Antoinette Schoenthaler

Background: Multilevel interventions targeting patients, providers, and health systems improve hypertension outcomes, but little is known about how well they are implemented in resource-limited settings. This study evaluated the implementation fidelity of ALTA, a multilevel approach for hypertension management, supported by practice facilitation, across six primary care practices in a Federally Qualified Health Center (FQHC). Methods: ALTA included five components delivered by FQHC staff. At the practice level, staff: (1) identified medication non-adherence in patients with uncontrolled hypertension and (2) referred them to a centralized nurse-led virtual team with a home blood pressure (BP) monitor. The nurse-led team: (3) delivered monthly health coaching based on home BP readings; (4) completed documentation in the electronic health record (EHR); and (5) monitored patients' BP and goals. We used a multi-method approach to assess implementation fidelity (adherence, dose, quality, responsiveness, differentiation) guided by Proctor's Implementation Outcome's Framework. Data sources included a structured EHR-embedded tool, narrative reports, and interviews over 12 months. ALTA adaptations were tracked using the Framework for Reporting Adaptations and Modifications–Expanded. Results: From 2022–2024, 124 staff across 6 FQHC sites implemented ALTA. Implementation adherence varied across sites. The identify component ranged from 49.0% to 57.9% across sites, while the refer component showed greater variability, ranging from 50.0% to 84.4%. Enrollment into health coaching also varied substantially, ranging from 20.8% to 88.5% across sites. The centralized nurse-led team delivered coaching to 91.4% of enrolled patients, completed all documentation, and conducted monitoring for 94.5%. Median implementation dose (patient exposure) was six coaching visits over 5.5 months, and 2.1 BP readings per week. Implementation quality was high, with comprehensive documentation of coaching visits. Six fidelity-consistent adaptations were made to improve feasibility, including simplifying the adherence screener and adding asynchronous training. Interviews ( n = 46) highlighted the need for team support, ongoing practice facilitation, and feedback on patients' progress as factors affecting fidelity. Conclusions: Implementation fidelity varied across the ALTA components and participating sites, with the greatest variability observed in delivery of the site-level components of identify and refer. Centralized components delivered by the nurse-led team (coach, document, monitor) were delivered consistently with high fidelity. Contextual factors like staffing and patient engagement shaped implementation success. Clinical Trial Registration: ClinicalTrials.gov , NCT03713515.

Health-related social needs (HRSN) screening has attracted growing investment in recent years. Despite rising HRSN screening rates, resource referral and receipt remain low. As a result, patient self-navigation models – in which patients independently identify, contact, access, and utilize resources to address HRSN with the support of tailored written resource navigation lists – are emerging, though they remain understudied, especially outside of pediatric care. In response, this study triangulated qualitative data from three informant groups involved in HRSN navigation in a large urban ambulatory care network, to improve understanding of determinants which shape HRSN self-navigation. For this qualitative study, semi-structured interviews were conducted from October 2023 - August 2024 with community health workers (CHWs; n = 8), physicians (n = 12), and community-based organization partners (n = 12) involved in HRSN screening, referral, or service delivery. All interviews were led by a trained qualitative researcher in English and were audio-recorded and transcribed verbatim. A team of analysts thematically analyzed the data, deductively grounding codebook development and theme generation in the Consolidated Framework for Implementation Research 2.0. All groups agreed that designing self-navigation tools is important for expanding reach, while acknowledging that doing so may only support a subset of patients. Patient-level barriers to HRSN self-navigation included fear of documentation status exposure, lack of time, limited digital literacy, and language discordance, while institutional barriers included physicians’ lack of time, training, and resources for HRSN navigation, as well as community health worker and community-based organization partners’ lack of capacity. Poor self-navigation experiences may produce unintended consequences (spread of misinformation; losing patients to navigation barriers), which hinder buy-in and resource connection. In response, all groups recommended offering optional, supplemental hands-on support for patients who need it to improve self-navigation effectiveness and equity. Following the push for enhanced HRSN screening, the demand for resource navigation far exceeds capacity. In response, participants expressed interest in designing and testing patient self-navigation tools, while also recognizing limitations and potential unintended consequences. Future research is needed to design and test equity-centered self-navigation tools for large urban health system contexts. The online version contains supplementary material available at 10.1186/s12913-026-14351-2.

