Search

Background: This paper explores the importance of electronic medical records (EMR) for predicting 30-day all-cause non-elective readmission risk of patients and presents a comparison of prediction performance of commonly used methods.Methods: The data are extracted from eight Advocate Health Care hospitals. Index admissions are excluded from the cohort if they are observation, inpatient admissions for psychiatry, skilled nursing, hospice, rehabilitation, maternal and newborn visits, or if the patient expires during the index admission. Data are randomly and repeatedly divided into fitting and validating sets for cross validations. Approaches including LACE, STEPWISE logistic, LASSO logistic, and AdaBoost, are compared with sample sizes varying from 2,500 to 80,000.Results: Our results confirm that LACE has moderate discrimination power with the area under receiver operating characteristic curve (AUC) around 0.65-0.66, which can be improved to 0.73-0.74 when additional variables from EMR are considered. These variables include Inpatient in the last six months, Number of emergency room visits or inpatients in the last year, Braden score, Polypharmacy, Employment status, Discharge disposition, Albumin level, and medical condition variables such as Leukemia, Malignancy, Renal failure with hemodialysis, History of alcohol substance abuse, Dementia and Trauma. When sample size is small (≤5000), LASSO is the best; when sample size is large (≥20,000), the predictive performance is similar. The STEPWISE method has a slightly lower AUC (0.734) comparing to LASSO (0.737) and AdaBoost (0.737). More than one half of the selected predictors can be false positives when using a single method and a single division of fitting/validating data.Conclusions: True predictors can be identified by repeatedly dividing data into fitting/validating subsets and referring the final model based on summarizing results. LASSO is a better alternative to the STEPWISE logistic regression, especially when sample size is not large. The evidence for adequate sample size can be explored by fitting models on gradually reduced samples. Our model comparison strategy is not only good for 30-day all-cause non-elective readmission risk predictions, but also applicable to other types of predictive models in clinical studies. [ABSTRACT FROM AUTHOR]

Background: The results of multivariable regression models are usually summarized in the form of parameter estimates for the covariates, goodness-of-fit statistics, and the relevant p-values. These statistics do not inform us about whether covariate information will lead to any substantial improvement in prediction. Predictive ability measures can be used for this purpose since they provide important information about the practical significance of prognostic factors. R²-type indices are the most familiar forms of such measures in survival models, but they all have limitations and none is widely used. Methods: In this paper, we extend the total gain (TG) measure, proposed for a logistic regression model, to survival models and explore its properties using simulations and real data. TG is based on the binary regression quantile plot, otherwise known as the predictiveness curve. Standardised TG ranges from 0 (no explanatory power) to 1 ('perfect' explanatory power). Results: The results of our simulations show that unlike many of the other R²-type predictive ability measures, TG is independent of random censoring. It increases as the effect of a covariate increases and can be applied to different types of survival models, including models with time-dependent covariate effects. We also apply TG to quantify the predictive ability of multivariable prognostic models developed in several disease areas. Conclusions: Overall, TG performs well in our simulation studies and can be recommended as a measure to quantify the predictive ability in survival models. [ABSTRACT FROM AUTHOR]

Background: Automated feature selection methods such as the Least Absolute Shrinkage and Selection Operator (LASSO) have recently gained importance in the prediction of quality-related outcomes as well as the risk-adjustment of quality indicators in healthcare. The methods that have been used so far, however, do not account for the fact that patient data are typically nested within hospitals. Methods: Therefore, we aimed to demonstrate how to account for the multilevel structure of hospital data with LASSO and compare the results of this procedure with a LASSO variant that ignores the multilevel structure of the data. We used three different data sets (from acute myocardial infarcation, COPD, and stroke patients) with two dependent variables (one numeric and one binary), on which different LASSO variants with and without consideration of the nested data structure were applied. Using a 20-fold sub-sampling procedure, we tested the predictive performance of the different LASSO variants and examined differences in variable importance. Results: For the metric dependent variable Duration Stay, we found that inserting hospitals led to better predictions, whereas for the binary variable Mortality, all methods performed equally well. However, in some instances, the variable importances differed greatly between the methods. Conclusion: We showed that it is possible to take the multilevel structure of data into account in automated predictor selection and that this leads, at least partly, to better predictive performance. From the perspective of variable importance, including the multilevel structure is crucial to select predictors in an unbiased way under consideration of the structural differences between hospitals. [ABSTRACT FROM AUTHOR]

Objective: To predict the influencing factors of neonatal pneumonia in pregnant women with diabetes mellitus using a Bayesian network model. By examining the intricate network connections between the numerous variables given by Bayesian networks (BN), this study aims to compare the prediction effect of the Bayesian network model and to analyze the influencing factors directly associated to neonatal pneumonia. Method: Through the structure learning algorithms of BN, Naive Bayesian (NB), Tree Augmented Naive Bayes (TAN), and k-Dependence Bayesian Classifier (KDB), complex networks connecting variables were presented and their predictive abilities were tested. The BN model and three machine learning models computed using the R bnlean package were also compared in the data set. Results: In constraint-based algorithms, three algorithms had different presentation DAGs. KDB had a better prediction effect than NB and TAN, and it achieved higher AUC compared with TAN. Among three machine learning modes, Support Vector Machine showed a accuracy rate of 91.04% and 67.88% of precision, which was lower than TAN (92.70%; 72.10%). Conclusion: KDB was applicable, and it can detect the dependencies between variables, identify more potential associations and track changes between variables and outcome. [ABSTRACT FROM AUTHOR]

