Applying machine learning to predict future adherence to physical activity programs

Demographics and socioeconomic determinants of health predict continued participation in a CT lung cancer screening program

PURPOSE

We developed machine learning (ML) models to assess demographic and socioeconomic status (SES) variables' value in predicting continued participation in a low-dose CT lung cancer screening (LCS) program.

MATERIALS AND METHODS

480 LCS subjects were retrospectively examined for the following outcomes: (#1) no follow-up (single LCS scan) vs. multiple follow-ups (220 and 260 subjects respectively) and (#2) absent or delayed (>1 month past the due date) follow-up vs timely follow-up (356 and 124 subjects respectively). We quantified the contributions of 14 socioeconomic, demographic, and clinical predictors to LCS adherence, and validated and compared prediction performances of multivariate logistic regression (MLR), support vector machine (SVM) and shallow neural network (NN) models.

RESULTS

For outcome #1, age, sex, race, insurance status, personal cancer history, and median household income were found to be associated with returning for follow-ups. For outcome #2, age, sex, race, and insurance status were significant predictor of absent/delayed LCS follow-up. Across 5-fold cross-validation, the MLR model achieved an average AUC of 0.732 (95% CI, 0.661-0.803) for outcome #1 and 0.633 (95% CI, 0.602-0.664) for outcome #2 and is the model with best predictive performance overall, whereas NN and SVM tended to overfit training data and fell short on testing data performance for either outcome.

CONCLUSIONS

We identified significant predictors of LCS adherence, and our ML models can predict which subjects are at higher risk of receiving no or delayed LCS follow-ups. Our results could inform data-driven interventions to engage vulnerable populations and extend the benefits of LCS.

Transforming the cardiometabolic disease landscape: Multimodal AI-powered approaches in prevention and management

The rise of artificial intelligence (AI) has revolutionized various scientific fields, particularly in medicine, where it has enabled the modeling of complex relationships from massive datasets. Initially, AI algorithms focused on improved interpretation of diagnostic studies such as chest X-rays and electrocardiograms in addition to predicting patient outcomes and future disease onset. However, AI has evolved with the introduction of transformer models, allowing analysis of the diverse, multimodal data sources existing in medicine today. Multimodal AI holds great promise in more accurate disease risk assessment and stratification as well as optimizing the key driving factors in cardiometabolic disease: blood pressure, sleep, stress, glucose control, weight, nutrition, and physical activity. In this article we outline the current state of medical AI in cardiometabolic disease, highlighting the potential of multimodal AI to augment personalized prevention and treatment strategies in cardiometabolic disease.

The Relationship Between Barriers to Physical Activity and Depressive Symptoms in Community-Dwelling Women

Background

Women are less physically active, report greater perceived barriers for exercise, and show higher levels of depressive symptoms. This contributes to high global disability. The relationship between perceived barriers for physical activity and depressive symptoms in women remains largely unexplored. The aims of this cross-sectional analysis were to examine the association between physical activity barriers and depressive symptoms, and identify types of barriers in physically inactive community-dwelling women.

Methods

Three hundred eighteen physically inactive women aged 25-65 years completed the Barriers to Being Active Quiz (BBAQ) developed by the Centers for Disease Control and Prevention, and the Center for Epidemiological Studies Depression Scale at the baseline visit of the mobile phone-based physical activity education trial. The BBAQ consists of six subscales (lack of time, social influence, lack of energy, lack of willpower, fear of injury, lack of skill, and lack of resources). We used multivariate regression analyses, correcting for sociodemographics.

Results

Higher physical activity barriers were associated with greater depressive symptoms scores (linear effect, estimate = 0.75, 95% confidence interval [CI]: 0.39-1.12, p < 0.001). This effect appeared to taper off for the higher barrier scores (quadratic effect, estimate: -0.02, 95% CI: -0.03 to -0.01, p = 0.002). Exploratory analyses indicated that these associations were most driven by the social influence (p = 0.027) and lack of energy subscales (p = 0.017).

Conclusions

Higher depression scores were associated with higher physical activity barriers. Social influence and lack of energy were particularly important barriers. Addressing these barriers may improve the efficacy of physical activity interventions in women with higher depressive symptoms. Future research should assess this in a randomized controlled trial.

Trial Registration ClinicalTrialsgov#

NCTO1280812 registered January 21, 2011.

