Ramai, Daryl

ABSTRACT Objective To comprehensively evaluate the validity of ICD-10-CM codes for both prevalent diagnoses and less common diseases, and to assess the performance of a large language model (LLM)-based system in validating these codes. Materials and Methods This retrospective study analyzed hospital admissions from the Medical Information Mart for Intensive Care (MIMIC-IV) database. We developed a validated LLM-based system using GPT-4o, refined through iterative prompt engineering, to assess ICD-10-CM code validity. We measured the PPV of ICD-10-CM codes, PPV of principal and secondary diagnoses, and the performance of an LLM-based system in code validation. Results Among 865,079 assigned codes, the PPV was 84.6% (95% CI, 84.5%-84.6%). Principal diagnoses had a PPV of 93.9% (95% CI, 93.7%-94.1%), while secondary diagnoses had a PPV of 83.8% (95% CI, 83.7%-83.9%). The LLM system demonstrated high performance in validating ICD codes, achieving 93.6% accuracy, 95.4% sensitivity and 85.2% specificity. Among correctly assigned secondary diagnoses, the majority (67.9%) represented historical or baseline conditions, while 32.1% reflected active conditions that deviated from baseline status; 22.3% of these emerged after hospital admission. PPV decreases with later diagnosis positions, with the largest decline occurring between principal and secondary diagnoses. Discussion and Conclusion In this large-scale evaluation, ICD-10-CM codes exhibited generally high accuracy, though variability existed by position and condition type. A validated LLM system performed comparably to physician review and offers a scalable means to improve coding accuracy. These findings support the potential for integrating LLM-based auditing into routine workflows to strengthen the quality of administrative and research data.

Abstract Background Studies evaluating the risk of post-polypectomy bleeding (PPB) after colorectal endoscopic mucosal resection (EMR) in patients with chronic kidney disease (CKD) and end-stage renal disease (ESRD) are limited. Methods This retrospective cohort study utilized the U.S. Collaborative Network to assess the risk of PPB after colorectal EMR in patients with CKD compared to controls. Using one-to-one propensity score matching (PSM), the primary outcome measured was PPB within 30 days after colorectal EMR. The PPB risk was further stratified by CKD severity: non-advanced CKD and advanced CKD. Results After PSM, each cohort included 7,460 patients. Overall, CKD was associated with increased risk of PPB following colorectal EMR (5.5% vs. 4%, odds ratio [OR] 1.38, 95% confidence interval [CI] 1.19–1.61, p < 0.001). The PPB risk was significantly higher in patients with advanced CKD (8.1% vs, 4.6%, OR 1.83, 95% CI 1.43–2.34, p < 0.001), while those with non-advanced CKD showed modest increase in risk of PPB (4.7% vs. 3.9%, OR 1.21, 95% CI 1.001–1.46, p = 0.049). Conclusion Patients with CKD had higher risk of PPB than patients without CKD, but absolute risk remained low. The PPB risk was notably increased in patients with advanced CKD. Optimizing patients with CKD, especially advanced CKD, before colorectal EMR and monitoring for post-procedure bleeding remains important.

INTRODUCTION: The impact of glucagon-like peptide-1 receptor agonists (GLP-1RAs) on colonoscopy bowel preparation is unclear. METHODS: We searched multiple databases for studies comparing bowel preparation quality between GLP-1RA users and the control group. Using a random-effects model, we calculated pooled odds ratios and mean differences with 95% confidence intervals (CIs). Inadequate bowel preparation (IBP) rate was the primary outcome, and Boston Bowel Preparation Scale mean score was the secondary outcome. RESULTS: Five studies with 10,833 patients (n = 5,434 GLP-1RAs vs n = 5,399 control) were included. The IBP rate was higher in the GLP-1RAs group compared with the control group (odds ratios 2.10, 95% CI 1.41–3.13, P = 0.0003). Furthermore, the mean Boston Bowel Preparation Scale score was lower in the GLP1-RAs group compared with the control group (mean differences −0.34; 95% CI −0.51, −0.17; P = 0.0001). DISCUSSION: Our findings suggest that using GLP-1RA increases the risk of IBP.

AbstractCholecystectomy is the standard treatment for acute cholecystitis, but it may not be suitable for all patients. For those who cannot undergo surgery, a percutaneous cholecystostomy tube (PCT) and ERCP-guided transpapillary gallbladder drainage are viable options. We aimed to perform a nationwide study to assess 30-day readmission rates, adverse events (AEs), and mortality rates in these two cohorts.We conducted a nationwide cohort study using data from the Nationwide Readmissions Database (NRD) from 2016 to 2019. We identified patients with acute cholecystitis during the index admission who underwent either PCT or ERCP-guided gallbladder drainage. Propensity score matching along with multivariable regression was used to compare cohorts.During the study period, 3,592 patients (average age 63.0 years) underwent endoscopic drainage, whereas 80,372 patients (average 70.8 years) underwent Interventional Radiology drainage. Utilizing multivariate Cox regression analysis, compared with ERCP, PCT had a higher risk for 30-day readmission (adjusted hazard ratio [aHR] 1.47; 95% confidence interval [CI] 1.27 to 1.71; P < 0.001). The PCT group had a significantly higher rate of readmission for acute cholecystitis compared with the ERCP group (2.72% vs 0.86%; P < 0.005). Cox proportional hazard ratio showed a 3.41-fold increased risk (95% CI 1.99 to 5.84) for readmission in the PCT group. ERCP was consistently associated with lower rates of post-procedural AEs compared with PCT including acute hypoxemic respiratory failure (P < 0.001), acute renal failure (P < 0.001), shock (P < 0.001), and need for blood transfusions (P < 0.001).Our nationwide analysis revealed that ERCP-guided gallbladder drainage should be the preferred approach for managing acute cholecystitis when unfit for surgery.

INTRODUCTION: Data on the risk of post-polypectomy bleeding after endoscopic mucosal resection (EMR) in patients with cirrhosis are limited. METHODS: This retrospective cohort study used the US Collaborative Network to assess post-polypectomy bleeding risk after colorectal EMR in patients with cirrhosis compared with controls. Using one-to-one propensity score matching, the primary outcome measured was bleeding within 30 days post-EMR. The bleeding risk was further stratified by cirrhosis severity (compensated and decompensated). RESULTS: After propensity score matching, each cohort included 2,381 patients. Patients with cirrhosis had a higher risk of post-polypectomy bleeding (odds ratio [OR] 1.38, 95% confidence interval [CI] 1.05–1.82, P = 0.02). However, endoscopic reinterventions for hemostasis were similar between cirrhosis and control cohorts (OR 1.20, 95% CI 0.71–2.05, P = 0.50). The post-polypectomy bleeding risk was significantly higher in patients with decompensated cirrhosis (OR 1.65, 95% CI 1.12–2.42, P = 0.01), while those with compensated cirrhosis showed no increased risk of post-polypectomy bleeding (OR 1.15, 95% CI 0.85–1.54, P = 0.37). DISCUSSION: The post-polypectomy bleeding risk was higher in patients with decompensated cirrhosis, while the bleeding risk was not increased in those with compensated disease. Optimizing patients with cirrhosis, especially with decompensated disease, before colorectal EMR and monitoring for post-procedure bleeding remain important. A conservative, supportive approach to managing bleeding in patients with cirrhosis may be as effective as in those without cirrhosis.