Category: Statistics

J Med Internet Res. 2026 Jul 20;28:e90023. doi: 10.2196/90023.

ABSTRACT

BACKGROUND: AI is rapidly transforming medical practice, with emerging applications in perioperative care and anesthesiology. However, the clinical implementation of AI-assisted decision-making systems in anesthetic management remains challenging and requires comprehensive evaluation.

OBJECTIVE: This study aimed to assess the performance and clinical applicability of an AI-assisted decision support system (ZW-AA-001) for general anesthesia management by comparing its decisions with those of experienced anesthesiologists across 6 medical centers.

METHODS: A multicenter retrospective study was conducted using perioperative data from 1008 patients who underwent elective noncardiac surgeries under total intravenous anesthesia. The AI system’s recommendations for anesthetic and hemodynamic medication adjustments were compared with anesthesiologists’ decisions. Key outcomes included decision concordance rates, temporal performance, and consistency across centers. Advanced statistical methods, including prevalence-adjusted and bias-adjusted κ (PABAK) and Gwet’s first-order agreement coefficient (AC1), were used to evaluate agreement metrics.

RESULTS: The study included 1008 patients, with a median age of 50 (IQR 37-59) years and female predominance (619/1008, 61.4%). During anesthesia maintenance, the AI system demonstrated moderate overall decision agreement with anesthesiologists (73.3%, 95% CI 72.4%-74.3%). Analysis of center-specific data revealed generally consistent performance across all 6 centers. For propofol management, high concordance was observed in dosage adjustment decisions (91.1%, 95% CI 90.4%-91.8%), though fair agreement was found in adjustment direction (68.6%, 95% CI 67.3%-69.8%). The AI system showed significantly faster decision-making time compared to anesthesiologists for the adjustment of propofol (pseudomedian difference -77.5, 95% CI -79.5 to -75.5 seconds; P<.001). Although esmolol-related decisions showed a numerically higher concordance of 71.4%, this was not statistically significant (PABAK=0.429; P=.21; AC1=0.622; P=.09). Decisions for atropine, ephedrine, and urapidil demonstrated substantially lower agreement (17.4%-29.8%). The AI system recommended hemodynamic interventions more frequently than anesthesiologists across all medications.

CONCLUSIONS: The AI-assisted decision support system demonstrated varying levels of concordance with anesthesiologists in managing surgery, with the highest agreement observed in propofol administration. However, the system’s lower agreement in hemodynamic medication management highlights the need for further optimization and validation. These findings underscore the potential of AI systems to enhance anesthetic decision-making, improve efficiency, and address workforce challenges in anesthesiology. Future research should focus on expanding the system’s capabilities, validating its performance in prospective clinical trials, and ensuring its integration into diverse clinical settings.

PMID:42475676 | DOI:10.2196/90023

J Med Internet Res. 2026 Jul 20;28:e85403. doi: 10.2196/85403.

ABSTRACT

BACKGROUND: Health literacy is crucial for pregnancy outcomes; yet, 15%-44% of pregnant women have low health literacy, negatively affecting maternal and fetal health. While pregnancy apps are increasingly used to support it, the psychological mechanisms through which they influence behavior remain unclear. Understanding these mechanisms is essential for designing effective interventions. Cognitive load theory (CLT) and the integrated model of cognitive-affective learning with media (ICALM) provide theoretical frameworks for understanding these digital interventions in pregnancy’s emotionally intense context.

OBJECTIVE: This study aimed to understand the psychological and social-cognitive pathways through which engagement with pregnancy apps may enhance health literacy and behaviors.

METHODS: This cross-sectional survey study included 998 pregnant and early postpartum women aged 18-45 years from the United States and analyzed the data using Bayesian Network (BN) modeling informed by CLT and ICALM. Participants were stratified into Pregnancy+ (Royal Philips) users (n=362), other pregnancy app users (n=507), and nonapp users (n=129). We measured digital engagement, emotions, health literacy levels, support levels and sources, and health behaviors, including physical activity, nutrition, and stress management. The BN model underwent systematic refinement through bootstrap resampling (n=1000) to evaluate arc strength, followed by k-fold cross-validation (k=5) to assess model generalizability and predictive performance.

