Computer categorization of spirometry data using statistical methods

A statistical method for evaluating spirometry has been developed and proposed as a more acceptable criteria for categorizing airway obstruction and chest restriction disorders. The lower limit of normal for a predicted value was determined by calculating 95% confidence interval (CI) for each spirometric measurement. By choosing multiples of 95% CI, categories of severity were established. The impact from implementing the statistical method was measured by performing a retrospective analysis of 1390 patients present to the LDS Hospital Pulmonary Laboratory. Comparisons of the new statistical and the older Intermountain Thoracic Society (ITS) criteria were performed to identify which patients would be affected. It was found that more old and short subjects tend to be categorized as normal (or with a less severe categorization) with the statistical method then with the ITS criteria. The opposite was found for young and tall patients. Therefore, the patients most affected were those located at extremes of age and height. Another purpose of this study was to comare the FEV1/FVC% with the FE25-75%, FEV1, FEV3, FEV3/FVC%, and FEF25-75/FVC for their ability to detect abnormal airway obstruction. In all cases the FEV1/FVC% detected more abnormal cases.