Leveraging Machine Learning to Develop Digital Engagement Phenotypes of Users in a Digital Diabetes Prevention Program: Evaluation Study

BACKGROUND

Digital diabetes prevention programs (dDPPs) are effective "digital prescriptions" but have high attrition rates and program noncompletion. To address this, we developed a personalized automatic messaging system (PAMS) that leverages SMS text messaging and data integration into clinical workflows to increase dDPP engagement via enhanced patient-provider communication. Preliminary data showed positive results. However, further investigation is needed to determine how to optimize the tailoring of support technology such as PAMS based on a user's preferences to boost their dDPP engagement.

OBJECTIVE

This study evaluates leveraging machine learning (ML) to develop digital engagement phenotypes of dDPP users and assess ML's accuracy in predicting engagement with dDPP activities. This research will be used in a PAMS optimization process to improve PAMS personalization by incorporating engagement prediction and digital phenotyping. This study aims (1) to prove the feasibility of using dDPP user-collected data to build an ML model that predicts engagement and contributes to identifying digital engagement phenotypes, (2) to describe methods for developing ML models with dDPP data sets and present preliminary results, and (3) to present preliminary data on user profiling based on ML model outputs.

METHODS

Using the gradient-boosted forest model, we predicted engagement in 4 dDPP individual activities (physical activity, lessons, social activity, and weigh-ins) and general activity (engagement in any activity) based on previous short- and long-term activity in the app. The area under the receiver operating characteristic curve, the area under the precision-recall curve, and the Brier score metrics determined the performance of the model. Shapley values reflected the feature importance of the models and determined what variables informed user profiling through latent profile analysis.

RESULTS

We developed 2 models using weekly and daily DPP data sets (328,821 and 704,242 records, respectively), which yielded predictive accuracies above 90%. Although both models were highly accurate, the daily model better fitted our research plan because it predicted daily changes in individual activities, which was crucial for creating the "digital phenotypes." To better understand the variables contributing to the model predictor, we calculated the Shapley values for both models to identify the features with the highest contribution to model fit; engagement with any activity in the dDPP in the last 7 days had the most predictive power. We profiled users with latent profile analysis after 2 weeks of engagement (Bayesian information criterion=-3222.46) with the dDPP and identified 6 profiles of users, including those with high engagement, minimal engagement, and attrition.

CONCLUSIONS

Preliminary results demonstrate that applying ML methods with predicting power is an acceptable mechanism to tailor and optimize messaging interventions to support patient engagement and adherence to digital prescriptions. The results enable future optimization of our existing messaging platform and expansion of this methodology to other clinical domains.

TRIAL REGISTRATION

ClinicalTrials.gov NCT04773834; https://www.clinicaltrials.gov/ct2/show/NCT04773834.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

RR2-10.2196/26750.

Predictors of sustained physical activity: behaviour, bodily health, and the living environment

This study examined the determinants of sustained physical activity. Eighty-four participants undertook a 7-weeks walking regime (i.e., a 1-h biometrically-monitored walk, at least 5 days/week), with bioelectrical impedance (BIA) and total cholesterol capillary blood measurements performed before and after programme. To investigate behavioural habit formation, 7 weeks after walking termination, all participants were interviewed and (health) re-tested. Data were modelled with an artificial neural network (ANN) cascading algorithm. Our results highlight the successful prediction of continued physical activity by considering one's physical fitness state, the environmental living context, and risk for cardiovascular disease. Importantly, those artificial neural network models also taking body mass index (BMI) and blood cholesterol as predictors excel at predicting walking continuation (i.e., predictions with 93% predictability). These results are first to highlight the type and importance of available physiological drivers in maintaining a sustained physical activity regime such as walking. They are discussed within the framework of habit formation and the nowadays health and/or wellbeing focus.

Classification of patients with chronic disease by activation level using machine learning methods

Patient Activation Measure (PAM) measures the activation level of patients with chronic conditions and correlates well with patient adherence behavior, health outcomes, and healthcare costs. PAM is increasingly used in practice to identify patients needing more support from the care team. We define PAM levels 1 and 2 as low PAM and investigate the performance of eight machine learning methods (Logistic Regression, Lasso Regression, Ridge Regression, Random Forest, Gradient Boosted Trees, Support Vector Machines, Decision Trees, Neural Networks) to classify patients. Primary data collected from adult patients (n=431) with Diabetes Mellitus (DM) or Hypertension (HT) attending Family Health Centers in Istanbul, Turkey, is used to test the methods. [Formula: see text] of patients in the dataset have a low PAM level. Classification performance with several feature sets was analyzed to understand the relative importance of different types of information and provide insights. The most important features are found as whether the patient performs self-monitoring, smoking and exercise habits, education, and socio-economic status. The best performance was achieved with the Logistic Regression algorithm, with Area Under the Curve (AUC)=0.72 with the best performing feature set. Alternative feature sets with similar prediction performance are also presented. The prediction performance was inferior with an automated feature selection method, supporting the importance of using domain knowledge in machine learning.

