Abstract
BACKGROUND: Pulmonary function testing (PFT) is an important tool in the diagnosis and management of most respiratory conditions, and appropriate interpretation of test results is a fundamental component of the final report. As part of developing a structured approach to interpretation of PFT results, we wished to characterize primary reasons for referral for testing in a range of PFT laboratories.
METHODS: Four PFT laboratories (3 public, 1 private) using similar PFT databases participated. Reasons for performance of PFTs were extracted from the databases and analyzed. Over 5,000 consecutive tests were evaluated from each lab.
RESULTS: Identifiable reason for referral was found in 83% of 24,602 test results and categorized. The major categories were follow-up of known respiratory disease (53% of 20,332 tests), investigation of specific symptoms (18%), possible specific lung disease (13%), possible induced lung disease (5%), investigation of lung function in known other diseases (5%), and other miscellaneous reasons (5%). Testing in known disease and/or assessing for PFT change was the primary reason for testing in 60% of tests performed. These data highlight the predominance of ongoing assessment of pulmonary function and the importance of access to previous test results to provide clinically useful test reports. They also emphasize the need for having valid criteria describing what constitutes a real clinical change in the various PFT parameters.
CONCLUSIONS: We have found that the majority of PFTs are performed to follow disease progress or response to treatment. This has implications with interpretation of test results and the clinical utility of PFT.
Introduction
Pulmonary function testing (PFT) provides a cornerstone for diagnosis and management of most respiratory conditions, and its clinical utility is recognized by the great rate of increase in its use in Australia over recent decades.1 Informed interpretation of PFT results is an important component of the report; however, PFT interpretation has been consistently demonstrated to be disturbingly variable.2–4
The approach to PFT interpretation recommended by the American Thoracic Society and European Respiratory Society5 provides a basis for standardization of interpretation, which, if widely adopted, may allow improvements in inter-subject interpretation concordance. Although one of the key components recommended in PFT interpretation involves answering the clinical question that prompted the tests to be performed,5 in our experience this aspect of interpretation is commonly overlooked. As part of a quality improvement initiative aimed at improving PFT interpretation, we considered it important to characterize the indications for performance of PFTs. Despite the clinical importance of PFTs, there is little information in the literature investigating indications for performance of PFTs in practice.
The aim of this study was to document the most common reasons for performance of PFTs in a range of adult PFT laboratories. This knowledge would provide a snapshot of clinical indications for PFT performance and could be used to aid the development of a systematic approach to PFT interpretation, and thereby assist in improving PFT laboratory practices.
QUICK LOOK
Current knowledge
Pulmonary function testing is a critical component of diagnosis and management of ambulatory respiratory disease and determination of disability in occupational lung disease. Referral of patients to pulmonary function laboratories is not well documented.
What this paper contributes to our knowledge
The majority of referrals for pulmonary function testing were to follow disease progress or response to current treatment. These findings highlight the importance of defining an “important change” in lung function and quality control of pulmonary function laboratories.
Methods
We interrogated the databases of 3 publically funded, university-affiliated PFT laboratories and one smaller private laboratory to extract the indication for performance of PFTs. These data had been manually transcribed from the referral letter/form into each laboratory's database system by the respiratory scientist at the time of testing. To ensure appropriate representation from each laboratory, a minimum of 5,000 consecutive tests in patients over 18 years of age were retrieved from each laboratory database. Apart from limiting the dataset to adults, no further selection criteria were applied. The dataset therefore does not represent individual subjects undergoing testing, but individual PFT sessions. Consecutive test records were retrieved in reverse chronological order from November 2010.
All 4 laboratories accepted referrals for testing from all sources: those external to their respective institutions (including from primary care), as well as from specialist physicians within their institutions. Each laboratory was capable of performing at least spirometry, carbon monoxide transfer factor, plethysmographic lung volumes and maximal respiratory pressures. Apart from date of test, patient age, and reason for referral, no further patient information was retrieved from the databases. The project was approved by our local human research ethics committee.
Analysis of reasons for referral was performed by categorizing into 6 broad groups and 22 subgroups. These groups were determined during the analysis phase of the study as it became clear what groupings would be appropriate to best represent the data. All groupings were made using spreadsheet sorting routines (Excel 2003, Microsoft, Redmond, Washington).
Results
A total of 24,602 PFT referrals were evaluated, with over 5,000 from each of the 4 laboratories. Table 1 shows the number of test reports that were assessed from each of the laboratories involved in the study. Overall, in 17% of cases no clear indication for testing could be determined, due to poor or absent clinical notes.
Referral Patterns for Each of the 4 Respiratory Laboratories
The reasons for testing in the remaining 20,332 referrals could be sorted into 6 major groups (Table 2), with follow-up in known respiratory disease being the most common reason for testing (53% of the 20,332). An analysis of this group of referrals revealed that 95% were identified as being for follow-up testing where tests had previously been performed in that testing laboratory. Testing for investigation of specific symptoms (18%) and searching for a specific lung disease (13%) were the next major groupings for reasons for referral. Table 2 also shows the major subcategories within each group.
Major Categories and Subcategories for Reasons for Referral for Pulmonary Function Testing
With regard to inter-laboratory comparisons, there were some differences in the relative proportions for referral; however, the relative rankings of the top 3 reasons for referral were identical in each laboratory (data not shown).