Background: Wearable accelerometers, which continuously record physical activity metrics, are commonly used in mobile health–enabled cardiac rehabilitation (mHealth-CR). The association between adherence to accelerometer use during mHealth-CR and improvement in clinical outcomes, such as functional capacity, is understudied. The emergence of artificial intelligence (AI) technology provides novel opportunities to investigate accelerometry use patterns in relation to mHealth-CR outcomes. Objective: In this study, we sought to use an AI clustering framework to identify distinct behavioral phenotypes of adherence to accelerometer use. We then aimed to quantify the association of these adherence phenotypes with functional capacity improvements in older adults undergoing mHealth-CR. Methods: We analyzed data from the RESILIENT (Rehabilitation at Home Using Mobile Health in Older Adults After Hospitalization for Ischemic Heart Disease) trial, the largest randomized clinical study to date comparing mHealth-CR versus usual care in older adults (aged ≥65 years). Intervention arm participants were instructed to wear a Fitbit accelerometer for the 3-month study duration. Adherence to accelerometer use was quantified as overall adherence (percentage of days worn) via k-means clustering AI-derived measures and compared with changes in 6-minute walk distance (6-MWD), adjusted for demographic and clinical covariates. Results: Among 271 participants with a mean age of 71 years (SD 8), of whom 198 (73%) were male, accelerometers were worn for an average of 76 days (95% confidence limits 73,78) over 3 months. Adjusted analyses showed a weak association between days of wear and improvement in 6-MWD, with every 30 additional days associated with an 11-meter improvement (P=.08). Our k-means clustering framework identified adherence phenotypes at two resolutions: low resolution (k=2 clusters) and high resolution (k=8 clusters). The consistently high adherence cluster trended toward a 24.6-meter improvement in 6-MWD compared to the low and declining adherence clusters (n=39; 95% CI 0.7-49.9; P=.06). The 8-cluster phenotyping revealed a richer set of adherence patterns, with the consistently high adherence cluster in this analysis having a 38.5-meter (95% CI 2.2-74.7; P=.04) improvement in 6-MWD than the low adherence cluster, as well as greater average daily steps over the 3-month intervention (mean 7518, SD 3415 vs mean 4800, SD 2920 steps; P=.008). Conclusions: A time-series AI clustering framework identified a range of behavioral phenotypes representing different degrees of adherence to accelerometer use. Regression analysis identified a weak association between the higher adherence phenotype and functional capacity improvement in older adults undergoing mHealth-CR. Our AI-derived accelerometry adherence phenotypes may offer a new approach to tailor mHealth-CR regimens to individual patients, potentially leading to better outcomes in this high-risk population. Trial Registration: ClinicalTrials.gov NCT03978130; https://clinicaltrials.gov/study/NCT03978130 International Registered Report Identifier (IRRID): RR2-10.2196/32163

The integration of generative AI (GenAI) in patient communication presents benefits and challenges. This retrospective observational study analyzed EHR audit logs to assess how 75 healthcare professionals (HCPs) utilized AI generated drafts for patient messages from October 2023 to August 2024 at a large health system in New York City. Overall utilization was low (19.4%), though prompt refinements improved usage (from 12% to 20%), particularly among physicians. GenAI drafts were generated for all messages, including 80% that received no response, adding to the review burden and potentially undermining efficiency. Text analysis showed HCPs preferred concise, information-rich drafts, with role-based differences: physicians favored shorter drafts, while clinical support staff preferred more empathetic responses. AI generated drafts reduced message turnaround time by 6.76% despite a marginal increase in required steps (InBasket actions). These findings highlight the need for targeted GenAI deployment strategies, better aligned with clinician workflows and optimized draft generation for improved efficiency.

