Development and Validation of an Obstetric Comorbidity Risk Score for Clinical Use

Hepatocellular Carcinoma in Metabolic Dysfunction-Associated Steatotic Liver Disease

Importance

In the US, hepatocellular carcinoma (HCC) has been the most rapidly increasing cancer since 1980, and metabolic dysfunction-associated steatotic liver disease (MASLD) is expected to soon become the leading cause of HCC.

Objective

To develop a prediction model for HCC incidence in a cohort of patients with MASLD.

Design, Setting, and Participants

This prognostic study was conducted among patients aged at least 18 years with MASLD, identified using diagnosis of MASLD using International Classification of Diseases, Ninth Revision (ICD-9) or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) diagnosis codes; natural language processing of radiology imaging report text, which identified patients who had imaging evidence of MASLD but had not been formally diagnosed; or the Dallas Steatosis Index, a risk equation that identifies individuals likely to have MASLD with good precision. Patients were enrolled from Kaiser Permanente Northern California, an integrated health delivery system with more than 4.6 million members, with study entry between January 2009 and December 2018, and follow-up until HCC development, death, or study termination on September 30, 2021. Statistical analysis was performed during February 2023 and January 2024.

Exposure

Data were extracted from the electronic health record and included 18 routinely measured factors associated with MASLD.

Main Outcome and Measures

The cohort was split (70:30) into derivation and internal validation sets; extreme gradient boosting was used to model HCC incidence. HCC risk was divided into 3 categories, with the cumulative estimated probability of HCC 0.05% or less classified as low risk; 0.05% to 0.09%, medium risk; and 0.1% or greater, high risk.

Results

A total of 1 811 461 patients (median age [IQR] at baseline, 52 [41-63] years; 982 300 [54.2%] female) participated in the study. During a median (range) follow-up of 9.3 (5.8-12.4) years, 946 patients developed HCC, for an incidence rate of 0.065 per 1000 person-years. The model achieved an area under the curve of 0.899 (95% CI, 0.882-0.916) in the validation set. At the medium-risk threshold, the model had a sensitivity of 87.5%, specificity of 81.4%, and a number needed to screen of 406. At the high-risk threshold, the model had a sensitivity of 78.4%, a specificity of 90.1%, and a number needed to screen of 241.

Conclusions and Relevance

This prognostic study of more than 1.8 million patients with MASLD used electronic health record data to develop a prediction model to discriminate between individuals with and without incident HCC with good precision. This model could serve as a starting point to identify patients with MASLD who may need intervention and/or HCC surveillance.

Maternal Characteristics and Rates of Unexpected Complications in Term Newborns by Hospital

Importance

The Joint Commission Unexpected Complications in Term Newborns measure characterizes newborn morbidity potentially associated with quality of labor and delivery care. Infant exclusions isolate relatively low-risk births, but unexpected newborn complications (UNCs) are not adjusted for maternal factors that may be associated with outcomes independently of hospital quality.

Objective

To investigate the association between maternal characteristics and hospital UNC rates.

Design, Setting, and Participants

This cohort study was conducted using linked 2016 to 2018 New York City birth and hospital discharge datasets among 254 259 neonates at low risk (singleton, ≥37 weeks, birthweight ≥2500 g, and without preexisting fetal conditions) at 39 hospitals. Logistic regression was used to calculate unadjusted hospital-specific UNC rates and replicated analyses adjusting for maternal covariates. Hospitals were categorized into UNC quintiles; changes in quintile ranking with maternal adjustment were examined. Data analyses were performed from December 2022 to July 2023.

Main Outcomes and Measures

UNCs were classified according to Joint Commission International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) criteria. Maternal preadmission comorbidities, obstetric factors, social characteristics, and hospital characteristics were ascertained.

