Explore the vast range of titles available.

The objective of this study was to design and develop a predictive model for 30-day risk of hospital readmission using machine learning techniques. The proposed predictive model was then validated with the two most commonly used risk of readmission models: LACE index and patient at risk of hospital readmission (PARR). The study cohort consisted of 180,118 admissions with 22,565 (12.5%) of actual readmissions within 30 days of hospital discharge, from 01 Jan 2015 to 31 Dec 2016 from two Auckland-region hospitals. We developed a machine learning model to predict 30-day readmissions using the model types XGBoost, Random Forests, and Adaboost with decision stumps as a base learner with different feature combinations and preprocessing procedures. The proposed model achieved the F1-score (0.386 ± 0.006), sensitivity (0.598 ± 0.013), positive predictive value (PPV) (0.285 ± 0.004), and negative predictive value (NPV) (0.932 ± 0.002). When compared with LACE and PARR(NZ) models, the proposed model achieved better F1-score by 12.7% compared with LACE and 23.2% compared with PARR(NZ). The mean sensitivity of the proposed model was 6.0% higher than LACE and 41% higher than PARR(NZ). The mean PPV was 15.9% and 14.6% higher than LACE and PARR(NZ) respectively. We presented an all-cause predictive model for 30-day risk of hospital readmission with an area under the receiver operating characteristics (AUROC) of 0.75 for the entire dataset.

Large language models (LLMs) such as ChatGPT-4, LLaMA-3.1, Gemini-1.5, DeepSeek-R1, and OpenAI-O3 have shown promising potential in health care, particularly for clinical reasoning and decision support. However, their reliability across critical tasks like diagnosis, medical coding, and risk prediction has received mixed reviews, especially in real-world settings without task-specific training. This study aims to evaluate and compare the zero-shot performance of reasoning and nonreasoning LLMs in three essential clinical tasks: (1) primary diagnosis generation, (2) ICD-9 (International Classification of Diseases, Ninth Revision) medical code prediction, and (3) hospital readmission risk stratification. The goal is to assess whether these models can serve as general-purpose clinical decision support tools and to identify gaps in current capabilities. Using the Medical Information Mart for Intensive Care-IV dataset, we selected a random cohort of 300 hospital discharge summaries. Prompts were engineered to include structured clinical content from 5 note sections: chief complaints, past medical history, surgical history, laboratories, and imaging. Prompts were standardized and zero-shot, with no model fine-tuning or repetition across runs. All model interactions were conducted through publicly available web user interfaces, without using application programming interfaces, to simulate real-world accessibility for nontechnical users. We incorporated rationale elicitation into prompts to evaluate model transparency, especially in reasoning models. Ground-truth labels were derived from the primary diagnosis documented in clinical notes, structured ICD-9 codes from diagnosis, and hospital-recorded readmission frequencies for risk stratification. Performance was measured using F1-scores and correctness percentages, and comparative performance was analyzed statistically. Among nonreasoning models, LLaMA-3.1 achieved the highest primary diagnosis accuracy (n=255, 85%), followed by ChatGPT-4 (n=254, 84.7%) and Gemini-1.5 (n=237, 79%). For ICD-9 prediction, correctness dropped significantly across all models: LLaMA-3.1 (n=128, 42.6%), ChatGPT-4 (n=122, 40.6%), and Gemini-1.5 (n=44, 14.6%). Hospital readmission risk prediction showed low performance in nonreasoning models: LLaMA-3.1 (n=124, 41.3%), Gemini-1.5 (n=122, 40.7%), and ChatGPT-4 (n=99, 33%). Among reasoning models, OpenAI-O3 outperformed in diagnosis (n=270, 90%) and ICD-9 coding (n=136, 45.3%), while DeepSeek-R1 performed slightly better in the readmission risk prediction (n=218, 72.6% vs O3's n=212, 70.6%). Despite improved explainability, reasoning models generated verbose responses. None of the models met clinical standards across all tasks, and performance in medical coding remained the weakest area across all models. Current LLMs exhibit moderate success in zero-shot diagnosis and risk prediction but underperform in ICD-9 code generation, reinforcing findings from prior studies. Reasoning models offer marginally better performance and increased interpretability, with limited reliability. Overall, statistical analysis between the models revealed that OpenAI-O3 outperformed the other models. These results highlight the need for task-specific fine-tuning and need human-in-the-loop checking. Future work will explore fine-tuning, stability through repeated trials, and evaluation on a different subset of deidentified real-world data with a larger sample size.

