About

Sarah Jane Ratcliffe, Ph.D., is an Adjunct Professor of Biostatistics in the Department of Biostatistics and Epidemiology at the University of Pennsylvania. Her educational background includes a B.Sc. (Hons) in Mathematics from the University of Technology, Sydney, Australia, obtained in 1996, and a Ph.D. in Statistics from Macquarie University, Sydney, Australia, completed in 2000. Her research expertise centers on biostatistics, with a focus on the development and application of statistical models for medical and health-related data. Her work includes the creation of models for ordinal outcomes with informative dropouts, analysis of multi-level correlated data, and the estimation of branching curves in the presence of subject-specific random effects. She has contributed to the field through various publications and collaborations, advancing statistical methodologies for longitudinal data, survival analysis, and quality indicators in healthcare.

Research topics

• Internal medicine
• Medicine
• Demography
• Materials science
• Intensive care medicine
• Family medicine
• Biology
• Ecology
• Environmental health
• Virology
• Emergency medicine

Selected publications

• Implementation of a pragmatic pathway to ensure high-quality processing of heterogeneous remote health sensing data

  Smart Health · 2026-03-07

  article
  PublisherDOI

• Short- and long-term metabolic exposure data as predicators of coronary microvascular dysfunction in a positron emission tomography myocardial perfusion imaging (PET-MPI) cohort with near concurrent angiography

  Biomarkers · 2026-03-06

  article

  BACKGROUND: Coronary microvascular disease (CMD) is defined by impaired myocardial stress flow reactivity and is associated with worse cardiovascular outcomes. Studying CMD is complicated by the overlap of its risk factors and patient-important cardiovascular sequelae with those of epicardial atherosclerotic disease. Published studies have not yet used longitudinal data to investigate the time dependencies of dynamic processes like obesity in their effects on microvascular health. METHODS AND RESULTS: = 0.048). Parallel modelling using single-time-point metabolomics data generated comparable results, suggesting that simplified assessments may be used as valid surrogates for repeated-measurement data in this setting.

  PublisherDOI

• A randomized controlled trial of artificial intelligence-based analytics for clinical deterioration

  Scientific Reports · 2026-02-05

  articleOpen access

  This pragmatic randomized controlled trial aimed to assess the effect of a passive display of artificial intelligence (AI)-based predictive analytics on hours free of clinical deterioration events among medical and surgical patients in an acute care cardiology medical-surgical ward. 10,422 inpatient visits were randomly assigned by cluster to the intervention group of a display of risk trajectories or to a control group of usual medical care. The trial was undertaken on an 85-bed inpatient cardiology and cardiac surgery ward of an academic hospital with a substantial implementation and education plan. This was a passive display with no specific response mandated. The primary analysis compared events of clinical deterioration (death, emergent ICU transfer, emergent endotracheal intubation, cardiac arrest, or emergent surgery) and compared mortality 21 days after admission. Patients with a large spike in risk score had, on average, twice the length of hospital stay (6.8 compared to 3.4 days). There was no change in the primary outcome between groups. Among those who had a clinical event, there were more event-free hours in the intervention/display-on group compared to the standard-of-care/display-off, but this did not reach statistical significance. Clinicians chose to transfer 11% of patients into or out of display beds, a censoring event removing them from the analysis, thereby undermining aspects of the randomized nature of the study. Predictive analytics monitoring incorporating continuous cardiorespiratory monitoring and displays of risk trajectories coupled with an education plan did not improve patient outcomes. While necessary to conduct the study, the pragmatic design allowed for significant movement towards intervention/displayed beds for sicker patients. Design considerations in the future must focus on understanding clinicians' interpretation, care processes, and communication practices.Clinical trial registration number: NCT04359641 Registered 4/24/20.

