Political Research Associates

Preventable readmissions represent a significant opportunity to improve quality and reduce healthcare costs, with approximately 26% of Medicare medicine readmissions considered preventable. However, evidence on the effectiveness of post-discharge interventions at scale remains mixed, and implementing evidence-based practices consistently across large, diverse health systems is a challenge. To address these concerns, the Mass General Brigham Population Health Services Organization (MGB PHSO) developed and implemented a novel, multidisciplinary, system-wide post-discharge intervention aimed at reducing 30-day readmissions within its Medicare Shared Savings Program (MSSP) Accountable Care Organization (ACO). It was hypothesized that standardizing delivery through a high-fidelity workflow would reduce readmissions. A standardized, multidisciplinary program was created involving: (1) a coordinator conducting chart review and obtaining records; (2) a pharmacist performing medication reconciliation; and (3) a registered nurse completing a post-discharge assessment. A prospective cohort study was conducted comparing the outcomes of patients at pilot intervention sites with those of a propensity-matched control group. The intervention cohort showed a directional reduction in 30-day readmission rates compared to the matched controls (13.5% vs. 16.3%, P = 0.07) but no significant difference in 30-day emergency department presentations. The intervention group also had a significantly higher rate of 14-day follow-up appointments (70.0% vs. 65.3%, P = 0.025). These findings support the effectiveness of a centralized, standardized post-discharge strategy for reducing readmissions within an ACO setting. This study demonstrates that structured, system-level interventions can improve care transitions and outcomes in value-based care models.

BackgroundPatients hospitalized with complicated urinary tract infections (cUTI) caused by multidrug-resistant uropathogens often need intravenous (IV) carbapenems, which can prolong inpatient (IP) stay and lead to increased adverse outcomes and resource utilization. This study aimed to determine the difference in duration between clinically ready for discharge (CRFD) status and actual hospital discharge date for hospitalized patients with cUTI treated with IV carbapenems.MethodsPatients aged ≥18 years who received IV carbapenem for cUTI with a positive urine culture between January 2019 –June 2024 were assessed retrospectively using structured electronic health records and unstructured clinical notes from the Mass General Brigham Research Patient Data Registry. Stringent (missing data based on the last observation of CRFD criterion carried forward) and lenient (missing data assumed to indicate CRFD criterion met) rule-based definitions for CRFD (Table 1), based on literature review and clinical guidelines and developed following a feasibility assessment of data availability and completeness, were used to determine when CRFD status was achieved.ResultsPatients (N=333) had a median (interquartile range [IQR]) age of 73.0 (63.0–83.0) years, 58.3% were female and 79.6% were White (Table 2). The median (IQR) duration of IP stay was 6.0 (4.0–9.0) days. Among patients tested, 74.1% (n=229/309) had extended-spectrum β-lactamase-producing Enterobacterales; 78.4% (n=196/250) and 54.7% (n=174/318) had isolates not-susceptible to fluoroquinolones and trimethoprim-sulfamethoxazole, respectively (Table 2). The stringent definition identified 35.1% of patients as CRFD before actual discharge, with a mean [median] difference between CRFD and actual discharge dates of 2.9 [2.0] days (range: 1.0–8.0 days). Using the lenient definition, 57.1% of patients were CRFD by a mean [median] of 3.2 [3.0] days (range: 1.0–10.0 days) before actual discharge.ConclusionThis study suggests that, even by stringent application of criteria, more than a third of patients with cUTI remained hospitalized longer than clinically necessary, possibly to receive IV carbapenem treatment. These findings emphasize the unmet need for new effective oral antibiotic treatments.Funding: GSK study 221816.DisclosuresSerena P. Koenig, MPH, MD, Brigham and Women's Hospital: Employee|GSK: Grant/Research Support Jeffrey J. Ellis, PharmD, MS, GSK: Employee|GSK: Stocks/Bonds (Public Company) Douglas Boettner, PhD, GSK: Employee|GSK: Stocks/Bonds (Public Company) Lindsey Parker, PharmD, GSK: Employee|GSK: Stocks/Bonds (Public Company) Rose Chang, ScD, Analysis Group: Employee|GSK: Grant/Research Support Louise Yu, MS, Analysis Group: Employee|GSK: Grant/Research Support Emily Reichert, MS, Analysis Group, Inc.: Employee|GSK: Grant/Research Support Joanne Chukwueke, MPH, Analysis Group, Inc.: Employee|GSK: Grant/Research Support Zheyi Cao, MS, Analysis Group, Inc.: Employee|GSK: Grant/Research Support Yichuan Grace Hseih, PhD, GSK: Grant/Research Support|Mass General Brigham: Employee Christopher Herrick, MBA, GSK: Grant/Research Support|Mass General Brigham: Employee Mei Sheng Duh, MPH, ScD, Analysis Group: Employee|GSK: Grant/Research Support Shawn N. Murphy, MD, PhD, GSK: Grant/Research Support|Mass General Brigham: Employee

