The CDHI team at the ETH WEV in Zurich


As a doctoral researcher at the Centre for Digital Health Interventions at ETH Zurich, my research focuses on using machine learning to predict and prevent user churn in digital health interventions.

ETH Profile | Project Website

Scientific contributions: GoogleScholar | ResearchGate | PubMed | ORCID | ETH Research Collection | Publons | ImpactStory | Academia| OSFHOME | Scopus |Semantic Scholar| Loop

Peer-review activities: npj Digital Medicine | BMC Digital Health | ACM Conference on Human Factors in Computing Systems (CHI 2023 & 2024) | Journal of Medical Internet Research (JMIR) | 18th International Conference on Business Informatics (WI23) | International Conference on Information Systems (ICIS 2024)

Teaching activities: Digital Health in Practice (University of Zurich, FS 2023) | Digital Therapeutics Project (University of St.Gallen, SS 2023) | Digital Health in Practice (University of Zurich, FS 2022) | Digital Health Project Course (University of St.Gallen, SS 2022) | Digital Health Project Course (ETH Zurich, FS 2021) | Digital Health Project Course (University of St.Gallen, SS 2021) | Digital Health Project Course (ETH Zurich, FS 2020)

Research Projects

Digital health interventions (DHIs) hold great potential in supporting patients and healthcare systems in dealing with the growing prevalence and economic costs of non-communicable diseases (NCDs) – the leading causes of death and disability worldwide. In particular, mobile health apps (mHealth apps) are considered an accessible and scalable solution for promoting behavior change among patients, improving health outcomes, and reducing healthcare costs. Despite their growing evidence base and availability, mHealth apps suffer from high dropout rates, with many users not adhering to them as intended. Understanding the barriers and facilitators to adherence is crucial to prevent dropouts and increase the effectiveness of digital health interventions. Identifying users at risk and preventing them from dropping out can also increase the efficiency and quality of DHIs, as many interventions are only beneficial if used regularly and for a prolonged period.

In this regard, my research has the following objectives:

  • Identifying factors that influence adherence and predict dropouts based on existing literature and real-world data from longitudinal studies of different health domains.
  • Identifying intervention components that effectively intervene in critical situations to prevent dropout
  • Developing and evaluating an early warning system that can predict and prevent dropouts.


Predicting early user churn in a public digital weight loss intervention

MAY 11, 2024 | Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI ’24)

Abstract: Digital health interventions (DHIs) offer promising solutions to the rising global challenges of noncommunicable diseases by promoting behavior change, improving health outcomes, and reducing healthcare costs. However, high churn rates are a concern with DHIs, with many users disengaging before achieving desired outcomes. Churn prediction can help DHI providers identify and retain at-risk users, enhancing the efficacy of DHIs. We analyzed churn prediction models for a weight loss app using various machine learning algorithms on data from 1,283 users and 310,845 event logs. The best-performing model, a random forest model that only used daily login counts, achieved an F1 score of 0.87 on day 7 and identified an average of 93% of churned users during the week-long trial. Notably, higher-dimensional models performed better at low false positive rate thresholds. Our findings suggest that user churn can be forecasted using engagement data, aiding in timely personalized strategies and better health results.

Read full-text →

Effective Behavior Change Techniques in Digital Health Interventions for the Prevention or Management of Noncommunicable Diseases: An Umbrella Review


Background: Despite an abundance of digital health interventions (DHIs) targeting the prevention and management of noncommunicable diseases (NCDs), it is unclear what specific components make a DHI effective.

Objective: This narrative umbrella review aimed to identify the most effective behavior change techniques (BCTs) in DHIs that address the prevention or management of NCDs.

Methods: Five electronic databases were searched for articles published in English between January 2007 and December 2022. Studies were included if they were systematic reviews or meta-analyses of DHIs targeting the modification of one or more NCD-related risk factors in adults. BCTs were coded using the Behavior Change Technique Taxonomy v1. Study quality was assessed using AMSTAR 2.

