented at the pre-test, post-test, and 3-month follow-up test.

Discussion

This is the first study evaluating a blended intervention promoting multiple health behaviors among Hong Kong community-dwelling older adults. If the effect of the blended intervention is superior to the traditional face-to-face group and the control group, it will enrich lifestyle intervention approaches and can be applied to older adults, helping them obtain health benefits. Furthermore, a better understanding of mechanisms will also have implications for theory-building.

Clinical trial registration

https://www.isrctn.com/ISRCTN32329348, ISRCTN32329348.

Telemedicine and Cancer Clinical Research: Opportunities for Transformation

ABSTRACT

Telemedicine represents an established mode of patient care delivery that has and will continue to transform cancer clinical research. Through telemedicine, opportunities exist to improve patient care, enhance access to novel therapies, streamline data collection and monitoring, support communication, and increase trial efficiency. Potential challenges include disparities in technology access and literacy, physical examination performance, biospecimen collection, privacy and security concerns, coverage of services by insurance, and regulatory considerations. Coupled with artificial intelligence, telemedicine may offer ways to reach geographically dispersed candidates for narrowly focused cancer clinical trials, such as those targeting rare genomic subsets. Collaboration among clinical trial staff, clinicians, regulators, professional societies, patients, and their advocates is critical to optimize the benefits of telemedicine for clinical cancer research.

Dataset size versus homogeneity: A machine learning study on pooling intervention data in e-mental health dropout predictions

Objective

This study proposes a way of increasing dataset sizes for machine learning tasks in Internet-based Cognitive Behavioral Therapy through pooling interventions. To this end, it (1) examines similarities in user behavior and symptom data among online interventions for patients with depression, social anxiety, and panic disorder and (2) explores whether these similarities suffice to allow for pooling the data together, resulting in more training data when prediction intervention dropout.

Methods

A total of 6418 routine care patients from the Internet Psychiatry in Stockholm are analyzed using (1) clustering and (2) dropout prediction models. For the latter, prediction models trained on each individual intervention's data are compared to those trained on all three interventions pooled into one dataset. To investigate if results vary with dataset size, the prediction is repeated using small and medium dataset sizes.

Results

The clustering analysis identified three distinct groups that are almost equally spread across interventions and are instead characterized by different activity levels. In eight out of nine settings investigated, pooling the data improves prediction results compared to models trained on a single intervention dataset. It is further confirmed that models trained on small datasets are more likely to overestimate prediction results.

Conclusion

The study reveals similar patterns of patients with depression, social anxiety, and panic disorder regarding online activity and intervention dropout. As such, this work offers pooling different interventions' data as a possible approach to counter the problem of small dataset sizes in psychological research.

Finding the Best Match - a Case Study on the (Text-)Feature and Model Choice in Digital Mental Health Interventions

With the need for psychological help long exceeding the supply, finding ways of scaling, and better allocating mental health support is a necessity. This paper contributes by investigating how to best predict intervention dropout and failure to allow for a need-based adaptation of treatment. We systematically compare the predictive power of different text representation methods (metadata, TF-IDF, sentiment and topic analysis, and word embeddings) in combination with supplementary numerical inputs (socio-demographic, evaluation, and closed-question data). Additionally, we address the research gap of which ML model types - ranging from linear to sophisticated deep learning models - are best suited for different features and outcome variables. To this end, we analyze nearly 16.000 open-text answers from 849 German-speaking users in a Digital Mental Health Intervention (DMHI) for stress. Our research proves that - contrary to previous findings - there is great promise in using neural network approaches on DMHI text data. We propose a task-specific LSTM-based model architecture to tackle the challenge of long input sequences and thereby demonstrate the potential of word embeddings (AUC scores of up to 0.7) for predictions in DMHIs. Despite the relatively small data set, sequential deep learning models, on average, outperform simpler features such as metadata and bag-of-words approaches when predicting dropout. The conclusion is that user-generated text of the first two sessions carries predictive power regarding patients' dropout and intervention failure risk. Furthermore, the match between the sophistication of features and models needs to be closely considered to optimize results, and additional non-text features increase prediction results.

Supplementary Information

The online version contains supplementary material available at 10.1007/s41666-023-00148-z.

Development and feasibility of a theory-guided and evidence-based physical activity intervention in pregnant women with high risk for gestational diabetes mellitus: a pilot clinical trial

BACKGROUND

Physical activity has been utilized as an effective strategy to prevent gestational diabetes mellitus (GDM). However, most pregnant women with high risk for GDM did not achieve the recommended physical activity level. Furthermore, relevant physical activity protocols have varied without theory-guided and evidence-based tailored to pregnant women with high risk for GDM. This study aimed to develop and pilot test a theory-guided and evidence-based physical activity intervention protocol for pregnant women with high risk for GDM.

METHODS

The study design was guided by the Medical Research Council Framework for Developing and Evaluating Complex Intervention (the MRC framework). The preliminary protocol for physical activity intervention was developed based on self-efficacy theory, research evidence identified from systematic reviews and clinic trials, stakeholder engagement, context, and economic considerations. The preliminary intervention protocol was validated through a content validity study by an expert panel of 10 experts. A single-blinded randomized controlled trial (RCT) was designed to test the feasibility and acceptability of the intervention.

RESULTS

The validity of the preliminary intervention protocol was excellent as consensus was achieved. The final 13 sessions of self-efficacy enhancing physical activity intervention protocol were developed, including knowledge education, exercise clinic visits and video, and group discussions with face-to-face and online blended sessions. In the feasibility study, 34 pregnant women with high risk for GDM were randomized for the intervention (n = 17) or the control group (n = 17). The recruitment and retention rates were 82.9% and 58.9%, respectively. Women in the intervention group had a lower incidence of GDM (26.7% vs. 36.5%) than the control group (P >0.05). All participants were satisfied with the intervention and agreed that the intervention was helpful.

CONCLUSIONS

The developed self-efficacy-enhancing physical activity intervention is a feasible and acceptable intervention for enhancing physical activity among pregnant women with high risk for GDM and is ready to be tested in a more extensive RCT study.