To explore the impact of an immersive virtual reality (VR) training module on infection prevention and control (IPC) knowledge and attitudes of healthcare personnel (HCP) and to demonstrate the use of VR for performance assessment in cleaning and disinfection of portable medical equipment (PME). Quasi-experimental study. Two academic medical centers and three long-term care facilities. HCP in clinical roles were recruited. Pilot sites trained participants on an immersive VR training module on PME cleaning and disinfection. Participants completed the VR module and pre- and post-knowledge and attitude assessment surveys, including a post-survey on the user experience of the VR module. Performance data were collected from the head-mounted displays (HMD) on the duration of the VR session, and participant performance including in-module task completion, hand hygiene compliance, PME disinfection percentage, and in-module quiz performance. Statistical significance and effect size were calculated using paired sample t-tests and Cohen's D for pre- and post-survey results. HMD data were analyzed using descriptive statistics. A total of 60 participants were recruited; 54 were included for analysis, with improvements in knowledge and attitudes post-training. Participant user experience was rated 50.19/55. HMD data demonstrated: 22-minute mean module duration, mean of 2.15/28 tasks not completed, mean of 2.56 missed hand hygiene opportunities, and 54% PME mean disinfection percentage, and varied performance on in-module quizzes. Immersive VR training may be effective in improving HCP knowledge and attitudes in IPC concepts. Performance data collected through VR training can evaluate learner performance and be used to target training for improvement.

The increasing integration of artificial intelligence (AI) tools into radiology has created an urgent need for clear, practical guidance on how to evaluate them. These tools, including computer-assisted detection and triage devices, hold promise for improving accuracy, efficiency, and workflow. However, their adoption into clinical practice requires rigorous evaluation to ensure safety, generalizability, and clinical value. Radiology has a strong foundation in diagnostic test assessment, and AI models represent an extension of this tradition, with new considerations in evaluation strategy, performance measurement, and study design. This article provides a structured primer on evaluating AI models across their development and deployment life cycle. It outlines key principles for internal and external testing, highlights performance metrics tailored to AI outputs-including classification, detection, segmentation, and continuous measures-and describes how to assess clinical impact with multireader multicase studies. Practical examples from radiology research, as well as updated reporting standards, are incorporated throughout. By translating core statistical concepts into radiology-specific guidance, this article aims to support radiologists, researchers, and reviewers in conducting and interpreting high-quality AI evaluation studies.

Racial inequities in substance use disorder treatment, and specifically buprenorphine treatment, contribute to overdose inequities. Bridge Clinics are a low-threshold, transitional treatment model that may improve treatment access. This report assesses the change in visit volume for Black and Hispanic patients and those with limited English proficiency (LEP) after the implementation of the Bridge Clinic expansion across 4 distinct sites. Between October 2021 and 2023, 3 hospital-based Bridge Clinics were expanded, and a community-based Bridge Clinic was established, all in regions with high racial and language inequities in overdose mortality. Implementation included creating Bridge Clinic performance measures, which were reviewed with clinic leadership monthly, developing a toolkit, and launching mobile services. We present Bridge Clinic visit volume aggregated by calendar year for Black-non Hispanic, Hispanic (any race), and LEP patients. We calculated the percent of total visits for each patient group and assessed the change over time. Comparing 2021 to 2023, total visits increased from 5323 to 10,350, and unique patients increased from 1893 to 3316. Annual visits increased from 437 to 1151 visits for Black patients; 566 to 1609 for Hispanic patients; and 96 to 265 for LEP patients. The percent of visits grew significantly for Black (8.21% vs. 10.24%, P<0.001), Hispanic (10.63% vs. 15.55%, P<0.001) and LEP patients (1.80% vs. 2.56%, P=0.003) from 2021 to 2023. Expanding low-barrier Bridge Clinics may increase substance use disorder (SUD) treatment visits for Black and Hispanic patients and those with LEP.

This study aimed to develop a human factors assessment for medication-related clinical decision support (CDS) based on a previously validated tool that assessed the integration of human factors principles in CDS, the instrument for evaluating human factors principles in medication-related decision support alerts (I-MeDeSA), and pilot it with 10 outpatient clinics across the United States.The human factors assessment was developed based on past validations of I-MeDeSA. Examples included changing the wording of questions and reformatting answer choices to check-box options, allowing for multiple answer choices. We also added a section about how clinicians resolved alerts. Clinics received a percentage score based on how well their CDS adhered to human factors principles. To take the assessment, testing teams at each clinic triggered a high-severity drug-drug interaction (DDI) alert, and then took the human factors assessment. This assessment was piloted in 10 outpatient clinics, each of which used a different commercial electronic health record (EHR) system.The final assessment included five sections and twelve questions related to aspects like the timing, visual aspect, severity, content, and actions within the DDI alert. The mean overall percentage score was 62%. The sections regarding the timing and visual aspects of the alert were ones where clinics' EHRs performed the best. However, in the "actions" section, 40% of the clinics could bypass high severity alerts without any safeguards in place.We found substantial variability in the integration of human factors principles in the design and delivery of DDI alerts among the outpatient clinics, and some lacked important medication safeguards. This assessment can be used by outpatient clinics for safety improvement initiatives.

