Population-based risk adjusted outcomes for out-of-hospital cardiac arrest

npj Cardiovascular Health · 2026-03-02

Abstract Out-of-hospital cardiac arrest (OHCA) impacts public health, with variable survival across the US. This study used a population-based risk adjustment model to understand factors influencing regional variability in OHCA survival to hospital discharge. We evaluated 202,406 OHCA cases from 2013-2015 Medicare Fee-For-Service claims across 205 hospital regions. A matched cohort from the Cardiac Arrest Registry to Enhance Survival (CARES) and Medicare claims was used to develop logistic regression models predicting survival. Standardized Incidence Ratios (SIRs) identified regions performing better or worse than expected. Of 205 regions, 101 (49.3%) demonstrated lower-than-expected risk-adjusted survival, while only 9 (4.4%) had higher-than-expected survival. Overperforming regions had smaller populations, higher proportions of residents aged 65 + , and more large hospitals (400+ beds). Hospitals with ≥100 beds were more likely in overperforming regions, while cardiac catheterization capability showed inverse association. These nationwide disparities highlight the need for targeted interventions and regionalized care approaches to improve survival rates.

The authors reply:

Critical Care Medicine · 2025-10-01

Hybrid Conjugates of Ibuprofen and 3,5‐Diarylidene‐4‐Piperidone: A New Avenue in Anti‐Inflammatory Drug Discovery

ChemMedChem · 2025-09-03 · 2 citations

Nonsteroidal anti-inflammatory drugs (NSAIDs) have been crucial in managing inflammation, pain, and fever since their introduction in 1897. Despite their widespread use, NSAIDs often face limitations due to gastrointestinal side effects from the nonselective inhibition of cyclooxygenase (COX) isoenzymes, COX-1 and COX-2. While selective COX-2 inhibitors reduce gastrointestinal toxicity, they come with increased cardiovascular risks. This study investigates the synthesis and biological evaluation of novel hybrid NSAID conjugates incorporating 3,5-diarylidene-4-piperidinone, ibuprofen, and amino acids. These hybrid molecules are designed to enhance anti-inflammatory efficacy while minimizing adverse effects. The synthesized compounds are evaluated for COX inhibition and their effects on inflammatory mediators such as interleukin-6, tumor necrosis factor-alpha, and nitric oxide. Computational studies, including molecular docking and ADME (absorption, distribution, metabolism, and excretion) analyses, are performed to clarify the mechanisms of action and to predict pharmacokinetic properties. The findings indicate that some hybrid conjugates display promising anti-inflammatory properties, necessitating further investigation for their potential therapeutic applications.

14: GEOGRAPHIC CLUSTERS IN SEPSIS MORTALITY AND THE ROLE OF TARGETED REGIONALIZATION: A SIMULATION STUDY

Critical Care Medicine · 2025-01-01 · 1 citations

Do Patient-Important Outcomes Differ by Care Setting? Findings from Semi-Structured Interviews with Individuals with Diabetes

Healthcare · 2025-12-01

Background: Patient-important outcomes (PIOs) reflect patient values and preferences. Prior studies have elicited a variety of PIOs for diabetes. However, no studies have examined whether, or how, PIOs differ across diabetes care settings. The purpose of this study was to compare the frequencies of PIOs derived from patients with diabetes in primary care (PC), acute care (emergency department (ED)), and post-acute care (post-hospital discharge (PHD)) settings within a large delivery system. Methods: This study was an analysis of 89 interviews with patients in PC, ED, and PHD settings. Participants had moderately to poorly controlled diabetes, defined as follows: presented to the ED with a diabetes-related problem, admitted to the hospital for a diabetes-related problem, or had at least two primary care measurements of hemoglobin A1c (HbA1c) > 7.5 in the prior year. A matrix analysis compared the frequencies of participants’ PIOs across the three settings. Results: Overall PIO frequencies were similar across care settings. PIOs fell into seven domains; all seven domains and 21 of the 26 PIOs were represented within each of the care settings. The most common PIOs included “be healthy”, “eat right”, and “reduce or get off medicines”. Conclusions: Participants identified similar PIOs in all care settings, indicating that recruitment from one or two care settings may often be sufficient for achieving saturation of PIOs. Furthermore, the results inform our understanding of patient priorities across the care continuum.