  PublisherOA PDFDOI

• A Pilot Study Testing the Clinical Congruency of Healthcare Providers to Use a Clinical Decision Support Tool to Assess Falls in Older Adults

  OBM Geriatrics · 2025-03-10

  articleOpen accessSenior author

  This research sought to determine feasibility for RNs to use the Post Fall Index™ (PFI) and to determine if an RN could identify underlying causal event factors for falls, would it be congruent with other providers (advanced practice nurse [APN], physician [MD])? PFI data from 23 falling residents of a nursing home were compiled into clinical vignettes and reviewed by experts for underlying causal event factors/fall sub-types. RNs used the PFI for one month in practice. The RN generated the most diagnoses; percent agreement was lower for RN: MD (between 37 to 87%) comparisons of fall sub-types versus APN: MD (between 57-87%). Significant agreement occurred between APN: MD for chronic problems (kappa = 0.060, p < 0.001) and equipment (p = 0.02), but not for RN: MD. RNs reported the PFI more precise. Although the PFI is feasible to use and an RN could identify underlying causes, percent agreement was higher for APN’s. Finding from this study indicate that three independent raters could generate similar fall related categories reinforcing a working assumption that clinical decision making for identifying specific fall related causal event factors maybe obtainable by multiple level providers when the correct tools are utilized.

  PublisherOA PDFDOI

• ACE-2-like enzymatic activity in COVID-19 convalescents with persistent pulmonary symptoms associated with immunoglobulin

  mBio · 2025-07-22

  articleOpen access

  Many difficult-to-understand clinical features characterize COVID-19 and post-acute sequelae of COVID-19 (PASC or long COVID [LC]). These can include blood pressure instability, hyperinflammation, coagulopathies, and neuropsychiatric complaints. The pathogenesis of these features remains unclear. The SARS-CoV-2 Spike protein receptor-binding domain (RBD) binds angiotensin converting enzyme 2 (ACE2) on the surface of host cells to initiate infection. We hypothesized that some people convalescing from COVID-19 may produce anti-RBD antibodies that resemble ACE2 sufficiently to have ACE2-like catalytic activity, that is, they are ACE2-like proteolytic abzymes that may help mediate the pathogenesis of COVID-19 and LC. In previous work, we showed that some people with acute COVID-19 had immunoglobulin-associated ACE2-like proteolytic activity, suggesting that some people with COVID-19 indeed produced ACE2-like abzymes. However, it remained unknown whether ACE2-like abzymes were seen only in acute COVID-19 or whether ACE2-like abzymes could also be identified in people convalescing from COVID-19. Here, we show that some people convalescing from COVID-19 attending a clinic for people with persistent pulmonary symptoms also have ACE2-like abzymes and that the presence of ACE2-like catalytic activity correlates with alterations in blood pressure in an exercise test. IMPORTANCE: Patients who have had COVID-19 can sometimes have troublesome symptoms, termed post-acute sequelae of COVID-19 (PASC) or long COVID (LC), which can include problems with blood pressure regulation, gastrointestinal problems, inflammation, blood clotting, and symptoms like "brain fog." The proximate causes for these problems are not known, which makes these problems difficult to treat definitively. We previously found that some acute COVID-19 patients make antibodies against SARS-CoV-2, the virus that causes COVID-19, that act like an enzyme, angiotensin converting enzyme 2 (ACE2). ACE2 normally helps regulate blood pressure and serves as the receptor for SARS-CoV-2 in the body. We show that patients convalescing from COVID-19 also make antibodies that act like ACE2 and that the presence of those antibodies correlates with problems in blood pressure regulation. The findings provide a new opening to potentially understanding the causes of LC, and so provide direction for the development of new treatments.

  PublisherDOI

• The use of memantine for prevention of Alzheimer's disease: Pilot feasibility study rationale and protocol

  Contemporary Clinical Trials Communications · 2025-01-01

  preprintOpen access

  Alzheimer’s disease (AD) affects over 6 million older adults in the Untiled States. Evidence suggests the neuropathology leading to the disorder begins decades earlier, calling for a preventative treatment that can be administered to at risk individuals. Memantine hydrochloride, an NMDA receptor antagonist, is a possible candidate for prophylactic treatment by diminishing excessive NMDA receptor activity. This is a 2-year, double-blind, randomized, placebo-controlled trial of memantine hydrochloride (1:1 randomization allocation using randomly permutated blocks of unequal size). Participants are APOε4 carriers slightly under the average age of AD symptom onset (50-65 years of age) with a positive family history of a first degree relative with AD. Amyloid PET scans are performed pre and post treatment. Cognitive assessments and physical and neurological examinations are completed at regular intervals throughout the feasibility trial. This study will assess the feasibility of the use of memantine hydrochloride for prevention of AD. The primary aim is to determine feasibility of participants who a) enrolled among those found eligible, and b) completed the study among those randomized to a study arm. Exploratory aims include examination of cognitive and safety assessments. Although not powered to determine efficacy, the study will provide direction on design elements needed for a Phase III clinical trial. No formal hypotheses are included in this feasibility trial. Clinical Trials NCT05063851

