Validation of the Recording of Acute Exacerbations of COPD in UK Primary Care Electronic Healthcare Records.

Liam Smeeth,Sara L Thomas,John R Hurst,Hana Müllerová,Kourtney Davis,Kieran J Rothnie,Sanjay Haresh Chotirmall,Jennifer K Quint

doi:10.1371/journal.pone.0151357

Abstract

BackgroundAcute Exacerbations of COPD (AECOPD) identified from electronic healthcare records (EHR) are important for research, public health and to inform healthcare utilisation and service provision. However, there is no standardised method of identifying AECOPD in UK EHR. We aimed to validate the recording of AECOPD in UK EHR.MethodsWe randomly selected 1385 patients with COPD from the Clinical Practice Research Datalink. We selected dates of possible AECOPD based on 15 different algorithms between January 2004 and August 2013. Questionnaires were sent to GPs asking for confirmation of their patients’ AECOPD on the dates identified and for any additional relevant information. Responses were reviewed independently by two respiratory physicians. Positive predictive value (PPV) and sensitivity were calculated.ResultsThe response rate was 71.3%. AECOPD diagnostic codes, lower respiratory tract infection (LRTI) codes, and prescriptions of antibiotics and oral corticosteroids (OCS) together for 5–14 days had a high PPV (>75%) for identifying AECOPD. Symptom-based algorithms and prescription of antibiotics or OCS alone had lower PPVs (60–75%). A combined strategy of antibiotic and OCS prescriptions for 5–14 days, or LRTI or AECOPD code resulted in a PPV of 85.5% (95% CI, 82.7–88.3%) and a sensitivity of 62.9% (55.4–70.4%).ConclusionUsing a combination of diagnostic and therapy codes, the validity of AECOPD identified from EHR can be high. These strategies are useful for understanding health-care utilisation for AECOPD, informing service provision and for researchers. These results highlight the need for common coding strategies to be adopted in primary care to allow easy and accurate identification of events.

Highlights

Chronic obstructive pulmonary disease (COPD) is a common, progressive disease characterised by airflow obstruction which is not fully reversible
A combined strategy of antibiotic and oral corticosteroids (OCS) prescriptions for 5–14 days, or lower respiratory tract infection (LRTI) or Acute Exacerbations of COPD (AECOPD) code resulted in a Positive predictive value (PPV) of 85.5% and a sensitivity of 62.9% (55.4–70.4%)
Using a combination of diagnostic and therapy codes, the validity of AECOPD identified from electronic healthcare records (EHR) can be high

Summary

Introduction

Chronic obstructive pulmonary disease (COPD) is a common, progressive disease characterised by airflow obstruction which is not fully reversible. Acute exacerbations of COPD (AECOPD) are important drivers of mortality[2, 3] and reduced quality of life[4] in COPD patients and as the second most common reason for emergency hospital admission[5], they are of great public health importance. Several studies[6,7,8] of AECOPD have been conducted in UK electronic healthcare records (EHR) which are becoming an increasingly important resource for evidence from real life research. Data from primary care are used by organisations such as Public Health England (PHE) to compare data on AECOPD incidence and management across localities and by clinical commissioning groups to inform delivery of care and design of services. Acute Exacerbations of COPD (AECOPD) identified from electronic healthcare records (EHR) are important for research, public health and to inform healthcare utilisation and service provision. We aimed to validate the recording of AECOPD in UK EHR

Objectives

Methods

Conclusion