RESULTS: In our study, pregnancy apps improve health literacy or behaviors, but not through direct pathways. Instead, they work through two indirect psychological pathways: (1) digital engagement related to positive emotions, which in turn is related to health literacy and stress management, and (2) digital engagement strengthened support networks, which in turn are related to health literacy. No direct pathways from digital engagement to health literacy or behaviors were retained. Pregnancy+ users demonstrated stronger adherence to these pathways compared with other app users and nonusers, by demonstrating more joyfulness (5.1 vs 4.7; P=.046), elevated support levels (4.1 vs 3.9; P=.03), and higher critical health literacy scores (8.4 vs 8.2 vs 7.7; P<.001). Pregnancy+ users also showed higher engagement frequencies (P<.001) and more frequently cited pregnancy apps as valuable information sources (78% vs 68%, P=.01) compared with users of other pregnancy apps.

CONCLUSIONS: This study reveals that pregnancy apps may enhance health literacy by fostering positive emotions and strengthening support networks. This finding challenges assumptions about digital health interventions and demonstrates that apps function as psychological support systems rather than mere information sources. The findings validate CLT and ICALM within pregnancy contexts, showing that emotionally supportive apps reduce irrelevant emotional processing that consumes cognitive capacity, thereby enhancing information retention. For practice, app developers should prioritize emotional support and social connection over content volume, while health care providers should recommend apps that enhance emotional well-being. This study demonstrates BN’s value for revealing complex psychological relationships in digital health contexts.

PMID:42475675 | DOI:10.2196/85403

JMIR Med Educ. 2026 Jul 20;12:e91345. doi: 10.2196/91345.

ABSTRACT

BACKGROUND: AI is transforming medicine by enhancing care, reducing administrative tasks, and facilitating research. AI also raises many concerns, including a lack of clinical context awareness, data dependence, and the absence of ethical judgment. As future practitioners, medical students must be prepared for these changes. Most studies assessing students’ attitudes and knowledge were conducted before AI became accessible and tailored to the needs of the population. Therefore, how medical students actually use AI remains largely unexplored.

OBJECTIVE: This study aimed to explore French medical students’ perceptions, attitudes, and use of AI.

METHODS: A mixed methods study was conducted in 2025 among French medical students in their clerkship year. An online survey included open-ended questions about the definition of AI and feelings toward AI, a Likert scale item to assess specific attitudes, and multiple-choice questions about student characteristics. Quantitative analysis was performed using Kruskal-Wallis tests, chi-square tests, and exploratory multivariable linear regression to examine associations between AI knowledge, attitudes, and student characteristics. Qualitative thematic analysis was conducted inductively on open-text responses regarding perceptions of AI, feelings, training expectations, and use.

RESULTS: Of the 1377 responses received, 1342 were included. Students had a median age of 23 (IQR 22-24) years and were predominantly in their fifth year. Only 5% (67/1342) provided a correct definition of AI, while 57.4% (770/1342) gave incorrect responses. Attitudes toward AI were generally positive, with a median score of 7 (IQR 5-8). Students with unknown AI definitions had significantly lower attitude scores (P=.02), although the magnitude of the difference was small (ε²=0.005). In multivariable analysis, belonging to the unknown AI definition category remained associated with a lower general attitude toward AI score compared with the incorrect category (regression coefficient β=-0.72, 95% CI -1.17 to -0.28; P=.001). Regarding education, 48.5% (651/1342) of students preferred AI training outside the formal curriculum. Qualitative analysis revealed 5 themes: representation, nuanced optimism, critical consideration, replacement, and AI use. Students describe AI as a robot, an improved search engine, or an unlimited data source. Their nuanced optimism blends enthusiasm for efficient patient care and the opportunity to focus more on the patient relationship, with concerns about dehumanization, energy costs, and skill regression. Critical consideration underscores distrust from ethical dilemmas and data security risks. Replacement concerns arise from shifting professional roles, though many believe human empathy remains irreplaceable. Regarding AI use, students highlight its potential for administrative aid, personalized training, and clinical support.

CONCLUSIONS: Medical students report generally positive but varied attitudes toward AI, despite having a limited understanding of its foundations. Ecological concerns, fears of skill loss, and ethical issues coexist with widespread self-directed use of AI for learning. These findings provide a descriptive baseline for future evaluations of AI-related training, particularly regarding critical appraisal, ethical issues, and self-directed AI use.

