About

Ellen McCreedy is an Associate Professor of Health Services, Policy, and Practice at the Brown University School of Public Health. Her research focuses on pragmatic evaluation of nonpharmaceutical interventions for managing neuropsychiatric symptoms in people living with dementia. Dr. McCreedy is currently leading an embedded pragmatic trial testing the effects of personalized music on agitation and antipsychotic use for nursing home residents with dementia, and a trial testing the effects of an enhanced advance care planning intervention on documentation of care wishes for people with dementia in assisted living centers. She has previously served as measurement lead for a trial testing the effect of tunable LED lighting on agitation and sleep for nursing home residents with dementia. Dr. McCreedy is also a Steering Committee member of the Technical and Data Core of the National Institute on Aging (NIA) IMbedded Pragmatic Alzheimer's disease (AD) and AD-Related Dementias (AD/ADRD) Clinical Trials (IMPACT) Collaboratory. She received her MPH in Global Health from the University of South Florida, her PhD in Health Services Research from the University of Minnesota, and completed a postdoctoral research fellowship at Brown University, Center for Gerontology and Healthcare Research.

Research topics

• Medicine
• Nursing
• Computer Science
• Medical emergency
• Database
• Gerontology
• World Wide Web
• Family medicine

Selected publications

• Age‐Friendly Health System Implementation in Outpatient Settings: A Systematic Review

  Journal of the American Geriatrics Society · 2026-02-16

  articleOpen access

  INTRODUCTION: The Age-Friendly Health Systems (AFHS) initiative aims to improve care for older adults through the "4Ms" framework: What Matters, Medication, Mentation, and Mobility. Despite national momentum and evidence within individual M domains, limited evidence guides outpatient AFHS implementation of the 4Ms as a set. The objective of this systematic review was to summarize the evidence of the impact of AFHS implementation in outpatient settings. METHODS: We searched Medline, EMBASE, CINAHL, Cochrane, and clinicaltrials.gov from 2015 to November 22, 2024. Comparative studies that implemented all 4Ms in outpatient settings were included. Risk of bias was assessed using questions derived from the Cochrane Risk of Bias tool for RCTs and the Risk of Bias In Non-randomized Studies-of Intervention tool for other study designs, and results were summarized using GRADE methodology. RESULTS: Twelve US-based studies met inclusion criteria. Overall, implementing AFHS interventions was associated with improved process measures across all 4Ms, though the effectiveness of specific implementation strategies could not be determined. Outcome and structural measures were infrequently reported. Study heterogeneity and poor reporting limited generalizability. DISCUSSION: Findings underscore the urgent need for standardized, outcomes-oriented AFHS measurement before policy and payment reforms, such as CMS's Age-Friendly Hospital Measure, are expanded into outpatient settings. To advance meaningful transformation, future research must prioritize implementation fidelity, outcome evaluation, and measures that reflect older adults' values and lived experiences.

  PublisherOA PDFDOI

• Feasibility of Care Coordination to Reduce Unnecessary Hospitalization For Assisted Living Residents

  Innovation in Aging · 2025-12-01

  articleOpen accessSenior author

  Abstract Almost half of Assisted Living Community (ALC) residents visit the emergency department (ED) yearly, experiencing more visits and longer stays compared to community-dwelling older adults. People living with Alzheimer’s disease and related dementias (ADRD) are at greater risk for delirium, falls, and accelerated decline associated with increased ED visits. Each transition provides an opportunity for care coordination and avoidance of unnecessary hospital admission. Bluestone Accountable Care Organization developed ED Early Response, a care coordination program. Care managers provide timely, structured information to ED providers via phone and fax within 120 minutes of ED registration. We assessed the feasibility of, and adherence to, the program. Between November 2023 and June 2024, we enrolled 1,376 patients with 1,989 eligible ED visits (mean: 1.4 visits per patient), 1,237 ED visits for patients with ADRD and 752 ED visits for patients without ADRD. Qualifying visits occurred during working hours (8 a.m. - 4 p.m.), with 82.5% of visits (n = 1,641) identified via electronic admission, discharge, and transfer notifications, and 17.5% of visits (n = 348) identified through direct communication between ALCs and care managers. Care managers successfully provided real-time information to ED providers for 44% of the eligible ED visits (547 of 1,237) for patients with ADRD, and for 40% of visits (304 of 752) for patients without ADRD. Care managers self-reported a hospital avoidance rate of 11%. While adherence was lower than anticipated, early structured communication could reduce unnecessary hospital admissions for ALC residents.

