Amanda L. Zaleski

Background: Readmission avoidance initiatives have been a priority for the Centers for Medicare & Medicaid Services for over a decade; however, interventions are often high-intensity, costly, and resource-intensive, and therefore, rarely scalable or sustainable. Large national payers are in a unique position to leverage data to identify members in real-time who are at high risk of readmission to prioritize the scaled delivery of tailored behavior change techniques to provide an educational intervention to modify health behaviors. Objective: This study aims to examine the impact of an informatics-driven, multichannel educational messaging campaign implemented to decrease 30- and 90-day acute inpatient readmissions and emergency department (ED) visits among Medicare Advantage members of a large national payer. Methods: A quality improvement initiative was designed and implemented to provide an evidence-based outreach campaign using human-centered design and behavior change principles to deliver multiple intervention functions, including timely, contextual, and relevant delivery of education, enablement, and persuasion, to reinforce health-promoting behaviors related to planned or unplanned inpatient admissions. Outcomes, including 30- and 90-day acute inpatient readmissions and ED visits, were retrospectively evaluated from Medicare Advantage members enrolled in a large national health plan residing across the United States between May 2020 and July 2022. Leveraging utilization management data, rules-based logic identified members (N=368,393) with a planned acute inpatient procedure (ie, preadmission) or discharged from an acute hospital stay (ie, postdischarge) within 15 days. Members were sequentially assigned to a standard (N=141,223) or an enhanced (N=227,470) messaging group, whereby the standard group received usual outreach and the enhanced group received an educational intervention via a messaging campaign deployed through multiple low-intensity communication channels (eg, text message, email, direct mail) in addition to standard outreach. Results: Members who received enhanced outreach had fewer relative 30-day acute inpatient readmissions (−4.1%, 95% CI −5.5% to −2.7%; P<.001) and ED visits (−3.4%, 95% CI −5.0% to −1.7%; P<.001) compared with members receiving standard outreach. Similarly, these findings persisted for relative 90-day outcomes such that members receiving enhanced outreach experienced fewer acute inpatient readmissions (−5.4%, 95% CI −6.5% to −4.3%; P<.001) and ED visits (−3.8%, 95% CI −5.0% to −2.5%; P<.001) compared with members receiving standard outreach messaging. Conclusions: Behavior change techniques deployed via educational interventions as low-intensity multi-channel outreach is an effective strategy to reduce avoidable 30- and 90-day inpatient readmissions and ED visits in recently discharged Medicare Advantage members (primarily >65 years).

BACKGROUND Readmission avoidance initiatives have been a priority for the Centers for Medicare & Medicaid Services for over a decade; however, interventions are often high-intensity, costly, and resource-intensive, and therefore, rarely scalable or sustainable. Large national payors are in a unique position to leverage data to identify members in real-time who are at high-risk of readmission to prioritize the scaled delivery of tailored behavior change techniques to provide an educational intervention to modify health behaviors. OBJECTIVE Readmission avoidance initiatives have been a priority for the Centers for Medicare & Medicaid Services for over a decade; however, interventions are often high-intensity, costly, and resource-intensive, and therefore, rarely scalable or sustainable. Large national payors are in a unique position to leverage data to identify members in real-time who are at high-risk of readmission to prioritize the scaled delivery of tailored behavior change techniques to provide an educational intervention to modify health behaviors. METHODS A quality improvement initiative was designed and implemented to provide an evidence-based outreach campaign using behavior change techniques to deliver multiple intervention functions, including education, enablement, and persuasion, to modify health behaviors related to planned or unplanned inpatient admissions. Outcomes including 30- and 90-day acute inpatient readmissions and ED visits were retrospectively evaluated from Medicare Advantage members enrolled in a large national health plan residing across the United States between May 2020 and July 2022. Leveraging utilization management data, rules-based logic identified members (N=368,393) with a planned acute inpatient procedure (i.e., pre-admission) or discharged from an acute hospital stay (i.e., post-discharge) within 15 days. Members were randomly assigned to a standard (N=141,223) or an enhanced (N=227,470) messaging group, whereby the standard group received usual outreach and the enhanced group received an educational intervention via a messaging campaign deployed through multiple low-intensity communication channels (e.g., text message, email, direct mail) in addition to standard outreach. RESULTS Members who received enhanced outreach had fewer relative 30-day acute inpatient readmissions (-4.1%, 95% CI: -5.5 to -2.7 %; P<.001) and ED visits (-3.4%, 95% CI: -5.0 to -1.7%; P<.001) compared to members receiving standard outreach. Similarly, these findings persisted for relative 90-day outcomes such that members receiving enhanced outreach experienced fewer acute inpatient readmissions (-5.4%, 95% CI: -6.5 to -4.3%; P<.001) and ED visits (-3.8%, 95% CI: -5.0 to -2.5%; P<.001) compared to members receiving standard outreach messaging. CONCLUSIONS Behavior change techniques deployed via educational interventions as low-intensity multi-channel outreach is an effective strategy to reduce avoidable 30- and 90-day inpatient readmissions and ED visits in recently discharged Medicare Advantage members (primarily >65 years).

