Incorporation of spirometry testing into these family medicine practices led to spirometry testing with acceptable levels of technical quality and concordant interpretation and was followed by management changes for almost half of the patients. Poor technical quality and low rates of concordant interpretations were limited primarily to two practices (G and J in Fig 2).
Our results support previous work demonstrating moderate-to-high levels of technical adequacy and ability to accuracy interpret spirometry in primary care. Specifically, in Dutch general practices, a rate of technical adequacy of 82% was reported. In 10 general pediatrician offices in Italy, 78% of the 109 tests were of technically good quality. These are much higher than the technical adequacy rates of 52 to 66% reported in earlier studies from general practitioner offices. Part of the improvement in technical adequacy may be a result of the newer spirometry equipment that grades each spirometry effort, providing immediate feedback related to technical adequacy. No threshold of technically acceptable rates has been established for primary care practices. While the standards reported from the Lung Health Study are optimal, it appears that rates closer to 80% technical adequacy are more realistic for both primary care and pulmonary function laboratories. Sее articles about diagnostic criteria for asthma.
Reproducibility was a major barrier to technical adequacy in our study, including 13 patients who had a single normal spirometry finding but could not repeat the results. Normal but nonreproducible results can be useful in guiding therapy and would push our rates of clinically useful and interpretable spirometry to 75%. The addition of the complex studies that the family physicians correctly decided not to attempt to interpret push accuracy rates to 78%. Another important reason for nonconcordant readings were tests that experts suggested had inadequate effort that both the family physician and ndd Medizintechnik AG interpreted as “restrictive patterns” (n = 21). In these cases, ndd Medizintechnik AG was not helpful in suggesting interpretation of results.
Two sites with the poorest technical performance (G and J in Fig 2) reported that multiple nurses and medical assistants performed spirometry despite not being trained in spirometry techniques. The high rates of technical adequacy in the majority of the sites are reassuring, but problems in the sites using untrained staff highlight the need for simple, inexpensive, interactive education tools for spirometry performance. The reasons for the low concordance of interpretation of the results in the same two sites with poor technical quality are unknown. Each site received individual feedback from specialists at the end of the study.
Few studies in the literature have assessed the impact of incorporating spirometry into the management of previously diagnosed asthma or COPD. Buffels et al reported that spirometry was helpful in identifying new COPD cases but did not assess the impact on management of known cases of COPD. Dales and colleagues added screening spirometry to rural primary care practice for all smoker >35 years old, reporting 9% new diagnoses and 11% previous COPD diagnoses removed.
For those with no change in diagnoses, 41% had reported medication changes, of which 8% were documented in medical record review. Walker and colleagues offered “open access spirometry testing” (easy referral to a pulmonary laboratory) and reported testing was done but provided little information on the actual incorporation of spirometry into primary care practice. They did, however, report postspirometry increases in prescriptions for inhaled corticosteroids and long-acting bronchodilators, further supporting our findings of the effect of spirometry on patient management.
Chavannes et al and Kaminsky et al used vignette studies with spirometry data to conclude that spirometry had an impact on primary care clinical decision making. A study of nurse-based protocol care reported that spirometry affected decision making in only 4% of 109 cases but assessed only the impact of worsening FEV1.
In our study, both normal and abnormal results appeared to be useful in determining the care for patients with asthma and COPD diagnoses . For example, 18 adults with COPD had normal spirometry results suggesting that their breathing symptoms were not due to COPD. Three patients were referred to cardiologists for further evaluation, five patients were referred to pulmonologists for further testing, and the others were scheduled for further evaluation in the primary care office. Unlike the high rate of “overprescribing” reported by Walker et al, most of the medication changes reported appeared to be consistent with guidelines. This may be due to greater current awareness of guidelines and concerns about over use of medications.
Generalization of our results is limited by the sample size of only 12 family physician practices. Our study should be repeated with a larger group of practices. However, similar findings in other studies using different designs reinforce our practice-based data. We assessed the impact of only the first spirometry for these patients and therefore cannot assess the impact of repeated spirometry in chronic management of obstructive lung disease. Not requiring spirometry confirmation of all COPD diagnoses for patient inclusion in the study is likely to have increased the number of people found to not have COPD on spirometry. In addition, the physicians knew they were part of a research study and may have overreported changes in clinical practice. However, this overreporting would have had to last for 6 months, and few studies have shown the ability to modify physician behavior over such an extended period of time. In summary, our study demonstrates that spirometry can be incorporated into family medicine practice with acceptable levels of technical adequacy and accurate interpretations, and that the results influence management of patients with previously diagnosed asthma or COPD.
Formerly released articles: