Objective To provide focused evaluation of predictive modeling of electronic medical record (EMR) data to predict 30 day hospital readmission. Design Systematic review. Data source Ovid Medline, Ovid Embase, CINAHL, Web of Science, and Scopus from January 2015 to January 2019. Eligibility criteria for selecting studies All studies of predictive models for 28 day or 30 day hospital readmission that used EMR data. Outcome measures Characteristics of included studies, methods of prediction, predictive features, and performance of predictive models. Results Of 4442 citations reviewed, 41 studies met the inclusion criteria. Seventeen models predicted risk of readmission for all patients and 24 developed predictions for patient specific populations, with 13 of those being developed for patients with heart conditions. Except for two studies from the UK and Israel, all were from the US. The total sample size for each model ranged between 349 and 1 195 640. Twenty five models used a split sample validation technique. Seventeen of 41 studies reported C statistics of 0.75 or greater. Fifteen models used calibration techniques to further refine the model. Using EMR data enabled final predictive models to use a wide variety of clinical measures such as laboratory results and vital signs; however, use of socioeconomic features or functional status was rare. Using natural language processing, three models were able to extract relevant psychosocial features, which substantially improved their predictions. Twenty six studies used logistic or Cox regression models, and the rest used machine learning methods. No statistically significant difference (difference 0.03, 95% confidence interval −0.0 to 0.07) was found between average C statistics of models developed using regression methods (0.71, 0.68 to 0.73) and machine learning (0.74, 0.71 to 0.77). Conclusions On average, prediction models using EMR data have better predictive performance than those using administrative data. However, this improvement remains modest. Most of the studies examined lacked inclusion of socioeconomic features, failed to calibrate the models, neglected to conduct rigorous diagnostic testing, and did not discuss clinical impact.

Purpose: Individuals with cerebral palsy (CP) are susceptible to early development of highburden medical conditions, which may place a considerable strain on health care resources. However, little is known about the prevalence of high-burden medical conditions or health care resource utilization among adults with CP. The purpose of this study was to determine the prevalence of high-burden medical conditions and health care resource utilization and costs among adults with CP, as compared to adults without CP. Patients and methods: Cross-sectional data from the 2016 Optum Clinformatics ® Data Mart, a de-identified nationwide claims database of beneficiaries from a single private payer in the US. ICD-10-CM diagnosis codes were used to identify all medical conditions among beneficiaries with and without CP who were between 18 and 64 years of age. Medical and outpatient pharmacy claims were used to identify annual all-cause health care resource utilization and health care costs as standardized reimbursement and patient out-of-pocket costs. Results: Adults with CP (n=5,555) had higher prevalence and odds of all medical conditions compared to adults without CP (OR=1.3-5.8; all P<0.05), except cancer (OR=1.1; 95% CI=0.9-1.3). Adults with CP had greater annual mean counts of all health care service types (eg, inpatient, emergency department) compared to adults without CP (all P<0.01). Adults with CP had higher unadjusted standardized reimbursement (mean difference=$16,288; cost ratio [CR]=3.0; 95% CI=2.9-3.1) and patient out-of-pocket (mean difference=$778; CR=1.7; 95% CI=1.6-1.7) costs compared to adults without CP. After adjusting for all prevalent medical conditions, adults with CP still had higher standardized reimbursement (CR=2.5; 95% CI=2.5-2.6) and patient out-of-pocket (CR=1.8; 95% CI=1.7-1.8) costs. Conclusion: Adults with CP have a higher prevalence of high-burden medical conditions, health care resource utilization, and health care costs compared to adults without CP. Study findings suggest the need for earlier screening strategies and preventive medical services to quell the disease and economic burden attributable to adults with CP.

No abstract

Aim To develop a new comorbidity index for adults with cerebral palsy (CP), the Whitney Comorbidity Index (WCI), which includes relevant comorbidities for this population and better predicts mortality than the Charlson Comorbidity Index (CCI) and Elixhauser Comorbidity Index (ECI). Method Data from the Optum Clinformatics Data Mart was used for this retrospective cohort study. Diagnosis codes were used to identify adults aged 18 years or older with CP (n=1511 females, n=1511 males; mean [SD; range] age=48y [19y 2mo; 18–89y]) and all comorbidities in the year 2014. The WCI was developed based on the comorbidities of the CCI and ECI and other relevant comorbidities associated with 2‐year mortality using Cox regression and competing risk analysis. The WCI was examined as unweighted (WCIunw) and weighted (WCIw). The model fit and discrimination (C‐statistic) of each index was assessed using Cox regression. Results Twenty‐seven comorbidities were included in the WCI; seven new comorbidities that were not part of the CCI or ECI were added. The WCIunw and WCIw showed a better model fit and discrimination for 1‐ and 2‐year mortality compared to the CCI and ECI. The WCIunw and WCIw were strong predictors for 1‐ and 2‐year mortality (C‐statistic [95% confidence interval] ranging from 0.81 [0.76–0.85] to 0.88 [0.82–0.94]). Interpretation The new WCI, designed to include clinically relevant comorbidities, provides a better model fit and discrimination of mortality for adults with CP. What this paper adds Common comorbidity indices exclude relevant comorbidities for adults with cerebral palsy (CP). A new comorbidity index for adults with CP was created by harmonizing clinical theory and data‐driven methods. The Whitney Comorbidity Index better predicted 1‐ and 2‐year mortality than other commonly used comorbidity indices.

OBJECTIVE To examine the association between hearing aids (HAs) and time to diagnosis of Alzheimer disease (AD) or dementia, anxiety or depression, and injurious falls among adults, aged 66 years and older, within 3 years of hearing loss (HL) diagnosis. DESIGN Retrospective cohort study. SETTING We used 2008 to 2016 national longitudinal claims data (based on office visit, inpatient, or outpatient healthcare encounters) from a large private payer. We used Kaplan‐Meier curves to examine unadjusted disease‐free survival and crude and adjusted Cox regression models to examine associations between HAs and time to diagnosis of three age‐related/HL‐associated conditions within 3 years of HL diagnosis. All models were adjusted for age, sex, race/ethnicity, census divisions, and prior diagnosis of cardiovascular conditions, hypertension, hypercholesterolemia, obesity, and diabetes. PARTICIPANTS The participants included 114 862 adults, aged 66 years and older, diagnosed with HL. MEASUREMENT Diagnosis of (1) AD or dementia; (2) depression or anxiety; and (3) injurious falls. INTERVENTION Use of HAs. RESULTS Large sex and racial/ethnic differences exist in HA use. Approximately 11.3% of women vs 13.3% of men used HAs (95% confidence interval [CI] difference = −0.024 to −0.016). Approximately 13.6% of whites (95% CI = 0.13‐0.14) vs 9.8% of blacks (95% CI = 0.09‐0.11) and 6.5% of Hispanics (95% CI = 0.06‐0.07) used HAs. The risk‐adjusted hazard ratios of being diagnosed with AD/dementia, anxiety/depression, and injurious falls within 3 years after HL diagnosis, for those who used HAs vs those who did not, were 0.82 (95% CI = 0.76‐0.89), 0.89 (95% CI = 0.86‐0.93), and 0.87 (95% CI = 0.80‐0.95), respectively. CONCLUSIONS Use of HAs is associated with delayed diagnosis of AD, dementia, depression, anxiety, and injurious falls among older adults with HL. Although we have shown an association between use of HAs and reduced risk of physical and mental decline, randomized trials are needed to determine whether, and to what extent, the relationship is causal. J Am Geriatr Soc 67:2362–2369, 2019