Background: Mobile health technology's impact on cardiovascular risk factor control is not fully understood. This study evaluates the association between interaction with a mobile health application and change in cardiovascular risk factors. Methods and Results: Participants with hypertension with or without dyslipidemia enrolled in a workplace‐deployed mobile health application‐based cardiovascular risk self‐management program between January 2018 and December 2022. Retrospective evaluation explored the influence of application engagement on change in blood pressure (BP), total cholesterol (TC), low‐density lipoprotein cholesterol (LDL‐C), and weight. Multiple regression analyses examined the influence of guideline‐based, nonpharmacological lifestyle‐based digital coaching on outcomes adjusting for confounders. Of 102 475 participants, 49.1% were women. Median age was 53 (interquartile range, 43–61) years, BP was 134 (interquartile range, 124–144)/84 (interquartile range, 78–91) mm Hg, TC was 183 (interquartile range, 155–212) mg/dL, LDL‐C was 106 (82–131) mg/dL, and body mass index was 30 (26–35) kg/m ² . At 2 years, participants with baseline systolic BP ≥140 mm Hg reduced systolic BP by 18.6 (SEM, 0.3) mm Hg. At follow up, participants with baseline TC ≥240 mg/dL reduced TC by 65.7 (SEM, 4.6) mg/dL, participants with baseline LDL‐C≥160 mg/dL reduced LDL‐C by 66.6 (SEM, 6.2) mg/dL, and participants with baseline body mass index ≥30 kg/m ² lost 12.0 (SEM, 0.3) pounds, or 5.1% of body weight. Interaction with digital coaching was associated with greater reduction in all outcomes. Conclusions: A mobile health application‐based cardiovascular risk self‐management program was associated with favorable reductions in BP, TC, LDL‐C, and weight, highlighting the potential use of this technology in comprehensive cardiovascular risk factor control.

Postexercise hypotension (PEH), or the immediate decrease in blood pressure (BP) lasting for 24 h following an exercise bout, is well-established; however, the influence of exercise training on PEH dynamics is unknown. This study investigated the reliability and time course of change of PEH during exercise training among adults with hypertension. PEH responders (n = 10) underwent 12 weeks of aerobic exercise training, 40 min/session at moderate-to-vigorous intensity for 3 d/weeks. Self-measured BP was used to calculate PEH before and for 10 min after each session. The intraclass correlation coefficient (ICC) and Akaike Information Criterion (AIC) determined PEH reliability and goodness-of-fit for each week, respectively. Participants were obese (30.6 ± 4.3 kg∙m⁻²), middle-aged (57.2 ± 10.5 years), and mostly men (60%) with stage I hypertension (136.5 ± 12.1/83.4 ± 6.7 mmHg). Exercise training adherence was 90.6 ± 11.8% with 32.6 ± 4.2 sessions completed. PEH occurred in 89.7 ± 8.3% of these sessions with BP reductions of 9.3 ± 13.1/3.2 ± 6.8 mmHg. PEH reliability was moderate (ICC ~0.6). AIC analysis revealed a stabilization of maximal systolic and diastolic BP reductions at 3 weeks and 10 weeks, respectively. PEH persisted throughout exercise training at clinically meaningful levels, suggesting that the antihypertensive effects of exercise training may be largely due to PEH. Further studies in larger samples and under ambulatory conditions are needed to confirm these novel findings.

Climate change is an alarming global threat to individual and public health. This commentary addresses how the field of climate health informatics can spearhead research and innovation and guide climate adaptation and mitigation efforts.

Background: Regular physical activity is critical for health and disease prevention. Yet, health care providers and patients face barriers to implement evidence-based lifestyle recommendations. The potential to augment care with the increased availability of artificial intelligence (AI) technologies is limitless; however, the suitability of AI-generated exercise recommendations has yet to be explored. Objective: The purpose of this study was to assess the comprehensiveness, accuracy, and readability of individualized exercise recommendations generated by a novel AI chatbot. Methods: A coding scheme was developed to score AI-generated exercise recommendations across ten categories informed by gold-standard exercise recommendations, including (1) health condition–specific benefits of exercise, (2) exercise preparticipation health screening, (3) frequency, (4) intensity, (5) time, (6) type, (7) volume, (8) progression, (9) special considerations, and (10) references to the primary literature. The AI chatbot was prompted to provide individualized exercise recommendations for 26 clinical populations using an open-source application programming interface. Two independent reviewers coded AI-generated content for each category and calculated comprehensiveness (%) and factual accuracy (%) on a scale of 0%-100%. Readability was assessed using the Flesch-Kincaid formula. Qualitative analysis identified and categorized themes from AI-generated output. Results: AI-generated exercise recommendations were 41.2% (107/260) comprehensive and 90.7% (146/161) accurate, with the majority (8/15, 53%) of inaccuracy related to the need for exercise preparticipation medical clearance. Average readability level of AI-generated exercise recommendations was at the college level (mean 13.7, SD 1.7), with an average Flesch reading ease score of 31.1 (SD 7.7). Several recurring themes and observations of AI-generated output included concern for liability and safety, preference for aerobic exercise, and potential bias and direct discrimination against certain age-based populations and individuals with disabilities. Conclusions: There were notable gaps in the comprehensiveness, accuracy, and readability of AI-generated exercise recommendations. Exercise and health care professionals should be aware of these limitations when using and endorsing AI-based technologies as a tool to support lifestyle change involving exercise.

Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Regular physical activity is critical for health and disease prevention. Yet, health care providers and patients face barriers to implement evidence-based lifestyle recommendations. The potential to augment care with the increased availability of artificial intelligence (AI) technologies is limitless; however, the suitability of AI-generated exercise recommendations has yet to be explored. Objective: The purpose of this study was to assess the comprehensiveness, accuracy, and readability of individualized exercise recommendations generated by a novel AI chatbot. Methods: A coding scheme was developed to score AI-generated exercise recommendations across ten categories informed by gold-standard exercise recommendations, including (1) health condition–specific benefits of exercise, (2) exercise preparticipation health screening, (3) frequency, (4) intensity, (5) time, (6) type, (7) volume, (8) progression, (9) special considerations, and (10) references to the primary literature. The AI chatbot was prompted to provide individualized exercise recommendations for 26 clinical populations using an open-source application programming interface. Two independent reviewers coded AI-generated content for each category and calculated comprehensiveness (%) and factual accuracy (%) on a scale of 0%-100%. Readability was assessed using the Flesch-Kincaid formula. Qualitative analysis identified and categorized themes from AI-generated output. Results: AI-generated exercise recommendations were 41.2% (107/260) comprehensive and 90.7% (146/161) accurate, with the majority (8/15, 53%) of inaccuracy related to the need for exercise preparticipation medical clearance. Average readability level of AI-generated exercise recommendations was at the college level (mean 13.7, SD 1.7), with an average Flesch reading ease score of 31.1 (SD 7.7). Several recurring themes and observations of AI-generated output included concern for liability and safety, preference for aerobic exercise, and potential bias and direct discrimination against certain age-based populations and individuals with disabilities. Conclusions: There were notable gaps in the comprehensiveness, accuracy, and readability of AI-generated exercise recommendations. Exercise and health care professionals should be aware of these limitations when using and endorsing AI-based technologies as a tool to support lifestyle change involving exercise.

Background: Mobile health (mHealth) technology holds great promise as an easily accessible and effective solution to improve population health at scale. Despite the abundance of mHealth offerings, only a minority are grounded in evidence-based practice, whereas even fewer have line of sight into population-level health care spending, limiting the clinical utility of such tools. Objective: This study aimed to explore the influence of a health plan–sponsored, wearable-based, and reward-driven digital health intervention (DHI) on health care spending over 1 year. The DHI was delivered through a smartphone-based mHealth app available only to members of a large commercial health plan and leveraged a combination of behavioral economics, user-generated sensor data from the connected wearable device, and claims history to create personalized, evidence-based recommendations for each user. Methods: This study deployed a propensity score–matched, 2-group, and pre-post observational design. Adults (≥18 years of age) enrolled in a large, national commercial health plan and self-enlisted in the DHI for ≥7 months were allocated to the intervention group (n=56,816). Members who were eligible for the DHI but did not enlist were propensity score–matched to the comparison group (n=56,816). Average (and relative change from baseline) medical and pharmacy spending per user per month was computed for each member of the intervention and comparison groups during the pre- (ie, 12 months) and postenlistment (ie, 7-12 months) periods using claims data. Results: Baseline characteristics and medical spending were similar between groups (P=.89). On average, the total included sample population (N=113,632) consisted of young to middle-age (mean age 38.81 years), mostly White (n=55,562, 48.90%), male (n=46,731, 41.12%) and female (n=66,482, 58.51%) participants. Compared to a propensity score–matched cohort, DHI users demonstrated approximately US $10 per user per month lower average medical spending (P=.02) with a concomitant increase in preventive care activities and decrease in nonemergent emergency department admissions. These savings translated to approximately US $6.8 million in avoidable health care costs over the course of 1 year. Conclusions: This employer-sponsored, digital health engagement program has a high likelihood for return on investment within 1 year owing to clinically meaningful changes in health-seeking behaviors and downstream medical cost savings. Future research should aim to elucidate health behavior–related mechanisms in support of these findings and continue to explore novel strategies to ensure equitable access of DHIs to underserved populations that stand to benefit the most.