[Validity and reliability of a screening instrument for the identification of extensive work-related problems in...

Work-related interventions have become an integral part of medical rehabilitation programmes on behalf of the German Pension Fund. Studies showed concordantly that work-related medical rehabilitation measures should focus on patients with extensive work-related problems. Therefore the identification of such problems has to be an integral part of effective work-related interventions in medical rehabilitation. Adequate screening instruments do exist, but publications regarding their psychometric quality across indications are still missing.

Based on the SIMBO-MSK, which is a screening instrument for identification of work-related medical rehabilitation needs in patients with musculoskeletal disorders (MSD), a generic screening instrument for chronic diseases, SIMBO-C, was developed and tested for reliability and validity.

For testing psychometric attributes a German Pension Fund sample consisting of rehabilitation patients with MSD (MSD, n=182), psychosomatic disorders (PSY, n=173) and internal diseases (INN, n=71) were recruited. They filled-in a questionnaire at admission and at the three-months follow-up. The test-retest reliability was examined by means of Intra-class Correlation Coefficient (ICC) und Percentage of Agreement (PA) in a subsample (n=91) of patients who had been assessed additionally two weeks before starting their rehabilitation programme. Spearman correlation coefficients were calculated to observe the relationship between SIMBO-C on the one hand and Physical and Emotional Role Function (MOS-36), health status (EQ5-D) and subjective prognosis of work ability (SPE) on the other hand. Predictive validity was tested in three data sources: information about work ability in the medical report upon discharge, self-report on work ability from the Work Ability Index (WAI) as well as socio-medical situation at follow-up. In this context multiple linear and logistic regression models adjusting for age, gender, form of rehabilitation, indication and work status at admission were used.

SIMBO-C was composed of the weighted sum of seven single items (range: 0-100). Extensive work-related problems were presented at a cut-off of > 30. Prevalence of extensive work-related problems was between 29% (MSK) and 43% (INN). With ICCs of 0.83 (PSY) up to 0.91 (INN) as well as PAs of 78% (INN) up to 90% (MSK), a high test-retest reliability across indications was found. Construct validity were moderate. Particularly, high effect sizes between patients with and without extensive work-related problems were found across indications. The prediction of future events was predominantly high. Statistically significant associations for self-reported work ability and cost-relevant events (claim for occupational rehabilitation, claim for early retirement, gradual reintegration, absence from work for more than six weeks at a stretch, unemployment after rehabilitation) could be accounted for, which partly demonstrated a "dose-response relationship".

SIMBO-C can be characterised as a valid and reliable instrument for identification of patients with extensive work-related problems. This study in particular outlines the ability of the SIMBO-C to make predictions about problematic developments, which qualifies it as an instrument to be implemented at an early point of time in the rehabilitation process.