Diagnosis Codes Research Articles

Increased risk of fracture among patients with iron overload: a population-based matched cohort study

Abstract Introduction Iron overloading disorders are associated with decreased bone mineral density. However, evidence on fracture risk is scarce. Therefore, we evaluated the risk of fracture associated with iron overload disorders, compared to matched controls. Methods Using THIN, a Cegedim database of UK general practice data, we identified patients >18 years with elevated iron (ferritin value >1000 µg/L) or an eligible diagnosis code for iron overloading disorders between 2010-2022. The first date of elevated iron or a diagnosis code defined the index date for iron overload patients, who were matched with to up to 10 controls. Time-varying confounder adjusted Cox proportional hazard models estimated the hazard ratios (HRs) and 95% confidence intervals. Analyses were stratified by osteoporotic fracture site (hip, vertebral, humerus, forearm), evidence of elevated serum ferritin at baseline (ferritin >1000 µg/L), and sex. Results We identified 20,264 eligible patients and 192,956 controls. Overall, there was a 55% increased risk of any fracture among iron overload patients (HR 1.55 [1.42-1.68]). Fracture risk was increased at all sites, with the highest risk observed for vertebral fractures (HR 1.97 [1.63-2.10]). Patients with ferritin >1000µg/L had a 91% increased risk of any fracture (HR 1.91 [1.73-2.16]) and a 2.5-fold increased risk of vertebral fractures (HR 2.51 [2.01-3.12]). There was no increased risk among patients without elevated serum ferritin at any site. Fracture risk was similar between sexes. Discussion This large population-based cohort study found a 55% increased risk of fracture associated with iron overload. The risk was highest among patients with laboratory confirmed iron overload, highlighting the importance for clinicians to consider initiating osteoporosis therapy in patients with serum ferritin >1000 µg/L to minimize fracture risk.

Retrospective Analysis of the Impact of Screening for Psoriatic Arthritis (PsA) with the Psoriasis Epidemiology Screening Tool (PEST) in Clinical Dermatology Practice

Introduction: Patients with plaque psoriasis (PSO) have an increased risk of developing PsA as the disease severity progresses. While it is estimated that 30% of patients with PSO have PsA, nearly 16% of patients with PSO may have undiagnosed PsA.1,2 Implementing the PEST questionnaire for PsA screening in dermatologic practice may increase early intervention by improving rates of timely diagnosis.3 This observational study aimed to assess the proportion of patients with PSO and no previous PsA diagnosis who screened positive on the PEST and were then diagnosed with PsA by a healthcare provider in the PEST deployment period within the Advanced Dermatology and Cosmetic Surgery (ADCS) specialty dermatology network. Procedure/Study: De-identified, real-world patient data [2018–2019 and 2021–2022] were sourced from electronic health records within the ADCS network of outpatient facilities across 14 states in the United States. Patients were included in the analysis if they received care at an ADCS facility where the PEST was deployed, had ≥1 encounter with a diagnosis code for PSO, and never had a diagnosis code for PsA before their first diagnosis code for PSO. For outcomes directly related to PEST evaluations, data from 1/1/2021–12/31/2022 (PEST deployment period) were used. For outcomes not directly related to PEST results, data from 1/1/2018–12/31/2019 (period before PEST deployment) were evaluated. Descriptive statistics, including frequencies, percentages, and 95% confidence intervals (CIs) for categorical variables, were used to characterize the patient cohorts by study period. Results: Among ADCS facilities using the PEST during the deployment period (44% of practices in the network), the final study population included 14,308 and 21,142 patients with PSO in the period before PEST deployment and during the PEST deployment period, respectively. In the PEST deployment cohort, 1,100 (5.2%; 95% CI: 4.9–5.5) patients were screened using the PEST. In the period before PEST deployment, 327 (2.3%; 95% CI: 2.0–2.5) patients with PSO and no previous PsA diagnosis were subsequently diagnosed with PsA. During the PEST deployment period, 449 (2.1%; 95% CI: 1.9–2.3) patients with PSO and no previous PsA diagnosis received a subsequent PsA diagnosis without PEST screening. Of all patients with PSO who screened positive for PsA on the PEST (score ≥3; n=296), 10.8% (n=32; 95% CI: 7.6–15.1) subsequently received a formal PsA diagnosis during the PEST deployment period. Conclusion: These results suggest that a meaningful proportion of patients with PsA may be underdiagnosed in the dermatology setting. Deployment of the PEST in the clinical dermatologic workflow may provide an opportunity to diagnose PsA earlier and improve patient outcomes.

