Because the New York Heart Association (NYHA) classification system categorizes patients based on subjective impression of the degree of functional compromise, a reduction in exercise might make a patient seem improved because the new lower level of ordinary activity produced fewer symptoms. To test this hypothesis, we studied three different sets of patients and compared their NYHA classes to their functional classes as determined by a new Specific Activity Scale (SAS) that is based on the metabolic equivalents of oxygen consumption required for activities the patient actually performs. Among ambulatory patients referred for exercise tests, the NYHA class was higher (i.e. indicated the patient was more limited) in 28% of patients and the SAS class was higher in 14% (p less than 0.001). Among patients interviewed at or near the time of catheterization for chest pain, the NYHA was higher in 20% and the SAS class was higher in 20% (p = NS). In both medically and surgically treated patients interviewed 1--3 yr after cardiac catheterization, the NYHA class was higher in only 4%, whereas the SAS class was higher in 28% (p less than 0.001). The SAS class was significantly more likely to be higher in patients who were not working full time and in patients who described their present activity level as sedentary or light. When the NYHA and SAS systems disagreed as to whether a patient was improved, SAS was significantly more likely to correlate with the patient's self-assessment. These findings suggest that some patients restrict their activity as their cardiac disease progresses; the resultant change in the definition of ordinary activity may reduce the apparent degree of cardiac compromise and thus give a false impression of improvement by NYHA criteria.