About

Dr. Debra Saliba holds the Anna and Harry Borun Chair in Geriatrics and Gerontology at UCLA. Her research focuses on various aspects of geriatrics, including nursing home quality, staffing, health outcomes, and disparities in care for older adults. She has contributed extensively to understanding transfer rates, self-efficacy among family caregivers in dementia care, and the impact of Medicare payment reforms on patient complexity and hospital admissions. Her work also addresses issues related to end-of-life care, racial and ethnic disparities, and the use of telehealth in older populations. Dr. Saliba's research aims to improve health and function for older adults through evidence-based practices and policy analysis.

Research topics

• Medicine
• Database
• Family medicine
• Gerontology
• Law
• Economic growth
• Psychiatry
• Business
• Actuarial science

Selected publications

• The role of declining therapy volumes in skilled nursing facility outcomes: a mediation analysis

  Health Affairs Scholar · 2026-02-01

  articleOpen access

  Introduction: Significant declines in therapy provision in skilled nursing facilities (SNFs) followed the 2019 implementation of Medicare's Patient-Driven Payment Model (PDPM) and the onset of the COVID-19 pandemic, raising concerns about effects on patient outcomes. Methods: Using Medicare fee-for-service claims and SNF assessment data from January 2018 through September 2021, we analyzed 3.5 million post-hospital SNF stays to assess whether changes in therapy volumes mediated changes in successful community discharge and 30-day hospital readmissions. Results: Average total physical, occupational, and speech therapy minutes per day declined from 122.2 before PDPM to 96.5 immediately after implementation and to 87.7 during the pandemic. Adjusted probabilities of successful community discharge rose modestly after PDPM but fell during COVID-19, while readmissions declined initially and then increased. Mediation analyses showed that reductions in therapy volumes were strongly associated with the declines in community discharge and increases in readmissions. These findings persisted for patients with dementia and moderate levels of functional impairment at admission; declining therapy volumes were associated with the observed worsening of discharge outcomes after PDPM implementation and during the pandemic. Conclusions: Results highlight therapy provision as a key modifiable policy target for improving post-acute outcomes and reducing rehospitalizations among older adults in SNFs.

  PublisherDOI

• What's in a Number: Do Transfer Rates Reflect Nursing Home Quality?

  Journal of the American Geriatrics Society · 2026-02-24

  article1st authorCorresponding

  BACKGROUND: Hospitalization rates from nursing homes (NHs) have gained traction as pragmatic quality measures that can be derived from claims data. However, claims-based hospitalization measures do not account for clinical complexity and the extent to which they reflect quality of care or quality of transfer decision making is unknown. We aim to examine agreement between a claims-based measure of potentially avoidable hospitalizations and expert clinician review of transfer decision making and care quality. METHODS: We randomly selected 252 hospital transfers across eight Veterans Administration (VA) NHs, known as Community Living Centers (CLCs). Eleven expert clinicians independently completed Structured Implicit Reviews (SIRs) of medical records to assess: (1) whether the transfer decision was appropriate (i.e., hospital was the lowest safe level-of-care given the resident's acute condition); (2) quality of care for evaluation or treatment of the acute change (adequate management of acute change), (3) quality of care for chronic conditions and preventing decline. We used VA Corporate Data Warehouse (CDW) data to determine a claims-based measure of potentially avoidable hospitalization. RESULTS: CDW data were available for 242 VA hospitalizations. The claims-based measure categorized 29 (12%) hospitalizations as potentially avoidable; only 2 of which matched the 20 SIR identified as inappropriate decisions to transfer. Furthermore, the claims-based measure flagged only 5 of 33 cases rated as inadequate treatment of acute decline and 6 of 17 rated as poor quality of chronic disease or preventive care. CONCLUSIONS: In a geographically diverse sample of CLC transfers, independent clinical experts' judgments of transfer decision appropriateness, quality of care for acute decline, and quality of chronic care differ from a claims-based potentially avoidable hospitalizations measure. Findings underscore the need for nuanced clinical consideration of hospitalization metrics for assessing quality and for understanding which aspects of care should be addressed to safely reduce NH transfers to hospitals.

  PublisherDOI

• <scp>JAGS</scp> Remains Committed to Scientific Principles That Support Better Health and Function for All Older Adults: Disparities, Diversity in Research, and Respectful Discourse

  Journal of the American Geriatrics Society · 2025-07-12

  article

  The authors declare no conflicts of interest.