Review of Completeness of Diagnosis Filling in Medical Record for Inpatients at Mental Hospital. Prof. HB. Saanin Padang

Hospitals are health care institutions that provide comprehensive individual health services that provide inpatient, outpatient, and emergency services. To realize this, hospitals need to organize medical record services to support the provision of health services. The important thing that must be considered by medical record personnel in maintaining the quality of medical records is the completeness of medical information related to the patient's medical history starting from the beginning of treatment until discharge from the hospital. The research aim that the results of this study can improve the quality of medical records and hospital accreditation. This research was conducted in June, 2024 at RSJ. Prof. HB. Saanin and continued with data processing at the Academy of Recorders and Health Information IRIS Padang. The type of research used in this study is descriptive research with quantitative methods and direct observation of medical record documents. Population of study was medical record documents for inpatients in in January - March 2024 totaling 600 medical record documents. The sample size using the Slovin formula was 99 samples. Result of research show that highest distribution of diagnoses based on ICD-10 at Prof. Hb Saanin Padang was paranoid schizophrenia with 35 patients, completeness of filling in the main diagnosis and diagnosis code in cases of mental disorders was 96%, Percentage of readability of diagnosis writing, where the writing of the main diagnosis and secondary diagnosis on the medical record is 100%. Keywords: Completeness, Medical record, diagnose.

Reoperations After Operatively and Non-Operatively Treated Periprosthetic Knee Fractures: A Nationwide Study on 1,931 Fractures After Primary Total Knee Arthroplasty.

Periprosthetic knee fracture (PPKF) following total knee arthroplasty (TKA) can be difficult to treat. A PPKF can be treated both operatively and non-operatively, and the treatment varies between fracture sites. This study aimed to assess the risk of reoperation according to the fracture site and treatment of the PPKF. This study is a nationwide cohort study using register data from 1997 to 2022. Cruciate-retaining or posterior-stabilized primary TKA were identified from the Danish Knee Arthroplasty Register (DKR). Subsequent PPKFs, fracture treatments, and later reoperations were identified through the International Classification of Diseases (ICD-10) diagnosis and procedure codes in the Danish National Patient Register. Additionally, indications for revision TKA (rTKA) in the DKR were used to identify PPKFs and the reason for reoperation. We included 1,931 PPKFs (1,494 femoral, 207 patellae, and 230 tibial) with an overall two-year reoperation risk of 20% (95% CI [confidence interval]: 18 to 23). Femoral PPKFs had a two-year reoperation risk of 16% (12 to 22) after non-operative treatment and 21% (18 to 24) after operative treatment. Patellar PPKFs had a two-year reoperation risk in five to 17 after non-operative treatment and 46% (30 to 69) after operative treatment. Tibial PPKFs had a two-year reoperation risk of 17% (11 to 27) after non-operative treatment, and 36% (25 to 53) after operative treatment. The overall two-year reoperation risk was 20% (18 to 23). Operative-treated PPKFs had a higher risk of reoperation across all fracture locations. The highest reoperation risk was found after operatively treated patella and tibial PPKFs (46 and 36%, respectively), and the lowest reoperation risk was found after non-operative treated patella PPKFs.

Using large administrative data for mining patients' trajectories for risk stratification: An example from urological diseases.

To identify latent clusters among urological patients by examining hospitalisation rate trajectories and their association with risk factors and outcome quality indicators. Victorian Admitted Episodes Dataset, containing information on all hospital admissions in Victoria from 2009 to 2019. The top twenty ICD-10 primary diagnosis codes in urology were used to select patients (n = 98,782) who were included in the study. Latent class trajectory modelling (LCTM) was used to cluster urological patient hospitalisation trajectories. Logistic regression was used to find baseline factors that influence cluster membership, the variables tested included comorbidities, baseline diagnosis codes, and socio-demographic factors. The analysis was further stratified into non-surgical procedures and surgical procedures. Five clusters of hospitalisation trajectories were identified based on clustering hospitalisation rates over time. Higher hospitalisation clusters were strongly associated with longer length of stay, higher readmission rates and higher complication rates. Higher-risk groups were strongly associated with comorbidities such as renal disease and diabetes. For surgical procedures, urological cancers (kidney, prostate and bladder cancer) and irradiation cystitis were associated with higher-risk groups. For non-surgical procedures, calculus of the bladder, urethral stricture and bladder neck obstruction were associated with higher-risk groups. For patients with two or more admissions, liver cardiovascular disease and being diagnosed with benign prostatic hyperplasia were also associated with higher risk groups. A novel statistical approach to cluster hospitalisation trajectories for urological patients was used to explore potential clusters of patient risks and their associations with outcome quality indicators. This study supports the observation that baseline comorbidities and diagnosis can be predictive of higher hospitalisation rates and, therefore, poorer health outcomes. This demonstrates that it is possible to identify patients at risk of developing complications, higher length of stay and readmissions by using baseline comorbidities and diagnosis from administrative data.

