Spirometer calibration checks: is 3.5% good enough?

BACKGROUND: Current standards for spirometry require daily calibration checks to come within 3.5% of the inserted volume but do not require evaluation of trends over time. We examined the current guidelines and candidate quality control rules to determine the best method for identifying spirometers with suboptimal performance.
METHODS: Daily calibration checks on seven volume spirometers recorded over 4 to 11 years were reviewed. Current guidelines and candidate quality control rules were applied to determine how well each detected suboptimal spirometer performance.
RESULTS: Overall, 98% of 7,497 calibration checks were within 3.5%. However, based on visual inspection of calibration check data plots, spirometers 3 and 5 demonstrated systematic sources of error, drift, and bias. The +/- 3.5% criteria did not identify these spirometers. The application of +/- 2% criteria identified these spirometers (9% out-of control values in spirometers 3 and 5 vs < 5% in other spirometers). A rule stipulating out-of-control conditions when four consecutive checks exceeded 1% deviation identified suboptimal spirometers (14% and 20% out-of-control values) but maintained low error detection rates in other spirometers (< or = 2%). Other candidate rules were less effective or required longer times to error detection.
CONCLUSIONS: The current recommendation that calibration checks come within +/- 3.5% of the inserted volume did not detect subtle errors. Alternative candidate rules were more effective in detecting errors and maintained low overall error-detection rates. Our findings emphasize the need for laboratories to systematically review calibration checks over time and suggest that more stringent guidelines for calibration checks may be warranted for volume spirometers. Although our general approach may also be appropriate for flow-type spirometers, the details are likely to differ since flow-type spirometers are a much more varied category of equipment.