Geographic Clusters in Sepsis Hospital Mortality and the Role of Targeted Regionalization

Critical Care Medicine · 2025-04-24 · 3 citations

OBJECTIVES: Sepsis is a severe condition associated with high mortality, and hospital performance is variable. The objective of this study was to develop geospatial sepsis clusters, identify sources of variation between clusters, and test the hypothesis that redistributing sepsis patients from low-performing hospitals to higher-performing hospitals within a cluster will improve sepsis outcomes. DESIGN, SETTING, AND PATIENTS: We conducted a cohort study of age-qualifying Medicare beneficiaries using administrative claims data from 2013 to 2015. We calculated risk-standardized mortality for hospitals then used a clustering algorithm to define geospatial cluster boundaries based on care-seeking and interhospital transfer patterns. Finally, we used simulation to model the effect of reallocating sepsis patients to higher-performing hospitals within the same cluster. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We included 1,125,308 patients, and they were grouped into 222 regional clusters. High-performing clusters were located largely in the Midwest, and they tended to be in less urban regions with smaller hospitals. In our simulation, the most impactful strategy was reassigning cases from the lowest-performing hospital in a cluster to the highest-performing hospital in the cluster, which was predicted to prevent 1705 deaths per year in the United States. This aggregate benefit was lower than the 5702 deaths predicted from reducing mortality by 1% absolute in hospitals in the lower half of the performance distribution. CONCLUSIONS: Geospatial clusters provide insight into regional approaches to system-based acute care. In a simulation study, targeted sepsis regionalization appears less effective than local performance improvement in reducing preventable sepsis deaths.

Using telematics data to evaluate safety policies: a case study of Chicago’s red-light camera programme

Injury Prevention · 2025-08-27

BACKGROUND: Mobile telematics offers a promising new data source for evaluating safety interventions, providing detailed information about driving behaviour and safety events. We examined whether telematics data could effectively evaluate the impact of red-light cameras on driver behaviour and crash risk. METHODS: We analysed mobile telematics data from over 770 000 users in Chicago to assess how the presence of a red-light camera at an intersection approach affected the likelihood of collisions and harsh braking. We matched intersection approaches with and without cameras on the number of lanes, speed limit, traffic volume and segment length. We used negative binomial regression models to evaluate the impact of cameras on collisions and harsh braking by time of day and season. FINDINGS: Harsh braking events occurred 24 times more frequently than collisions and showed remarkably similar patterns of association with environmental factors. Both showed higher frequency during rush hour (11% and 23% increases, respectively), lower at night (73% and 80% decreases) and increasing frequency with more lanes. These effects were consistent across seasons and time of day. Cameras reduced both collisions (25% reduction; 95% CI 15% to 34%) and harsh braking events (21% reduction; 95% CI 12% to 28%). INTERPRETATION: Telematics data show effects of cameras that are consistent with past evaluations. Furthermore, there was close correspondence between collision and harsh braking patterns. Together, these suggest that telematics-reported data provide a surrogate measure for road safety and can provide richer information for safety evaluation in settings where crash data are sparse, though the inability to distinguish injury severity remains a limitation.