PMID:42475670 | DOI:10.2196/91345

Neurology. 2026 Aug 11;107(3):e218297. doi: 10.1212/WNL.0000000000218297. Epub 2026 Jul 20.

ABSTRACT

BACKGROUND AND OBJECTIVES: Mild cognitive impairment (MCI) is a high-risk stage for progression to dementia. The SUPERBRAIN-MEET trial previously demonstrated cognitive benefits of a multidomain lifestyle intervention (MLI) delivered through face-to-face sessions and digital platforms. The aim of this study was to examine whether the cognitive effects of the MLI varied according to sociodemographic characteristics, digital readiness, baseline cognitive status, and vascular risk profiles, in a prespecified subgroup analysis of the SUPERBRAIN-MEET randomized controlled trial.

METHODS: SUPERBRAIN-MEET was a 24-week, multicenter, outcome assessor-blinded, parallel-group randomized controlled trial conducted across 17 hospitals in South Korea. Adults aged 60-85 years with MCI were randomly assigned (1:1) to the MLI or control group. Subgroups were defined based on sociodemographic factors (sex, age, education, and living status), digital readiness (internet access, frequency of digital device use), cognitive status (MMSE score), and vascular risk profiles (body mass index [BMI], obesity status, smoking status, systolic and diastolic blood pressure, serum total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, fasting plasma glucose, and cardiovascular comorbidity). The primary outcome was the change in the repeatable battery for the assessment of neuropsychological status (RBANS) total index score from baseline to 24 weeks. Effect modification was assessed using linear mixed-effects models with interaction terms and false discovery rate correction.

RESULTS: Of 300 randomized participants, 277 participants (mean age, 72.9 years; 69.3% female) were included in the modified intention-to-treat analysis (142 in the MLI group and 135 in the control group). MLI significantly improved RBANS total index scores compared with control at 24 weeks (between-group difference, 4.17 points; 95% CI 1.92-6.43). No subgroup interactions remained significant after false discovery rate correction, indicating consistent effects across subgroups.

DISCUSSION: In this subgroup analysis of the SUPERBRAIN-MEET trial, the cognitive benefits of MLI were not significantly modified by sociodemographic factors, digital readiness, baseline cognitive status, or vascular risk profiles. These findings should be interpreted with caution, given the exploratory nature of subgroup analyses and limited statistical power. The results support the generalizability and scalability of technology-assisted multidomain lifestyle interventions for dementia prevention in older adults with MCI.

TRIAL REGISTRATION INFORMATION: The trial was registered at ClinicalTrials.gov (NCT05023057; submitted August 20, 2021; first patient enrolled September 3, 2021; URL: clinicaltrials.gov/study/NCT05023057).

PMID:42475653 | DOI:10.1212/WNL.0000000000218297

Neurology. 2026 Aug 11;107(3):e218284. doi: 10.1212/WNL.0000000000218284. Epub 2026 Jul 20.

ABSTRACT

BACKGROUND AND OBJECTIVES: Degeneration of the central autonomic network with relative sparing of peripheral autonomic neurons underlies neurogenic orthostatic hypotension in patients with multiple system atrophy (MSA). Ampreloxetine, a novel, selective, norepinephrine (NE) reuptake inhibitor, allows once-daily dosing to precisely target residual peripheral autonomic neurons. Based on the hypothesis that patients with MSA would be most responsive and the substantial unmet need for symptomatic therapy in this population, an MSA subgroup analysis was prespecified.

METHODS: We conducted a run-in 4-week, parallel-group, randomized controlled trial (SEQUOIA), followed by a pivotal enriched randomized withdrawal (RW) trial with 16-week open-label treatment and 6 weeks of 1:1 RW (REDWOOD). Inclusion criteria for the MSA subgroup included (1) probable or possible MSA, (2) 3-minute orthostatic blood pressure (BP) fall >20/10 mm Hg, and (3) dizziness or lightheadedness score >4 points. Outcome measures included self-reported symptom burden captured on the 10-item OH Questionnaire (OHQ). Differences were analyzed using logistic regression and mixed-model repeated measures analysis.