  PublisherOA PDFDOI

• Engaging Assisted Living Residents in the Development of an Evaluation Plan for a Future Trial of an Exercise Program

  Innovation in Aging · 2025-12-01

  articleOpen accessSenior author

  Abstract Many trials are conducted without first assessing which outcomes are important to patients receiving the interventions. The goal of this work was to ask participants of an ongoing exercise program what they believed the most important impacts of the program were to better align outcomes for a future trial with participant priorities. The Rev6 program is a seated, 30-minute exercise class designed for older adults with varying cognitive abilities. The program is offered twice a week by a large assisted living community (ALC) in Rhode Island. We observed three Rev6 sessions and held a focus group with seven Rev6 participants. We also interviewed one ALC staff member responsible for delivering the Rev6 program and two corporate leaders. Three members of the research team analyzed the focus group and interview transcripts to identify themes. Themes related to physical health included staff- and participant-reported impact of the Rev6 program on balance and walking, strength, and pain. Themes related to mental health included staff- and participant-reported impact of the Rev6 program on “feeling better” and increased social connection. With input from ALC staff, researchers developed visual cue cards for each theme and met with nine program participants to vote on the most important theme to evaluate as an outcome measure for a future study. Four residents voted for balance and walking, three residents voted for feeling better, and two residents voted for pain as the most important impact of the program. This framework for evaluation codesign includes the voices of people with dementia.

  PublisherDOI

• Resident and Nursing Home Factors Associated With Adherence to a Personalized Music Intervention: Secondary Analyses From Music & MEmory: A Pragmatic TRial for Nursing Home Residents With ALzheimer’s Disease (METRIcAL)

  Journal of Aging Research · 2025-01-01

  articleOpen accessSenior authorCorresponding

  Objectives: Music offers a promising nonpharmacological alternative for managing agitation in people with Alzheimer’s disease and other dementias (ADRD). We report resident and nursing home (NH) characteristics associated with uptake of a personalized music intervention. Design: Post hoc analysis of a cluster‐randomized embedded pragmatic clinical trial (ePCT) involving delivering resident‐preferred music to manage agitated behaviors. Setting and Participants: A total of 463 residents with ADRD in 27 NHs randomized to receive the intervention. Methods: We obtained resident and NH characteristics from Minimum Data Set and Long‐Term Care FocUS data. In addition, we created a study‐specific engagement measure, which describes the proportion of enrolled residents in a given NH with any nursing staff use of the intervention. We used hierarchical models to estimate associations between resident and NH characteristics and (1) any exposure to the personalized music intervention and (2) minutes of music received per study day. Results: This post hoc analysis included 463 residents from 27 NHs (mean age: 80 years (standard deviation, SD: 12.2), 68.5% female, and 25.3% Black or African American). Resident characteristics associated with a greater likelihood of any exposure to the music included being Black or African American ( p = 0.02). NH characteristics were associated with greater likelihood of any exposure included higher quality star ratings ( p = 0.01) and nursing staff engagement with the intervention ( p = 0.01). Among those exposed to the music, younger residents ( p = 0.02), Black residents ( p = 0.03), and those with less health instability ( p = 0.03) received greater doses. Residents living in NHs with high nursing staff engagement also received higher doses ( p ≤ 0.001). Conclusions and Implications: Black race was associated with a greater probability of exposure and more use of a personalized music intervention, after controlling for NH quality. Nursing staff engagement with a personalized music intervention increased uptake. These findings are useful for future ePCTs of behavioral interventions in NHs. Trial Registration: Clinicaltrials.gov Identifier: NCT03821844

  PublisherOA PDFDOI

• Applying Iterative RE-AIM to Translate Evaluation Data Into Real-Time Adaptations

  Innovation in Aging · 2025-12-01

  articleOpen access

  Abstract Iterative RE-AIM provides a novel approach to evaluating interventions in real-time and translating those insights into actionable adaptations. Using the principles of iterative RE-AIM, we conducted a midpoint survey with the members of the coaching intervention team and the implementation evaluation team for 40Winks, a nursing home-based sleep trial. We convened a facilitated meeting with these two teams to discuss results and priorities, evaluate existing adaptations, review the effectiveness of the current intervention plan, propose future adaptations, and formulate a plan for implementation. The aggregated survey data were displayed and discussed by the two teams together as part of this process. While iterative RE-AIM outlines a process for evaluating an intervention’s adherence to the RE-AIM model, we applied this framework to evaluate the intervention’s components rather than the RE-AIM dimensions. This modification allowed us to bring all levels of the intervention and evaluation teams together to have a combined voice in the direction and evaluation of the intervention, and to make critical, real-time adaptations to enhance the effectiveness and reach of the intervention. This presentation will provide a roadmap of the process our team developed, guided by the principles of iterative RE-AIM, to display the evaluation survey and data and discuss strategies for implementing the results of this survey to lead real-time adaptations. This novel methodology allows the teams to learn from and act upon their own data in real time, allowing for iterative and systematic improvements in an intervention as a trial progresses.