Assessing Retrieval-Augmented Large Language Models for Medical Coding

NEJM AI · 2025-09-25 · 2 citations

Identifying Bias at Scale in Clinical Notes Using Large Language Models

Mayo Clinic Proceedings Digital Health · 2025-10-15 · 1 citations

Objective: To evaluate whether generative pretrained transformer (GPT)-4 can detect and revise biased language in emergency department (ED) notes, against human-adjudicated gold-standard labels, and to identify modifiable factors associated with biased documentation. Patients and Methods: We randomly sampled 50,000 ED medical and nursing notes from the Mount Sinai Health System (January 1, 2023, to December 31, 2023). We also randomly sampled 500 discharge notes from the Medical Information Mart for Intensive Care IV database. The GPT-4 flagged 4 types of bias: discrediting, stigmatizing/labeling, judgmental, and stereotyping. Two human reviewers verified model detections. We used multivariable logistic regression to examine associations between bias and health care utilization, presenting problems (eg, substance use), shift timing, and provider type. We then asked physicians to rate GPT-4's proposed language revisions on a 10-point scale. Results: The GPT-4 showed 97.6% sensitivity and 85.7% specificity compared with the human review. Biased language appeared in 6.5% (3229 of 50,000) of Mount Sinai notes and 7.4% (37 of 500) of Medical Information Mart for Intensive Care IV notes. In adjusted models, frequent health care utilization (adjusted odds ratio [aOR], 2.85; 95% CI, 1.95-4.17), substance use presentations (aOR, 3.09; 95% CI, 2.51-3.80), and overnight shifts (aOR, 1.37; 95% CI, 1.23-1.52) showed elevated odds of biased documentation. Physicians were more likely to include bias than nurses (aOR, 2.26; 95% CI, 2.07-2.46); GPT-4's recommended revisions received mean physician ratings above 9 of 10. Conclusion: The study showed that GPT-4 accurately detects biased language in clinical notes, identifies modifiable contributors to that bias, and delivers physician-endorsed revisions. This approach may help mitigate documentation bias and reduce disparities in care.

Disparities by Race and Ethnicity in Percutaneous Coronary Intervention

JAMA Network Open · 2025-09-18 · 4 citations

Importance: Hispanic and non-Hispanic Black patients with ST-segment elevation myocardial infarction (STEMI) are less likely than White non-Hispanic patients to receive guideline-recommended percutaneous coronary intervention (PCI). Research suggests disparities arise before and during STEMI treatment, but it is unclear when the largest disparities in PCI emerge. Objective: To assess when in the care process the largest disparities in PCI receipt occur in patients with STEMI presenting to an emergency department. Design, Setting, and Participants: This cross-sectional study evaluated adult patients with STEMI presenting to Florida hospitals from January 1, 2011, to December 31, 2021. Data were analyzed from June 29, 2023, to May 29, 2025. Exposure: Patient race and ethnicity. Main Outcomes and Measures: The main outcomes were presentation to PCI-capable hospitals, receipt of PCI if initially presenting to PCI-capable hospitals, transfer if initially presenting to non-PCI capable hospitals, and receipt of PCI at receiving hospital if transferred. Logistic regression was used to compare outcomes for patients with STEMI by race and ethnicity, controlling for payer, age, sex, weekend presentation, time of presentation, comorbidities, and hospital characteristics. Results: Among 139 629 patients with STEMI included in the analysis, 68.81% were male. Mean (SD) age was 64.4 (13.0) years. A total of 9.09% identified as Black, 15.17% as Hispanic, 70.56% as White, and 5.17% as other or missing race. In adjusted analyses, Black (-1.8 [95% CI, -2.6 to 1.1] percentage points [pp]) and Hispanic (-3.1 [95% CI, -3.7 to -2.4] pp) patients were less likely than White patients to present to PCI-capable hospitals (P < .001 for both). Among patients initially presenting to PCI-capable hospitals, Black patients were less likely to receive PCI than White patients (-8.6 [95% CI, -9.5 to -7.7] pp; P < .001). Among patients initially presenting to non-PCI-capable hospitals, Black (-4.0 [95% CI, -6.4 to -1.5] pp; P = .001) and Hispanic (-4.2 [95% CI, -6.3 to -2.0] pp; P < .001) patients were less likely to be transferred than White patients. Among transferred patients, Black patients were less likely to undergo PCI at the receiving hospital than White patients (-13.3 [95% CI, -16.6 to -9.9] pp; P < .001). Conclusions and Relevance: In this cross-sectional study examining racial and ethnic disparities in receipt of PCI for patients with STEMI, racial and ethnic disparities persisted throughout the care process. The largest magnitude of disparity was PCI receipt if transferred, but the disparity with the largest impact was PCI receipt when initially presenting to PCI-capable hospitals.