  PublisherDOI

• Change in <scp>HOS</scp> Effectiveness of Care Quality Among Medicare Advantage Enrollees Associated With the <scp>COVID</scp>‐19 Pandemic

  Journal of the American Geriatrics Society · 2025-05-20

  article

  The authors declare no conflicts of interest. Data S1. jgs19528-sup-0001-Tables. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

  PublisherDOI

• Telehealth Use by Home Health Agencies Before, During, and After <scp>COVID</scp>‐19

  Health Services Research · 2025-05-22 · 2 citations

  articleOpen access

  OBJECTIVE: To examine telehealth adoption and discontinuation by home health agencies (HHAs) during the COVID-19 pandemic in the context of telehealth pre-pandemic diffusion into the industry and its continued use once the pandemic abated. STUDY SETTING AND DESIGN: HHAs nationally, serving the most patients with dementia (averaging 33% of the agency's patients) were surveyed during October 2023 to November 2024. Key variables included the agency's adoption and discontinuation of specific telehealth technologies by year, the reasons for discontinuation, and the reasons for not adopting any telehealth technology, either before or during the pandemic. DATA SOURCES AND ANALYTIC SAMPLE: Data were collected via a web-based survey with telephone follow-ups. We received 791 responses (37% response-rate) and provide descriptive statistics of responses and graphics. PRINCIPAL FINDINGS: By 2019, prior to COVID-19, 183 (23%) of HHAs used telehealth, increasing to 446 (56%) by 2021. Growth occurred mainly in virtual visits. Of those HHAs adopting telehealth, 96 (19%) discontinued use later in the pandemic. Key concerns were about the appropriateness of the patient population and reimbursement. CONCLUSIONS: Patterns of adoption and discontinuation suggest that COVID-19 interrupted the innovation diffusion process of telehealth into home health. Telehealth's future will depend on information about cost-effectiveness and Medicare reimbursement policies.

  PublisherOA PDFDOI

• Impact of the Patient-Driven Groupings Model on Rehabilitation Services in Post-Acute Home Health Care

  Innovation in Aging · 2025-12-01

  articleOpen access

  Abstract Home health (HH) agencies offer skilled nursing and rehabilitation (physical therapy [PT], occupational therapy [OT], and speech language pathology [SLP]) services to facilitate patients’ transition back home following hospitalization. Implemented in 2020, the Patient-Driven Groupings Model (PDGM) removed previous incentives to provide rehabilitation visits, raising concerns about reduced access to these services. Using 100% Traditional Medicare HH claims and assessment data from 2018-2021, we examined changes in receipt of rehabilitation services following PDGM implementation. We modelled odds of receiving any PT, OT, and ST visits across 2,657,967 post-acute HH stays using logistic regression. We modelled number of PT, OT, and ST visits among those who received any visits using negative binomial regression. All models accounted for clustering of stays within agencies and controlled for patient sociodemographic and clinical characteristics, length of stay, agency characteristics, and county-level COVID-19 case rates as the pandemic emerged nearly simultaneously with PDGM implementation. While the adjusted odds of receiving PT, OT, and SLP services during post-acute HH stays were not significantly lower under PDGM compared to before PDGM, the number of visits received decreased significantly for PT (aIRR: 0.86; 95%CI: 0.85,0.86), OT (aIRR: 0.82; 95%CI: 0.82,0.83), and SLP (aIRR: 0.76, 95%CI: 0.75,0.77). This reduction on average equates to about one fewer visit each of PT, OT, and SLP per post-acute HH stay. Decreases in PT, OT, and SLP visits under PDGM may lead to worse rehabilitation-related outcomes for post-acute HH patients. Additional examination of the impact of reduced rehabilitation on patient outcomes is urgently needed.