  PublisherOA PDFDOI

• Home Health Initiation by Payor Type and Associated Outcomes among Medicare-Enrolled Veterans

  Innovation in Aging · 2025-12-01

  articleOpen access

  Abstract Background Medicare-enrolled Veterans may receive Veterans Health Administration (VA), Traditional Medicare (TM), or Medicare Advantage (MA) funded home health care (HH). Payor type influences rehospitalization and mortality, but does days to HH initiation by payor relate to rehospitalization, nursing home (NH) admission, and death? Methods Our retrospective cohort study evaluated 72,743 Medicare-enrolled Veterans discharged from VA medical centers (2017–2019) who initiated HH within 14 days for 30- and 90-day rehospitalization, death, and NH admission. We also evaluated timeliness of HH initiation (within vs. >2 days after discharge), relative risks by payor, and for Veterans at lower baseline mortality risk. Results Among Veterans starting HH within 2 days (n = 35,988), 34.8% received VA-paid HH, 56.3% TM, and 8.8% MA. Rehospitalization within 30 days was 18.5% for VA, 19.5% TM, and 19.7% MA; 90-day rates were 33.3%, 33.8%, and 33.5%, respectively. Thirty-day mortality was higher for VA-paid HH (3.3%) vs TM (2.9%) and MA (3.0%), and at 90 days (9.6% vs. 8.3% and 8.9%). NH admission at 30 days was similar. For HH initiation timeliness, VA-paid HH more often began ≤2 days (RRadj=0.95, 95% CI 0.93–0.98 for all Medicare; 0.95, 95% CI 0.92–0.99 for TM; 0.87, 95% CI 0.83–0.91 for MA). Conclusions Preliminarily, outcomes differ by payor type: VA-paid HH had lower rehospitalization but higher short-term mortality than TM, and faster initiation than Medicare-paid HH. These findings have policy relevance of payor-driven differences in access, quality, and outcomes. This novel study has implications for quality and access monitoring across payors.

  PublisherOA PDFDOI

• Recruiting and Consenting Community-Dwelling Adults With Dementia to Test Sleep Tracking Devices

  Innovation in Aging · 2025-12-01

  articleOpen accessSenior author

  Abstract Poor sleep is associated with decreased physical and cognitive function, poor mood, and increased falls. Understanding sleep patterns may help physicians better recommend behavioral programs for improving sleep. Researchers partnered with the PACE Organization of Rhode Island (PACE-RI) to recruit community-dwelling older adults living with dementia to pilot test sleep-tracking devices. PACE-RI provides healthcare to adults 55+ with complex medical conditions, with a focus on maintaining their independence. Participants in this study used a wearable sleep tracking device and a sleep tracking mat for three consecutive nights. Caregivers residing with participants completed a four-question sleep diary for each of the three nights. We encountered challenges recruiting up to 15 participants in this study, with only 5 participant-caregiver dyads enrolled over a one-year period. Key barriers to recruitment included: lack of Spanish language materials / bilingual recruiters, requirement to have a cohabitating caregiver, no internet service at home, lack of trust of researchers and/or technology, and concerns about researchers entering homes. At PACE-RI, approximately 45% of their clients have dementia, but most do not have a family caregiver living with them. In an effort to establish greater trust and connection with participants, the research team visited each participant three times (at consent / equipment setup, at equipment pickup, and after personalized findings were available) and called daily. We will share our one-page participant reports that summarize each participant’s sleep data using simple language and pictures. In this ongoing pilot, the follow-up visits and reports have increased snowball-based recruitment.

  PublisherOA PDFDOI

• Bluestone’s Emergency Department Early Response Program: Promoting High-Quality and Safe Transitions

  Innovation in Aging · 2025-12-01

  articleOpen accessSenior author

  Abstract Almost half of all Assisted Living Community (ALC) residents transfer to the emergency department (ED) each year. For people living with dementia (PLWD), providing a medical history can be a challenge, resulting in prolonged stays and risk of delirium or hospital acquired infections. Bluestone Physician Services developed a ED early response program in which complex care managers (CCM) receive an electronic Admission, Discharge, and Transfer (ADT) notification when one of their patients registers at an ED. If the notification occurs during business hours, the CCMs call the ED using a script and follow-up with a structured fax. We conducted semi-structured interviews with 12 CCMs involved in the program to assess the feasibility and acceptability of the program. There were five themes: patients with dementia and those on hospice were especially likely to benefit from the program; strengths of the program, including increased communication between Bluestone and ED providers; weakness of the program, including a lack of awareness of the program among variable ED staff and challenges with timing the call and fax to maximize benefit. The CCMs also shared some learnings and adaptations that increased contact rate over time and individual success stories. In this session, we will also provide tips on how to integrate real-time ADT notifications into CCM workflows. Next steps include examining the ED provider, patient and caregiver perceptions of the program, quantifying the impact of the program on utilization, and analyzing how the program would perform in other health care settings.