Portable medical equipment (PME) is inconsistently cleaned and disinfected, resulting in contamination that increases the risk of healthcare-associated infections. A virtual reality PME cleaning and disinfection training module was designed and tested at multiple healthcare facilities. Barriers identified during an initial phase led to improvements in the second phase.

The Brain Care Score (BCS) was developed in partnership with patients and practitioners to convey actionable knowledge to individuals everywhere that can motivate change in health-related behaviors and thereby reduce the risk of dementia, stroke, and late-life depression (LLD). Because diseases outside the brain share modifiable risk factors with dementia, stroke, and LLD, we investigated the associations of the BCS with other common age-related diseases, including cardiovascular disease (CVD) and cancer. Among all UK Biobank (UKB) participants with complete BCS data, we performed Cox proportional hazard regression analyses between the BCS at baseline and incident CVD (ischemic heart disease, stroke, and heart failure) and the three most common cancer types (lung, colorectal, and breast cancer), adjusted for sex and stratified by age. Among 416370 UKB participants (mean age: 57 years; 54% female), 33944 cases of CVD (8.8%) and 16090 cases of cancer (4.0%) were identified over a median follow-up of 12.5 years. A 5-point higher BCS at baseline was associated with a lower incidence of CVD (hazard ratio [HR]: 0.57 [95% confidence interval {95% CI}: 0.55-0.59]) and lower incidence of the three most common cancer types (HR: 0.69 [95% CI: 0.66-0.72]). A higher BCS at baseline is associated with a lower incidence of CVD and three cancer types. Although developed specifically as an actionable tool to guide individuals in reducing their risk of common age-related brain diseases, we show that it may also offer ancillary benefits, providing a single place to start for guiding individuals toward improving their chances of healthy aging more generally.

Limited data exist comparing the effectiveness and safety of rivaroxaban and warfarin in patients with right ventricular dysfunction (RVD), a common acute pulmonary embolism (PE) complication. To assess the effectiveness and safety of rivaroxaban compared with warfarin among patients with PE and RVD. Adult patients newly prescribed rivaroxaban or warfarin during PE-related hospitalization with evidence of RVD were identified from Mass General Brigham's Research Patient Data Registry database (January 2013-May 2023). Outcomes included time-to-first recurrent venous thromboembolism (VTE) and time-to-first major bleeding event. The proportion of international normalized ratio (INR) measurements within therapeutic range (INR: 2-3) while on warfarin was described. Kaplan-Meier analysis described event rates at 6-month intervals up to 36 months, which were compared using hazard ratios, 95% CIs, and P values from Cox proportional hazards models. Overall, 246 rivaroxaban and 315 warfarin users were included (mean age, 63 years; female: 53%). Median time of treatment was 270 and 235 days for rivaroxaban and warfarin users, respectively; 50.9% of INR measurements among warfarin users were within therapeutic range. Rivaroxaban was associated with significantly lower risk of VTE recurrence than warfarin at all-time points, including 41% lower risk at 36 months (20.4% vs 30.3%; hazard ratio [95% CI], 0.59 [0.38, 0.92]). There was no significant difference in risk of major bleeding between cohorts up to 36 months of treatment (8.2% vs 13.6%). Rivaroxaban was associated with lower risk of recurrent VTE compared with warfarin, without a significant difference in risk of major bleeding.

ObjectiveThis study aimed to examine user actions within a clinical decision support (CDS) alert addressing hypertension (HTN) in chronic kidney disease (CKD).MethodsA pragmatic randomized controlled trial of a CDS alert for primary care patients with CKD and uncontrolled blood pressure included prechecked default orders for medication initiation or titration, basic metabolic panel (BMP), and nephrology electronic consult (e-consult). We examined each type of action and calculated percentages of placed and signed orders for subgroups of firings.ResultsThere were firings for medication initiation (813) and medication titration (430), and every firing also included orders for nephrology e-consult (1,243) and BMP (1,243). High rates of override (59.6%) and deferral (14.6%) were observed, and CDS-recommended orders were only signed about one-third of the time from within the alert. The percentage of orders that were signed after being placed within the alert was higher for medication initiation than for medication titration (33 vs. 12.0% for angiotensin-converting enzyme inhibitors [ACEi] and 38.8 vs. 14% for angiotensin II receptor blockers [ARBs]). Findings suggest that users are hesitant to commit to immediate action within the alert.ConclusionEvaluating user interaction within alerts reveals nuances in physician preferences and workflow that should inform CDS alert design. This study is registered with the Clinicaltrials.gov Trial Registration (identifier: NCT03679247).