  PublisherOA PDFDOI

• An Elusive Ideal—Why a Pragmatic “Home Time” Quality Measure Remains Hard to Define

  Journal of the American Geriatrics Society · 2025-07-17 · 4 citations

  editorialOpen accessSenior author

  “Days at home,” “facility-free days,” “excess days of acute care,” and “healthcare contact days” are increasingly common measures used to evaluate the outcomes of healthcare interventions [1-4]. These home time measures have broad appeal, and researchers have used them as outcomes in a range of studies [5-8]. These measures are also gaining traction in policy circles. For example, the Centers for Medicare & Medicaid Services (CMS) include home time performance measures in alternative payment models such as the Accountable Care Organization (ACO) Realizing Equity, Access, and Community Health (REACH). Part of the reason for the increased use of these outcomes is their intuitive appeal. We as clinicians think most patients value being at home, and many believe it is cheaper for the health care system—a rare win-win. In addition, home time measures are viewed as “pragmatic” because they can be derived from existing administrative claims data [9]. However, these measures have not been subjected to careful evaluation to understand the degree to which they are truly person-centered [1, 10]. In this issue of JAGS, Van Houtven and colleagues describe [11] part of a larger effort to address this gap and develop a universal person-centered quality of life measure, based on time spent at home as compared to time spent in healthcare settings. This study is noteworthy in its focus on anchoring days at home to older adults' quality of life. The authors convened a panel of Veterans Affairs (VA) leadership, clinician scientists, and researchers and presented the panel with data from a mixed methods study that combined quantitative analyses and qualitative work with older Veterans surrounding their perceptions of home time versus the value of health care contact [12, 13]. Delphi panel methodology was employed here because the team's prior work revealed discordance between prior qualitative and quantitative findings about how stays in different healthcare settings (e.g., hospital, emergency department, and skilled nursing facilities) affected patients' quality of life. To reconcile these past contradictions, the Delphi panel was charged with: (1) determining whether to include, in addition to hospital and post-acute facilities, the emergency department as a setting in time away from home; (2) creating relative “weights” to capture varying effects of stays in different healthcare settings on quality of life; and (3) establishing the timeframe over which to most meaningfully assess days away from home. Perhaps the most striking finding of this careful approach was that—like the prior mixed methods study—this panel of experts was unable to come to consensus. While 75% of panelists rated as acceptable the inclusion of emergency department (ED) stays in time away from home, panelists had significant disagreement on establishing relative weights or an appropriate timeframe. Even the rating of ED visits was conditional, with some panelists recommending limiting inclusion to ED visits that led to inpatient stays and the most endorsed response being an acceptance of ED inclusion “with reservations.” This thoughtful evaluation by Van Houtven et al. points out that the enthusiasm for home time measures may be racing ahead of the evidence behind them. Initial evaluations suggest these measures may be relatively insensitive to interventions, perhaps because even in populations with multimorbidity, most older adults spend most of their days at home [2]. The Van Houtven et al. paper therefore raises a critically important question: is it possible (or even the right goal) to derive a pragmatic, person-centered quality of life measure using administrative data? History suggests even well-meaning efforts in this regard frequently meet with failure. An individual's experiences, preferences, and needs will affect their perceived quality of life. This complex interplay changes over time, while, unfortunately, our assessment of older adult's goals too often remains static. Claims-based metrics have limited ability to account for patient factors such as access to a caregiver, living arrangements, cognitive and functional limitations, symptom burden, social roles, beliefs, and socioeconomic factors that impact health and healthcare. This information is not readily available in administrative claims databases, but it is needed to understand both quality and the lens through which a patient views their care [14]. Ultimately, endpoints like “days at home” may be best considered a pragmatic endpoint as part of a bundle of relevant outcomes, rather than a single clear “good” to be achieved. How, then, do we capture this type of nuance? An effective universal measure would need to allow for personalization. Van Houtven and colleagues suggest additional methods to capture the variation in patient preferences, such as using discrete choice experiments to inform weighting across time frames and settings. This approach would require obtaining input from patient populations across multiple care settings and at multiple time points—a challenging feat. The Age-Friendly Health System's focus on capturing What Matters regularly in clinical encounters provides another opportunity, but to our knowledge no systems have been developed to ensure this measure is used longitudinally across care settings and over time [15]. Instead of using an easily calculated proxy measure, we believe refocusing on longitudinal capture and use of patient-reported outcome measures is a worthy goal. If CMS initiatives like the Guiding an Improved Dementia Experience (GUIDE) model [16] are successful, we may see greater clinical implementation of patient and caregiver-centered outcome measures, but to date it is not standard widespread practice to collect such surveys or document them in a way that facilitates easy assessment of outcomes [17]. Major practical challenges to overcome with obtaining patient and caregiver reports include potential low response rates, missing data, and potential to add to clinical demands. However, significant costs are also associated with claims-based metrics and with building, documenting, and extracting required data elements in electronic health records [18]. Investing in infrastructure to reliably and validly capture patient-reported quality of life and goal-concordant care measures could improve the efficiency of obtaining quality measures. If efficiency is viewed as the potential gain in meaningful data relative to health system and provider costs, such assessments could be as, if not more, cost effective for health care providers than claims-based metrics. The Veterans Affairs (VA) Clinical Assessment Reporting and Tracking (CART) Program is one example of this type of infrastructure. The CART Program, a quality assessment initiative for percutaneous coronary intervention, captures measures through the Patient Reported Health Status Assessment (PROST) system. These automated assessments are collected during routine clinical care and embedded into the electronic health record, reducing the burden of data capture, while potentially enhancing clinical decision making and individualized care [19]. There still may be value in using administrative claims data to assess some, but not all, care metrics. Healthcare costs and utilization, among other metrics, are still meaningful to payers and health systems. Though costs and use may not be perfect indicators of care quality, it remains reasonable to include these measures in a more comprehensive assessment of care. One could envision a portfolio of measures by which to capture the impact of a medical intervention or care model. For example, CMS has incorporated both patient-reported outcome performance measures and claims-based care utilization metrics to evaluate programs such as GUIDE and the Comprehensive Care for Joint Replacement Model [20]. A multidimensional approach to quality of life and home-time assessment may be the best way to capture what matters in the patient experience while also meeting the needs of policymakers and researchers. While home time may be an appealing and easily quantifiable metric, Van Houtven's careful work reminds us to be circumspect in its use, recognizing that it may not capture the nuances of patient preferences, variability in the meaning of “home,” and the complex social and medical factors that shape an older adult's experience. Without meaningful integration with other patient-reported outcomes, a home time metric will oversimplify and risk inaccurately conveying the experience it aims to reflect. As we strive to create useful measures, perhaps home time should be viewed as one piece of a broad framework for evaluating care, rather than as a universal proxy measure for quality of life. K.O., R.E.B., and D.S. contributed equally to manuscript preparation. The authors declare no conflicts of interest. This publication is linked to a related editorial by Van Houtven et al. To view this article, visit https://doi.org/10.1111/jgs.19506.

