Background Previous studies have questioned the validity and interpretation of spirometry undertaken in primary care. Knowing that data are accurate is important as many respiratory diseases are diagnosed and managed in primary care. Additionally researchers use data entered into electronic health records both as single measurements and to track changes in lung function over time. We aimed to determine whether spirometry undertaken in primary care for patients suspected to have COPD was of sufficient quality, and was correctly interpreted.
Methods As part of previous studies to validate the recording of COPD diagnosis and exacerbations of COPD in the clinical practice research datalink (CPRD) we obtained additional information from GPs which included spirometry traces. In this subset, a respiratory physician assessed spirometry traces for: 1) quality and 2) diagnostic interpretation. We used logistic regression to assess predictors of GPs interpretation of spirometric traces with the outcome of COPD diagnosis confirmed by respiratory physician adjudication of spirometry traces as correct and age, sex and previous record for asthma as covariates.
Results We obtained spirometry traces for 306 patients, of which 221 (72.2%) were conducted in primary care. 96.5% of traces were of adequate quality such that a valid interpretation could be made. Of those traces which were of adequate quality and conducted in primary care, and in whom a GP diagnosis of COPD had been made (N = 218), 73.4% showed obstruction, suggestive of COPD (Table 1). There was some evidence that correct interpretation of spirometry (either as obstructive, restrictive or normal) was influenced by a previous asthma diagnosis (OR 0.49, 95% CI 0.26–0.93). There was no evidence that correct interpretation was modified by age (OR 0.98, 0.96–1.01) or sex (OR 1.28, 0.69–2.38).
Conclusions Spirometry is performed in primary care to a high standard. Interpretation in patients with suspected COPD in primary care is moderate. Efforts should be made to improve spirometry interpretation for high quality patient care, and for research. As quality of spirometry measurements were high, researchers could use actual recorded values of FEV1 and FVC, however should exercise caution with using interpretation of spirometry values documented in primary care records.