TRIAL REGISTRATION

The study was registered on 4 February 2022 (ChiCTR2200056355) by the Chinese Clini Trial Registry (CHiCTR).

Predicting Disengagement to Better Support Outcomes in a Web-Based Weight Loss Program Using Machine Learning Models: Cross-Sectional Study

BACKGROUND

Engagement is key to interventions that achieve successful behavior change and improvements in health. There is limited literature on the application of predictive machine learning (ML) models to data from commercially available weight loss programs to predict disengagement. Such data could help participants achieve their goals.

OBJECTIVE

This study aimed to use explainable ML to predict the risk of member disengagement week by week over 12 weeks on a commercially available web-based weight loss program.

METHODS

Data were available from 59,686 adults who participated in the weight loss program between October 2014 and September 2019. Data included year of birth, sex, height, weight, motivation to join the program, use statistics (eg, weight entries, entries into the food diary, views of the menu, and program content), program type, and weight loss. Random forest, extreme gradient boosting, and logistic regression with L1 regularization models were developed and validated using a 10-fold cross-validation approach. In addition, temporal validation was performed on a test cohort of 16,947 members who participated in the program between April 2018 and September 2019, and the remaining data were used for model development. Shapley values were used to identify globally relevant features and explain individual predictions.

RESULTS

The average age of the participants was 49.60 (SD 12.54) years, the average starting BMI was 32.43 (SD 6.19), and 81.46% (39,594/48,604) of the participants were female. The class distributions (active and inactive members) changed from 39,369 and 9235 in week 2 to 31,602 and 17,002 in week 12, respectively. With 10-fold-cross-validation, extreme gradient boosting models had the best predictive performance, which ranged from 0.85 (95% CI 0.84-0.85) to 0.93 (95% CI 0.93-0.93) for area under the receiver operating characteristic curve and from 0.57 (95% CI 0.56-0.58) to 0.95 (95% CI 0.95-0.96) for area under the precision-recall curve (across 12 weeks of the program). They also presented a good calibration. Results obtained with temporal validation ranged from 0.51 to 0.95 for area under a precision-recall curve and 0.84 to 0.93 for area under the receiver operating characteristic curve across the 12 weeks. There was a considerable improvement in area under a precision-recall curve of 20% in week 3 of the program. On the basis of the computed Shapley values, the most important features for predicting disengagement in the following week were those related to the total activity on the platform and entering a weight in the previous weeks.

CONCLUSIONS

This study showed the potential of applying ML predictive algorithms to help predict and understand participants' disengagement with a web-based weight loss program. Given the association between engagement and health outcomes, these findings can prove valuable in providing better support to individuals to enhance their engagement and potentially achieve greater weight loss.

Acceptability and preliminary effectiveness of a single-arm 12-week digital behavioral health intervention in patients with knee osteoarthritis

BACKGROUND

Digital health interventions may improve osteoarthritis (OA) management. This study evaluated the acceptability and preliminary effectiveness of a multimodal digital nutrition, exercise, and mindfulness intervention in adults with knee OA.

METHODS

Adults with advanced knee OA and an orthopaedic referral were invited to self-enroll in a pragmatic 12-week single-arm intervention. OA-focused nutrition and exercise resources were delivered weekly by email, and secondary components accessed on-demand (web-platform, webinars, and nutrition consultation). Acceptability was assessed by qualitative interview data and completion rates. Preliminary effectiveness on clinical outcomes was assessed by change in health-related quality of life, well-being, mindfulness, self-efficacy, and interest in total knee arthroplasty (TKA) between baseline and 12-weeks.

RESULTS

N = 102 patients self-enrolled (73.5% female, age 64 ± 7 years, body mass index 32.9 ± 7.3 kg/m²); n = 53 completed the 12-week intervention (71.7% female, age 65 ± 7 years, body mass index 33.4 ± 6.3 kg/m²). Acceptability was demonstrated by positive perceptions of tailored intervention resources. In study completers, health-related quality of life components of pain and physical functioning domains improved at 12-weeks [change in SF36 4.4 (95%CI 0.2-8.6), p = 0.016, and 6.7 (95%CI 2.7-10.7), p < 0.001, respectively]. Self-efficacy for managing daily activities improved [change in PROMIS T-score 4.4 (95%CI 2.8-6.0), p < 0.001].

CONCLUSION

A 12-week digital multimodal intervention for knee OA was acceptable to patients and shows preliminary effectiveness in improving self-efficacy, aspects of quality of life, and decreasing interest in TKA. Digital behavioral interventions for knee OA may be an acceptable approach to improve patient outcomes and OA self-management while potentially reducing utilization of costly health system resources.

Days between sessions predict attrition in text-based internet intervention of Binge Eating Disorder

BACKGROUND

The number of days between treatment sessions is often overlooked as a predictor of attrition in psychotherapy. In text-based Internet interventions, days between sessions may be a simple yet powerful predictor of attrition.

OBJECTIVE

We hypothesized that a larger number of days between sessions increased the likelihood of attrition among participants with Binge Eating Disorder (BED) in a 12-session Internet-based cognitive behavioral therapy (iCBT) program. Participants could work on the sessions whenever convenient for them and received written support from a psychologist.

MATERIAL AND METHODS

We compared 201 adult participants with mild to moderate BED (85 non-completers and 116 completers) on the number of days between sessions to predict attrition rates.

