Athreya, Arjun P

ABSTRACT Consumer‐grade wearables, including smartwatches, fitness trackers, and smart rings, are accessible tools for studying pharmacologic responses in real‐time, monitoring participant health in decentralized clinical trials, optimizing outcome measures, and delivering digital interventions that promote behavior change. However, incorporating consumer wearables into clinical studies is not as straightforward as simply purchasing devices and providing them to participants. It requires engaging with a complex engineering ecosystem with multiple investigator decisions on device selection, associated data management practices, and patient/participant engagement and technical support. This tutorial, based on experience running large clinical trials using consumer wearables, outlines the wearables market landscape across varying form factors and manufacturers as well as the data management software products supporting wearable use. It delineates the major decision processes involved in designing studies that incorporate wearables and presents a structured framework to guide the selection of devices and supporting data tools according to study objectives.

ABSTRACT Adherence in digital health studies with extended observation periods (≥ 12 months) is limited, and participant retention considerably reduces with time. The US Food and Drug Administration has issued guidelines for improving participant engagement, adherence, and diversity in digital health studies combined with decentralized procedures. A decentralized digital health study on well‐being was designed with protocolized procedures to study the feasibility of participant engagement and technology support to facilitate adherence (wearing the smartwatch ≥ 70% of time) sustained over a 12‐month period. At the end of the study, participants were asked about their ease of participation and free‐response questions about how wearing the smartwatches impacted their physical wellness. An inductive thematic analysis (ITA) was performed to assess themes of those responses and association with adherence. A total of 298 participants were recruited between 2022 and 2023 ( n = 129 in Cohort A in October 22, n = 169 in Cohort B in April 23), with 23% non‐white participants accrued. Among the 298 participants accrued, 273 (92% of accrued participants) completed the 12‐month study with an average overall adherence of 77.4% (SD = 32.64) wear‐time across 12 months. Median adherence of participants whose responses exemplified an ITA theme encompassing perceived behavior changes in sleep and physical activity was higher than those who did not have a response exemplifying that theme. Conversely, those expressing perceived discomfort or intrusiveness of the smartwatch had a statistically lower adherence. These results highlight the crucial roles of technology support and robust engagement efforts to enable sustained adherence over extended follow‐up periods in decentralized digital health studies.

Abstract Pharmacogenomic (PGx) biomarkers integrated using machine learning can be embedded within the electronic health record (EHR) to provide clinicians with individualized predictions of drug treatment outcomes. Currently, however, drug alerts in the EHR are largely generic (not patient‐specific) and contribute to increased clinician stress and burnout. Improving the usability of PGx alerts is an urgent need. Therefore, this work aimed to identify principles for optimal PGx alert design through a health‐system‐wide, mixed‐methods study. Clinicians representing multiple practices and care settings ( N = 1062) in urban, rural, and underserved regions were invited to complete an electronic survey comparing the usability of three drug alerts for citalopram, as a case study. Alert 1 contained a generic warning of pharmacogenomic effects on citalopram metabolism. Alerts 2 and 3 provided patient‐specific predictions of citalopram efficacy with varying depth of information. Primary outcomes included the System's Usability Scale score (0–100 points) of each alert, the perceived impact of each alert on stress and decision‐making, and clinicians' suggestions for alert improvement. Secondary outcomes included the assessment of alert preference by clinician age, practice type, and geographic setting. Qualitative information was captured to provide context to quantitative information. The final cohort comprised 305 geographically and clinically diverse clinicians. A simplified, individualized alert (Alert 2) was perceived as beneficial for decision‐making and stress compared with a more detailed version (Alert 3) and the generic alert (Alert 1) regardless of age, practice type, or geographic setting. Findings emphasize the need for clinician‐guided design of PGx alerts in the era of digital medicine.

AbstractObjectivesInterpatient variability in bipolar I depression (BP‐D) symptoms challenges the ability to predict pharmacotherapeutic outcomes. A machine learning workflow was developed to predict remission after 8 weeks of pharmacotherapy (total score of ≤8 on the Montgomery Åsberg Depression Rating Scale [MADRS]).MethodsSupervised machine learning models were trained on data from BP‐D patients treated with olanzapine (N = 168) and were externally validated on patients treated with olanzapine/fluoxetine combination (OFC; N = 131) and lamotrigine (LTG; N = 126). Top predictors were used to develop a prognosis rule informing how many symptoms should change and by how much within 4 weeks to increase the odds of achieving remission.ResultsAn AUC of 0.76 (NIR:0.59; p = 0.17) was established to predict remission in olanzapine‐treated subjects. These trained models achieved AUCs of 0.70 with OFC (NIR:0.52; p < 0.03) and 0.73 with LTG (NIR:0.52; p < 0.003), demonstrating external replication of prediction performance. Week‐4 changes in four MADRS symptoms (reported sadness, reduced sleep, reduced appetite, and concentration difficulties) were top predictors of remission. Across all pharmacotherapies, three or more of these symptoms needed to improve by ≥2 points at Week‐4 to have a 65% chance of achieving remission at 8 weeks (OR: 3.74, 95% CI: 2.45–5.76; p < 9.3E‐11).ConclusionMachine learning strategies achieved cross‐trial and cross‐drug replication in predicting remission after 8 weeks of pharmacotherapy for BP‐D. Interpretable prognoses rules required only a limited number of depressive symptoms, providing a promising foundation for developing simple quantitative decision aids that may, in the future, serve as companions to clinical judgment at the point of care.