Causal Analysis of Back Belts to Prevent Low Back Pain

• Activity Number: 279
• Type: Contributed
• Date/Time: Wednesday, August 14, 2002 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology*
• Author(s): James Wassell*+
• Affiliation(s): Centers for Disease Control and Prevention
• Keywords: Injury Epidemiology ; Nonrandomized Trial ; Propensity Scores ; Occupational Health ; Prospective Study ; Back Belts

Abstract

In a prospective study of the effect of back belts to prevent low back pain, 4,203 material handlers, who did not report a history of previous back pain in a baseline telephone interview, completed a six-month follow-up telephone interview. Belt wearing was self-determined, but influenced by store policy, for which there was poor compliance. Classification tree methods were used to define ten groups of workers with similar propensity to wear back belts. The effect of self-reported belt wearing on self-reported incident low back pain was evaluated within these ten strata. Stratum-specific risk differences, odds ratios, and logistic regression estimates of the effects of back belts were used to calculate weighted estimates and weighted estimates of variance. There was little evidence to suggest any lack of homogeneity of the effects of belt-wearing across strata. Given this propensity stratum adjustment for nonrandom treatment assignment, there was no difference in the incidence of low back pain between those who reported wearing back belts "usually everyday," and those who reported "never" to a question about belt-wearing habits.