RESULTS

Mixed model binomial logistic regression showed that non-completers spent significantly more days between sessions across the first four treatment sessions (1-4) when controlling for age, gender, and intake measures of BMI, BED, overall health status (EQ VAS), and depression symptoms (MDI) (OR = 1.042, p < .001). Age (OR = 0.976, p < .001) and EQ VAS (OR = 0.984, p < .001) were also significant. The risk of attrition increased by 4.2 % for each additional day participants spent completing a session.A receiver operating characteristic (ROC) curve analysis showed that classification accuracy increased across sessions from 61.1 % in session 1 and 65.7 % in session 2 to 68.8 % in session 3 and 73.2 % in session 4. The optimal cut-off point in session 4 was 17.5 days, which detected 60.4 % of non-completers (sensitivity) and 78.4 % of completers (specificity).An exploratory repeated measures of ANOVA of days between sessions showed a significant within-subjects effect, where both non-completers and completers spent more days between sessions as they progressed from sessions 1 through 4 (F = 20.54, df = 3, p < .001). There was no interaction effect, suggesting that the increase in slope did not differ between non-completers and completers.

CONCLUSIONS

Participants spending more days between sessions are at increased risk of dropping out of treatment. This may have important implications for identifying measures to reduce attrition, e.g., intensifying interventions through automated reminders or therapist messages. Our findings may have important transdiagnostic implications for text-based Internet interventions. Further studies should investigate the predictive value of days between sessions in other diagnoses.

Perspectives on Promoting Physical Activity Using eHealth in Primary Care by Health Care Professionals and Individuals With Prediabetes and Type 2 Diabetes: Qualitative Study

BACKGROUND

The trend of an exponential increase in prediabetes and type 2 diabetes (T2D) is projected to continue rising worldwide. Physical activity could help prevent T2D and the progression and complications of the disease. Therefore, we need to create opportunities for individuals to acquire the necessary knowledge and skills to self-manage their chronic condition through physical activity. eHealth is a potential resource that could facilitate self-management and thus improve population health. However, there is limited research on users' perception of eHealth in promoting physical activity in primary care settings.

OBJECTIVE

This study aims to explore the perspectives of health care professionals and individuals with prediabetes and T2D on eHealth to promote physical activity in primary care.

METHODS

A qualitative approach was applied using focus group discussions among individuals with prediabetes or T2D (14 participants in four groups) and health care professionals (10 participants in two groups). The discussions were audio-recorded and transcribed verbatim. Qualitative content analysis was used inductively to code the data.

RESULTS

Three main categories emerged: utility, adoption process, and accountability. The utility of eHealth was described as a motivational, entertaining, and stimulating tool. Registration of daily medical measurements and lifestyle parameters in a cohesive digital platform was recognized as a potential resource for strengthening self-management skills. The adoption process includes eHealth to increase the accessibility of care and personalize the support of physical activity. However, participants stated that digital technology might only suit some and could increase health care providers' administrative burden. Accountability refers to the knowledge and skills to optimize eHealth and ensure data integrity and security.

CONCLUSIONS

People with prediabetes and T2D and health care professionals positively viewed an integration of eHealth technology in primary care to promote physical activity. A cohesive platform using personal metrics, goal-setting, and social support to promote physical activity was suggested. This study identified eHealth illiteracy, inequality, privacy, confidentiality, and an increased workload on health care professionals as factors of concern when integrating eHealth into primary care. Continuous development of eHealth competence was reported as necessary to optimize the implementation of eHealth technology in primary care.

Can a Single Variable Predict Early Dropout From Digital Health Interventions? Comparison of Predictive Models From Two Large Randomized Trials

BACKGROUND

A single generalizable metric that accurately predicts early dropout from digital health interventions has the potential to readily inform intervention targets and treatment augmentations that could boost retention and intervention outcomes. We recently identified a type of early dropout from digital health interventions for smoking cessation, specifically, users who logged in during the first week of the intervention and had little to no activity thereafter. These users also had a substantially lower smoking cessation rate with our iCanQuit smoking cessation app compared with users who used the app for longer periods.

OBJECTIVE

This study aimed to explore whether log-in count data, using standard statistical methods, can precisely predict whether an individual will become an iCanQuit early dropout while validating the approach using other statistical methods and randomized trial data from 3 other digital interventions for smoking cessation (combined randomized N=4529).

METHODS

Standard logistic regression models were used to predict early dropouts for individuals receiving the iCanQuit smoking cessation intervention app, the National Cancer Institute QuitGuide smoking cessation intervention app, the WebQuit.org smoking cessation intervention website, and the Smokefree.gov smoking cessation intervention website. The main predictors were the number of times a participant logged in per day during the first 7 days following randomization. The area under the curve (AUC) assessed the performance of the logistic regression models, which were compared with decision trees, support vector machine, and neural network models. We also examined whether 13 baseline variables that included a variety of demographics (eg, race and ethnicity, gender, and age) and smoking characteristics (eg, use of e-cigarettes and confidence in being smoke free) might improve this prediction.

RESULTS

The AUC for each logistic regression model using only the first 7 days of log-in count variables was 0.94 (95% CI 0.90-0.97) for iCanQuit, 0.88 (95% CI 0.83-0.93) for QuitGuide, 0.85 (95% CI 0.80-0.88) for WebQuit.org, and 0.60 (95% CI 0.54-0.66) for Smokefree.gov. Replacing logistic regression models with more complex decision trees, support vector machines, or neural network models did not significantly increase the AUC, nor did including additional baseline variables as predictors. The sensitivity and specificity were generally good, and they were excellent for iCanQuit (ie, 0.91 and 0.85, respectively, at the 0.5 classification threshold).

CONCLUSIONS

Logistic regression models using only the first 7 days of log-in count data were generally good at predicting early dropouts. These models performed well when using simple, automated, and readily available log-in count data, whereas including self-reported baseline variables did not improve the prediction. The results will inform the early identification of people at risk of early dropout from digital health interventions with the goal of intervening further by providing them with augmented treatments to increase their retention and, ultimately, their intervention outcomes.