Transforming Estonian health data to the Observational Medical Outcomes Partnership (OMOP) Common Data Model: lessons learned

Objective

To describe the reusable transformation process of electronic health records (EHR), claims, and prescriptions data into Observational Medical Outcome Partnership (OMOP) Common Data Model (CDM), together with challenges faced and solutions implemented.

Materials and Methods

We used Estonian national health databases that store almost all residents' claims, prescriptions, and EHR records. To develop and demonstrate the transformation process of Estonian health data to OMOP CDM, we used a 10% random sample of the Estonian population (n = 150 824 patients) from 2012 to 2019 (MAITT dataset). For the sample, complete information from all 3 databases was converted to OMOP CDM version 5.3. The validation was performed using open-source tools.

Results

In total, we transformed over 100 million entries to standard concepts using standard OMOP vocabularies with the average mapping rate 95%. For conditions, observations, drugs, and measurements, the mapping rate was over 90%. In most cases, SNOMED Clinical Terms were used as the target vocabulary.

Discussion

During the transformation process, we encountered several challenges, which are described in detail with concrete examples and solutions.

Conclusion

For a representative 10% random sample, we successfully transferred complete records from 3 national health databases to OMOP CDM and created a reusable transformation process. Our work helps future researchers to transform linked databases into OMOP CDM more efficiently, ultimately leading to better real-world evidence.

Predicting Adherence to Behavior Change Support Systems Using Machine Learning: Systematic Review

BACKGROUND

There is a dearth of knowledge on reliable adherence prediction measures in behavior change support systems (BCSSs). Existing reviews have predominately focused on self-reporting measures of adherence. These measures are susceptible to overestimation or underestimation of adherence behavior.

OBJECTIVE

This systematic review seeks to identify and summarize trends in the use of machine learning approaches to predict adherence to BCSSs.

METHODS

Systematic literature searches were conducted in the Scopus and PubMed electronic databases between January 2011 and August 2022. The initial search retrieved 2182 journal papers, but only 11 of these papers were eligible for this review.

RESULTS

A total of 4 categories of adherence problems in BCSSs were identified: adherence to digital cognitive and behavioral interventions, medication adherence, physical activity adherence, and diet adherence. The use of machine learning techniques for real-time adherence prediction in BCSSs is gaining research attention. A total of 13 unique supervised learning techniques were identified and the majority of them were traditional machine learning techniques (eg, support vector machine). Long short-term memory, multilayer perception, and ensemble learning are currently the only advanced learning techniques. Despite the heterogeneity in the feature selection approaches, most prediction models achieved good classification accuracies. This indicates that the features or predictors used were a good representation of the adherence problem.

CONCLUSIONS

Using machine learning algorithms to predict the adherence behavior of a BCSS user can facilitate the reinforcement of adherence behavior. This can be achieved by developing intelligent BCSSs that can provide users with more personalized, tailored, and timely suggestions.

Identifying Predictors of Adherence to the Physical Activity Goal: A Secondary Analysis of the SMARTER Weight Loss Trial

INTRODUCTION/PURPOSE

Research is needed to inform tailoring supportive strategies for promoting physical activity (PA) in the context of behavioral treatment of obesity. We aimed to identify baseline participant characteristics and short-term intervention response predictors associated with adherence to the study-defined PA goal in a mobile health (mHealth) weight loss trial.

METHODS

A secondary analysis was conducted of a 12-month weight loss trial (SMARTER) that randomized 502 adults with overweight or obesity to either self-monitoring of diet, PA, and weight with tailored feedback messages ( n = 251) or self-monitoring alone ( n = 251). The primary outcome was average adherence to the PA goal of ≥150 min·wk -1 of moderate- and vigorous-intensity aerobic activities (MVPA) from Fitbit Charge 2™ trackers over 52 wk. Twenty-five explanatory variables were considered. Machine learning methods and linear regression were used to identify predictors of adherence to the PA goal.