  PublisherDOI

• Splenic MARCO ⁺ marginal zone macrophages regulate rapid production of MCP-1 and KC but are dispensable for alloantibody generation in response to stored RBCs in a murine model

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-02-15

  preprintOpen access

  ABSTRACT BACKGROUND Alloimmunization to transfused red blood cells (RBCs) remains a significant clinical problem. However, the cells that initiate immune responses to transfused RBCs remain incompletely characterized. Recently published work has identified splenic marginal zone B (MZB) cells as being critically required for the production of anti-RBC alloantibodies in response to RBCs. In infectious models, MZB cell activation has been shown to depend on a unique population of marginal zone macrophages (MZMs). We hypothesized that MZMs would capture stored RBCs and present them to MZBs, and ultimately MZMs would be required for generation of anti-RBC alloantibodies in response to stored RBC transfusion. STUDY DESIGN AND METHODS Stored GFP + murine RBCs were utilized to determine the splenic localization and erythrophagocytosis by splenic macrophage populations. To determine the functional impact of MZMs, we compared LXRα-KO mice, which have been reported to lack MZMs, with wild type mice. Both innate and adaptive immune responses to stored HOD allogenic RBC transfusion were measured in LXRα-KO and wild type mice. RESULTS RBC storage leads to a significant increase in the phagocytosis of transfused RBCs by splenic MZMs. LXRα-KO mice demonstrated a lack of MZMs and had significantly decreased rapid phase production of cytokines MCP-1 and KC, but similar levels of IL-6. Surprisingly, anti-RBC alloantibody levels were unaffected by the absence of splenic MZMs. CONCLUSIONS Splenic MZMs are involved in the innate response to transfused stored HOD RBCs, contributing to both MCP-1 and KC cytokine production. However, MZMs are dispensable for anti-RBC alloantibody production.

  PublisherDOI

• Annual Costs and Nursing Care Interventions to Prevent Falls Over 1 Year in a Long Term Care Facility

  OBM Geriatrics · 2025-09-16

  articleOpen accessSenior author

  Evidence shows the use of interventions to prevent falls are costly to healthcare facilities. Using a sample of older adult patients who fell at least once during the intervention year of a three-year cohort study in one long term care nursing facility, at a continuing care community providing skilled nursing and assisted living, we provide detailed evidence of the number and costs of durable medical equipment and number and type of non- durable medical nursing care interventions utilized to prevent subsequent falls. This level of description can aid healthcare facilities and administrators in their plans to reduce recurrent falls among older adults.

  PublisherOA PDFDOI

• Economic impacts of artificial intelligence (AI)-based risk analytics for clinical deterioration

  Research Square · 2025-11-16

  preprintOpen access
  PublisherOA PDFDOI

• Analyzing alert patterns of artificial intelligence-based (AI) analytics: A visualization approach to engage clinicians and minimize alert fatigue (Preprint)

  2025-07-06

  preprint

  <sec> <title>UNSTRUCTURED</title> Despite the rapid proliferation of artificial-intelligence (AI)-based model development, few studies have prospectively implemented models in practice within the context of ongoing surveillance to predict clinical deterioration. To our knowledge, the concept of AI-based alert fatigue has not yet been introduced in this context. A necessary first step is to evaluate patterns of alerts to understand the system-level implications of using clinical deterioration alerts in practice and establish a process that enables clinicians to engage in setting alert thresholds. </sec>

  PublisherDOI

Recent grants

• NIH Grant R01GM104470

  NIH · $1.2M · 2018

• NIH Grant U01HD072906

  NIH · $5.5M · 2019

Frequent coauthors

• Thomas F. Floyd

  The University of Texas Southwestern Medical Center

  64 shared

• Anne Rentoumis Cappola

  54 shared

• Marc R. Blackman

  Veterans Health Administration

  53 shared

• Shalender Bhasin

  Brigham and Women's Hospital

  52 shared

• Jane A. Cauley

  University of Pittsburgh

  52 shared

• Joseph E. Bavaria

  University of Pennsylvania

  51 shared

• Joseph M. Zmuda

  University of Pittsburgh

  51 shared

• Rachel Weinstein

  Janssen (United States)

  51 shared