Abstract 4138397: Gender Difference in Prevalence and Survival of Pulmonary Hypertension: A Prospective Cohort Study of 502,173 UK Biobank Participants

Background: Research into pulmonary hypertension (PH) reveals significant sex-based disparities in long-term survival rates. Notably, women, despite being more susceptible to PH, often experience better survival outcomes than men—a phenomenon possibly linked to the "estrogen paradox." This study leverages the comprehensive UK Biobank to examine the role of sex in influencing these outcomes. By exploring differences in PH incidence and survival rates across genders, our research aims to provide a deeper understanding of the mechanisms at play and inform more effective, gender-specific analysis. Methods: This prospective cohort study utilized data from the UK Biobank, with baseline assessment between 2006 and 2010 and follow-up until November 30, 2023, for England and Scotland, and May 31, 2022, for Wales. Through diagnostic codes, medical records, and imaging studies, 3,296 PH cases and 497,580 cases without PH were identified . We followed them for the mortality rates over a period of 10 years. Survival analysis was conducted using Kaplan-Meier curves and Cox proportional hazards models to assess sex-specific survival differences, adjusting for confounders like age, BMI, ethnic, menopause, smoking status, and comorbidities. Results: During 7,978,349 person-years of follow-up, of 502,173 participants in this prospective cohort, 3,296 were diagnosed with PH, while 497,580 did not receive the PH diagnosis. The prevalence of PH in the male cohort and the female cohort were 0.76% (1743 of 228,991 men) and 0.57% (1553 of 273,182 women), respectively, with significant difference between the two groups (p<0.001). In the PH cohort, the mortality rates of patients were 47.06%. In the male patients with PH, the mortality rate was significantly higher than the female PH patients (51.78% vs. 41.53%, p<0.001). After stratifying women into menopause and no-menopause group, Kaplan-Meier curves showed that even the age of male group was significantly younger than the menopause group (p<0.001), the male still had worse survival situation. Conclusion: To conclude, the prevalence of PH in the male participants was significantly higher than the female participants. While the mortality rate in the male participants was also significantly higher than the female participants.

Abstract 4140118: Identification of Obstructive and Non-Obstructive Hypertrophic Cardiomyopathy Patients Using Natural Language Processing in a Large Integrated Healthcare System

Introduction: Accurately identifying and characterizing patients with hypertrophic cardiomyopathy (HCM) is critical for population management and care optimization. Research Question: To develop natural language processing (NLP) algorithms to identify and characterize obstructive (oHCM) and non-obstructive (nHCM) HCM patients directly from echocardiograms, and to compare with the presence or absence of HCM-related diagnosis codes. Methods: We developed and validated NLP algorithms to identify HCM from all adult (age≥18yrs) echocardiograms performed from 2010-2019 in Kaiser Permanente Northern CA (KPNC), capturing measures of any HCM, HCM subtype, hypertrophy subtype, septal and posterior LV wall thickness, resting and stress/Valsalva LVOT gradients, and systolic anterior motion. We developed a rules-based algorithm (following AHA/ACC criteria) to classify patients as having HCM, including oHCM or nHCM subtypes, and possible HCM (defined as wall thickness ≥2cm without other criteria meeting an HCM definition). We evaluated the presence of HCM-related ICD-9/10 diagnosis codes among patients classified as HCM/non-HCM from echocardiograms using NLP, and linked baseline demographics and clinical parameters from our integrated electronic medical record. Results: Among 472,405 adults with echocardiograms, we identified 2,892 patients with HCM based upon NLP-derived measures (all NLP measures achieved >95% positive predictive value and >95% negative predictive value), including 1,585 (55%) with oHCM, 1,145 (40%) with nHCM, and 162 (6%) which could not be classified (Figure). Among those 2,892 patients, 1,283 did not have any associated HCM ICD-9/10 diagnosis codes (Table). Among 469,513 patients with no identified HCM from NLP-based algorithms, HCM ICD-9/10 diagnosis codes existed in 1,567 patients (Table). We also identified 4,593 patients with possible HCM by NLP, only 4.5% of whom had an associated HCM code. Among confirmed HCM patients by NLP, oHCM patients were slightly older (66 vs 61 yrs), more likely female (53% vs 43%), had similar mean septal wall thickness (1.7cm vs 1.7cm), but were more likely to have a septal hypertrophy subtype (46% vs 28%) compared to nHCM patients. Conclusions: Echocardiogram-based NLP methods can improve the identification of and care for HCM patients. Many patients with possible HCM may be underdiagnosed, representing an opportunity for quality improvement.

New Diagnoses of Juvenile Idiopathic Arthritis Early in the COVID-19 Pandemic

Background/Objective Little is known about the rates of rheumatic disease diagnosis among children during the COVID-19 pandemic. We examined the impact of the pandemic on the diagnosis of juvenile idiopathic arthritis (JIA) in the United States. Methods We performed a historical cohort study using US commercial insurance data (2016–2021) to identify children aged <18 years without prior JIA diagnosis or treatment in the prior ≥12 months. New JIA diagnoses were identified using a combination of ICD-10-CM diagnosis codes, location, and timing of medical services. Crude rates with 95% confidence intervals (CIs) of JIA diagnosis per 100,000 enrolled children per quarter were estimated and stratified by age group, sex, region, JIA type, and uveitis. The incidence rate ratio (95% CI) for JIA diagnosis was estimated using Poisson regression, adjusted for various demographic variables. Results From 2018–2021, 643 children were diagnosed with JIA. Crude new JIA diagnoses per 100,000 children per quarter dropped from 2.62 (95% CI, 2.39–2.87) prepandemic to 1.94 (95% CI, 1.66–2.25) during the pandemic. Declines in JIA diagnosis were more apparent in the US Northeast and West regions and among children aged 6–11 years. After adjustment for covariates, JIA diagnoses fell by 30% during the pandemic compared with the prior 3 years (IRR, 0.70; 95% CI, 0.59–0.83). Conclusions Compared with the prepandemic period, JIA was diagnosed 30% less often during the early pandemic among commercially insured children in the United States. More research is needed to understand the underlying reasons for these changes in JIA diagnosis and more recent trends.

Abstract 4139199: Quantifying the Burden of High Polygenic Risk for Cardiometabolic Disease in a US-based Healthcare System Biobank

Background: Healthcare systems are increasingly collecting genotype data on patients. Polygenic risk scores (PRS), which quantify inherited risk of a given disease, have the potential to leverage this data to identify high risk individuals. While PRS for single traits have often been evaluated, the same data can be used to calculate PRSs across traits and evaluate the overall burden of high cardiometabolic polygenic risk using a single test. Methods: We examined the association of cardiometabolic PRS from the Polygenic Score Catalog among 236,393 All of Us (AoU) participants for 6 target traits of interest, using all available PRS for coronary artery disease (CAD; 57 PRS), atrial fibrillation (AF; 30 PRS), heart failure (HF; 9 PRS), type 2 diabetes mellitus (T2DM; 133 PRS), venous thromboembolism (VTE; 9 PRS), and thoracic aortic aneurysm (TAA; 2 PRS). PRSMix is a tool that incorporates information across several PRS for a given trait to improve prediction accuracy for a target population. We generated a PRSMix for each trait and externally benchmarked its performance against individual PRS in the Mass General Brigham Biobank (MGBB). Phenotypes were ascertained using hospital diagnosis and procedural codes. Logistic regression was performed, adjusting for age, sex, and the first 10 principal components of inferred genetic similarity. Results: Of the 53,306 genotyped MGBB participants, 55.6% are female, mean enrollment age is 53 ± 17 years. Each trait-specific PRSMix demonstrated generally stronger associations with the target trait compared to any individual contributing score. This included trait-specific PRSMix associations (OR/SD, [95%CI]) of 2.07 [1.99-2.15] for CAD, 1.88 [1.82-1.95] for AF, 1.47 [1.43-1.51] for HF, 2.07 [2.00-2.14] for T2DM, 1.43 [1.39-1.47] for VTE, and 1.40 [1.34-1.46] for TAA. Findings were consistent across inferred genetically similar ancestry groups, with stronger PRSMix associations for CAD detected in the EUR group (2.14 [2.05-2.22]) compared to the AFR group (1.43 [1.21-1.70]). Approximately 37% of people in MGBB have 3-fold greater risk of at least one cardiometabolic disease, with 22% at risk for CAD, 14% for T2DM, 13% for Afib, 3% for VTE, 2% for HF and <1% for TAA. Conclusion: We developed PRSMix scores for cardiometabolic traits in AoU and validated them in MGBB. We identified 37% of biobank participants with at least 3-fold greater disease risk associated with high polygenic scores for at least one cardiometabolic outcome.