RESULTS

The sample ( N = 502) was mostly female (80%), White (82%) with the average age of 45 ± 14.4 yr and body mass index of 33.7 ± 4.0 kg·m -2 . Machine learning methods identified PA goal adherence for the first week as the most important predictor of long-term PA goal adherence. In the parsimonious linear regression model, higher PA goal adherence for the first week, greater PA FB messages opened, older age, being male, higher education, being single and not having obstructive sleep apnea were associated with higher long-term PA goal adherence.

CONCLUSIONS

To our knowledge, this is the first study using machine learning approaches to identify predictors of long-term PA goal adherence in a mHealth weight loss trial. Future studies focusing on facilitators or barriers to PA among young and middle-age adults and women with low PA goal adherence are warranted.

Implementation and User Evaluation of an eHealth Technology Platform Supporting Patients With Cardiovascular Disease in Managing Their Health After a Cardiac Event: Mixed Methods Study

BACKGROUND

eHealth technology can help patients with cardiovascular disease adopt and maintain a healthy lifestyle by supporting self-management and offering guidance, coaching, and tailored information. However, to support patients over time, eHealth needs to blend in with their needs, treatment, and daily lives. Just as needs can differ between patients, needs can change within patients over time. To better adapt technology features to patients' needs, it is necessary to account for these changes in needs and contexts of use.

OBJECTIVE

This study aimed to identify and monitor patients' needs for support from a web-based health management platform and how these needs change over time. It aimed to answer the following research questions: "How do novice and more advanced users experience an online health management platform?" "What user expectations support or hinder the adoption of an online health management platform, from a user perspective?" and "How does actual usage relate to user experiences and adoption?"

METHODS

A mixed methods design was adopted. The first method involved 2 rounds of usability testing, followed by interviews, with 10 patients at 0 months (round 1) and 12 patients at 6 months (round 2). In the second method, log data were collected to describe the actual platform use.

RESULTS

After starting cardiac rehabilitation, the platform was used frequently. The patients mentioned that they need to have an incentive, set goals, self-monitor their health data, and feel empowered by the platform. However, soon after the rehabilitation program stopped, use of the platform declined or patients even quit because of the lack of continued tailored or personalized advice. The reward system motivated them to log data, but most participants indicated that being healthy should be the main focus, not receiving gifts. A web-based platform is flexible, accessible, and does not have any obligations; however, it should be implemented as an addition to regular care.

CONCLUSIONS

Although use of the platform declined in the longer term, patients quitting the technology did not directly indicate that the technology was not functioning well or that patients no longer focused on achieving their values. The key to success should not be user adherence to a platform but adherence to healthy lifestyle habits. Therefore, the implementation of eHealth should include the transition to a stage where patients might no longer need support from a technology platform to be independently and sustainably adherent to their healthy lifestyle habits. This emphasizes the importance of conducting multi-iterative evaluations to continuously monitor whether and how patients' needs and contexts of use change over time. Future research should focus on how this transition can be identified and monitored and how these insights can inform the design and implementation of the technology.

Machine Learning in Nutrition Research

Data currently generated in the field of nutrition are becoming increasingly complex and high-dimensional, bringing with them new methods of data analysis. The characteristics of machine learning (ML) make it suitable for such analysis and thus lend itself as an alternative tool to deal with data of this nature. ML has already been applied in important problem areas in nutrition, such as obesity, metabolic health, and malnutrition. Despite this, experts in nutrition are often without an understanding of ML, which limits its application and therefore potential to solve currently open questions. The current article aims to bridge this knowledge gap by supplying nutrition researchers with a resource to facilitate the use of ML in their research. ML is first explained and distinguished from existing solutions, with key examples of applications in the nutrition literature provided. Two case studies of domains in which ML is particularly applicable, precision nutrition and metabolomics, are then presented. Finally, a framework is outlined to guide interested researchers in integrating ML into their work. By acting as a resource to which researchers can refer, we hope to support the integration of ML in the field of nutrition to facilitate modern research.

