Online Program

A Population-Based Study of Associations Between Attainment of Incentivized Primary Care Indicators and Emergency Hospital Admissions Among Those with Type 2 Diabetes in England (309831)

*Laura H Gunn, University of North Carolina at Charlotte & Imperial College London 
Azeem Majeed, Department of Primary Care and Public Health, Imperial College London 
Ailsa J McKay, Imperial College London 
German Molina, Independent Researcher 
Mariam Molokhia, Department of Population Health Sciences, King's College London 
Jonathan Valabhji, NHS England and NHS Improvement; St. Mary's Hospital; Imperial College London 
Eszter P Vamos, Department of Primary Care and Public Health, Imperial College London 

Keywords: Multivariate Poisson Models, Multivariate Cox Proportional Hazards Models, Propensity Score Matching, Type 2 Diabetes, Incentivized Care, Clinical Practice Research Datalink

The United Kingdom Clinical Practice Research Datalink, which contains longitudinal primary care data for 60 million patients, 16 million of whom are currently registered patients, was used to identify a population-based historical cohort of 84,441 adults with type 2 diabetes in England across 330 primary care practices between 2010-2017. These patients also had linked Hospital Episode Statistics and Office for National Statistics data available. We examined associations between attainment of Quality and Outcomes Framework (QOF) and National Diabetes Audit (NDA) primary care indicators and diabetes, cardiovascular, and all-cause emergency hospital admissions. Multivariate Poisson and Cox proportional hazards models, adjusting for sociodemographic, disease-related, lifestyle, and comorbidity factors, were fitted using propensity score matching. All-cause admission rates were lower among those who met the HbA1c and cholesterol indicators (incidence rate ratio (IRR)=0.91; 95% CI 0.89-0.92) and 0.87; 95% CI 0.86-0.89, respectively), with similar findings for diabetes and cardiovascular admissions. Patients achieving the blood pressure target had lower cardiovascular admission rates (IRR=0.98; 95% CI 0.96-0.99). Meeting all nine NDA processes was significantly associated with reductions in all types of emergency admissions by between 22-26%. Meeting the HbA1c or cholesterol QOF indicators, or completing 7-9 NDA processes, was also associated with a delay in all three types of emergency admissions. Out-of-hospital diabetes care optimization is needed to improve implementation of core interventions and reduce unplanned hospital admissions.