Background: Our population is ageing and in 2050 more than one out of five people will be 60 years or older; 80% of whom will be living in a low-and-middle income country. Living longer does not entail living healthier; however, there is not a widely accepted measure of healthy ageing hampering policy and research. The World Health Organization defines healthy ageing as the process of developing and maintaining functional ability that will enable well-being in older age. We aimed to create a healthy ageing index (HAI) in a subset of six low-and-middle income countries, part of the 10/66 study, by using items of functional ability and intrinsic capacity.Methods: The study sample included residents 65-years old and over (n = 12,865) from catchment area sites in Cuba, Dominican Republic, Peru, Venezuela, Mexico and Puerto Rico. Items were collected by interviewing participants or key informants between 2003 and 2010. Two-stage factor analysis was employed and we compared one-factor, second-order and bifactor models. The psychometric properties of the index, including reliability, replicability, unidimensionality and concurrent convergent validity as well as measurement invariance per ethnic group and gender were further examined in the best fit model.Results: The bifactor model displayed superior model fit statistics supporting that a general factor underlies the various items but other subdomain factors are also needed. The HAI indicated excellent reliability (ω = 0.96, ωΗ = 0.84), replicability (H = 0.96), some support for unidimensionality (Explained Common Variance = 0.65) and some concurrent convergent validity with self-rated health. Scalar measurement invariance per ethnic group and gender was supported.Conclusions: A HAI with excellent psychometric properties was created by using items of functional ability and intrinsic capacity in a subset of six low-and-middle income countries. Further research is needed to explore sub-population differences and to validate this index to other cultural settings. [ABSTRACT FROM AUTHOR]

Background: Patient-Reported Outcome Measures (PROMs) have been proposed for benchmarking health care quality across hospitals, which requires extensive case-mix adjustment. The current study's aim was to develop and compare case-mix models for mortality, a functional outcome, and a patient-reported outcome measure (PROM) in ischemic stroke care.Methods: Data from ischemic stroke patients, admitted to four stroke centers in the Netherlands between 2014 and 2016 with available outcome information (N = 1022), was analyzed. Case-mix adjustment models were developed for mortality, modified Rankin Scale (mRS) scores and EQ-5D index scores with respectively binary logistic, proportional odds and linear regression models with stepwise backward selection. Predictive ability of these models was determined with R-squared (R2) and area-under-the-receiver-operating-characteristic-curve (AUC) statistics.Results: Age, NIHSS score on admission, and heart failure were the only common predictors across all three case-mix adjustment models. Specific predictors for the EQ-5D index score were sex (β = 0.041), socio-economic status (β = - 0.019) and nationality (β = - 0.074). R2-values for the regression models for mortality (5 predictors), mRS score (9 predictors) and EQ-5D utility score (12 predictors), were respectively R2 = 0.44, R2 = 0.42 and R2 = 0.37.Conclusions: The set of case-mix adjustment variables for the EQ-5D at three months differed considerably from the set for clinical outcomes in stroke care. The case-mix adjustment variables that were specific to this PROM were sex, socio-economic status and nationality. These variables should be considered in future attempts to risk-adjust for PROMs during benchmarking of hospitals. [ABSTRACT FROM AUTHOR]

Background: In the absence of a consensus on definition and measurement of healthy ageing, we created a healthy ageing index tallying with the functional ability framework provided by the World Health Organization. To create this index, we employed items of functional ability and intrinsic capacity. The current study aims to establish the predictive validity and discrimination properties of this healthy ageing index in settings in Latin American, part of the 10/66 cohort.Methods: Population-based cohort studies including 12,865 people ≥65 years old in catchment areas of Cuba, Dominican Republic, Venezuela, Mexico and Peru. We employed latent variable modelling to estimate the healthy ageing scores of each participant. We grouped participants according to the quintiles of the healthy ageing score distribution. Cox's proportional hazard models for mortality and sub-hazard (competing risks) models for incident dependence (i.e. needing care) were calculated per area after a median of 3.9 years and 3.7 years, respectively. Results were pooled together via fixed-effects meta-analysis. Our findings were compared with those obtained from self-rated health.Results: Participants with lowest levels, compared to participants with highest level of healthy ageing, had increased risk of mortality and incident dependence, even after adjusting for sociodemographic and health conditions (HR: 3.25, 95%CI: 2.63-4.02; sub-HR: 5.21, 95%CI: 4.02-6.75). Healthy ageing scores compared to self-rated health had higher population attributable fractions (PAFs) for mortality (43.6% vs 19.3%) and incident dependence (58.6% vs 17.0%), and better discriminative power (Harrell's c-statistic: mortality 0.74 vs 0.72; incident dependence 0.76 vs 0.70).Conclusion: These results provide evidence that our healthy ageing index could be a valuable tool for prevention strategies as it demonstrated predictive and discriminative properties. Further research in other cultural settings will assist moving from a theoretical conceptualisation of healthy ageing to a more practical one. [ABSTRACT FROM AUTHOR]