Follow-up studies 6 to 24 months after an inpatient rehabilitation treatment were conducted using 5 samples: 68 male and 89 female patients matched according to age and diagnosis, 63 male myocardial infarction patients and 81 male patients with psychophysiological-psychosomatic disorders matched according to age, and 75 male myocardial infarction patients. Rehabilitation outcome at follow-up was defined by patient's outcome rating, frequency of physical complaints, number of sickness reports during the time of follow-up, and use of tranquillizers or analgesics. To predict outcome, variables recorded at the beginning of the treatment were used in multiple stepwise regression analyses: medical data (ergometric performance, physiological measures), psychological data (emotional lability, frequency of physical complaints, mood) socioeconomic data, and living habits. Substantial differences in the prediction of outcome variables were observed ranging from 29% of the explained variance (number of sickness reports) to 59% (frequency of physical complaints). The best single predictors were the same variables measured at an earlier stage, but in most of the analyses up to six predictors significantly contributed to the multiple correlation coefficients. Many of the predictors were specific for the different patient samples but several were replicated. Important predictors were mood, frequency of physical complaints, disturbed sleep, extra jobs, physical activity, age, former inpatient treatments, and performance during ergometric exercise. No systematic differences emerged between the samples in the prediction of a given outcome variable. The results are discussed within the framework of the illness behaviour concept.
KEY WORDS:Rehabilitation, follow-up study, coronary heart disease, psychophysiological disorders, illness behaviour, multiple stepwise regression.