BACKGROUND: Electronic medical data (EMD) from electronic health records of general practice computer systems have enormous research potential, yet many variables are unreliable. OBJECTIVE: The aim of this study was to compare selected data variables from general practice EMD with a reliable, representative national dataset (Bettering the Evaluation and Care of Health (BEACH)) in order to validate their use for primary care research. METHOD: EMD variables were compared with encounter data from the nationally representative BEACH program using χ2 tests and robust 95% confidence intervals to test their validity (measure what they reportedly measure). The variables focused on for this study were patient age, sex, smoking status and medications prescribed at the visit. RESULTS: The EMD sample from six general practices in the Illawarra region of New South Wales, Australia, yielded data on 196,515 patient encounters. Details of 90,553 encounters were recorded in the 2013 BEACH dataset from 924 general practitioners. No significant differences in patient age ( p = 0.36) or sex ( p = 0.39) were found. EMD had a lower rate of current smokers and higher average scripts per visit, but similar prescribing distribution patterns. CONCLUSION: Validating EMD variables offers avenues for improving primary care delivery and measuring outcomes of care to inform clinical practice and health policy.
BACKGROUND: Electronic medical data (EMD) from electronic health records of general practice computer systems have enormous research potential, yet many variables are unreliable. OBJECTIVE: The aim of this study was to compare selected data variables from general practice EMD with a reliable, representative national dataset (Bettering the Evaluation and Care of Health (BEACH)) in order to validate their use for primary care research. METHOD: EMD variables were compared with encounter data from the nationally representative BEACH program using χ2 tests and robust 95% confidence intervals to test their validity (measure what they reportedly measure). The variables focused on for this study were patient age, sex, smoking status and medications prescribed at the visit. RESULTS: The EMD sample from six general practices in the Illawarra region of New South Wales, Australia, yielded data on 196,515 patient encounters. Details of 90,553 encounters were recorded in the 2013 BEACH dataset from 924 general practitioners. No significant differences in patient age ( p = 0.36) or sex ( p = 0.39) were found. EMD had a lower rate of current smokers and higher average scripts per visit, but similar prescribing distribution patterns. CONCLUSION: Validating EMD variables offers avenues for improving primary care delivery and measuring outcomes of care to inform clinical practice and health policy.
Entities:
Keywords:
data quality; electronic health records; general practice; health information management; primary health care
Authors: Alex Kitsos; Gregory M Peterson; Matthew D Jose; Masuma Akter Khanam; Ronald L Castelino; Jan C Radford Journal: J Prim Care Community Health Date: 2019 Jan-Dec
Authors: Julie Gordon; Helena Britt; Graeme C Miller; Joan Henderson; Anthony Scott; Christopher Harrison Journal: Int J Environ Res Public Health Date: 2022-02-09 Impact factor: 3.390