  PublisherOA PDFDOI

• Emergency Department Care Coordination Program for Assisted Living Residents With Dementia

  JAMA Network Open · 2025-08-11

  articleOpen accessSenior author

  Importance: Care transitions to the emergency department (ED) from assisted living centers (ALCs) for residents may include incomplete or inaccurate information during transfer. These transitions can be especially difficult for vulnerable populations, including persons living with dementia (PLWD). Objective: To assess perceptions of complex care managers (CCMs) implementing a care coordination program designed to improve communication for transfers from ALCs to the ED. Design, Setting, and Participants: This qualitative study analyzed semistructured video conference interviews with CCMs in February 2024. The ED early response program was available through a physician services group (Bluestone Physician Services), which provides care to residents in ALCs in Florida, Minnesota, and Wisconsin for patients in its accountable care organization. The physician services group identified the CCMs based on scheduling convenience. Data were analyzed in March and April 2024 using directed content analysis. Exposure: The ED early response program included electronic notification to the physician services group CCMs when a patient registered at an ED. CCMs then communicated via fax and telephone with the ED staff to provide key clinical information. Main Outcomes and Measures: CCM-perceived strengths and weaknesses of the program. Results: Of 22 total CCMs, 12 participated in this study (employed as a CCM for a median [IQR] of 2 [1-3] years; 12 [100%] female) and identified populations that they perceived to especially benefit from the program, including PLWD, patients in hospice, and patients living in group homes. CCMs shared how they communicated with various ED staff, including nurses and physicians, and that receptivity varied among staff. Strengths of the program include CCM advocacy for patients and program adaptability. CCMs described areas of opportunity as lack of education about the program among ED staff and lack of 24-hour coverage for CCMs. Overall, the perception shared by the CCMs was that the program positively affected both the ED experience for patients and the facilitation of goal-concordant care. Conclusions and Relevance: In this qualitative study of a care coordination intervention, CCMs advocated for their patients remotely by filling information gaps, particularly for PLWD and patients in hospice, and perceived that the intervention was associated with improved patient care. CCMs also identified key areas for improvement, such as to increase ED staff awareness of the program and to expand program hours. This care coordination intervention may provide an opportunity to address gaps in care for individuals living in ALCs who present to the ED.

  PublisherOA PDFDOI

• A Bayesian Framework for Latent Compliance Modeling in Cluster Randomized Trials with One-Sided Noncompliance

  ArXiv.org · 2025-09-18

  preprintOpen access

  In pragmatic cluster randomized controlled trials (PCRCTs), healthcare providers are randomized while both providers and patients may deviate from the assigned intervention. In many PCRCTs, cluster-level implementation is measured using multiple continuous metrics, while individual compliance is recorded as a binary indicator. Standard complier average causal effect (CACE) estimands focus on individual-level compliance and do not account for heterogeneity in implementation across clusters. When intervention uptake is shaped by both provider- and patient-level processes, it is of scientific interest to characterize how effects vary across these sources of compliance. We propose a Bayesian framework for PCRCTs with one-sided binary noncompliance at the individual level and one-sided partial compliance at the cluster level. The method uses a latent mixture model to summarize heterogeneity in cluster-level implementation based on baseline characteristics and observed implementation measures, and links these latent implementation types to individual compliance and outcomes through a joint model. Because compliance is only observed in treated clusters, the model imputes unobserved compliance behavior for clusters and individuals assigned to control. The framework enables estimation of finite- and super-population intent-to-treat (ITT) and CACE estimands, both marginally and within latent implementation types. We apply the method to the METRIcAL trial, a pragmatic cluster randomized study evaluating a personalized music intervention for nursing home residents with dementia. The analysis illustrates how accounting for implementation heterogeneity and individual compliance can provide insights beyond standard ITT analyses.}{Causal inference; Principal stratification; Complier average causal effect; Cluster randomized trials; Noncompliance; Bayesian methods; Latent variable models; Interference.

  PublisherOA PDFDOI

Frequent coauthors

• Vincent Mor

  Providence College

  126 shared

• Rosa Baier

  Providence College

  92 shared

• Kali S. Thomas

  Johns Hopkins University

  78 shared

• Robert L Kane

  52 shared

• James L. Rudolph

  Brown University

  44 shared

• Roee Gutman

  44 shared

• A. Lynn Snow

  Tuscaloosa VA Medical Center

  38 shared

• Susan L. Mitchell

  Harvard University

  37 shared