Background Readmissions are costly and inconvenient for patients, and occur frequently in hepatopancreatobiliary (HPB) surgery practice. Readmission prediction tools exist, but most have not been designed or tested in the HPB patient population. Methods Pancreatectomy and hepatectomy operation–specific readmission models defined as subspecialty readmission risk assessments (SRRA) were developed using clinically relevant data from merged 2014-15 ACS NSQIP Participant Use Data Files and Procedure Targeted datasets. The two derived procedure-specific models were tested along with 6 other readmission models in institutional validation cohorts in patients who had pancreatectomy or hepatectomy, respectively, between 2013 and 2017. Models were compared using area under the receiver operating characteristic curves (AUC). Results A total of 16,884 patients (9169 pancreatectomy and 7715 hepatectomy) were included in the derivation models. A total of 665 patients (383 pancreatectomy and 282 hepatectomy) were included in the validation models. Specialty-specific readmission models outperformed general models. AUC characteristics of the derived pancreatectomy and hepatectomy SRRA (pancreatectomy AUC=0.66, hepatectomy AUC=0.74), modified Readmission After Pancreatectomy (AUC=0.76), and modified Readmission Risk Score for hepatectomy (AUC=0.78) outperformed general models for readmission risk: LOS/2 + ASA integer-based score (pancreatectomy AUC=0.58, hepatectomy AUC=0.66), LACE Index (pancreatectomy AUC=0.54, hepatectomy AUC=0.62), Unplanned Readmission Nomogram (pancreatectomy AUC=0.52, hepatectomy AUC=0.55), and institutional ARIA (pancreatectomy AUC=0.46, hepatectomy AUC=0.58). Conclusion HPB readmission risk models using 30-day subspecialty-specific data outperform general readmission risk tools. Hospitals and practices aiming to decrease readmissions in HPB surgery patient populations should use specialty-specific readmission reduction strategies.

Fibrinogen, a biomarker of thrombosis and inflammation, is related to a high risk for cardiovascular diseases. However, studies on the prognostic value of blood fibrinogen concentrations for heart failure (HF) patients are few and controversial. We performed a retrospective analysis among acute or deteriorating chronic HF patients admitted to a hospital in Sichuan, China, between 2016 and 2019, integrating electronic health care records and external outcome data (N = 1532). During 6 months of follow-up, 579 HF patients were readmitted within 6 months, and 46 of them died. Surprisingly, we found an inverted U-shaped association of blood fibrinogen levels with risk of readmission within 6 months but not with risk of death within 6 months. It was found that HF patients had the highest risk for readmission within 6 months after reaching the turning point for blood fibrinogen (2.4 g/L). In HF patients with low fibrinogen levels < 2.4 g/L, elevated fibrinogen concentrations were still significantly associated with a higher risk for readmission within 6 months [OR = 2.3, 95% CI (1.2, 4.6); P  = 0.014] after controlling for relevant covariates. There was no significant association between blood fibrinogen and readmission within 6 months [(OR = 1.0, 95% CI (0.9, 1.1); P  = 0.675] in HF patients with high fibrinogen (> 2.4 g/L). The effect difference for the two subgroups was significant ( P  = 0.014). However, we did not observe any association between blood fibrinogen and death within 6 months stratified by the turning point, and the effect difference for the stratification was not significant ( P  = 0.380). We observed an inverted U-shaped association between blood fibrinogen and rehospitalization risk in HF patients for the first time. Additionally, our results did not support that elevated blood fibrinogen was related to increased death risk after discharge.