Effects of Face-to-Face and eHealth Blended Interventions on Physical Activity, Diet, and Weight-Related Outcomes among Adults: A Systematic Review and Meta-Analysis

An increasing number of studies are blending face-to-face interventions and electronic health (eHealth) interventions to jointly promote physical activity (PA) and diet among people. However, a comprehensive summary of these studies is lacking. This study aimed to synthesize the characteristics of blended interventions and meta-analyze the effectiveness of blended interventions in promoting PA, diet, and weight-related outcomes among adults. Following the PRISMA guidelines, PubMed, SPORTDiscus, PsycINFO, Embase, and Web of Science were systematically searched to identify eligible articles according to a series of inclusion criteria. The search was limited to English language literature and publication dates between January 2002 and July 2022. Effect sizes were calculated as standardized mean difference (SMD) for three intervention outcomes (physical activity, healthy diet, and weight-related). Random effect models were used to calculate the effect sizes. A sensitivity analysis and publication bias tests were conducted. Of the 1561 identified studies, 17 were eligible for the systematic review. Studies varied in participants, intervention characteristics, and outcome measures. A total of 14 studies were included in the meta-analyses. There was evidence of no significant publication bias. The meta-analyses indicated that the blended intervention could lead to a significant increase in walking steps (p < 0.001), total PA level (p = 0.01), and diet quality (p = 0.044), a significant decrease in energy intake (p = 0.004), weight (p < 0.001), BMI (p < 0.001), and waist circumferences (p = 0.008), but had no influence on more moderate-to-vigorous physical activity (MVPA) or fruit and vegetable intake among adults, compared with a control group. The study findings showed that blended interventions achieve preliminary success in promoting PA, diet, and weight-related outcomes among adults. Future studies could improve the blended intervention design to achieve better intervention effectiveness.

Predictors of Dropout in a Digital Intervention for the Prevention and Treatment of Depression in Patients With Chronic Back Pain: Secondary Analysis of Two Randomized Controlled Trials

Background

Depression is a common comorbid condition in individuals with chronic back pain (CBP), leading to poorer treatment outcomes and increased medical complications. Digital interventions have demonstrated efficacy in the prevention and treatment of depression; however, high dropout rates are a major challenge, particularly in clinical settings.

Objective

This study aims to identify the predictors of dropout in a digital intervention for the treatment and prevention of depression in patients with comorbid CBP. We assessed which participant characteristics may be associated with dropout and whether intervention usage data could help improve the identification of individuals at risk of dropout early on in treatment.

Methods

Data were collected from 2 large-scale randomized controlled trials in which 253 patients with a diagnosis of CBP and major depressive disorder or subclinical depressive symptoms received a digital intervention for depression. In the first analysis, participants’ baseline characteristics were examined as potential predictors of dropout. In the second analysis, we assessed the extent to which dropout could be predicted from a combination of participants’ baseline characteristics and intervention usage variables following the completion of the first module. Dropout was defined as completing <6 modules. Analyses were conducted using logistic regression.

Results

From participants’ baseline characteristics, lower level of education (odds ratio [OR] 3.33, 95% CI 1.51-7.32) and both lower and higher age (a quadratic effect; age: OR 0.62, 95% CI 0.47-0.82, and age²: OR 1.55, 95% CI 1.18-2.04) were significantly associated with a higher risk of dropout. In the analysis that aimed to predict dropout following completion of the first module, lower and higher age (age: OR 0.60, 95% CI 0.42-0.85; age²: OR 1.59, 95% CI 1.13-2.23), medium versus high social support (OR 3.03, 95% CI 1.25-7.33), and a higher number of days to module completion (OR 1.05, 95% CI 1.02-1.08) predicted a higher risk of dropout, whereas a self-reported negative event in the previous week was associated with a lower risk of dropout (OR 0.24, 95% CI 0.08-0.69). A model that combined baseline characteristics and intervention usage data generated the most accurate predictions (area under the receiver operating curve [AUC]=0.72) and was significantly more accurate than models based on baseline characteristics only (AUC=0.70) or intervention usage data only (AUC=0.61). We found no significant influence of pain, disability, or depression severity on dropout.

Conclusions

Dropout can be predicted by participant baseline variables, and the inclusion of intervention usage variables may improve the prediction of dropout early on in treatment. Being able to identify individuals at high risk of dropout from digital health interventions could provide intervention developers and supporting clinicians with the ability to intervene early and prevent dropout from occurring.

Enrollment and retention of participants in remote digital health studies: a scoping review and framework proposal (Preprint)

Background

Objective

Methods

Results

Conclusions

Factors Influencing Adherence to mHealth Apps for Prevention or Management of Noncommunicable Diseases: Systematic Review

Background

Mobile health (mHealth) apps show vast potential in supporting patients and health care systems with the increasing prevalence and economic costs of noncommunicable diseases (NCDs) worldwide. However, despite the availability of evidence-based mHealth apps, a substantial proportion of users do not adhere to them as intended and may consequently not receive treatment. Therefore, understanding the factors that act as barriers to or facilitators of adherence is a fundamental concern in preventing intervention dropouts and increasing the effectiveness of digital health interventions.

Objective

This review aimed to help stakeholders develop more effective digital health interventions by identifying factors influencing the continued use of mHealth apps targeting NCDs. We further derived quantified adherence scores for various health domains to validate the qualitative findings and explore adherence benchmarks.

Methods

A comprehensive systematic literature search (January 2007 to December 2020) was conducted on MEDLINE, Embase, Web of Science, Scopus, and ACM Digital Library. Data on intended use, actual use, and factors influencing adherence were extracted. Intervention-related and patient-related factors with a positive or negative influence on adherence are presented separately for the health domains of NCD self-management, mental health, substance use, nutrition, physical activity, weight loss, multicomponent lifestyle interventions, mindfulness, and other NCDs. Quantified adherence measures, calculated as the ratio between the estimated intended use and actual use, were derived for each study and compared with the qualitative findings.

Results

The literature search yielded 2862 potentially relevant articles, of which 99 (3.46%) were included as part of the inclusion criteria. A total of 4 intervention-related factors indicated positive effects on adherence across all health domains: personalization or tailoring of the content of mHealth apps to the individual needs of the user, reminders in the form of individualized push notifications, user-friendly and technically stable app design, and personal support complementary to the digital intervention. Social and gamification features were also identified as drivers of app adherence across several health domains. A wide variety of patient-related factors such as user characteristics or recruitment channels further affects adherence. The derived adherence scores of the included mHealth apps averaged 56.0% (SD 24.4%).