Abstract 4118402: National Burden, Trends, and Racial Disparities in Cardiovascular Mortality Among Incarcerated Individuals in the United States from 2000-2019

Background: Cardiovascular disease is the leading cause of death in the U.S. Mass incarceration is a national public health issue, with the US having the highest incarceration rate in the world. However, there is a paucity of data on cardiovascular mortality among incarcerated individuals, or how the racialization of mass incarceration contributes to cardiovascular health inequities. The aim of this study was to evaluate the burden of cardiovascular mortality, trends in cardiovascular deaths, as well as racial differences in cardiovascular death and receipt of care among those incarcerated in U.S. state prisons from 2001-2019. Methods: We utilized the database “Mortality in Correctional Institutions: ICD-10 Diagnosis Codes for Natural Deaths Occurring in State Prison or Local Jail Custody, 2000-2019” from the United States Bureau of Justice Statistics. This database collects inmate death records from each of the nation’s 50 state prison systems. We extracted all deaths related to cardiovascular causes to evaluate the epidemiology of individuals with cardiovascular death, trends in cardiovascular death rates, and differences between racial groups. Results: From 2000-2019, there were a total of 18227 (28.0%) deaths secondary to cardiovascular causes among those incarcerated in US state prisons. The rate of CV death increased from 71.3 per 100,000 persons in 2001 to 94 per 100,000 persons in 2019. Black individuals who died while incarcerated had higher rates of death related to a cardiovascular cause than White individuals (30.9% vs. 27.8%, p<0.0001). Black individuals had higher rates of death due to ischemic heart disease (23.2% vs 21.9%) and hypertensive-related cardiovascular deaths (3.71% vs. 2.96%) as compared to white counterparts (p<0.001). Black incarcerated individuals who died from a cardiovascular cause were less likely to be evaluated by a medical provider in prison (78.6% vs. 81.9%; p<0.0001), were less likely to have a diagnostic test performed (65.3% vs. 69%, p<0.0001), were less likely to receive medications (68.1% vs 72.6%, p<0.001), as well as treatment other than medications (49.1% vs 52.8%, p <0.0001) for the condition which caused death as compared to White counterparts. Conclusions: Deaths due to cardiovascular causes are high among patients incarcerated in US state prisons. Combined with decarceral efforts to address mass incarceration, strategies to ensure quality and equitable cardiovascular care for this population are needed.

Stroke Risk Following Nonarteritic Anterior Ischemic Optic Neuropathy.

The association between nonarteritic anterior ischemic optic neuropathy (NAION) and an increased risk of stroke has been a subject of debate. However, multinational studies on this topic are scarce. To evaluate the short-term and long-term stroke risk after NAION compared with a matched control group. This global, retrospective, population-based cohort study used aggregated electronic health records from January 1, 2004, through March 19, 2024, sourced from the Global Collaborative Network of TriNetX, which includes data from over 152 million patients across 17 countries. Patients in the study were followed up for a maximum duration of 10 years. Patients with NAION and age-related cataract were included in the analysis. Those with stroke before the diagnosis of NAION and age-related cataract were excluded. Propensity score matching was applied to balance age, sex, race, ethnicity, comorbidities, and medication use. International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) diagnosis code for NAION or age-related cataract. The primary outcome was the relative risk (RR) of stroke (ICD-10 code I60-63) in the NAION cohort vs the matched controls. Multivariable logistic regression analyses were applied to identify potential clinical factors associated with stroke within the NAION cohort. A total of 89 811 patients were identified in both the NAION (mean [SD] age, 57.2 [18.5] years; 38 678 men [43.1%]) and control (mean [SD] age, 57.0 [17.9] years; 40 014 men [44.6%]) cohorts after matching. The NAION cohort demonstrated a significantly higher all-stroke risk at all time points: 1 month (RR, 5.04; 95% CI, 4.41-5.78), 3 months (RR, 3.79; 95% CI, 3.40-4.21), 1 year (RR, 2.50; 95% CI, 2.32-2.70), 5 years (RR, 1.54; 95% CI, 1.45-1.63), and 10 years (RR, 1.33; 95% CI, 1.23-1.43). Sensitivity analysis in patients without comorbidities similarly revealed a significantly increased all-stroke risk across all intervals: 1 month (RR, 7.55; 95% CI, 4.74-12.03), 3 months (RR, 6.70; 95% CI, 4.48-10.04), 1 year (RR, 3.96; 95% CI, 2.94-5.34), 5 years (RR, 2.85; 95% CI, 2.18-3.72), and 10 years (RR, 1.68; 95% CI, 1.25-2.26). Among all the clinical factors of interest, only hypertension was consistently associated with all subtypes of stroke following NAION. This cohort study of patients with NAION found a significantly elevated risk of stroke compared with matched controls, independently of comorbidities. These findings underscore the importance of regular stroke workups following the onset of NAION.

