Using Linked Health Service Data in The Evaluation of Innovative Models of Healthcare

Main Article Content

Rebecca Kwok-yee Pang
Velandai Srikanth
Gary Braun
Fergus McGee
Belinda Berry
Iain Edwards
Ruth Azzopardi
Nadine Andrew

Abstract

Introduction
Ageing populations place considerable burden on healthcare systems. Innovative methods to implement and evaluate new care models to reduce unwarranted hospital presentations, are needed.


Objectives and Approach
To use linked Electronic Medical Record (EMR) data to evaluate a pilot community-based model of care to reduce hospital presentations in a cohort of patients at-risk of hospital representations following an acute admission. Patients admitted to a metropolitan hospital with a non-surgical condition and
identified as being at-risk of readmission using a state-wide risk algorithm, were eligible to receive a 30-day care navigation model of care. The intervention group were matched to similar eligible patients, who received usual care using propensity score matching. Linked data were obtained from the EMR based Data Warehouse to provide information on subsequent readmissions and community-based health service contacts. Readmission rates were compared at 30, 60, and 90-days post-discharge using Cox Proportional Hazards Regression, adjusted for confounders and weighted by the propensity score. Descriptive analyses were used to compare demographics and healthcare utilisations.


Results
Data from 412,971 potentially eligible patients were extracted. Sixty-five received the intervention of which 63 were matched to 280 controls (aged 51-91 years, 54% female, 63% lived alone). At 30-days post-discharge (during intervention delivery) an 80% reduction in readmission rate in the intervention group was observed (Hazard Ratio [HR] 0.20; 95%Confidence Interval (CI) [0.01, 4.20]). At 60-days (HR 0.48; 95%CI [0.03, 8.05]) and 90-days (HR 0.83; 95%CI [0.06, 11.36]), readmission rate reductions decreased to 52% and 17% respectively. In the intervention group, 68.9% received an out-patient appointment compared to 40% of controls (P<0.001) and 7.7% received allied health compared to none of the controls (p=0.003) within 90-days post-discharge.


Conclusion/Implications
Results have informed hospital-wide service implementation thereby demonstrating the value of linked EMR data and advanced statistical methods in the evaluation of real-world healthcare.

Article Details

How to Cite
Pang, R. K.- yee, Srikanth, V., Braun, G., McGee, F., Berry, B., Edwards, I., Azzopardi, R. and Andrew, N. (2020) “Using Linked Health Service Data in The Evaluation of Innovative Models of Healthcare”, International Journal of Population Data Science, 5(5). doi: 10.23889/ijpds.v5i5.1528.

Most read articles by the same author(s)