Can valid and practical risk-prediction or casemix adjustment models, including adjustment for comorbidity, be generated from English hospital administrative data (Hospital Episode Statistics)? A national observational study

Abstract

Background NHS hospitals collect a wealth of administrative data covering accident and emergency (A&E) department attendances, inpatient and day case activity, and outpatient appointments. Such data are increasingly being used to compare units and services, but adjusting for risk is difficult. Objectives To derive robust risk-adjustment models for various patient groups, including those admitted for heart failure (HF), acute myocardial infarction, colorectal and orthopaedic surgery, and outcomes adjusting for available patient factors such as comorbidity, using England’s Hospital Episode Statistics (HES) data. To assess if more sophisticated statistical methods based on machine learning such as artificial neural networks (ANNs) outperform traditional logistic regression (LR) for risk prediction. To update and assess for the NHS the Charlson index for comorbidity. To assess the usefulness of outpatient data for these models. Main outcome measures Mortality, readmission, return to theatre, outpatient non-attendance. For HF patients we considered various readmission measures such as diagnosis-specific and total within a year. Methods We systematically reviewed studies comparing two or more comorbidity indices. Logistic regression, ANNs, support vector machines and random forests were compared for mortality and readmission. Models were assessed using discrimination and calibration statistics. Competing risks proportional hazards regression and various count models were used for future admissions and bed-days. Results Our systematic review and empirical analysis suggested that for general purposes comorbidity is currently best described by the set of 30 Elixhauser comorbidities plus dementia. Model discrimination was often high for mortality and poor, or at best moderate, for other outcomes, for example c = 0.62 for readmission and c = 0.73 for death following stroke. Calibration was often good for procedure groups but poorer for diagnosis groups, with overprediction of low risk a common cause. The machine learning methods we investigated offered little beyond LR for their greater complexity and implementation difficulties. For HF, some patient-level predictors differed by primary diagnosis of readmission but not by length of follow-up. Prior non-attendance at outpatient appointments was a useful, strong predictor of readmission. Hospital-level readmission rates for HF did not correlate with readmission rates for non-HF; hospital performance on national audit process measures largely correlated only with HF readmission rates. Conclusions Many practical risk-prediction or casemix adjustment models can be generated from HES data using LR, though an extra step is often required for accurate calibration. Including outpatient data in readmission models is useful. The three machine learning methods we assessed added little with these data. Readmission rates for HF patients should be divided by diagnosis on readmission when used for quality improvement. Future work As HES data continue to develop and improve in scop...

Regional variation in hospitalisation and mortality in heart failure: comparison of England and Lombardy using multistate modelling

“…Most studies of readmissions ignore the competing risk of death or account for it by using a combined endpoint of death or readmission [10], which is unsatisfactory, not least because death and readmission are far from being of equal importance. Other modelling options include hurdle models and resource buckets, which can be useful for some questions [11] but are not flexible enough to show progression of the underlying chronic disease. Standard Cox and marginal models are inappropriate for the same reason as they ignore the serial nature of admissions [12]; extensions of these can deal with the clustering of patients within hospitals and multiple events per patient but not the serial nature of those events, or they can account for the competing risk of death under certain assumptions but not the repeat events.…”

Risk Modelling Framework for Emergency Hospital Readmission, Using Hospital Episode Statistics Inpatient Data

“…The inpatient In terms of data quality, there are parameters, associations and timeframes that are missing from HES due to confidentiality, different practices and limitations of defined fields. For instance, inconsistency in data due to changes in policies, care services and facilities, and large missing attendance records in A&E records [13].…”

Variations in mortality across the week following emergency admission to hospital: linked retrospective observational analyses of hospital episode data in England, 2004/5 to 2013/14

“…21 We adjusted for patient characteristics using age on admission, gender, ethnicity and socioeconomic deprivation. We linked patients' residential postcodes to the IMD at LSOA level and divided the deprivation scores into quintiles.…”

“…We first estimated each patient's risk of mortality within 30 days to categorise patients into five quintiles based on their predicted risk of mortality. We used a logistic regression model that included risk adjustment for interaction between gender and age, ethnicity, primary diagnosis (based on ICD-10 classification and defined according to SHMI-grouped clinical classifications software), 19 comorbidities measured using Elixhauser conditions, [20][21][22] source of admission (home or another hospital provider or institution), deprivation in the patient's area of residence (categorised in quintiles), admitting hospital and month of admission. l We then examined how the estimated risk of mortality varied by time of arrival at A&E and by mode of arrival.…”