Multiple imputation for completion of a national clinical audit dataset

Area of Science:

• Cardiology
• Biostatistics
• Health Services Research

Background:

• The Myocardial Ischaemia National Audit Project (MINAP) collects heart attack data from 234 hospitals in England and Wales.
• MINAP monitors hospital performance against government targets for myocardial infarction (heart attack) care.
• Missing data in observational databases like MINAP can introduce bias into statistical analyses.

Purpose of the Study:

• To reduce the impact of missing data in the MINAP database using multiple imputation.
• To provide a methodology for implementing an imputation scheme for complex health datasets.
• To generate multiply completed MINAP datasets for more efficient and reliable patient care analysis.

Main Methods:

• Multiple imputation techniques were employed to handle missing data points within the MINAP dataset.
• A detailed description of the implemented imputation scheme is provided.
• The study focuses on creating usable, complete datasets for various analytical purposes.

Main Results:

• The application of multiple imputation successfully addressed missing data challenges in the MINAP register.
• Multiply completed datasets were generated, suitable for a range of statistical analyses.
• The methodology facilitates more accurate inferences regarding patient care and outcomes.

Conclusions:

• Multiple imputation is an effective strategy for mitigating bias caused by missing data in the Myocardial Ischaemia National Audit Project.
• The developed approach enhances the utility of MINAP data for analyzing and improving coronary care.
• This work supports MINAP's objective of providing high-quality data for patient care research and evaluation.