Abstract 4123373: Estimating obesity-attributable cardiovascular disease burden in the United States: a Bayesian network analysis

Background: Obesity is a known contributor to the development of cardiovascular disease (CVD) and increases the risk of subsequent mortality. To fully understand the burden of obesity, accurate information regarding its effects on CVD is necessary. Current estimates are limited by a focus on direct effects, use of non-representative data sources, and a reliance on inaccurate diagnosis codes. Aim: To determine the size of CVD burden attributable to obesity in the US and identify the mechanisms through which obesity affects CVD. Methods: This study used data collected from US adults (aged ≥35 years) participating in the National Health and Nutrition Examination Survey (NHANES) from 1999 to 2020 and US Census projections. NHANES data were longitudinally adapted, and participants were excluded if they had missing data for weight history, risk factors, or a given outcome of interest. Bayesian networks (BNs) mapping out the relationship between risk factors and each outcome were constructed to estimate current CVD risk and expected CVD risk in the absence of obesity. Obesity-attributable CVD burden was then calculated for each outcome, and the direct and indirect pathways through which changes in CVD risk occur were characterized. A subgroup analysis was conducted in non-smoking participants aged ≤65 years for whom obesity is likely the primary CVD risk factor. Results: The final sample included data from approximately 29,000 participants. Of these, 1286 (4.3%) had a history of myocardial infarction (MI), 1283 (4.4%) had a history of stroke, 1092 (3.7%) had a history of heart failure (HF), and 1992 (6.7%) had a history of CVD (composite of MI, stroke, angina, coronary heart disease, and HF). Results from the fitted BN indicate that 21% of MI, 16% of stroke, 38% of HF, and 19% of CVD cases were attributable to obesity, with ≥80% of this burden due to the effect of obesity on intermediate risk factors (hypertension, type 2 diabetes, and prior cardiovascular events). In the subgroup analysis, the obesity-attributable CVD burden was approximately twice as high for non-smokers aged ≤65 years, with 39% of MI, 32% of stroke, 65% of HF, and 38% of CVD cases attributable to obesity. Conclusion: Through a sophisticated network modeling approach, we generated estimates of obesity-attributable CVD burden in the US. Results highlight the importance of obesity as a CVD risk factor, particularly in younger non-smokers, and provide insight into the nature of this relationship.

Abstract 4119094: Claims-Based versus Registry-Reported Stroke Events in the NCDR Left Atrial Appendage Occlusion Registry

Background: Administrative claims data are increasingly used for post-marketing surveillance of drugs and medical devices, though data reliability remains uncertain. Given their associated morbidity and mortality, stroke events are particularly important to monitor for cardiovascular therapies such as transcatheter left atrial appendage occlusion (LAAO). However, there are limited data comparing stroke events in contemporary claims data to events reported in large nationwide registries with comprehensive data collection. Aim: To compare stroke events in Medicare claims data with adjudicated registry-reported events in the National Cardiovascular Data Registry Left Atrial Appendage Occlusion (LAAO) Registry. Methods: LAAO Registry data were linked with 2016-2022 Medicare inpatient claims. Primary diagnosis ICD-10 codes from inpatient hospitalizations were compared to adjudicated registry-reported ischemic and hemorrhagic stroke events occurring after discharge with up to two years of follow-up. In multiple secondary analyses, primary and secondary diagnosis codes were counted as events in claims data and patients with prior history of stroke were also excluded. Results: The cohort included 71,043 adults ≥65 years old in the LAAO Registry with inpatient Medicare data. Following LAAO hospitalization discharge, there were 1405 claims-based and 1308 registry-reported total stroke events. Sensitivity and positive predictive value (PPV) of claims for identifying registry-reported events were 60.8% and 50.5% for ischemic stroke (kappa statistic 0.55), 42.7% and 50.5% for hemorrhagic stroke (kappa 0.46), and 58.1% and 54.1% for all stroke (kappa 0.55), respectively. Inclusion of secondary diagnosis codes improved sensitivity of claims (70.3% for ischemic stroke, 53.9% for hemorrhagic stroke and 67.0% for all stroke), but PPV decreased (44.8% for ischemic stroke, 34.8% for hemorrhagic stroke and 47.8% for all stroke). Specificity and negative predictive values of claims data were >98% for all stroke types. Findings were similar after excluding patients with prior history of stroke ( Table ). Conclusions: Among Medicare patients in the LAAO Registry, administrative claims data had moderate agreement with adjudicated registry-reported stroke events. Claims data may not reliably capture all stroke events in real-world practice, and additional event ascertainment methods may be needed to ensure accurate assessment of stroke in post-marketing surveillance efforts.

Identifying Hemolytic Disease of the Fetus and Newborn within a Large Integrated Health Care System.