BACKGROUND Wearable accelerometers, which continuously record physical activity metrics, are commonly used in mobile health–enabled cardiac rehabilitation (mHealth-CR). The association between adherence to accelerometer use during mHealth-CR and improvement in clinical outcomes, such as functional capacity, is understudied. The emergence of artificial intelligence (AI) technology provides novel opportunities to investigate accelerometry use patterns in relation to mHealth-CR outcomes. OBJECTIVE In this study, we sought to use an AI clustering framework to identify distinct behavioral phenotypes of adherence to accelerometer use. We then aimed to quantify the association of these adherence phenotypes with functional capacity improvements in older adults undergoing mHealth-CR. METHODS We analyzed data from the RESILIENT (Rehabilitation at Home Using Mobile Health in Older Adults After Hospitalization for Ischemic Heart Disease) trial, the largest randomized clinical study to date comparing mHealth-CR versus usual care in older adults (aged ≥65 years). Intervention arm participants were instructed to wear a Fitbit accelerometer for the 3-month study duration. Adherence to accelerometer use was quantified as overall adherence (percentage of days worn) via k-means clustering AI-derived measures and compared with changes in 6-minute walk distance (6-MWD), adjusted for demographic and clinical covariates. RESULTS Among 271 participants with a mean age of 71 years (SD 8), of whom 198 (73%) were male, accelerometers were worn for an average of 76 days (95% confidence limits 73,78) over 3 months. Adjusted analyses showed a weak association between days of wear and improvement in 6-MWD, with every 30 additional days associated with an 11-meter improvement (P=.08). Our k-means clustering framework identified adherence phenotypes at two resolutions: low resolution (k=2 clusters) and high resolution (k=8 clusters). The consistently high adherence cluster trended toward a 24.6-meter improvement in 6-MWD compared to the low and declining adherence clusters (n=39; 95% CI 0.7-49.9; P=.06). The 8-cluster phenotyping revealed a richer set of adherence patterns, with the consistently high adherence cluster in this analysis having a 38.5-meter (95% CI 2.2-74.7; P=.04) improvement in 6-MWD than the low adherence cluster, as well as greater average daily steps over the 3-month intervention (mean 7518, SD 3415 vs mean 4800, SD 2920 steps; P=.008). CONCLUSIONS A time-series AI clustering framework identified a range of behavioral phenotypes representing different degrees of adherence to accelerometer use. Regression analysis identified a weak association between the higher adherence phenotype and functional capacity improvement in older adults undergoing mHealth-CR. Our AI-derived accelerometry adherence phenotypes may offer a new approach to tailor mHealth-CR regimens to individual patients, potentially leading to better outcomes in this high-risk population. CLINICALTRIAL ClinicalTrials.gov NCT03978130; https://clinicaltrials.gov/study/NCT03978130 INTERNATIONAL REGISTERED REPORT RR2-10.2196/32163

Background/Objectives: The aim of this secondary analysis is to determine the baseline characteristics that are associated with a higher likelihood of weight-loss success in a personalized nutrition intervention. Methods: Data were analyzed in adults with abnormal glucose metabolism and obesity from a 6-month behavioral counseling randomized clinical trial. Participants were randomized to two calorie-restricted diets: a low-fat diet (Standardized) or a personalized nutrition diet leveraging a machine learning algorithm (Personalized). The gradient boosting machine method was used to determine the baseline variables (i.e., age, weight-loss self-efficacy) that predicted successful weight loss (≥5%) at 6 months in each study arm separately, using repeated five-fold cross-validation with 100 repetitions. Results: A total of 155 participants (Personalized: n = 84 vs. Standardized: n = 71) contributed data (mean [standard deviation]: age, 59 [10] y; 66.5% female; 56.1% White; body mass index (BMI), 33.4 [4.6] kg/m²). In both arms, higher baseline self-efficacy for weight loss was a predictor of weight-loss success. Participants with a higher BMI (p < 0.0001) in the Standardized arm and those who were older (p < 0.0001) in the Personalized arm were more likely to achieve successful weight loss. Conclusions: Future weight-loss interventions may consider providing tailored behavioral support for individuals based on weight-loss self-efficacy, BMI, and age.

Background: Few studies assess the association between county‐level incarceration rates and stroke death, and none test mechanisms. We examine the link between county imprisonment rates and stroke death, testing pathways and identifying racial disparities. Methods: In a cross‐sectional design, we regressed stroke death onto imprisonment rates, adjusting for poverty, racial composition, education, unemployment, insurance, and violent crime. Using bootstrap methodology, we tested mediation through sports/recreational facilities and food environment, mental health provider and primary care physician (PCP) access, and community mental distress. Data spanned 4 years (2016–2019) and included 2260 counties. Results: Adjusted models indicated a 0.08 (95% CI, 0.05–0.10) increase in stroke deaths for every 1‐unit change in imprisonment rate. This association was mediated by food environment (indirect effect, 0.006 [95% CI, 0.000–0.014]), primary care physician access (indirect effect, 0.002 [95% CI, 0.000–0.006]), and mental distress (indirect effect, 0.014 [95% CI, 0.007–0.022]). Counties in the top versus bottom quintile of imprisonment rates had 86.26% larger Black populations and 23.46% smaller White populations. Counties in the top versus bottom quintile of stroke death had 88.94% larger Black populations and 16.19% smaller White populations. Conclusions: Our results complement evidence that living in high‐jail‐incarceration counties contributes to stroke death and associated racial disparities. We provide new evidence on prison incarceration rates and the pathways underpinning this association. County‐level imprisonment rates and the identified mechanisms represent avenues for further research into how stroke death and disparities may be mitigated.

Remote patient monitoring (RPM) has been shown to support adults with treated but uncontrolled hypertension (HTN) outside the clinic setting, yielding significant benefits in the treatment and control of blood pressure (BP). Despite its proven efficacy and recommendation as guideline-concordant care, adoption of RPM is suboptimal, particularly among marginalized populations, who face structural barriers to HTN control. A barrier to equitable adoption among marginalized populations is the lack of digital inclusivity in the design and deployment of RPM. Digitally inclusive tools consider factors such as affordability, access, digital literacy, and skills. To address this challenge, the authors describe a digitally inclusive model of RPM for HTN management within the Family Health Centers (FHCs) at NYU Langone, a federally qualified health center (FQHC) that serves more than 110,000 patients each year. The model uses protocols from the Target: BP initiative in combination with team-based care and digitally inclusive strategies to improve HTN control. Specifically, care teams work collaboratively to identify patients with uncontrolled HTN and order RPM using electronic health record–embedded clinical decision support; provide patients with free home BP monitors and training in accurate BP measurement; deliver language-concordant health coaching and optimize the antihypertensive regimen via a virtual high-risk clinic (VHRC); and monitor patient progress through shared communications. Patients also receive support from community health workers (CHWs) to address digital barriers and unmet social needs. The authors present utilization and preliminary outcome data of their model, involving 429 patients who were enrolled in RPM and the VHRC across five FHC practices between January 1, 2022, and December 31, 2023. Enrolled patients attended a mean of 4.9 (standard deviation [SD]: 0.5) visits with a nurse practitioner for medication adjustments and counseling; 5.7 (SD: 0.5) health coaching visits with a nurse; and 1 visit (SD: 0.2) with a CHW for digital and social needs over a mean of 5.7 months (SD: 0.8). Enrolled patients sent a mean number of 65 BP readings (SD: 96.4) over their period of participation. On average, enrolled patients exhibited a −13.5/−8.0 mmHg reduction from their enrollment date to the date that they were discharged from the VHRC (approximately 5.7 months). This is in comparison to a −0.5/+0.6 mmHg change in mean BP exhibited by patients with uncontrolled HTN not enrolled in the Advancing Long-term Improvements in Hypertension Outcomes through a Team-based Care Approach (ALTA) program and receiving care at the practices during the same period (n=2,843). Across the practices, BP control had also increased from the pre-ALTA baseline period (January 1, 2021, to December 31, 2021) of 68.44%–82.99%, by the end of December 31, 2023, among all patients with uncontrolled HTN. While the implementation of this digitally inclusive RPM model has shown success in a large FQHC that cares for a diverse population of patients, there remain digital inequity barriers that must be addressed at the policy level to ensure this efficacious approach reaches all patients.