A Machine-Learning Based Approach for Predicting Older Adults' Adherence to Technology-Based Cognitive Training

Adequate adherence is a necessary condition for success with any intervention, including for computerized cognitive training designed to mitigate age-related cognitive decline. Tailored prompting systems offer promise for promoting adherence and facilitating intervention success. However, developing adherence support systems capable of just-in-time adaptive reminders requires understanding the factors that predict adherence, particularly an imminent adherence lapse. In this study we built machine learning models to predict participants' adherence at different levels (overall and weekly) using data collected from a previous cognitive training intervention. We then built machine learning models to predict adherence using a variety of baseline measures (demographic, attitudinal, and cognitive ability variables), as well as deep learning models to predict the next week's adherence using variables derived from training interactions in the previous week. Logistic regression models with selected baseline variables were able to predict overall adherence with moderate accuracy (AUROC: 0.71), while some recurrent neural network models were able to predict weekly adherence with high accuracy (AUROC: 0.84-0.86) based on daily interactions. Analysis of the post hoc explanation of machine learning models revealed that general self-efficacy, objective memory measures, and technology self-efficacy were most predictive of participants' overall adherence, while time of training, sessions played, and game outcomes were predictive of the next week's adherence. Machine-learning based approaches revealed that both individual difference characteristics and previous intervention interactions provide useful information for predicting adherence, and these insights can provide initial clues as to who to target with adherence support strategies and when to provide support. This information will inform the development of a technology-based, just-in-time adherence support systems.

Dynamic Physical Activity Recommendation Delivered through a Mobile Fitness App: A Deep Learning Approach

Regular physical activity has a positive impact on our physical and mental health. Adhering to a fixed physical activity regimen is essential for good health and mental wellbeing. Today, fitness trackers and smartphone applications are used to promote physical activity. These applications use step counts recorded by accelerometers to estimate physical activity. In this research, we performed a two-level clustering on a dataset based on individuals’ physical and physiological features, as well as past daily activity patterns. The proposed model exploits the user data with partial or complete features. To include the user with partial features, we trained the proposed model with the data of users who possess exclusive features. Additionally, we classified the users into several clusters to produce more accurate results for the users. This enables the proposed system to provide data-driven and personalized activity planning recommendations every day. A personalized physical activity plan is generated on the basis of hourly patterns for users according to their adherence and past recommended activity plans. Customization of activity plans can be achieved according to the user’s historical activity habits and current activity objective, as well as the likelihood of sticking to the plan. The proposed physical activity recommendation system was evaluated in real time, and the results demonstrated the improved performance over existing baselines.

Data-driven Interpretable Policy Construction for Personalized Mobile Health

To promote healthy behaviors, many mobile health applications provide message-based interventions, such as tips, motivational messages, or suggestions for healthy activities. Ideally, the intervention policies should be carefully designed so that users obtain the benefits without being overwhelmed by overly frequent messages. As part of the HeartSteps physical-activity intervention, users receive messages intended to disrupt sedentary behavior. HeartSteps uses an algorithm to uniformly spread out the daily message budget over time, but does not attempt to maximize treatment effects. This limitation motivates constructing a policy to optimize the message delivery decisions for more effective treatments. Moreover, the learned policy needs to be interpretable to enable behavioral scientists to examine it and to inform future theorizing. We address this problem by learning an effective and interpretable policy that reduces sedentary behavior. We propose Optimal Policy Trees + (OPT+), an innovative batch off-policy learning method, that combines a personalized threshold learning and an extension of Optimal Policy Trees under a budget-constrained setting. We implement and test the method using data collected in HeartSteps V2/V3. Computational results demonstrate a significant reduction in sedentary behavior with a lower delivery budget. OPT+ produces a highly interpretable and stable output decision tree thus enabling theoretical insights to guide future research.

A vision for the future of wearable sensors in spine care and its challenges: narrative review

OBJECTIVE

This review aimed to: (I) provide a brief overview of some topical areas of current literature regarding applications of wearable sensors in the management of low back pain (LBP); (II) present a vision for a future comprehensive system that integrates wearable sensors to measure multiple parameters in the real world that contributes data to guide treatment selection (aided by artificial intelligence), uses wearables to aid treatment support, adherence and outcome monitoring, and interrogates the response of the individual patient to the prescribed treatment to guide future decision support for other individuals who present with LBP; and (III) consider the challenges that will need to be overcome to make such a system a reality.

BACKGROUND

Advances in wearable sensor technologies are opening new opportunities for the assessment and management of spinal conditions. Although evidence of improvements in outcomes for individuals with LBP from the use of sensors is limited, there is enormous future potential.

METHODS

Narrative review and literature synthesis.

CONCLUSIONS

Substantial research is underway by groups internationally to develop and test elements of this system, to design innovative new sensors that enable recording of new data in new ways, and to fuse data from multiple sources to provide rich information about an individual's experience of LBP. Together this system, incorporating data from wearable sensors has potential to personalise care in ways that were hitherto thought impossible. The potential is high but will require concerted effort to develop and ultimately will need to be feasible and more effective than existing management.