This study aims to identify hemolytic disease of the fetus and newborn (HDFN) pregnancies using electronic health records (EHRs) from a large integrated health care system. A retrospective cohort study was performed among pregnant patients receiving obstetrical care at Kaiser Permanente Southern California health care system between January 1, 2008, and June 30, 2022. Using structured (diagnostic/procedural codes, medication, and laboratory records) and unstructured (clinical notes analyzed via natural language processing) data abstracted from EHRs, we extracted HDFN-specific "indicators" (maternal positive antibody test and abnormal antibody titer, maternal/infant HDFN diagnosis and blood transfusion, hydrops fetalis, infant intravenous immunoglobulin [IVIG] treatment, jaundice/phototherapy, and first administrated Rho[D] Immune Globulin) to identify potential HDFN pregnancies. Chart reviews and adjudication were then performed on select combinations of indicators for case ascertainment. HDFN due to ABO alloimmunization alone was excluded. The HDFN frequency and proportion of each combination were fully analyzed. Among the 464,711 eligible pregnancies, a total of 136 pregnancies were confirmed as HDFN pregnancies. The percentage of the HDFN-specific indicators ranged from 0.02% (infant IVIG treatment) to 34.53% (infant jaundice/phototherapy) among the eligible pregnancies, and 32.35% (infant IVIG treatment) to 100% (maternal positive antibody test) among the 136 confirmed HDFN pregnancies. Four combination groups of four indicators, four combination groups of five indicators, and the unique combination of six indicators showed 100% of HDFN pregnancies, while 80.88% of confirmed HDFN pregnancies had the indicator combination of maternal positive antibody test, maternal/infant HDFN diagnosis, and infant jaundice/phototherapy. We successfully identified HDFN pregnancies by leveraging a combination of medical indicators extracted from structured and unstructured data that may be used in future pharmacoepidemiologic studies. Traditional indicators (positive antibody test results, high titers, and clinical diagnosis codes) alone did not accurately identify HDFN pregnancies, highlighting an unmet need for improved practices in HDFN coding. · A case ascertainment method was developed to identify HDFN from structured and unstructured data.. · The method used in this study may be used in future pharmacoepidemiologic studies.. · The study highlighted an unmet need for improved practices in HDFN coding..

Use of Artificial Intelligence to Predict Outcomes in Mild Aortic Valve Stenosis

Abstract Background and Aims Aortic stenosis (AS) is a common and progressive disease which, if left untreated, results in increased morbidity and mortality. Monitoring and follow-up care can be challenging due to significant variability in disease progression. This study aimed to develop machine learning models to predict the risks of disease progression and mortality in patients with mild aortic stenosis. Methods A comprehensive database including 9611 patients with serial transthoracic echocardiograms was collected from a single institution across 3 clinical sites. The dataset included parameters from echocardiograms, electrocardiograms, laboratory values, and diagnosis codes. Data from a single clinical site was preserved as an independent test group. Machine learning models were trained to identify progression to severe stenosis and all-cause mortality and tested in their performance for endpoints at 2 and 5 years. Results In the independent test group, the AS progression model differentiated those with progression to severe AS within 2 and 5 years with an area under the curve (AUC) of 0.86 for both. The feature of greatest importance was aortic valve mean gradient, followed by other valve hemodynamic measurements including valve area and dimensionless index. The mortality model identified those with mortality within 2 years and 5 years with an AUC of 0.84 and 0.87, respectively. Smaller reduced-input validation models had similarly robust findings. Conclusions Machine learning models can be used in patients with mild AS to identify those at high risk for disease progression and mortality. Implementation of such models may facilitate real-time, patient-specific follow-up recommendations.

Assessing acute kidney injury risk after COVID vaccination and infection in a large cohort study

Acute kidney injury (AKI) has been noticed after both COVID-19 vaccination and infection, affecting risk-benefit evaluations and vaccine hesitancy. We conducted a large-scale N3C cohort study to compare AKI incidence following COVID-19 vaccination and infection. Participants from December 2020 to August 2023 were divided into two groups based on their initially observed COVID-19 antigen exposure: COVID-19 vaccination group (n = 2,953,219) and COVID-19 infection group (n = 3,616,802). AKI was defined by diagnostic codes and serum creatinine changes within a 30 day follow-up window after exposure. The absolute risk of AKI was 0.66% in the vaccination group versus 4.88% in the infection group. After adjusting for various confounders, COVID-19 infection was associated with a significantly higher risk of AKI than COVID-19 vaccination (aHR = 10.31, P < 0.001). Our study reveals that COVID-19 vaccination is associated with a significant lower AKI risk compared to COVID-19 infection.

