Table 1

Quality criteria applied to each article

  1. Purpose of study clearly stated (aetiology versus prognosis)

  2. Main features of study population described

  3. Response rate 80% at baseline and reported at follow up

  4. Psychosocial data collected using validated instrument

  5. Data on physical workload collected

  6. Clear, reproducible definition of low back pain

  7. Low back pain data collected for at least one year

  8. Multivariate analysis including confounder control

  9. Outcome measures (OR, RR) reported with 95% confidence intervals