Long-term Maintenance of a Successful Occupational Sun Safety Intervention

Abstract

Skin cancer prevention specialists should attempt to reduce ultraviolet radiation exposure (UVR) among outdoor workers, as pointed out in a recent review.1 To address this need, Project SUNWISE, a randomized controlled sun safety intervention trial with Southern Californian United States Postal Service letter carriers, was conducted from 2001 through 2004. The intervention consisted of providing free sunscreen, free wide-brim hats, a series of 6 brief onsite educational sessions, and sun safety prompts. The primary outcome measures were validated questionnaire items that asked participants how often they had used specific sun protection strategies over the past 5 workdays while delivering mail.2, 3 They were administered at baseline and 3, 12, 24, and 36 months post-baseline. The 5 response options ranged from “never” to “always”; we considered “always” as “consistent use” in analyses. At the 2-year follow-up evaluation, participants at the intervention postal stations had significantly higher rates of consistent sunscreen use and wide-brim hat use than those at control stations.2 Details about the study procedures and sample characteristics have been published previously.2 Immediately following the 2-year evaluation, control station participants received the free items and 3 of the education sessions (i.e., introduction and protections strategies, sun safety for eyes, and recap/encouragement to maintain sun safety practices). At the intervention stations, we continued to provide free sunscreen during that year. This paper describes the behavioral outcomes at the 3-year follow-up evaluation. Evaluation cohort retention rates (of those completing questionnaires) from the 2- to 3- year follow-ups were 93.2% (927/994) for the intervention group and 94.4% (1130/1196) for the control group. The trends over 3 years in two key outcomes—consistent use of sunscreen and wide-brim hats--were analyzed using generalized linear mixed models treating 3 months, 1 year, 2 years, and 3 years as a set of repeated measures on each postal worker. In addition, we adjusted for postal workers nested within post offices within a multi-level model; the intraclass correlation coefficients for sunscreen and hats were .015 and .067, respectively. All analyses were adjusted for the baseline level of the corresponding outcome variable. For each analysis, we tested 1) the time by condition interaction, to determine if the intervention effect remained constant over time, and 2) the condition main effect. Consistent sunscreen use levels for intervention participants at the 2- and 3- year follow-ups were 39.2% and 38.3%, respectively, and for control participants, they were 26.3% and 34.3%, respectively. Wide-brim hat levels for intervention participants during these periods were 40.0% and 43.8%, respectively, and for control participants, they were 22.3% and 33.0%, respectively. Results of the analyses for each of these outcomes showed significant condition by time interaction effects (ps<0.0001), indicating that the differences between the conditions were changing over time. The table shows these patterns. For both behaviors, the change in the odds ratios from year 2 to year 3 is not due to a drop among intervention participants but rather to an increase among control group participants. We are encouraged that intervention effects were maintained at least one year after the program ended, and that once the control participants received the intervention, their consistent use of sun protection increased substantially. Table Results of generalized linear mixed-model analyses a: Patterns of odds ratios over three years of follow-up