Background Hospital readmissions among older adults are associated with progressive functional worsening, increased institutionalization and mortality. Aim Identify the main predictors of readmission in older adults. Methods We examined readmission predictors in 777 hospitalized subjects (mean age 84.40 ± 6.77 years) assessed with Comprehensive Geriatric Assessment (CGA), clinical, anthropometric and biochemical evaluations. Comorbidity burden was estimated by Charlson Comorbidity Index (CCI). Median follow-up was 365 days. Results 358 patients (46.1%) had a second admission within 365 days of discharge. Estimated probability of having a second admission was 0.119 (95%C.I. 0.095–0.141), 0.158 (95%C.I. 0.131–0.183), and 0.496 (95%C.I. 0.458–0.532) at 21, 30 and 356 days, respectively. Main predictors of readmission at 1 year were length of stay (LOS) > 14 days ( p  < 0.001), albumin level < 30 g/l ( p 0.018), values of glomerular filtration rate (eGFR) < 40 ml/min ( p  < 0.001), systolic blood pressure < 115 mmHg ( p  < 0.001), CCI ≥ 6 ( p  < 0.001), and cardiovascular diagnoses. When the joint effects of selected prognostic variables were accounted for, LOS > 14 days, worse renal function, systolic blood pressure < 115 mmHg, higher comorbidity burden remained independently associated with higher readmission risk. Discussion Selected predictors are associated with higher readmission risk, and the relationship evolves with time. Conclusions This study highlights the importance of performing an accurate CGA, since defined domains and variables contained in the CGA (i.e., LOS, lower albumin and systolic blood pressure, poor renal function, and greater comorbidity burden), when combined altogether, may offer a valid tool to identify the most fragile patients with clinical and functional impairment enhancing their risk of unplanned early and late readmission.

Background: Effectively predicting and reducing readmission in long-term home care (LTHC) is challenging. We proposed, validated, and evaluated a risk management tool that stratifies LTHC patients by LACE predictive score for readmission risk, which can further help home care providers intervene with individualized preventive plans. Method: A before-and-after study was conducted by a LTHC unit in Taiwan. Patients with acute hospitalization within 30 days after discharge in the unit were enrolled as two cohorts (Pre-Implement cohort in 2017 and Post-Implement cohort in 2019). LACE score performance was evaluated by calibration and discrimination (AUC, area under receiver operator characteristic (ROC) curve). The clinical utility was evaluated by negative predictive value (NPV). Results: There were 48 patients with 87 acute hospitalizations in Pre-Implement cohort, and 132 patients with 179 hospitalizations in Post-Implement cohort. These LTHC patients were of older age, mostly intubated, and had more comorbidities. There was a significant reduction in readmission rate by 44.7% (readmission rate 25.3% vs. 14.0% in both cohorts). Although LACE score predictive model still has room for improvement (AUC = 0.598), it showed the potential as a useful screening tool (NPV, 87.9%; 95% C.I., 74.2–94.8). The reduction effect is more pronounced in infection-related readmission. Conclusion: As real-world evidence, LACE score-based risk management tool significantly reduced readmission by 44.7% in this LTHC unit. Larger scale studies involving multiple homecare units are needed to assess the generalizability of this study.

To evaluate the effectiveness of long-acting injectable antipsychotics (LAI-a) in reducing the 90-day and annual readmission rates in schizophrenia inpatients. We conducted a cross-sectional study and included 180 adult patients with psychotic disorders discharged from 2018 to 2019 at a state psychiatric hospital. Descriptive statistics were used to measure the differences between the readmit and nonreadmit cohorts. Logistic regression model was used to measure the odds ratio (OR) for 90-day and annual readmission and was controlled for potential readmission risk factors. A lower proportion of patients receiving LAI-a were readmitted within 90-day (28.6%) and 1-year (32.4%) periods. Patients receiving LAI-a had lower odds of association for 90-day (OR 0.36, 95% confidence intervals [CI] 0.139-0.921) and annual (OR 0.35, 95% CI 0.131-0.954) readmissions compared to those discharged on oral antipsychotics. A higher proportion of inpatients who received fluphenazine LAI had 90-day (25%) and annual (18.2%) readmissions compared to other LAI-a. Utilization of LAI-a in patients with psychotic disorders can decrease both 90-day and annual psychiatric readmissions by 64% to 65%. Physicians should prefer LAI-a to reduce the readmission rate and improve the quality of life, and decrease the healthcare-related financial burden.

