From the abstract:
Coarse exposure assessment and assignment is a common issue facing epidemiological studies of shift work. Such measures ignore a number of exposure characteristics that may impact on health, increasing the likelihood of biased effect estimates and masked exposure–response relationships. To demonstrate the impacts of exposure assessment precision in shift work research, this study investigated relationships between work schedule and depression in a large survey of Canadian nurses.
The Canadian 2005 National Survey of the Work and Health of Nurses provided the analytic sample (n = 11450). Relationships between work schedule and depression were assessed using logistic regression models with high, moderate, and low-precision exposure groupings. The high-precision grouping described shift timing and rotation frequency, the moderate-precision grouping described shift timing, and the low-precision grouping described the presence/absence of shift work. Final model estimates were adjusted for the potential confounding effects of demographic and work variables, and bootstrap weights were used to generate sampling variances that accounted for the survey sample design.
The high-precision exposure grouping model showed the strongest relationships between work schedule and depression, with increased odds ratios [ORs] for rapidly rotating (OR = 1.51, 95% confidence interval [CI] = 0.91–2.51) and undefined rotating (OR = 1.67, 95% CI = 0.92–3.02) shift workers, and a decreased OR for depression in slow rotating (OR = 0.79, 95% CI = 0.57–1.08) shift workers. For the low- and moderate-precision exposure grouping models, weak relationships were observed for all work schedule categories (OR range 0.95 to 0.99).
Findings from this study support the need to consider and collect the data required for precise and conceptually driven exposure assessment and assignment in future studies of shift work and health. Further research into the effects of shift rotation frequency on depression is also recommended.