Results: Eighty-five articles, spanning 12 health domains and comprising over 865,000 individual participants, were included in the review. We found evidence that DHIs are effective in improving health outcomes for patients with cardiovascular disease, cancer, type 2 diabetes, and asthma, and health-related behaviors including physical activity, sedentary behavior, diet, weight management, medication adherence, and abstinence from substance use. There was strong evidence to suggest that credible source, social support, prompts and cues, graded tasks, goals and planning, feedback and monitoring, human coaching and personalization components increase the effectiveness of DHIs targeting the prevention and management of NCDs.

Conclusions: This review identifies the most common and effective BCTs used in DHIs, which warrant prioritization for integration into future interventions. These findings are critical for the future development and upscaling of DHIs and should inform best practice guidelines.

Read full-text →

Factors Influencing Adherence to mHealth Apps among Adults with Non-communicable Disease: A 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.

Read full-text →

Engagement With a Mobile Phone–Based Life Skills Intervention for Adolescents and Its Association With Participant Characteristics and Outcomes: Tree-Based Analysis


Background: Mobile phone-delivered life skills programs are an emerging and promising way to promote mental health and prevent substance use among adolescents, but little is known about how adolescents actually use them.

Objective: The aim of this study is to determine engagement with a mobile phone-based life skills program and its different components, as well as the associations of engagement with adolescent characteristics and intended substance use and mental health outcomes.

Methods: We performed secondary data analysis on data from the intervention group (n=750) from a study that compared a mobile phone-based life skills intervention for adolescents recruited in secondary and upper secondary school classes with an assessment-only control group. Throughout the 6-month intervention, participants received 1 SMS text message prompt per week that introduced a life skills topic or encouraged participation in a quiz or individual life skills training or stimulated sharing messages with other program participants through a friendly contest. Decision trees were used to identify predictors of engagement (use and subjective experience). The stability of these decision trees was assessed using a resampling method and by graphical representation. Finally, associations between engagement and intended substance use and mental health outcomes were examined using logistic and linear regression analyses.

Results: The adolescents took part in half of the 50 interactions (mean 23.6, SD 15.9) prompted by the program, with SMS text messages being the most used and contests being the least used components. Adolescents who did not drink in a problematic manner and attended an upper secondary school were the ones to use the program the most. Regarding associations between engagement and intended outcomes, adolescents who used the contests more frequently were more likely to be nonsmokers at follow-up than those who did not (odds ratio 0.86, 95% CI 0.76-0.98; P=.02). In addition, adolescents who read the SMS text messages more attentively were less likely to drink in a problematic manner at follow-up (odds ratio 0.43, 95% CI 1.29-3.41; P=.003). Finally, participants who used the program the most and least were more likely to increase their well-being from baseline to 6-month follow-up compared with those with average engagement (βs=.39; t586=2.66; P=.008; R2=0.24).

Conclusions: Most of the adolescents participating in a digital life skills program that aimed to prevent substance use and promote mental health engaged with the intervention. However, measures to increase engagement in problem drinkers should be considered. Furthermore, efforts must be made to ensure that interventions are engaging and powerful across different educational levels. First results indicate that higher engagement with digital life skills programs could be associated with intended outcomes. Future studies should apply further measures to improve the reach of lower-engaged participants at follow-up to establish such associations with certainty.

Read full-text →

Fact Sheet: Success Factors of mHealth Applications


Abstract: How can mobile health apps support prevention and treatment in NCD, mental health and addiction? Set goals - observe behavior - provide feedback: These key features characterize successful mHealth apps. These are the main findings of a comprehensive literature study conducted by ETH Zurich and the Universities of St. Gallen and Zurich on behalf of the FOPH (BAG). For continuous use and behavioral change, it is essential that mHealth applications take individual needs into account, remind users automatically, are intuitively designed, function smoothly, and offer the possibility of personal support.

Read full-text →

Literature Review on Behavior Change Through mHealth Applications.


Background: mHealth applications open up a wide range of possibilities for individualized prevention, the promotion of protective behaviors, and the self-management of non-communicable diseases. At the same time, their development and maintenance is significantly more complex compared to browser-based eHealth applications, and users often find it difficult to select suitable apps. While evaluation frameworks already exist for general quality criteria such as data protection, design, usability, or security, the conditions necessary to achieve behavioral change among users through mHealth applications have not yet been systematically researched and summarized. Within two separate literature studies, the present work investigated and identified (1) usage enhancement techniques and (2) behavior change techniques that should be considered in the design and development of mHealth applications and on the basis of which it is also possible to develop a set of criteria for evaluating health apps for users.