Objectives We aimed to evaluate the utility of an Observation Unit (OU) in management of acute exacerbations of chronic obstructive pulmonary disease (AECOPD) and to identify the clinical characteristics of patients readmitted within 30-days for AECOPD following index admission to the OU or inpatient floor from the OU. Methods This is a retrospective observational study of patients admitted from January to December 2017 for AECOPD to an OU in an urban-based tertiary care hospital. Primary outcome was rate of 30-day readmission after admission for AECOPD for patients discharged from the OU versus inpatient service after failing OU management. Regression analyses were used to define risk factors. Results 163 OU encounters from 92 unique patients were included. There was a lower readmission rate (33%) for patients converted from OU to inpatient care versus patients readmitted after direct discharge from the OU (44%). Patients with 30-day readmissions were more likely to be undomiciled, with history of congestive heart failure (CHF), pulmonary embolism (PE), or had previous admissions for AECOPD. Patients with >6 annual OU visits for AECOPD had higher rates of substance abuse, psychiatric diagnosis, and prior PE; when these patients were excluded, the 30-day readmission rate decreased to 13.5%. Conclusion Patients admitted for AECOPD with a history of PE, CHF, prior AECOPD admissions, and socioeconomic deprivation are at higher risk of readmission and should be prioritized for direct inpatient admission. Further prospective studies should be conducted to determine the clinical impact of this approach on readmission rates.

Background The identification of patients at high risk of unplanned readmission is an important component of discharge planning strategies aimed at preventing unwanted returns to hospital. The aim of this study was to investigate the factors associated with unplanned readmission in a Sydney hospital. We developed and compared validated readmission risk scores using routinely collected hospital data to predict 7-day, 30-day and 60-day all-cause unplanned readmission. Methods A combination of gradient boosted tree algorithms for variable selection and logistic regression models was used to build and validate readmission risk scores using medical records from 62,235 live discharges from a metropolitan hospital in Sydney, Australia. Results The scores had good calibration and fair discriminative performance with c-statistic of 0.71 for 7-day and for 30-day readmission, and 0.74 for 60-day. Previous history of healthcare utilization, urgency of the index admission, old age, comorbidities related to cancer, psychosis, and drug-abuse, abnormal pathology results at discharge, and being unmarried and a public patient were found to be important predictors in all models. Unplanned readmissions beyond 7 days were more strongly associated with longer hospital stays and older patients with higher number of comorbidities and higher use of acute care in the past year. Conclusions This study demonstrates similar predictors and performance to previous risk scores of 30-day unplanned readmission. Shorter-term readmissions may have different causal pathways than 30-day readmission, and may, therefore, require different screening tools and interventions. This study also re-iterates the need to include more informative data elements to ensure the appropriateness of these risk scores in clinical practice.

Background This study aims to examine the relationship between the serum uric acid-to-high density lipoprotein cholesterol ratio (UHR) and short-term readmission risk in patients with heart failure. Methods In this retrospective cohort study, we analyzed clinical data from Zigong Fourth People's Hospital (December 2016 to June 2019) using multivariate logistic regression, restricted cubic splines, and subgroup analyses to examine the independent association between UHR and short-term readmission risk. Results The analysis included 1795 participants, of whom 58.4% were female and 53.7% were aged 60–79 years. Our study showed that an increase of one unit in the UHR was associated with a 275% increase in 28-day readmission risk (OR = 3.75, 95% CI = 1.56–9.01, P  = 0.003) after adjusting for potential confounders. The 28-day readmission risk was markedly elevated in the high UHR group (Q 4 ) compared to the low UHR group (OR = 2.38, 95% CI = 1.27–4.45, P  = 0.007). The relationship between UHR and 3-month readmission risk showed a similar trend. RCS revealed a significant nonlinear relationship between UHR and readmission risk (28-day P for nonlinearity = 0.038, 3-month P for nonlinearity = 0.018), with inflection points at 2.30 for 28 days and 3 months. Subgroup analyses showed that the association between UHR and readmission risk was stronger among octogenarians and males. Sensitivity analyses confirmed the robustness of our findings. Conclusions UHR demonstrated a nonlinear association with short-term readmission risk in heart failure, with an inflection point at 2.30. This association was more pronounced in octogenarians and male patients.