Does attrition bias longitudinal population-based studies on back pain?

The representativeness of the sample is consecutively reduced because of differential attrition over the different measurement points. Despite this, bias due to attrition has a marginal impact on the point estimates of virtually all back pain related outcomes.