ATS Pulmonary Function Laboratory Manual

chapter  20

6.4. Use caution when extrapolating equations for patients of ages or heights not covered in data gener- ated by the reference set. 6.5. Consider the ethnic origin of the subject being tested. 7. Upper and lower limits of reference ranges (1) 7.1. Reference ranges may be based on either upper or lower 5th percentiles or 95% confidence intervals. These can also be expressed as standard deviation (Z) scores (i.e., the 5th percentile corresponds to a Z = −1.645). Percentiles provide appropriate limits for reference ranges regardless of the distribution of the values. They do require a relatively large sample size (at least 120 subjects per grouping) to provide accurate estimates of the reference range limits. Confidence intervals may be accurate with smaller sample sizes but should only be used when the distribution of the measured values is Gaussian. 7.2. The limits of reference ranges are inherently variable; so interpret results close to limits with caution. 7.3. A fixed percent of predicted for the lower limit of normal (e.g., 80%) may be misleading and varies according to the parameter being considered (i.e., LLN for FEV 1 may be approximately 80%, but LLN for FEF 25-75% may be much lower) and subject characteristics including age and gender. 8. Cite the source (i.e., author and year) of reference values for each test on the final report (1). If a mathemati- cal adjustment for ethnic group is used, specify the percentage. Define the lower and/or upper limits of the reference range for each lung function parameter. 1.2. Begin the interpretation with a statement about test quality (1, 11). 1.2.1. If test quality is poor, the interpretation should deal with the effect of test quality problems on the results. 1.3. Consider the clinical question asked. 1.4. Use a conservative approach when suggesting a diagnosis based on PF tests alone. Diagnostic state- ments usually require clinical information in addition to lung function data. 1.5. When possible, the interpretation should reflect an understanding of the statistical distribution of measured results in disease states and the prior probability of the disease. 1.6. False-positive results increase as the number of indices used in the interpretation increases. 1.7. The primary indices to be used for spirometric interpretation are VC (FVC) or slow VC, FEV 1 , and FEV 1 /VC. 1.8. An individual familiar with pediatric lung function should interpret PF tests performed on children. 1.9. The individual(s) performing the interpretation should have documented training and competency verification. If more than one individual performs interpretation of test results, a system should be employed to ensure consistency of interpretation. 2. Consider the effect of race and ethnic group on test results (1, 8). The use of race and ethnicity in interpret- ing lung function tests presents significant challenges (12) that should be understood by those interpreting tests. 2.1. For African Americans, the actual values for total lung capacity (TLC), FEV 1 , and FVC may be about 12% lower than Caucasians with the same standing height (1). 2.2. For African Americans, the actual values for functional residual capacity (FRC) and residual volume (RV) may be 7% lower (1). Interpretation of Laboratory Data 1. General principles and guidelines (1) 1.1.

235