Development of a computer-aided text message platform for user engagement with a digital Diabetes Prevention Program: a case study

Digital Diabetes Prevention Programs (dDPP) are novel mHealth applications that leverage digital features such as tracking and messaging to support behavior change for diabetes prevention. Despite their clinical effectiveness, long-term engagement to these programs remains a challenge, creating barriers to adherence and meaningful health outcomes. We partnered with a dDPP vendor to develop a personalized automatic message system (PAMS) to promote user engagement to the dDPP platform by sending messages on behalf of their primary care provider. PAMS innovates by integrating into clinical workflows. User-centered design (UCD) methodologies in the form of iterative cycles of focus groups, user interviews, design workshops, and other core UCD activities were utilized to defined PAMS requirements. PAMS uses computational tools to deliver theory-based, automated, tailored messages, and content to support patient use of dDPP. In this article, we discuss the design and development of our system, including key requirements and features, the technical architecture and build, and preliminary user testing.

Tailored physical activity behavior change interventions: challenges and opportunities

A physically active lifestyle provides innumerable benefits; yet, few individuals are physically active enough to reap those benefits. Tailored physical activity interventions may address low rates of physical activity by offering individualized strategies that consider a person's characteristics, needs, preferences, and/or context, rather than the traditional one-size-fits-all approach. However, the tailoring methodology is in its nascency, and an understanding of how best to develop such interventions is needed. In this commentary, we identify future directions to enhance the impact of tailored interventions designed to increase physical activity participation. A multi-country collaborative was established to review the literature and discuss an agenda for future research. Two overarching research opportunities are suggested for improving the development of tailored, behavioral physical activity interventions: (a) optimize the engagement of diverse knowledge users in intervention co-design and (b) examine ethical considerations that may impact the use of technology to support tailored physical activity delivery. Specifically, there is a need for better reporting and evaluation of knowledge user involvement alongside targeting diversity in the inclusion of knowledge users. Furthermore, while technology boasts many opportunities to increase the scale and precision of interventions, examinations of how it impacts recipients' experiences of and participation in tailored interventions are needed to ensure the benefits of technology use outweigh the risks. A better understanding of these research areas will help ensure that the diverse needs of individuals are met, technology is appropriately used to support tailoring, and ultimately it improves the effectiveness of tailored physical activity interventions.

Smart wearable devices in cardiovascular care: where we are and how to move forward

Technological innovations reach deeply into our daily lives and an emerging trend supports the use of commercial smart wearable devices to manage health. In the era of remote, decentralized and increasingly personalized patient care, catalysed by the COVID-19 pandemic, the cardiovascular community must familiarize itself with the wearable technologies on the market and their wide range of clinical applications. In this Review, we highlight the basic engineering principles of common wearable sensors and where they can be error-prone. We also examine the role of these devices in the remote screening and diagnosis of common cardiovascular diseases, such as arrhythmias, and in the management of patients with established cardiovascular conditions, for example, heart failure. To date, challenges such as device accuracy, clinical validity, a lack of standardized regulatory policies and concerns for patient privacy are still hindering the widespread adoption of smart wearable technologies in clinical practice. We present several recommendations to navigate these challenges and propose a simple and practical 'ABCD' guide for clinicians, personalized to their specific practice needs, to accelerate the integration of these devices into the clinical workflow for optimal patient care.

The potential of artificial intelligence in enhancing adult weight loss: a scoping review

OBJECTIVE

To present an overview of how artificial intelligence (AI) could be used to regulate eating and dietary behaviours, exercise behaviours and weight loss.

DESIGN

A scoping review of global literature published from inception to 15 December 2020 was conducted according to Arksey and O'Malley's five-step framework. Eight databases (CINAHL, Cochrane-Central, Embase, IEEE Xplore, PsycINFO, PubMed, Scopus and Web of Science) were searched. Included studies were independently screened for eligibility by two reviewers with good interrater reliability (k = 0·96).

