WELL-BEING AND FUTURE SICK-LEAVE - MULTIVARIATE ANALYSES WITH REGARD TO PRECEDING SICK-LEAVE

Consistent associations between self-rated well-being and future sick- leave have previously been noted in a selected group of repeated short -term sick-leavers. The object was to retest those associations after ruling out expected influences of preceding sick-leave. Hypothetically the well-being-sick-leave linkages initially observed would thereby o verlap with the behavioural conformity over time. Accordingly, they wo uld possibly be erased in a hierarchical stepwise regression analysis. The study group comprised 61 females and 62 males with diffuse reason s for their high repeated sick-leave. Instead of having to present a d octor's certificate on every new sick-leave occasion, they chose to ta ke part in a support programme. Altogether 8 hierarchical multiple reg ression analyses were run with sick-leave occasions and days the first and second years after the contact as separate dependent variables. I n the first step, long-term behavioural proneness to sickness absence was evaluated, i.e. the correlations between the sick-leave parameters the year before and each of the 2 years after the programme were comp uted. in the second step, the possible additional impact of well-being ascertained at entrance into as well as exit from the clinical contac t was established. Sick-leave correlated strongly over the study perio ds as expected. Yet, fairly consistent associations between well-being and future sick-leave remained. The well being parameters accounted f or another 4-8% of the entire Variance in 5 of the 8 regression analys es performed. That increment corresponded to between 25 and 100% of th e proportion initially explained by preceding sick-leave. Thus the ind ependent role of perceived self-image for the long-term inclination to sickness absence in the current 'risk' group was underscored. This wa s particularly so since the influence of several other background and job-related factors have previously proven to be empirically negligibl e. Moreover, similar prospective correlations have been seen in other study groups.