RESULTS: Seventy-three patients with MSA entered the program (mean age 63 years old [range: 43-80], 52% male). Both SEQUOIA and REDWOOD did not meet their primary endpoints. In REDWOOD, 40 (61%) fulfilled enrichment criteria and were randomized. After 16-week open-label, OHQ symptom assessment (OHSA) composite domain scores improved 2.6 ± (SD = 2.1) points from pretreatment. The proportion of participants with treatment failure at week 6 of RW treatment period was 40% in the placebo arm and 15% in the ampreloxetine arm (p = 0.11). In secondary endpoints, at week 6 of RW, symptoms remained stable in the ampreloxetine group, but worsened on placebo (mean difference OHSA composite: -1.6 points ± 0.5; p = 0.0056, minimal clinically important worsening = 0.7-1.1 points). Standing for a short time favored ampreloxetine (-2.0 points ± 0.8; p = 0.015). Standing BP remained unchanged from open-label in the ampreloxetine group (systolic: 5.6 ± 4.1; diastolic: 3.7 ± 2.9 [SE] mm Hg) but fell after placebo withdrawal (systolic: -10.0 ± 4.5; diastolic: -6.0 ± 3.1 mm Hg). The catecholamine profile was consistent with NE transporter inhibition. There were no observed increases in supine BP.

DISCUSSION: In a prespecified subgroup analysis of MSA participants in the REDWOOD trial, patients randomized to placebo worsened, whereas those who were randomized to treatment maintained their open-label level of function.

TRIAL REGISTRATION INFORMATION: REDWOOD trial, NCT03829657; first submitted to registry January 10, 2019; first participant enrolled February 22, 2019. SEQUOIA trial, NCT03750552; first submitted to registry November 20, 2018; first participant enrolled January 24, 2019. See ClinicalTrials.gov for full-protocol and statistical analysis plan.

CLASSIFICATION OF EVIDENCE: This study provides Class III evidence that in patients with MSA who had symptomatic benefit on orthostatic hypotension with ampreloxetine, there was no difference in the odds of treatment failures between those maintained on ampreloxetine and those withdrawn to placebo.

PMID:42475649 | DOI:10.1212/WNL.0000000000218284

Am J Phys Med Rehabil. 2026 Jul 21. doi: 10.1097/PHM.0000000000003083. Online ahead of print.

ABSTRACT

OBJECTIVE: The aim of this study is to systematically evaluate and summarize the evidence related to NMES-assisted rehabilitation of patients with ankle injuries.

DESIGN: A systematic search of eight Chinese and English databases was conducted to identify randomized controlled trials investigating NMES-assisted rehabilitation for ankle injuries. Studies were screened and relevant information extracted from included publications. Meta-analysis was performed using statistical software RevMan 5.3 and STATA 17.0.

RESULTS: A total of 7010 articles were retrieved from the literature, and 23 studies were finally included. Meta-analysis showed that the NMES group had the best results in terms of CAIT, VAS, AOFOS, clinical efficacy, area of motion ellipse, valgus relative peak moment value, dorsiflexion relative peak moment value, plantarflexion relative peak moment value, step speed, step length, single-support phase and double-support phase; for the total stability index, the improvement was not significantly different from that of the control group. The meta-analysis results revealed no significant evidence of publication bias.

CONCLUSION: The results of this review suggest that NMES may have some potential role in assisting rehabilitation of ankle injuries, and that a longer course of NMES may have better efficacy.

PMID:42475641 | DOI:10.1097/PHM.0000000000003083

Ann Surg. 2026 Jul 20. doi: 10.1097/SLA.0000000000007152. Online ahead of print.

ABSTRACT

OBJECTIVE: To evaluate the impact of CYP2A6 genetic polymorphisms on survival outcomes in patients with pancreatic ductal adenocarcinoma (PDAC) receiving perioperative S-1 therapy.

BACKGROUND: S-1, an oral fluoropyrimidine containing tegafur, is widely used for perioperative treatment of PDAC in Japan. Tegafur requires metabolic activation to 5-fluorouracil, primarily by CYP2A6. Given the high prevalence of reduced-function CYP2A6 variants in East Asian populations, their clinical significance in PDAC remains unclear.

METHODS: This retrospective study included 145 patients with PDAC who underwent curative resection and received adjuvant S-1, including 78 patients who had received S-1-containing neoadjuvant therapy. Germline CYP2A6 variants were identified using whole-exome sequencing of peripheral blood samples. Decreased-function alleles (*4, *7, *9) and the rare nonsense allele *48 were evaluated. Patients were classified into poor, intermediate, and normal metabolizer groups. Survival outcomes were analyzed using Kaplan-Meier and Cox regression analyses.