RESULTS

Sixty-six out of 5573 potential studies were included, representing more than 2031 participants. Three tenets of self-regulation were identified - self-monitoring (n 66, 100 %), optimisation of goal setting (n 10, 15·2 %) and self-control (n 10, 15·2 %). Articles were also categorised into three AI applications, namely machine perception (n 50), predictive analytics only (n 6) and real-time analytics with personalised micro-interventions (n 10). Machine perception focused on recognising food items, eating behaviours, physical activities and estimating energy balance. Predictive analytics focused on predicting weight loss, intervention adherence, dietary lapses and emotional eating. Studies on the last theme focused on evaluating AI-assisted weight management interventions that instantaneously collected behavioural data, optimised prediction models for behavioural lapse events and enhance behavioural self-control through adaptive and personalised nudges/prompts. Only six studies reported average weight losses (2·4-4·7 %) of which two were statistically significant.

CONCLUSION

The use of AI for weight loss is still undeveloped. Based on the current study findings, we proposed a framework on the applicability of AI for weight loss but cautioned its contingency upon engagement and contextualisation.

Agent-Oriented Goal Models in Developing Information Systems Supporting Physical Activity Among Adolescents: Literature Review and Expert Interviews

Background

Information and communication technologies (ICTs) are becoming increasingly popular in supporting the fight against low physical activity (PA) levels among adolescents. However, several ICT solutions lack evidence-based content. Therefore, there is a need to identify important features that have the potential to efficiently and consistently support the PA of adolescents using ICT solutions.

Objective

This study aims to create evidence-based models of requirements for ICT solutions supporting PA by combining scientific evidence from literature and health experts. In addition, we test the suitability of agent-oriented goal models in this type of modeling process.

Methods

A literature search of PubMed, Web of Science, and Scopus databases was conducted to identify evidence-based functional, quality, and emotional goals that have previously been proven to be relevant in supporting PAs among youth using ICT solutions. The identified goals were presented in the form of goal models. These models were used to collaborate with health experts to receive their input on the topic and suggestions for improvement. The initial goal models were improved based on the feedback from the experts.

Results

The results indicated that agent-oriented goal modeling is a suitable method for merging information from the literature and experts. One strength of agent-oriented goal models is that they present emotional requirements together with quality and functional requirements. Another strength is the possibility of presenting results from a literature review in a systematic manner and using them thereafter in the communication process with stakeholders. Agent-oriented goal models that were created were easy to understand for health experts without previous experience in requirements engineering, which facilitates and supports collaboration with nontechnical stakeholders.

Conclusions

The proposed agent-oriented goal models effectively merged information from scientific literature and experts in the field and presented early functional, quality, and emotional requirements in a holistic and coherent manner. We believe that the created models have high potential to help requirements engineers and developers to provide more efficient ICT solutions that support PA among adolescents in the future.

Differences in objectively measured daily physical activity patterns related to depressive symptoms in community dwelling women - mPED trial

Physical activity (PA) is an effective depression treatment. However, knowledge on how variation in day-to-day PA relates to depression in women is lacking. The purposes of this study were to 1) compare overall objectively measured baseline daily steps and duration of moderate to vigorous PA (MVPA) and 2) examine differences in steps and MVPA on days of the week between women aged 25–65 years, who were physically inactive, with high and low depressive symptoms, enrolled in a run-in period of the mobile phone based physical activity education (mPED) trial. The Center for Epidemiological Studies Depression Scale was used to categorize low/high depressive symptom groups. We used linear mixed-effects models to examine the associations between steps and MVPA and depression-status overall and by day of the week, adjusting for selected demographic variables and their interactions with day of the week. 274 women were included in the final analysis, of which 58 had high depressive symptoms. Overall physical activity levels did not differ. However, day of the week modified the associations of depression with MVPA (p = 0.015) and daily steps (p = 0.08). Women with high depression were characterized by reduced activity at the end of the week (Posthoc: Friday: 791 fewer steps, 95% CI: 73–1509, p = 0.03; 8.8 lower MVPA, 95% CI: 2.16–15.5, p = 0.0098) compared to women with low depression, who showed increased activity. Day of the week might be an important target for personalization of physical activity interventions. Future work should evaluate potential causes of daily activity alterations in depression in women.

A Clinical Prediction Model of Medication Adherence in Hypertensive Patients in a Chinese Community Hospital in Beijing

BACKGROUND

Hypertension remains a global health problem. Since, there is a significant positive correlation between antihypertensive medication adherence and blood pressure control, it is therefore of great importance to elucidate the determinants of adherence to antihypertensive medications among hypertensive patients.

