Development of an expert system for the interpretation of serial peak expiratory flow measurements in the diagnosis of occupational asthma

If asthma is due to work exposures there must be a relation between these e xposures and the asthma. Asthma causes airway hyperresponsiveness and obstr uction; the obstruction can be measured with portable meters, which usually measure peak expiratory flow, or sometimes forced expiratory volume in 1 s econd (FEV1). These can be measured serially (for instance 2 hourly) over s everal weeks at and away from work. Once occupational asthma develops, the asthma will be induced by many non-specific triggers common to non-occupati onal asthma. The challenge is to identify changes in peak expiratory flow d ue to work among other nonoccupational causes. Standard statistical tests h ave been found to be insensitive or non-specific, principally because of th e variable period for deterioration to occur after exposure, and the someti mes prolonged time for recovery to occur, such that days away from work may initially have lower measurements than days at work. A computer assisted d iagnostic aid (Oasys) has been developed to separate occupational from non- occupational causes of airflow obstruction. Oasys-2 is based on a discrimin ant analysis, and achieved a sensitivity of 75% and a specificity of at lea st 94%; therefore peak expiratory flow monitoring combined with Oasys-2 ana lysis is better to confirm than to exclude occupational asthma. A neural ne twork version in development has improved on this. Both have been based on expert interpretation of peak flow measurements plotted as daily maximum, m ean, and minimum, with the first reading at work taken as the first reading of the day. Oasys has been evaluated with independent criteria against mea surements made in a wide range of occupational situations. Oasys is suffici ently developed to be the initial method for the confirmation, although les s so for exclusion of occupational asthma.