  PublisherOA PDFDOI

• Glucagon-Like Peptide-1 Receptor Agonists and Incidence of Dementia Among Older Adults With Type 2 Diabetes

  Annals of Internal Medicine · 2025-07-21 · 9 citations

  article

  BACKGROUND: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have been shown to decrease blood glucose levels, promote weight loss, and prevent cardiovascular events. However, evidence is limited regarding their effect on dementia, although emerging observational studies, some with serious methodological limitations, have suggested large reductions in dementia associated with GLP-1RAs that may not be entirely causally related. OBJECTIVE: To compare the effect of GLP-1RAs versus dipeptidyl peptidase-4 inhibitors (DPP4is) as second-line therapy for type 2 diabetes on risk for dementia among older adults. DESIGN: Target trial emulation. SETTING: United States from January 2016 to December 2020. PARTICIPANTS: Medicare fee-for-service beneficiaries aged 66 years or older with diabetes who used metformin and did not have dementia at baseline and initiated GLP-1RAs or DPP4is between January 2017 and December 2018. MEASUREMENTS: Onset of dementia was defined as 1 year before the date of a new dementia diagnosis. Risks were calculated at 30 months in GLP-1RA and DPP4i groups matched in a 1:2 ratio on an estimated propensity score and compared via ratios and differences. RESULTS: Among 2418 patients initiating GLP-1RAs and 4836 matched patients initiating DPP4is, the mean age was 71 years, and 55% were female. Over a median follow-up of 1.9 years, the outcome occurred in 96 patients in the GLP-1RA group and 217 in the DPP4i group. The estimated risk difference at 30 months was -0.93 (95% CI, -2.33 to 0.23) percentage points, and the estimated risk ratio was 0.83 (95% CI, 0.61 to 1.05). The estimated risk ratios were 0.64 (95% CI, 0.46 to 0.93) and 1.22 (95% CI, 0.74 to 1.66) among those younger than 75 years and aged 75 years or older, respectively. LIMITATIONS: Potential residual confounding (no data on body mass index, glycemic control, or duration of diabetes), outcome misclassification, and short follow-up. CONCLUSION: Among older adults with diabetes, no clear evidence was found that the incidence of dementia differed overall between patients using GLP-1RAs versus DPP4is. Under conventional statistical criteria, an effect of GLP-1RAs between a 39% decrease and a 5% increase in risk for dementia was highly compatible with the data, although estimates differed by age. Randomized trials are needed to quantify the effect of GLP-1RAs on dementia. PRIMARY FUNDING SOURCE: Gregory Annenberg Weingarten, GRoW @ Annenberg.