Objective: The first sub-study investigated which techniques used in mHealth apps on NCDs, mental health, and addiction influence user adherence. The second substudy examined the effect of behavior change techniques on intended behavior change.

Methods: In sub-study 1, a systematic literature review of existing primary studies was conducted to identify relevant techniques for increasing usage. In a first step, techniques were identified that improved usage adherence within the primary studies. In a second step, the influence of additional factors on usage adherence was investigated, such as the characteristics of the target population or the way the application was delivered. In a third step, for each primary study, usage adherence was calculated as the quotient of intended and actual usage to provide a reference value within the different health domains and to identify mHealth applications with high usage adherence. In sub-study 2, to identify relevant behavior change techniques, a systematic overview (English: Overview or Umbrella Review) of existing systematic reviews was conducted. Relevant scientific articles for both substudies were identified through systematic searches of electronic literature databases. Relevant information from each article was extracted and analyzed.

Results: The literature search for substudy 1 yielded a total of 2862 potentially relevant articles, of which 99 were relevant to this review and were analyzed in more detail. Techniques with a positive impact on app use were presented separately for the 7 health domains, with the following 3 techniques identified as relevant to all health domains: (1) personalization or tailoring the content of the mHealth app to the individual needs of the user, (2) reminders in the form of individualized push notifications, (3) a user-friendly app design, and technical stability. 5 Usage adherence derived from the primary studies averaged 56.0% and was highest for lifestyle interventions aimed at changing multiple behaviors simultaneously (60.1%) and lowest for mHealth apps aimed at reducing substance use (46.1%). Further, quantitative analysis revealed a positive correlation between adherence to use and level of personalized care during the intervention. For the NCD self-management domain, there was a significant positive correlation between adherence to use and the average age of study participants. The literature search for substudy 2 yielded a total of 615 potentially relevant articles, of which 66 were relevant to this review and were analyzed in more detail. For the area of NCD self-management, the effectiveness of exclusively app-based programs is predominantly mixed or still unclear, with the exception of diabetes management apps. Key behavior change techniques in NCD self-management are goals that can be individualized as much as possible regarding the targeted behavior (e.g., taking medications), self-monitoring of behavior (e.g., via diary function in the app), and feedback on behavior (e.g., graphical representation regarding the achievement or non-achievement of the behavioral goal). Guidance from a real-life professional appears to be an important component of effective digital programs to support chronic disease management. Evidence on the effectiveness of app-based programs for dietary behavior change is also still mixed, with dietary changes, such as increasing fruit and vegetable consumption, being achieved more frequently than reductions in the amount of energy consumed. The mHealth applications used to date predominantly use behavior change techniques that have also proven effective in traditional individual and group counseling sessions to change dietary behavior: Individual goal setting, behavioral monitoring and feedback, and social support. The extent to which other techniques, such as self-image change or social comparison, are effective cannot be answered based on the data to date. The effectiveness of apps to increase physical activity is now well established scientifically, with sicker and vulnerable populations benefiting particularly. Again, setting individual activity goals, monitoring them, and providing feedback on their achievement play a key role. The involvement of a real professional does not seem necessary in these programs. In contrast, programs that use data automatically collected by the system (e.g., via motion sensor) for individualization are more effective. In the case of apps for weight reduction and for changing several behaviors at the same time (so-called lifestyle interventions), which usually also aim at weight reduction by promoting physical activity and healthy eating, the effectiveness is mixed. Key components include the use of multiple and interactive behavior change techniques, particularly for goal setting, as well as behavioral monitoring and feedback. In mental health improvement programs, elements of cognitive behavioral therapy have proven effective in reducing anxiety and depression via the Internet or app. Similar to self-management programs for NCDs, in-person guidance from a professional seems to help with effectiveness. In addition to self-monitoring of behavior, changes in cognitive processes (e.g., increasing positive thoughts, cognitive flexibility, perceived control) and skills (e.g., using mindfulness skills or cognitive-behavioral techniques) represent key mechanisms of action. Evidence on the effectiveness of app programs to reduce alcohol use in the general population has been mixed, with some positive studies but also many without significant results. Successful programs are characterized in particular by offering users practical, easy-to-implement advice on how to replace alcohol consumption and solve problems; this should come from a source perceived as credible. Evidence on the effectiveness of tobacco cessation apps has also been quite heterogeneous. In reviews of primarily Internet-based programs, various techniques were associated with effectiveness: setting specific behavioral goals and action planning, advice on problem solving and health consequences of smoking, weighing the pros and cons of quitting smoking, and also social and medication support.