METHODS

Hereby, we retrospectively analyzed the medical records of a hypertensive cohort recruited from a community hospital in Beijing, China, to investigate the factors affecting adherence to antihypertensive medications using decision trees. In addition, all data were assigned into a training set (75%) and testing set (25%) by the random number seed method to build and validate a compliance predictive model. We identified that how many times patients became nonadherent to antihypertensive medications in the year before the first prescription, types of antihypertensive drugs used in the year before the first prescription, body weight, smoking history, total number of hospital visits in the past year, total number of days of medication use in the year before enrollment, age, total number of outpatient follow-ups in the year after the first prescription, and concurrent diabetes greatly affected the compliance to antihypertensive medications.

RESULTS

The compliance predictive model we built showed a 0.78 sensitivity and 0.69 specificity for the prediction of the compliance to antihypertensive medications, with an area under the representative operating characteristics curve of 0.810.

CONCLUSIONS

Our data provide new insights into the improvements of the compliance to antihypertensive medications, which is beneficial for the management of hypertension, and the compliance predictive model may be used in community-based hypertension management.

Artificial Intelligence Chatbot Behavior Change Model for Designing Artificial Intelligence Chatbots to Promote Physical Activity and a Healthy Diet: Viewpoint

BACKGROUND

Chatbots empowered by artificial intelligence (AI) can increasingly engage in natural conversations and build relationships with users. Applying AI chatbots to lifestyle modification programs is one of the promising areas to develop cost-effective and feasible behavior interventions to promote physical activity and a healthy diet.

OBJECTIVE

The purposes of this perspective paper are to present a brief literature review of chatbot use in promoting physical activity and a healthy diet, describe the AI chatbot behavior change model our research team developed based on extensive interdisciplinary research, and discuss ethical principles and considerations.

METHODS

We conducted a preliminary search of studies reporting chatbots for improving physical activity and/or diet in four databases in July 2020. We summarized the characteristics of the chatbot studies and reviewed recent developments in human-AI communication research and innovations in natural language processing. Based on the identified gaps and opportunities, as well as our own clinical and research experience and findings, we propose an AI chatbot behavior change model.

RESULTS

Our review found a lack of understanding around theoretical guidance and practical recommendations on designing AI chatbots for lifestyle modification programs. The proposed AI chatbot behavior change model consists of the following four components to provide such guidance: (1) designing chatbot characteristics and understanding user background; (2) building relational capacity; (3) building persuasive conversational capacity; and (4) evaluating mechanisms and outcomes. The rationale and evidence supporting the design and evaluation choices for this model are presented in this paper.

CONCLUSIONS

As AI chatbots become increasingly integrated into various digital communications, our proposed theoretical framework is the first step to conceptualize the scope of utilization in health behavior change domains and to synthesize all possible dimensions of chatbot features to inform intervention design and evaluation. There is a need for more interdisciplinary work to continue developing AI techniques to improve a chatbot's relational and persuasive capacities to change physical activity and diet behaviors with strong ethical principles.

Statistical modeling can determine what factors are predictive of appropriate follow-up in patients presenting with incidental pulmonary nodules on CT

PURPOSE

To assess the performance of statistical modeling in predicting follow-up adherence of incidentally detected pulmonary nodules (IPN) on CT, based on patient variables (PV), radiology report related variables (RRRV) and physician-patient communication variables (PPCV).

METHODS

200 patients with IPN on CT were retrospectively identified and randomly selected. PV (age, gender, smoking status, ethnicity), RRRV (nodule size, patient context, whether follow-up recommendations were provided) and PPCV (whether referring physician documented IPN and ordered follow-up on the electronic medical record) were recorded. Primary outcome was whether patients received appropriate follow-up within +/- 1 month of the recommended time frame. Statistical methods included logistic regression and machine learning (K-nearest neighbors and support vector machine).

RESULTS

Adherence was low, with or without recommendations provided in the radiology report (23.4 %-27.4 %). Whether the referring physician ordered follow-up was the dominant predictor of adherence in all models. The following variables were statistically significant predictors of whether referring physician ordered follow-up: recommendations provided in the radiology report, smoking status, patient context and nodule size (FDR logworth of respectively 21.18, 11.66, 2.35, 1.63, p < 0.05). Prediction accuracy varied from 72 % (PV) to 93 % (PPCV, all variables).

CONCLUSION

PPCV are the most important predictors of adherence. Amongst all variables, patient context, smoking status, nodule size, and whether the radiologist provided follow-up recommendations in the report were all statistically significant predictors of patient follow-up adherence, supporting the utility of statistical modeling for analytics, quality assurance and optimization of outcomes related to IPN.