RESULTS: Decreased-function CYP2A6 genotypes were identified in 82 patients (56.6%), including 6 poor metabolizers (4.1%). These patients had a higher frequency of lymph node metastasis than those with normal genotypes (78.0% vs. 52.4%, P=0.001). Recurrence-free survival (RFS) was significantly shorter (median, 19.2 vs. 32.3 mo; P=0.040), whereas overall survival (OS) was shorter but not statistically significant (47.9 vs. 62.2 mo; P=0.104). In multivariable analysis, CYP2A6 poor metabolizer status was independently associated with OS (HR, 3.17; 95% CI, 1.06-9.46; P=0.039). Among the 78 patients who received S-1-containing neoadjuvant therapy, both OS and RFS were significantly worse in those with decreased-function genotypes.

CONCLUSIONS: Decreased CYP2A6 functional genotypes were associated with poorer oncological outcomes in patients with PDAC receiving perioperative S-1 therapy, suggesting that the CYP2A6 genotype may serve as a biomarker for optimizing treatment strategies.

PMID:42475636 | DOI:10.1097/SLA.0000000000007152

JCO Clin Cancer Inform. 2026 Jul-Sep;10(3):e2500385. doi: 10.1200/CCI-25-00385. Epub 2026 Jul 20.

ABSTRACT

PURPOSE: Guidelines recommend germline testing in advanced prostate cancer (PCa) to inform treatment and personal/familial cancer risk, yet Black patients are less likely than White patients to complete testing. Artificial intelligence (AI) tools are increasingly used in pretest education to improve access but remain unevaluated for racial bias.

MATERIALS AND METHODS: We developed ProGene, a secure, generative AI chatbot for PCa germline testing education in Black patients. We prompted ProGene with seven questions about testing types, personal benefits, family benefits, drawbacks, logistics, costs, and privacy. Each question was asked nine times across three patient vignettes (Black, non-Hispanic White, race-agnostic), in triplicate. Two blinded reviewers assessed responses across five domains: (1) comprehensiveness (0%-100%), (2) accuracy (presence/absence of inaccuracies), (3) readability (grade level via Simple Measure of Gobbledygook [SMOG] and Flesch-Kincaid), (4) actionability (0%-100% via Patient Education Materials Assessment Tool), and (5) quality (1-18 via DISCERN-AI). Outcomes were compared by race and question using two-sample t-tests or Wilcoxon rank-sum tests for continuous measures and chi-square or proportion tests for categorical measures; ANOVA was used for question-level comparisons.

RESULTS: The mean comprehensiveness was 67% and did not vary by race, although responses for question-1 (genetic testing types) were less comprehensive for Black versus race-agnostic vignette (60% v 93%, P < .01). Inaccuracies appeared in 32% of responses, primarily related to sample collection and cost/insurance, and did not vary by race. The mean readability were 10th (SMOG) and 13th (Flesch-Kincaid) grades, actionability 92%, and DISCERN score 14/18 (good quality); none varied by race.

CONCLUSION: We identified no statistically significant racial disparities across the five evaluation domains in ProGene. Responses were generally good quality and actionable. Although AI may facilitate equity in PCa genetic education, deficiencies in comprehensiveness, accuracy, and readability highlight the need for refinement and/or human oversight with implementation.

PMID:42475635 | DOI:10.1200/CCI-25-00385

J Am Coll Health. 2026 Jul 20:1-11. doi: 10.1080/07448481.2026.2693666. Online ahead of print.