  PublisherDOI

• Medicare Part D: Major Shifts With the Inflation Reduction Act and a Way Forward

  Journal of the American Geriatrics Society · 2025-01-07 · 3 citations

  articleOpen accessSenior author

  The Inflation Reduction Act (IRA) of 2022 introduced major changes in the Part D benefit that aim to improve medication access and correct several of the financial misalignments in the current Part D benefit. The changes address financial obligations of Medicare beneficiaries, the federal government, Part D plan sponsors (i.e., insurance companies), and drug manufacturers. The changes include new brand and biologic manufacturer obligations to beneficiaries eligible for the low-income subsidy. Effects on the drug supply chain and stakeholder behaviors remain to be seen but current financial arrangements inform likely responses. Currently, the Pharmacy Benefit Managers (PBMs) nestled between the plans, drug manufacturers, and pharmacies heavily influence manufacturers' list prices and squeeze community pharmacies. With the IRA restructuring of Part D, plans are likely to interject more administrative obstacles before beneficiaries can obtain higher-cost therapies, while drug manufacturers might alter their patient assistance programs. Manufacturers have already begun to change their assistance programs at many safety net pharmacies. Many Medicare beneficiaries who relied on these deeply discounted medications will face significant late enrollment penalties if they do enroll in Part D plans, creating a major barrier to participation. Providers and policymakers should understand the Part D changes and leverage the skills of pharmacists to support community and team-based care that improves access to medications and ensures that medications are doing more good than harm.

  PublisherOA PDFDOI

• Self-efficacy change among diverse family caregivers in dementia care

  The Journals of Gerontology Series B · 2025-11-13

  article

  OBJECTIVES: To determine whether changes in caregiver self-efficacy (beliefs about one's ability to manage dementia-related problems and access help) differed by caregiver race and ethnicity across all participants enrolled in a large pragmatic trial of comprehensive dementia care. METHODS: In the Dementia Care Study (D-CARE), community-dwelling older adults with dementia and their unpaid family caregivers were randomized to receive usual care, community-based, or health system-based comprehensive dementia care. Caregiver self-efficacy was assessed at baseline, 6 months, and 18 months using a 4-item scale (range 4-20, higher scores indicate greater self-efficacy). Among all caregivers, we compared least squares means for overall change in self-efficacy from baseline across racial and ethnic groups (Latino, non-Latino Black, and non-Latino White). RESULTS: Among 2,126 dementia caregivers, 205 self-identified as Latino, 247 as non-Latino Black, and 1,674 as non-Latino White. Mean baseline caregiver self-efficacy scores were between 13.1 and 13.6 for all racial and ethnic groups. The least squares mean for self-efficacy change was between +1.53 and +1.66 from the baseline for all racial and ethnic groups. Caregiver self-efficacy change did not differ significantly by caregiver race and ethnicity. DISCUSSION: Black, Latino, and White dementia caregivers reported similar improvements in caregiver self-efficacy after participating in a comprehensive dementia care trial. Personalized aspects of comprehensive dementia care appear to address the needs of diverse caregiver populations. CLINICAL TRIALS REGISTRATION: NCT03786471.

  PublisherDOI

Frequent coauthors

• David A. Ganz

  147 shared

• Joel E. Streim

  146 shared

• Joseph G. Ouslander

  Florida Atlantic University

  114 shared

• John F. Schnelle

  114 shared

• Neil S. Wenger

  97 shared

• Maria Orlando Edelen

  Brigham and Women's Hospital

  91 shared

• Laurence Z. Rubenstein

  86 shared

• Shinya Ishii

  70 shared