Literature Collection

INTRODUCTION: Early detection and intervention are crucial for reducing the impacts of depression and associated healthcare costs. Few studies have used electronic health records (EHR) and machine learning (ML) with a longitudinal design to predict depression onset. We developed and validated ML algorithms using EHR to identify patients at high risk for the onset of diagnosis-based major depressive disorder (MDD) in primary care settings. METHODS: Using a prognostic modeling approach with retrospective cohort study design, we identified patient visits in primary care settings for individuals aged ≥18 years from the Accelerating Data Value Across a National Community Health Center Network Clinical Research Network 2015-2021 data. We measured 267 features at six-month intervals starting six months prior to the first encounter. We developed algorithms using Least Absolute Shrinkage and Selection Operator (LASSO), random forest, and XGBoost with 10-fold cross validation. Using hold-out testing data, we measured prediction performance (e.g., C-statistics), stratified patients into decile risk subgroups, and assessed model biases. RESULTS: Among eligible 1,965,399 individuals (mean age = 43.52 ± 16.04 years; male = 35%; African American = 20%) with 4,985,280 person-periods, the MDD onset rate was 1% during the study period. XGBoost performed similarly to other models and had the fewest predictors, (C-statistic = 0.763, 95% CI = [0.760, 0.767]). XGBoost had a 66.78% sensitivity, 74.19% specificity, and 2.55% positive predictive value at the balanced threshold identified using Youdan Index. The top three risk decile subgroups captured ∼70% of MDD cases, without significant racial or sex biases. CONCLUSIONS: An ML algorithm using EHR data can effectively identify individuals at high risk of depression onset within the subsequent six months, without exacerbating racial or sex biases, providing a valuable tool for targeted early interventions.

BACKGROUND: Although mobile health apps integrated with Internet of Things-enabled devices are increasingly used to satisfy the growing needs for home-based older adult care resulting from rapid population aging, their effectiveness is constrained by 3 key challenges: a focus on specific functions rather than on holistic and integrated support, absence of a solid theoretical framework for development, and a lack of personalized, real-time feedback to address diverse care needs. To overcome these limitations, we developed a knowledge-based clinical decision support system using mobile health technology-an intelligent and integrated older adults care model (SMART system). OBJECTIVE: This study aims to systematically outline the development process and outcomes of a knowledge base and trigger rules for the SMART system. METHODS: Our study adopted a user-centered approach guided by the nursing process and intervention mapping (IM) framework. We first identified older adult care needs through semistructured, in-depth interviews. Guided by the nursing process and informed by guidance from the World Health Organization's Integrated Care for Older People and World Health Organization International Classification of Functioning, Disability, and Health, along with the North American Nursing Diagnosis Association-I nursing diagnosis, we then determined care problems along with their underlying causes and risk factors and diagnostic criteria. Building on these findings, we applied the first 3 steps of the intervention mapping framework to formulate corresponding long-term and short-term care objectives, select appropriate evidence-based interventions, and match practical implementation approaches, which were grounded in rigorous evidence derived from systematic literature reviews, clinical guidelines, and expert insights. We also developed a set of trigger rules to link abnormalities in older adults with corresponding care problems and interventions in the SMART knowledge base. RESULTS: The semistructured in-depth interviews identified 5 types of care needs-daily life care, health care, external support, social participation, and self-development-which formed the foundation of the SMART knowledge base. Based on this, we identified 138 care problems, each with associated causes and risk factors and diagnostic criteria. The objective matrix comprised 138 long-term and 195 short-term care objectives. Guided by 15 expert-defined selection criteria, we then selected 450 evidence-based interventions, each paired with at least 1 feasible and practical implementation approach. Additionally, we developed diagnostic rules to match the assessment data with relevant care problems and their causes and risk factors and intervention trigger rules to formulate personalized interventions based on individual characteristics, ensuring tailored care aligned with specific care objectives. CONCLUSIONS: This study outlines the development process and outcomes of the SMART knowledge base and trigger rules. The study methodology offers theoretical support for developing knowledge bases and trigger rules of similar clinical decision support systems for home-based older adult care.

There is a growing movement to integrate behavioral health specialists into primary care settings in order to better manage patients' health behaviors. Group interventions in healthcare settings can provide services to multiple individuals simultaneously; however, the participants' experiences taking part in these activities and the logistics of integrating them into clinical settings are largely under-studied. This article describes the development and implementation of a novel group intervention for health behavior change, The Kickstart Health Program, which integrates components of cognitive, behavioral, acceptance, and experiential therapies. Participant feasibility, acceptability, experiences, and treatment course were assessed. Acceptability among a small sample of attendees was high, and initial data on behavior change suggest there were benefits to patients who attended the program. Increases in mindfulness practice and decreases in exercise barriers from baseline to 10-week follow-up were detected as were improvements in overall perceived health and well-being. Participants expressed that the program was acceptable and successful at helping them reach their individual health goals; however, enrollment barriers negatively impacted the feasibility of the program in regard to attendance. Modification to the enrollment process such as embedding referrals into the electronic medical record, encouraging spouse or family co-enrollment, and peer coaching may address these barriers. The Kickstart Health Program has the potential to improve health behaviors and paves the way for unique studies of dissemination and implementation of efficacious behavioral health interventions into real-world healthcare settings.