Conclusions

This study contributes to the scarce scientific evidence on factors that positively or negatively influence adherence to mHealth apps and is the first to quantitatively compare adherence relative to the intended use of various health domains. As underlying studies mostly have a pilot character with short study durations, research on factors influencing adherence to mHealth apps is still limited. To facilitate future research on mHealth app adherence, researchers should clearly outline and justify the app’s intended use; report objective data on actual use relative to the intended use; and, ideally, provide long-term use and retention data.

The Feasibility of a Web-Based Educational Lifestyle Program for People With Multiple Sclerosis: A Randomized Controlled Trial

Background

Modifiable lifestyle factors are important to aid people with multiple sclerosis in the self-management of their disease. Current self-management programs are limited by their face-to-face mode of delivery but there is immense potential with the internet to deliver these programs effectively.

Objective

The aims of this study are to assess the feasibility of a digitalized educational lifestyle self-management program for people with MS.

Methods

In this randomized controlled trial, people with MS were randomly allocated to participate in a 6-week tailored web-based educational lifestyle program or 6-week generic standard-care educational course, and were blinded to their allocation. Participants were recruited through multiple sclerosis (MS) Societies in four countries: Australia, New Zealand, Canada, and the United States. The primary outcome was to assess acceptability of the program defined as percentage completion of all modules at 6-weeks post-course commencement. Secondary outcomes included evaluating participant responses to the follow-up survey across three domains: accessibility, learnability, and desirability.

Results

Thirty-five participants from Australia, Canada, New Zealand, and the US completed the baseline survey and were randomized. Four participants were deemed ineligible due to incomplete baseline data; therefore, nine out of 15 and eight out of 16 participants completed 100% of the course in the intervention and standard-care arm courses, respectively.

Conclusions

This study found that this web-based educational lifestyle program is a feasible means of delivering educational content to people with MS via the internet according to our a priori targets of >40% of participants in the intervention arm, and >25% in the control arm to completing 100% of the course. It is therefore appropriate to evaluate this intervention further in a large, randomized controlled trial.

Trial registration

This study was prospectively registered with the Australian New Zealand Clinical Trials Registry (ID: ACTRN12621000245897).

Mitigating Issues With/of/for True Personalization

A common but false perception persists about the level and type of personalization in the offerings of contemporary software, information systems, and services, known as Personalization Myopia: this involves a tendency for researchers to think that there are many more personalized services than there genuinely are, for the general audience to think that they are offered personalized services when they really are not, and for practitioners to have a mistaken idea of what makes a service personalized. And yet in an era, which mashes up large amounts of data, business analytics, deep learning, and persuasive systems, true personalization is a most promising approach for innovating and developing new types of systems and services—including support for behavior change. The potential of true personalization is elaborated in this article, especially with regards to persuasive software features and the oft-neglected fact that users change over time.

Weight Loss Trajectories in Healthy Weight Coaching: Cohort Study

Background

As global obesity prevalence continues to increase, there is a need for accessible and affordable weight management interventions, such as web-based programs.

Objective

This paper aims to assess the outcomes of healthy weight coaching (HWC), a web-based obesity management program integrated into standard Finnish clinical care.

Methods

HWC is an ongoing, structured digital 12-month program based on acceptance and commitment therapy. It includes weekly training sessions focused on lifestyle, general health, and psychological factors. Participants received remote one-on-one support from a personal coach. In this real-life, single-arm, prospective cohort study, we examined the total weight loss, weight loss profiles, and variables associated with weight loss success and program retention in 1189 adults (963 women) with a BMI >25 kg/m² among participants of the program between October 2016 and March 2019. Absolute (kg) and relative (%) weight loss from the baseline were the primary outcomes. We also examined the weight loss profiles, clustered based on the dynamic time-warping distance, and the possible variables associated with greater weight loss success and program retention. We compared different groups using the Mann-Whitney test or Kruskal-Wallis test for continuous variables and the chi-squared test for categorical variables. We analyzed changes in medication using the McNemar test.

Results

Among those having reached the 12-month time point (n=173), the mean weight loss was 4.6% (SE 0.5%), with 43% (n=75) achieving clinically relevant weight loss (≥5%). Baseline BMI ≥40 kg/m² was associated with a greater weight loss than a lower BMI (mean 6.6%, SE 0.9%, vs mean 3.2%, SE 0.6%; P=.02). In addition, more frequent weight reporting was associated with greater weight loss. No significant differences in weight loss were observed according to sex, age, baseline disease, or medication use. The total dropout rate was 29.1%. Dropouts were slightly younger than continuers (47.2, SE 0.6 years vs 49.2, SE 0.4 years; P=.01) and reported their weight less frequently (3.0, SE 0.1 entries per month vs 3.3, SE 0.1 entries per month; P<.001).

Conclusions

A comprehensive web-based program such as HWC is a potential addition to the repertoire of obesity management in a clinical setting. Heavier patients lost more weight, but weight loss success was otherwise independent of baseline characteristics.

Effectiveness of eHealth and mHealth Interventions Supporting Children and Young People Living With Juvenile Idiopathic Arthritis: Systematic Review and Meta-analysis

BACKGROUND

Juvenile idiopathic arthritis (JIA) management aims to promote remission through timely, individualized, well-coordinated interdisciplinary care using a range of pharmacological, physical, psychological, and educational interventions. However, achieving this goal is workforce-intensive. Harnessing the burgeoning eHealth and mobile health (mHealth) interventions could be a resource-efficient way of supplementing JIA management.

OBJECTIVE

This systematic review aims to identify the eHealth and mHealth interventions that have been proven to be effective in supporting health outcomes for children and young people (aged 1-18 years) living with JIA.