Results

Among 254 259 singleton births at 37 weeks or later who were at low risk (125 245 female [49.3%] and 129 014 male [50.7%]; 71 768 births [28.2%] to Hispanic, 47 226 births [18.7%] to non-Hispanic Asian, 42 682 births [16.8%] to non-Hispanic Black, and 89 845 births [35.3%] to non-Hispanic White mothers and 2738 births [1.0%] to mothers with another race or ethnicity), 148 393 births (58.4%) were covered by Medicaid and 101 633 births (40.0%) were covered by commercial insurance. The 2016 to 2018 cumulative UNC incidence in New York City hospitals was 37.1 UNCs per 1000 births. Infants of mothers with preadmission risk factors had increased UNC risk; for example, among mothers with vs without preeclampsia, there were 104.4 and 35.8 UNCs per 1000 births, respectively. Among hospitals, unadjusted UNC rates ranged from 15.6 to 215.5 UNCs per 1000 births and adjusted UNC rates ranged from 15.6 to 194.0 UNCs per 1000 births (median [IQR] change from adjustment, 1.4 [-4.7 to 1.0] UNCs/1000 births). The median (IQR) change per 1000 births for adjusted vs unadjusted rates showed that hospitals with low (<601 deliveries/year; -2.8 [-7.0 to -1.6] UNCs) to medium (601 to <954 deliveries/year; -3.9 [-7.1 to -1.9] UNCs) delivery volume, public ownership (-3.6 [-6.2 to -2.3] UNCs), or high proportions of Medicaid-insured (eg, ≥90.72%; -3.7 [-5.3 to -1.9] UNCs), Black (eg, ≥32.83%; -5.3 [-9.1 to -2.2] UNCs), or Hispanic (eg, ≥6.25%; -3.7 [-5.3 to -1.9] UNCs) patients had significantly decreased UNC rates after adjustment, while rates increased or did not change in hospitals with the highest delivery volume, private ownership, or births to predominantly White or privately insured individuals. Among all 39 hospitals, 7 hospitals (17.9%) shifted 1 quintile comparing risk-adjusted with unadjusted quintile rankings.

Conclusions and Relevance

In this study, adjustment for maternal case mix was associated with small overall changes in hospital UNC rates. These changes were associated with performance assessment for some hospitals, and these results suggest that profiling on this measure should consider the implications of small changes in rates for hospitals with higher-risk obstetric populations.

Identifying Predictors of Homelessness Among Adults in a Large Integrated Health System in Northern California

Introduction Homelessness contributes to worsening health and increased health care costs. There is little published research that leverages rich electronic health record (EHR) data to predict future homelessness risk and inform interventions to address it. The authors' objective was to develop a model for predicting future homelessness using individual EHR and geographic data covariates. Methods This retrospective cohort study included 2,543,504 adult members (≥ 18 years old) from Kaiser Permanente Northern California and evaluated which covariates predicted a composite outcome of homelessness status (hospital discharge documentation of a homeless patient, medical diagnosis of homelessness, approved medical financial assistance application for homelessness, and/or "homeless/shelter" in address name). The predictors were measured in 2018-2019 and included prior diagnoses and demographic and geographic data. The outcome was measured in 2020. The cohort was split (70:30) into a derivation and validation set, and logistic regression was used to model the outcome. Results Homelessness prevalence was 0.35% in the overall sample. The final logistic regression model included 26 prior diagnoses, demographic, and geographic-level predictors. The regression model using the validation set had moderate sensitivity (80.4%) and specificity (83.2%) for predicting future cases of homelessness and achieved excellent classification properties (area under the curve of 0.891 [95% confidence interval = 0.884-0.897]). Discussion This prediction model can be used as an initial triage step to enhance screening and referral tools for identifying and addressing homelessness, which can improve health and reduce health care costs. Conclusions EHR data can be used to predict chance of homelessness at a population health level.

Prediction of Emergency Cesarean Section Using Machine Learning Methods: Development and External Validation of a Nationwide Multicenter Dataset in Republic of Korea

This study was a multicenter retrospective cohort study of term nulliparous women who underwent labor, and was conducted to develop an automated machine learning model for prediction of emergent cesarean section (CS) before onset of labor. Nine machine learning methods of logistic regression, random forest, Support Vector Machine (SVM), gradient boosting, extreme gradient boosting (XGBoost), light gradient boosting machine (LGBM), k-nearest neighbors (KNN), Voting, and Stacking were applied and compared for prediction of emergent CS during active labor. External validation was performed using a nationwide multicenter dataset for Korean fetal growth. A total of 6549 term nulliparous women was included in the analysis, and the emergent CS rate was 16.1%. The C-statistics values for KNN, Voting, XGBoost, Stacking, gradient boosting, random forest, LGBM, logistic regression, and SVM were 0.6, 0.69, 0.64, 0.59, 0.66, 0.68, 0.68, 0.7, and 0.69, respectively. The logistic regression model showed the best predictive performance with an accuracy of 0.78. The machine learning model identified nine significant variables of maternal age, height, weight at pre-pregnancy, pregnancy-associated hypertension, gestational age, and fetal sonographic findings. The C-statistic value for the logistic regression machine learning model in the external validation set (1391 term nulliparous women) was 0.69, with an overall accuracy of 0.68, a specificity of 0.83, and a sensitivity of 0.41. Machine learning algorithms with clinical and sonographic parameters at near term could be useful tools to predict individual risk of emergent CS during active labor in nulliparous women.