Discussion
These data reveal the primary indication for the majority of PFTs performed in these adult laboratories is to document lung function in a patient with previously diagnosed respiratory disease, and most of these referrals were in patients where previous PFTs had been performed within the testing laboratory. The implication from these data is that a large proportion of testing is being performed to provide objective assessment of change in lung function. While standardized testing techniques and effective laboratory quality assurance procedures are clearly important for this task, access to previous PFT data and reference benchmarks on clinically important change are essential to allow appropriate interpretation of results. This raises 2 important issues. First, the mechanism for providing immediate access to previous test results is critical. While the development of computer systems and interfacing of PFT systems into computer networks has progressed substantially over recent times, communication interfacing protocols (such as the HL7 health system interfacing standard) are not completely defined for all tests of respiratory function. Further refinements to these standards are needed, and PFT system manufacturers should be encouraged to provide open access to their database systems to further facilitate ease of access to previous test results. Access to all available test results, and not just the most recent, are critical to assist in monitoring for change in lung function since inter-session comparisons may not reveal real longer-term trends.
The second important issue raised by our finding that a large proportion of PFTs are performed to assess change in lung function is that clinically useful interpretation requires evidence about what defines a clinically important change for all measures of lung function. Test-retest reproducibility defines the change that can be detected with confidence, and the American Thoracic Society/European Respiratory Society guidelines focus on identifying when test results are outside the expected variation.5 However, there is evidence that smaller changes than these in FEV1 and FVC are associated with perceived changes in symptoms,6 and with mortality.6,7 Overall there are scant published data about what constitutes a clinically important change in respiratory parameters, particularly for non-spirometric measures that are commonly performed, such as transfer factor and static lung volumes. Until these are identified, it will remain difficult to provide appropriate and clinically relevant interpretation of a large fraction of PFTs.
These data have also drawn our attention to questioning the practice of routinely assessing complete lung function when it is not clear what a real change in a measure such as total lung capacity involves. The clinical utility of these more complex measures of lung function, over and above simple spirometry, remains to be identified. There are very few studies investigating and correlating clinical status and patient symptoms with contemporaneous, multiple markers of lung function to ascertain the value of more complex testing. An exception is the inspiratory capacity, a non-routinely assessed but simply determined marker of hyperinflation, where the magnitude of a clinically important change has been described in COPD.8
Searching for a particular reason for symptoms or for a specific respiratory disease was the reason for referral in approximately one third of cases. Interpretative strategies that focus on suggested likely diagnoses (for example, the algorithmic approach suggested in the American Thoracic Society/European Respiratory Society guidelines),5 are clearly relevant here. However, our data indicate that this approach may not be required for the majority of lung function tests performed.
The breakdown into subcategories of reason for referral provides unique and interesting data (see Table 2). The relative contributions of asthma and COPD in the various subcategories reflect the high prevalence of these diseases in contemporary adult Australians, and are similar to previous studies in Kuwait9 and Nigeria.10 The most common symptoms prompting referral were dyspnea and cough, as documented elsewhere.9 The high incidence of cough for investigation, at over one third of this subcategory, is not surprising, given that cough is the most common symptom for presentation for general practitioner consultation.11 The relatively high referral rate for patients with sleep and/or obesity-related disorders reflects the increasing incidence of these disorders in the general community,12 and might be expected to continue to increase in the future.
No identifiable reason for performance of tests could be determined in approximately 1 in 6 cases (17%). This makes informed interpretation problematic and is likely to compromise the clinical utility of the interpretative comment, if not the complete test report. While further analysis of this subset of the data could not be done, it could be assumed that a large portion of these test requests were performed for follow-up purposes. If this is the case, then significantly more than 60% of all tests in our analysis were performed in patients with previously diagnosed respiratory disease.
Ongoing monitoring of lung function is clearly an important part of clinical management and is recommended for management of most common lung diseases.13–15 However, there is recent evidence that patient outcomes are not improved with regular spirometry follow-up in the primary-care setting in asthma16,17 and in COPD.17 While it is beyond the scope of our study to determine the value and clinical utility of regular PFT follow-up in known disease, or whether unnecessary ordering of PFTs is occurring, our findings suggest that these issues may be important in evaluating the way in which PFTs are utilized.
Conclusions
In summary, these data provide a snapshot of why PFTs are ordered in a range of adult lung function laboratories. The finding that most testing is performed in known disease to monitor progress highlights the fundamental need for access to previous lung function when interpreting laboratory results, and also raises our awareness of the paucity of evidence describing what constitutes an important change in lung function. This is particularly so with regard to tests of lung function other than spirometry. Until such data are available, the interpretation of PFTs, and therefore its clinical utility, will be compromised.
Footnotes
- Correspondence: Jeff Pretto Doc Hlth Sc, Respiratory and Sleep Medicine, John Hunter Hospital, Lookout Road, New Lambton Heights, New South Wales 2350 Australia. E-mail: jeff.pretto{at}hnehealth.nsw.gov.au.
The authors have disclosed no conflicts of interest.
Dr Pretto presented a version of this paper at the European Respiratory Society Annual Congress, held September 24–28, 2011, in Amsterdam, The Netherlands, and at the Thoracic Society of Australia and New Zealand Annual Scientific Meeting, held April 20–23, 2012, in Canberra, Australia.
- Copyright © 2013 by Daedalus Enterprises