METHODS

We systematically searched 15 databases (2018-2021). Studies were eligible if they considered children and young people (aged 1-18 years) diagnosed with JIA, an eHealth or mHealth intervention, any comparator, and health outcomes related to the used interventions. Independently, 2 reviewers screened the studies for inclusion and appraised the study quality using the Downs and Black (modified) checklist. Study outcomes were summarized using a narrative, descriptive method and, where possible, combined for a meta-analysis using a random-effects model.

RESULTS

Of the 301 studies identified in the search strategy, 15 (5%) fair-to-good-quality studies met the inclusion criteria, which identified 10 interventions for JIA (age 4-18.6 years). Of these 10 interventions, 5 (50%) supported symptom monitoring by capturing real-time data using health applications, electronic diaries, or web-based portals to monitor pain or health-related quality of life (HRQoL). Within individual studies, a preference was demonstrated for real-time pain monitoring over recall pain assessments because of a peak-end effect, improved time efficiency (P=.002), and meeting children's and young people's HRQoL needs (P<.001) during pediatric rheumatology consultations. Furthermore, 20% (2/10) of interventions supported physical activity promotion using a web-based program or a wearable activity tracker. The web-based program exhibited a moderate effect, which increased endurance time, physical activity levels, and moderate to vigorous physical activity (standardized mean difference [SMD] 0.60, SD 0.02-1.18; I²=79%; P=.04). The final 30% (3/10) of interventions supported self-management development through web-based programs, or apps, facilitating a small effect, reducing pain intensity (SMD -0.14, 95% CI -0.43 to 0.15; I²=53%; P=.33), and increasing disease knowledge and self-efficacy (SMD 0.30, 95% CI 0.03-0.56; I²=74%; P=.03). These results were not statistically significant. No effect was seen regarding pain interference, HRQoL, anxiety, depression, pain coping, disease activity, functional ability, or treatment adherence.

CONCLUSIONS

Evidence that supports the inclusion of eHealth and mHealth interventions in JIA management is increasing. However, this evidence needs to be considered cautiously because of the small sample size, wide CIs, and moderate to high statistical heterogeneity. More rigorous research is needed on the longitudinal effects of real-time monitoring, web-based pediatric rheumatologist-children and young people interactions, the comparison among different self-management programs, and the use of wearable technologies as an objective measurement for monitoring physical activity before any recommendations that inform current practice can be given.

Exploring the relationship between the usability of a goal-oriented mobile health application and non-usage attrition in patients with multimorbidity: A blended data analysis approach

Background

Mobile health applications are increasingly used to support the delivery of health care services to a variety of patients. Based on data obtained from a pragmatic trial of the electronic Patient Reported Outcome (ePRO) app designed to support goal-oriented care primary care, this study aims to (1) examine how patient-reported usability changed over the one-year intervention period, and (2) explore participant attrition rate of the electronic Patient Reported Outcome app over one year study period.

Methods

We performed a secondary analysis of 44 older adults with complex chronic needs enrolled in the electronic Patient Reported Outcome-digital health intervention. App usage and attrition were measured using device-generated usage logs; usability was measured using the patient-reported post-study system usability questionnaire collected at 3, 6, 9, and 12 months. Research memos were used to interpret potential contextual contributing factors to patients' overall usage and usability score pattern. A data triangulation method of both quantitative and qualitative data was used to analyze and interpret study findings.

Results

While there was gradual attrition in the use of the ePRO app, patients' usability scores remained consistent throughout the study period. Qualitative memos suggested patients' encounters with technical difficulties and relationship dynamics with primary providers influenced patients' adherence to the ePRO app.

Conclusion

This study highlights that the patient-provider relationship is a key determining factor that influences complex patients' continued engagement with a Mobile health app. The finding calls attention to the measurement of usability of a Mobile health app, its impact on attrition, and contributing factors that influence patients' attrition. Trial registration: Clinicaltrials.gov Identified NCT02917954.

Exercise Amounts and Short- to Long-Term Weight Loss: Psychological Implications for Behavioral Treatments of Obesity

Purpose: Although exercise is typically included in behavioral weight-loss programs, the amount associated with meaningful short- to long-term weight reduction required investigation. Indirect paths between exercise-associated psychological changes and weight loss might be more relevant than the direct effect of exercise on energy expenditure-related weight loss in deconditioned/obese individuals. Method: Sedentary women with obesity (N = 97; M_age = 47.2 years) participated in a year-long cognitive-behavioral weight-loss treatment that emphasized building self-regulatory skills to maintain exercise in advance of transferring those skills to controlled eating. Results: There was a significant increase in exercise (metabolic equivalents/week or leisure score index; LSI), and significant improvements in mood, self-regulation for exercise, and self-regulation for eating from baseline to Months 6, 12, and 24. There were 5.9%, 5.8%, and 5.8% reductions in weight, respectively. Completion of 15-20 LSI did not significantly differ from greater amounts on associated weight losses except for the rare occurrence of ≥ 30 LSI over the full 24-month study period. There were significant bivariate relationships between completion of ≥ 15 LSI and weight loss over 6, 12, and 24 months. Within serial mediation analyses assessing changes from baseline-Months 6 and 24, there were significant indirect paths from ≥ 15 LSI→self-regulation for exercise→self-regulation for eating→weight loss, and ≥ 15 LSI→negative mood→self-regulation for exercise→self-regulation for eating→weight loss. Those paths were not significant when baseline-Month 12 changes were entered. Conclusion: Findings suggested the value of even manageable exercise amounts because of their association with psychosocial correlates of weight loss, and informed behavioral obesity treatments.

Effects of online stigma-reduction programme for people experiencing mental health conditions: A systematic review and meta-analysis

Despite the increased awareness of mental health-related issues, people experiencing mental health conditions have continued to face stigmatization worldwide. The literature on help-seeking behaviours has frequently highlighted the development of self-stigma because of public stigma and emphasized the need to address public stigmatization faced by them. Given the increasing acceptance of digital services in recent years, this systematic review aimed to examine the effects of online and face-to-face anti-stigma interventions in reducing public stigma towards people experiencing mental health conditions. A search was conducted on the Cochrane Library, CINAHL, PubMed, Embase, PsycInfo, and ProQuest from inception of the databases to October 2020. Studies were included in this review if they have explored: (i) public stigmatization towards people of all ages with different types of mental health conditions; (ii) online interventions; and (iii) face-to-face interventions. Nine studies were included in this review, of which only five were included in the meta-analysis as the remaining four had incomplete data. The meta-analysis included an aggregate of 1203 participants while the four excluded studies included 713 participants. Results revealed that online interventions performed favourably with small effect sizes in comparison to face-to-face, wait-list control, and no-intervention groups. Results from the studies excluded from the meta-analysis also found a significant reduction of public stigmatization with online interventions. Such findings provide insightful evidence for the effectiveness of online interventions in reducing public stigmatization. Hence, mental health organizations and groups can consider adopting online interventions suitable for their target audience and type of mental health conditions.

Predicting Success of a Digital Self-Help Intervention for Alcohol and Substance Use With Machine Learning

Background

Digital self-help interventions for reducing the use of alcohol tobacco and other drugs (ATOD) have generally shown positive but small effects in controlling substance use and improving the quality of life of participants. Nonetheless, low adherence rates remain a major drawback of these digital interventions, with mixed results in (prolonged) participation and outcome. To prevent non-adherence, we developed models to predict success in the early stages of an ATOD digital self-help intervention and explore the predictors associated with participant's goal achievement.

Methods

We included previous and current participants from a widely used, evidence-based ATOD intervention from the Netherlands (Jellinek Digital Self-help). Participants were considered successful if they completed all intervention modules and reached their substance use goals (i.e., stop/reduce). Early dropout was defined as finishing only the first module. During model development, participants were split per substance (alcohol, tobacco, cannabis) and features were computed based on the log data of the first 3 days of intervention participation. Machine learning models were trained, validated and tested using a nested k-fold cross-validation strategy.

Results

From the 32,398 participants enrolled in the study, 80% of participants did not complete the first module of the intervention and were excluded from further analysis. From the remaining participants, the percentage of success for each substance was 30% for alcohol, 22% for cannabis and 24% for tobacco. The area under the Receiver Operating Characteristic curve was the highest for the Random Forest model trained on data from the alcohol and tobacco programs (0.71 95%CI 0.69-0.73) and (0.71 95%CI 0.67-0.76), respectively, followed by cannabis (0.67 95%CI 0.59-0.75). Quitting substance use instead of moderation as an intervention goal, initial daily consumption, no substance use on the weekends as a target goal and intervention engagement were strong predictors of success.

Discussion

Using log data from the first 3 days of intervention use, machine learning models showed positive results in identifying successful participants. Our results suggest the models were especially able to identify participants at risk of early dropout. Multiple variables were found to have high predictive value, which can be used to further improve the intervention.

Factors associated with adherence to a web-based alcohol intervention among college students

Abstract This study aimed to evaluate the association between student characteristics and recruitment strategies in the adherence of college students to a web-based alcohol intervention. Participants were 46,329 Brazilian students aged from 18 to 30, who consumed alcohol during the past three months. Three recruitment strategies were implemented: open invitations, and personally-addressed invitations with or without non-monetary incentives. We evaluated the educational, sociodemographic, motivational, and alcohol consumption effects on adherence using logistic regression models. Women (aOR = 1.09 [1.04; 1.14]), students with higher income (aOR = 1.32 [1.21; 1.45]), and more motivated students (aOR = 1.04 [1.03; 1.05]) were more adherent to the intervention, as well as those reporting binge drinking (aOR = 1.26 [1.19; 1.33]) and alcohol hazardous use (aOR = 1.11 [1.05; 1.18]). The use of incentives was the main factor associated with adherence (aOR = 3.69 [2.46; 5.55]). Our results may help the development of future web-based interventions related to alcohol use.

Leveraging technology to move more and sit less

One of the major changes in the updated physical activity (PA) guidelines is the recommendation for adults to simply move more and sit less throughout the day. This recommendation comes during a time of proliferation and advancement of personal health technologies that allow adults greater access to interventions to increase PA. Wearable activity monitors provide direct feedback of activity levels allowing users to reach PA targets throughout the day. Gamification of these and other devices can engage users and sustain their motivation to increase PA, along with the formation of social networks through social media platforms. This review will discuss and present an overview of current technologies that can be leveraged to increase PA in adults. Specific attention will be paid to wearable activity monitors, gamification and social network platforms that can help adults increase and sustain their PA levels to improve their overall health.

Developing a Process for the Analysis of User Journeys and the Prediction of Dropout in Digital Health Interventions: Machine Learning Approach

Background

User dropout is a widespread concern in the delivery and evaluation of digital (ie, web and mobile apps) health interventions. Researchers have yet to fully realize the potential of the large amount of data generated by these technology-based programs. Of particular interest is the ability to predict who will drop out of an intervention. This may be possible through the analysis of user journey data—self-reported as well as system-generated data—produced by the path (or journey) an individual takes to navigate through a digital health intervention.

Objective

The purpose of this study is to provide a step-by-step process for the analysis of user journey data and eventually to predict dropout in the context of digital health interventions. The process is applied to data from an internet-based intervention for insomnia as a way to illustrate its use. The completion of the program is contingent upon completing 7 sequential cores, which include an initial tutorial core. Dropout is defined as not completing the seventh core.

Methods

Steps of user journey analysis, including data transformation, feature engineering, and statistical model analysis and evaluation, are presented. Dropouts were predicted based on data from 151 participants from a fully automated web-based program (Sleep Healthy Using the Internet) that delivers cognitive behavioral therapy for insomnia. Logistic regression with L1 and L2 regularization, support vector machines, and boosted decision trees were used and evaluated based on their predictive performance. Relevant features from the data are reported that predict user dropout.

Results

Accuracy of predicting dropout (area under the curve [AUC] values) varied depending on the program core and the machine learning technique. After model evaluation, boosted decision trees achieved AUC values ranging between 0.6 and 0.9. Additional handcrafted features, including time to complete certain steps of the intervention, time to get out of bed, and days since the last interaction with the system, contributed to the prediction performance.