Conclusions: Technologies for personalization and individualization of content are central to high app usage and effectiveness. Users should be able to set personally relevant behavioral goals and monitor their degree of realization over time through the app. Interactive functions that take into account the degree of goal achievement as well as characteristics of the person and the context are particularly suitable. Regular reminders by the app that take into account the individual's availability and need for interaction represent an essential prerequisite for using these central techniques for goal setting, behavioral observation and feedback over a longer period of time. In addition to these automated features, opportunities for personal guidance and social support, especially for apps used in clinical groups, form an essential element for their use and effectiveness. Technical stability and user-friendly app design are also relevant for regular use. Overall, research on promising usage enhancement techniques as well as behavior change techniques for mHealth apps is still poorly advanced. The underlying studies often have a pilot character, and the implementation of the techniques and operationalization of the results is very inconsistent. Since mHealth apps usually use multiple techniques to increase usage and change behavior, causal statements about individual techniques are hardly possible. This will require more controlled and experimental studies in the future. The recommended techniques for individualized goal setting, behavioral monitoring, feedback, reminders, and social support also represent basic elements of current health behavior models and proven cognitive-behavioral therapy interventions. Their integration into mHealth applications provides a solid foundation. In order to optimize them, however, new techniques whose full potential can only be exploited through digital technologies should also be tested and reviewed in the future.

Read full-text →

The Potential of Mobile Apps for Improving Asthma Self-Management: A Review of Publicly Available and Well-Adopted Asthma Apps

AUGUST 2, 2017 | JMIR mHealth and uHealth

Background: Effective disease self-management lowers asthma's burden of disease for both individual patients and health care systems. In principle, mobile health (mHealth) apps could enable effective asthma self-management interventions that improve a patient's quality of life while simultaneously reducing the overall treatment costs for health care systems. However, prior reviews in this field have found that mHealth apps for asthma lack clinical evaluation and are often not based on medical guidelines. Yet, beyond the missing evidence for clinical efficacy, little is known about the potential apps might have for improving asthma self-management.

Objective: The aim of this study was to assess the potential of publicly available and well-adopted mHealth apps for improving asthma self-management.

Methods: The Apple App store and Google Play store were systematically searched for asthma apps. In total, 523 apps were identified, of which 38 apps matched the selection criteria to be included in the review. Four requirements of app potential were investigated: app functions, potential to change behavior (by means of a behavior change technique taxonomy), potential to promote app use (by means of a gamification components taxonomy), and app quality (by means of the Mobile Application Rating Scale [MARS]).

Results: The most commonly implemented functions in the 38 reviewed asthma apps were tracking (30/38, 79%) and information (26/38, 68%) functions, followed by assessment (20/38, 53%) and notification (18/38, 47%) functions. On average, the reviewed apps applied 7.12 of 26 available behavior change techniques (standard deviation [SD]=4.46) and 4.89 of 31 available gamification components (SD=4.21). Average app quality was acceptable (mean=3.17/5, SD=0.58), whereas subjective app quality lied between poor and acceptable (mean=2.65/5, SD=0.87). Additionally, the sum scores of all review frameworks were significantly correlated (lowest correlation: r36=.33, P=.04 between number of functions and gamification components; highest correlation: r36=.80, P<.001 between number of behavior change techniques and gamification components), which suggests that an app's potential tends to be consistent across review frameworks.

Conclusions: Several apps were identified that performed consistently well across all applied review frameworks, thus indicating the potential mHealth apps offer for improving asthma self-management. However, many apps suffer from low quality. Therefore, app reviews should be considered as a decision support tool before deciding which app to integrate into a patient's asthma self-management. Furthermore, several research-practic

Read full-text →
Get in touch →