Grace Joshy1, Rosemary J Korda2, Walter P Abhayaratna3, Kay Soga2, Emily Banks4. 1. National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT, Australia grace.joshy@anu.edu.au. 2. National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT, Australia. 3. Medical School, College of Medicine, Biology and the Environment, Australian National University, Canberra, ACT, Australia. 4. National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT, Australia; The Sax Institute, Sydney, NSW, Australia.
Abstract
UNLABELLED: Objectives and importance of the study: Routine hospital administrative data provide an important source of information about cardiovascular disease (CVD) for health statistics reporting, health services and research. How such conditions are classified and grouped has implications for the use and interpretation of these data. International Classification of Diseases (ICD) diagnosis codes from hospital data collections are often used to classify CVD, but there is little published evidence on the most appropriate ways to use these codes to categorise CVD in a way that maximises the usefulness of hospital data for reporting and research. In particular, ICD codes under 'Diseases of the circulatory system' (I00-I99) are often grouped together into a general CVD category. However, this category is heterogeneous and combines common severe atherosclerotic and thrombotic CVDs (such as myocardial infarction and pulmonary embolism) with common, less severe and pathologically dissimilar conditions (such as varicose veins and haemorrhoids). In addition, hospital data collections contain a range of data fields, including those relating to primary and additional diagnoses and those relating to procedures. All of these have the potential to contribute valuable information on CVD. This paper proposes a pragmatic approach to using ICD diagnosis codes and procedure codes to capture major atherosclerotic and arteriovenous thromboembolic and related CVD. METHODS: We reviewed the ICD diagnosis codes and procedure codes and developed an algorithm for classifying and categorising major CVD diagnoses. This approach was then applied to linked hospitalisation data from individuals participating in the 45 and Up Study, a cohort study of 267 153 New South Wales residents aged 45 and over, to investigate the implications of the proposed approach for quantifying CVD. RESULTS: Large differences were observed in the numbers of events in grouped CVD outcomes, depending on the methods used. CONCLUSIONS: In cases where the reporting and research interest relates to incident disease, it may be appropriate to prioritise specific disease categories and pathological homogeneity.
UNLABELLED: Objectives and importance of the study: Routine hospital administrative data provide an important source of information about cardiovascular disease (CVD) for health statistics reporting, health services and research. How such conditions are classified and grouped has implications for the use and interpretation of these data. International Classification of Diseases (ICD) diagnosis codes from hospital data collections are often used to classify CVD, but there is little published evidence on the most appropriate ways to use these codes to categorise CVD in a way that maximises the usefulness of hospital data for reporting and research. In particular, ICD codes under 'Diseases of the circulatory system' (I00-I99) are often grouped together into a general CVD category. However, this category is heterogeneous and combines common severe atherosclerotic and thrombotic CVDs (such as myocardial infarction and pulmonary embolism) with common, less severe and pathologically dissimilar conditions (such as varicose veins and haemorrhoids). In addition, hospital data collections contain a range of data fields, including those relating to primary and additional diagnoses and those relating to procedures. All of these have the potential to contribute valuable information on CVD. This paper proposes a pragmatic approach to using ICD diagnosis codes and procedure codes to capture major atherosclerotic and arteriovenous thromboembolic and related CVD. METHODS: We reviewed the ICD diagnosis codes and procedure codes and developed an algorithm for classifying and categorising major CVD diagnoses. This approach was then applied to linked hospitalisation data from individuals participating in the 45 and Up Study, a cohort study of 267 153 New South Wales residents aged 45 and over, to investigate the implications of the proposed approach for quantifying CVD. RESULTS: Large differences were observed in the numbers of events in grouped CVD outcomes, depending on the methods used. CONCLUSIONS: In cases where the reporting and research interest relates to incident disease, it may be appropriate to prioritise specific disease categories and pathological homogeneity.
Authors: Rosemary J Korda; Wei Du; Cathy Day; Karen Page; Peter S Macdonald; Emily Banks Journal: BMC Health Serv Res Date: 2017-03-21 Impact factor: 2.655
Authors: Emily Banks; Grace Joshy; Rosemary J Korda; Bill Stavreski; Kay Soga; Sam Egger; Cathy Day; Naomi E Clarke; Sarah Lewington; Alan D Lopez Journal: BMC Med Date: 2019-07-03 Impact factor: 8.775