Background: We evaluated the accuracy, limitations and potential sources of improvement in the clinical utility of the administrative dataset for acute medicine admissions.
Methods: Accuracy of clinical coding in 8888 patient discharges following an emergency medical hospital admission to a teaching hospital and a district hospital over 3 years was ascertained by a coding accuracy audit team in respect of the primary and secondary diagnoses, morbidities and financial variance.
Results: There was at least one change to the original coding in 4889 admissions (55%) and to the primary diagnosis of at least one finished consultant episodes of 1496 spells (16.8%). There were significant changes in the number of secondary diagnoses and the Charlson morbidity index following the audit. Charlson score increased in 8.2% and decreased in 2.3% of patients. An income variance of £816 977 (+5.0%) or £91.92 per patient was observed.
Conclusions: The importance and applications of coded healthcare big data within the NHS is increasing. The accuracy of coding is dependent on high-fidelity information transfer between clinicians and coders, which is prone to subjectivity, variability and error. We recommend greater involvement of clinicians as part of multidisciplinary teams to improve data accuracy, and urgent action to improve abstraction and clarity of assignment of strategic diagnoses like pneumonia and renal failure.
Keywords: clinical coding; health informatics; outcomes management; patient safety.
© The Author 2015. Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved. For permissions, please e-mail: email@example.com.