ABSTRACT

Objective: This study aimed to measure the effect of a Social Cognitive Theory (SCT)-based, urban gardening, cooking, and nutrition education intervention on health behavior mediators and fruit and vegetable intake in food-insecure college students. Methods: In a 6-week, randomized controlled trial, 89 food-insecure students at a southeastern university (86% female, 59% Hispanic), randomized into a controlled 107 (n = 42) or intervention group (n = 47), attended once a week, cooking/gardening sessions, that incorporated nutrition education, recipes, and at home garden materials. Outcome measures were assessed at baseline and post intervention and included anthropometrics, health behavior mediators such as nutrition knowledge (Dickson-Spillmann Consumer Oriented Nutrition Knowledge Survey), Dewar’s SCT construct subscales: self-efficacy, outcome expectations, reciprocal determinism, and fruit and vegetable intake using the National Cancer Institute’s Fruit and Vegetable Screener. Linear regressions and ANOVA’s were conducted. Results: Intervention participants increased fruit and vegetable intake (F = 19.19, p < .001) and health behavior mediators: self-efficacy (F = 17.65, p < .001), reciprocal determinism (F = 4.16, p = .045), An increase in reciprocal determinism was significantly associated with an increase in self-efficacy (β = .392, t = 2.66, p = .011). Intervention mediators did not have a significant effect on change in fruit and vegetable intake (F = . 580, p = .679, R2 = . 061). Conclusions: An SCT-based cooking and gardening nutrition program tailored to food-insecure college students improved health behavior mediators and increased fruit and vegetable intake. A nutrition education program that utilizes a hands-on, interactive approach may be an effective intervention to promote fruit and vegetable intake and health behaviors among food insecure college students.

PMID:42475621 | DOI:10.1080/07448481.2026.2693666

J Med Internet Res. 2026 Jul 20;28:e93347. doi: 10.2196/93347.

ABSTRACT

BACKGROUND: Patients with atrial fibrillation (AF) face significant bleeding risks, particularly those receiving oral anticoagulation; however, existing risk scores such as HAS-BLED and ORBIT demonstrate limited predictive accuracy. Pulse pressure (PP), calculated as the difference between systolic blood pressure (SBP) and diastolic blood pressure, is a noninvasive marker of arterial stiffness that has been associated with cardiovascular outcomes. However, PP has not been evaluated as a predictor of bleeding in this population.

OBJECTIVE: This study evaluated whether elevated PP independently predicts major bleeding events, overall and by subtype, in patients with AF after adjusting for established clinical risk factors.

METHODS: We conducted a retrospective cohort study using electronic health records from REACHnet, a PCORnet-affiliated clinical data network in Louisiana. A total of 4935 adults (mean age 63.7, SD 11.0 y; n=1606, 32.5% female) with AF between 2010 and 2019 were included via consecutive sampling of all eligible patients. PP was derived from outpatient blood pressure measurements closest to AF diagnosis and analyzed in tertiles (low: <46, middle: 46-62, high: >62 mm Hg) and continuously per 10 mm Hg. The primary outcome was time to the first bleeding event, a composite of gastrointestinal bleeding, intracranial hemorrhage, and other clinically significant bleeding, identified using ICD-9/ICD-10 codes. Kaplan-Meier survival curves with log-rank testing were used for univariable analysis. Multivariable Cox proportional hazards regression was adjusted for age, sex, race, comorbidities, medications, and the ORBIT score. A sensitivity analysis applied multivariable logistic regression additionally incorporating SBP. Statistical significance was set at P<.05.

RESULTS: Over a 5-year follow-up, 677 out of 4935 (13.7%) patients experienced a bleeding event (intracranial hemorrhage: n=60, 1.2%; gastrointestinal bleeding: n=195, 4.0%; and other bleeding: n=149, 3.0%). Gastrointestinal bleeding differed significantly across PP tertiles (P=.007). Kaplan-Meier analysis confirmed lower gastrointestinal bleeding-free survival in the highest tertile (log-rank P=.004). No significant differences were observed for intracranial (P=.08), other (P=.58), or composite bleeding (P=.22). In multivariable Cox regression, each 1 mm Hg increase in PP was independently associated with a 1.4% higher gastrointestinal bleeding risk (hazard ratio 1.014, 95% CI 1.001-1.028; P=.04), approximately 15% per 10 mm Hg. This remained significant after adjusting for SBP and ORBIT score (odds ratio 1.013/mm Hg, 95% CI 1.001-1.025; P=.03), while SBP was not independently significant (P=.13).

CONCLUSIONS: PP independently predicts gastrointestinal bleeding risk in patients with AF beyond established clinical risk factors and validated bleeding risk scores. Unlike prior investigations that examined SBP or diastolic blood pressure components in isolation, this is the first study to identify PP as a predictor of gastrointestinal bleeding in this population. As a readily available, low-cost hemodynamic parameter derived from routine clinical measurements, PP could enhance existing risk stratification tools and inform more personalized bleeding risk management strategies in patients with AF.

PMID:42475619 | DOI:10.2196/93347