Introduction The primary goal in the rehabilitation of patients with chronic aspecific low back pain is return to work. Return to work was not reported in Tulder’s review (1997), who found strong evidence for the effectiveness of manipulation, back schools and exercise therapy for chronic low back pain especially for short-term effects. Van der Weide’s review (1997) investigated vocational outcome after treatment of patients with CLBP based on RCT until December 1996. No evidence was found for the effectiveness of behavioural treatment, exercise, manipulation or back school. Since then, several studies regarding RCT reporting vocational outcome were performed. It remains questionable whether return to work can be improved. Therefore a new review is performed.

Method Literature search: MEDLINE, Cochrane Clinical Trial Register, EMBASE, PSYCLIT, cross reference tracing. The search combined the key word ‘low back pain’ with one of the following terms: absenteeism, (re)employment, sick leave, return to work, sickness absence, occupational disability and employment status. Criteria for inclusion of studies were: randomised clinical trial, subacute or chronic aspecific low back pain (> 4 weeks), subjects at working age, outcome return to work or sick days, conservative intervention. Based on prior reviews passive treatment modalities were excluded. To assess methodological quality internal validity was rated on a 10 point scale (1 = concealment of treatment allocation, 2 = blinded assessment, 3 = blinding of patients, 4 = blinding of therapist, 5 = intention to treat analysis, 6 = timing of outcome assessment comparable, 7 = withdrawal/drop out rate, 8 = cointervention avoided or equal, 9 = compliance, 10 = outcome measurements relevant). Four levels of evidence were defined. Level 1, Strong evidence: multiple relevant high quality RCTs. Level 2, Moderate evidence: One relevant, high quality RCT and one or more relevant, low quality RCTs. Level 3, Limited evidence: one relevant, high quality RCT or multiple relevant, low quality RCTs. Level 4, No evidence: only one relevant, low quality RCT, no relevant RCTs or contradictory outcomes.

Results More then 900 publications were identified. 120 were further examined. So far 14 RCTs fulfilled the inclusion criteria. Results on methodological quality and treatment effectiveness will be presented at the EULAR Conference in Prague.