Conclusions

The results support the feasibility and potential of analyzing user journey data to predict dropout. Theory-driven handcrafted features increased the prediction performance. The ability to predict dropout at an individual level could be used to enhance decision making for researchers and clinicians as well as inform dynamic intervention regimens.

Challenges in Acceptance and Compliance in Digital Health Assessments During Pregnancy: Prospective Cohort Study

BACKGROUND

Pregnant women are increasingly using mobile apps to access health information during the antenatal period. Therefore, digital health solutions can potentially be used as monitoring instruments during pregnancy. However, a main factor of success is high user engagement.

OBJECTIVE

The aim of this study was to analyze engagement and factors influencing compliance in a longitudinal study targeting pregnant women using a digital health app with self-tracking.

METHODS

Digitally collected data concerning demographics, medical history, technical aspects, and mental health from 585 pregnant women were analyzed. Patients filling out ≥80% of items at every study visit were considered to be highly compliant. Factors associated with high compliance were identified using logistic regression. The effect of a change in mental and physical well-being on compliance was assessed using a one-sample t test.

RESULTS

Only 25% of patients could be considered compliant. Overall, 63% left at least one visit blank. Influential variables for higher engagement included higher education, higher income, private health insurance, nonsmoking, and German origin. There was no relationship between a change in the number of physical complaints or depressive symptoms and study dropout.

CONCLUSIONS

Maintaining high engagement with digital monitoring devices over a long time remains challenging. As cultural and socioeconomic background factors had the strongest influence, more effort needs to be directed toward understanding the needs of patients from different demographic backgrounds to ensure high-quality care for all patients. More studies need to report on compliance to disclose potential demographic bias.

Beating cancer-related fatigue with the Untire mobile app: Results from a waiting-list randomized controlled trial

Objective

This waiting‐list randomized controlled trial examined the effectiveness of a self‐management mHealth app in improving fatigue and quality of life (QoL) in cancer patients and survivors.

Methods

Persons with cancer‐related fatigue (CRF) were recruited across four English speaking countries, via social media, and randomized into intervention (n = 519) and control (n = 280) groups. Whereas the intervention group received immediate access to the Untire app, the control group received access only after 12‐weeks. Primary outcomes fatigue severity and interference, and secondary outcome QoL were assessed at baseline, 4, 8, and 12‐weeks. We ran generalized linear mixed models for all outcomes to determine the effects of app access (yes/no), over 12‐weeks, following the intention‐to‐treat principle.

Results

Compared with the control group, the intervention group showed significantly larger improvements in fatigue severity (d = 0.40), fatigue interference (d = 0.35), and overall QoL on average (d = 0.32) (P's < .01), but not for overall QoL in the past week (P = .07). Sensitivity analyses indicated that participants with medium or high app use benefited most when compared with nonusers and control participants (P's ≤ .02). The intervention effect on fatigue interference was slightly stronger in younger participants (≤56 vs. >56). Effects did not depend on education and cancer status. Reliable change analyses indicated that significantly more people showed full recovery for fatigue in the intervention vs the control group (P's = .02).

Conclusions

The Untire app can be an effective mHealth solution for cancer patients and survivors with moderate to severe CRF.

Critical Predictors for the Early Detection of Conversion From Unipolar Major Depressive Disorder to Bipolar Disorder: Nationwide Population-Based Retrospective Cohort Study

BACKGROUND

Unipolar major depressive disorder (MDD) and bipolar disorder are two major mood disorders. The two disorders have different treatment strategies and prognoses. However, bipolar disorder may begin with depression and could be diagnosed as MDD in the initial stage, which may later contribute to treatment failure. Previous studies indicated that a high proportion of patients diagnosed with MDD will develop bipolar disorder over time. This kind of hidden bipolar disorder may contribute to the treatment resistance observed in patients with MDD.

OBJECTIVE

In this population-based study, our aim was to investigate the rate and risk factors of a diagnostic change from unipolar MDD to bipolar disorder during a 10-year follow-up. Furthermore, a risk stratification model was developed for MDD-to-bipolar disorder conversion.

METHODS

We conducted a retrospective cohort study involving patients who were newly diagnosed with MDD between January 1, 2000, and December 31, 2004, by using the Taiwan National Health Insurance Research Database. All patients with depression were observed until (1) diagnosis of bipolar disorder by a psychiatrist, (2) death, or (3) December 31, 2013. All patients with depression were divided into the following two groups, according to whether bipolar disorder was diagnosed during the follow-up period: converted group and nonconverted group. Six groups of variables within the first 6 months of enrollment, including personal characteristics, physical comorbidities, psychiatric comorbidities, health care usage behaviors, disorder severity, and psychotropic use, were extracted and were included in a classification and regression tree (CART) analysis to generate a risk stratification model for MDD-to-bipolar disorder conversion.

RESULTS

Our study enrolled 2820 patients with MDD. During the follow-up period, 536 patients were diagnosed with bipolar disorder (conversion rate=19.0%). The CART method identified five variables (kinds of antipsychotics used within the first 6 months of enrollment, kinds of antidepressants used within the first 6 months of enrollment, total psychiatric outpatient visits, kinds of benzodiazepines used within one visit, and use of mood stabilizers) as significant predictors of the risk of bipolar disorder conversion. This risk CART was able to stratify patients into high-, medium-, and low-risk groups with regard to bipolar disorder conversion. In the high-risk group, 61.5%-100% of patients with depression eventually developed bipolar disorder. On the other hand, in the low-risk group, only 6.4%-14.3% of patients with depression developed bipolar disorder.

CONCLUSIONS

The CART method identified five variables as significant predictors of bipolar disorder conversion. In a simple two- to four-step process, these variables permit the identification of patients with low, intermediate, or high risk of bipolar disorder conversion. The developed model can be applied to routine clinical practice for the early diagnosis of bipolar disorder.