Neighborhood socioeconomic status and postpartum depression among commercial health insurance enrollees: a retrospective cohort study

BackgroundPostpartum depression (PPD) is a prevalent psychological condition. Although the effect of obstetrical and maternal complications on PPD are well described, the impact of neighborhood socioeconomic status (SES) on PPD is relatively unexplored.ObjectivesThis study examined the relationship between neighborhood SES score and PPD.MethodsA summary SES measure for each U.S. zip code was constructed using income, education, and occupational 2021 Census data and linked to national commercial claims for 2017–2023. PPD status using diagnosis codes at outpatient and inpatient visits, and prescription drug use 3, 6, 9, and 12 months postpartum, were determined. Multivariate analysis controlled for potential confounders.ResultsPPD prevalence in commercial claims was 11.48%. Patients with PPD had higher rates of obstetrical (OR: 1.555, p < .0001) and maternal complications (OR: 1.145, p < .0001), and more lifestyle risk factors (OR: 1.113, p < .0001). Comorbidity scores were higher for patients with PPD. Controlling for age and clinical factors, living in a disadvantaged neighborhood was associated with an increased incidence of PPD (OR: 1.137, p < .0001).LimitationsClaims data may include potential inaccurate coding of diagnoses/procedures. Clinical information is limited to conditions and treatments defined by ICD-10-CM codes. Area-based SES measures inevitably misclassify people on both ends of the socioeconomic spectrum (this misclassification is random; direction of bias is known).ConclusionsThe inverse and significant effect of area-based high SES on PPD rates demonstrates that preventive efforts may require interventions focusing on both the patient and the lived environment.

Can ChatGPT Support Clinical Coding Using the ICD-10-CM/PCS?

Introduction: With the growing development and adoption of artificial intelligence in healthcare and across other sectors of society, various user-friendly and engaging tools to support research have emerged, such as chatbots, notably ChatGPT. Objective: To investigate the performance of ChatGPT as an assistant to medical coders using the ICD-10-CM/PCS. Methodology: We conducted a prospective exploratory study between 2023 and 2024 over 6 months. A total of 150 clinical cases coded using the ICD-10-CM/PCS, extracted from technical coding books, were systematically randomized. All cases were translated into Portuguese (the native language of the authors) and English (the native language of the ICD-10-CM/PCS). These clinical cases varied in complexity levels regarding the quantity of diagnoses and procedures, as well as the nature of the clinical information. Each case was input into the 2023 ChatGPT free version. The coding obtained from ChatGPT was analyzed by a senior medical auditor/coder and compared with the expected results. Results: Regarding the correct codes, ChatGPT’s performance was higher by approximately 29 percentage points between diagnoses and procedures, with greater proficiency in diagnostic codes. The accuracy rate for codes was similar across languages, with rates of 31.0% and 31.9%. The error rate in procedure codes was substantially higher than that in diagnostic codes by almost four times. For missing information, a higher incidence was observed in diagnoses compared to procedures of slightly more than double the comparative rates. Additionally, there was a statistically significant excess of codes not related to clinical information, which was higher in procedures and nearly the same value in both languages under study. Conclusion: Given the ease of access to these tools, this investigation serves as an awareness factor, demonstrating that ChatGPT can assist the medical coder in directed research. However, it does not replace their technical validation in this process. Therefore, further developments of this tool are necessary to increase the quality and reliability of the results.

Association Between Follow-Up Visit Timing After A Concussion and Subsequent Care Seeking in Children and Youth: A Population-Based Study in British Columbia.

To assess the relationship between follow-up visit timing and occurrence of the first subsequent health care seeking visit. The province of British Columbia, Canada. A total of 21029 children and youth who were diagnosed with an initial concussion from January 1, 2016, to December 31, 2017. These data were obtained from Population Data BC. A retrospective, descriptive correlational study. Follow-up timing was measured categorically as timely (4weeks), delayed (1-3months), or no follow-up; the occurrence of a subsequent health care visit beyond 3months postinjury was measured up to 12months at 3-month intervals (ie, 4-6, 7-9, and 10-12months). These variables were measured using diagnostic codes for concussion, post-concussion syndrome, and the 17 concussion symptoms. After controlling for sociodemographic characteristics, having a delayed follow-up, relative to timely follow-up, was associated with higher odds of a subsequent health care seeking visit at 4 to 6months (odds ratio [OR]=2.68; confidence interval [CI], 2.08-3.47), 7 to 9months (OR=1.71; CI, 1.21-2.40), and 10 to 12months (OR=1.67; CI, 1.13-2.48). In contrast to having a delayed follow-up, having no follow-up, relative to timely follow-up, was associated with not having a subsequent health care seeking visit at 4 to 6months (OR=0.57; CI, 0.48-0.67) and 7 to 9months (OR=0.79; CI, 0.66-0.96), respectively. Follow-up visit timing after a concussion in children and youth is associated with subsequent health care seeking. Greater efforts are needed to investigate the importance of initial follow-up visit timing, as they may play an integral role in injury prevention and symptom management following injury.

Association of advance care planning with hospital use and costs at the end of life: a population-based retrospective cohort study

ObjectiveTo investigate associations between the availability and timing of digitally available advance care planning (ACP) documents and hospital use and costs during the last 6 months of life.DesignRetrospective population-based cohort...