S09-1 What measure to use for monitoring and research on precarious employment? comparison of the multidimensional employment precariousness scale and the temporary employment indicator

Abstract

Objective Temporary employment is the most frequently used indicator of precarious employment for both monitoring and epidemiologic research. However, concerns have been raised regarding its capacity to correctly identify subjects, given temporary jobs are heterogeneous and not always precarious, and there is a spill-over of employment precariousness onto permanent jobs, traditionally considered the unexposed group. The aim of this study is to assess accuracy of temporary employment as indicator of precarious employment using sensitivity and specificity analysis. Methods Analysis were performed on data from a sample of salaried workers with a formal contract (n = 3691) from the 2009–2010 Chilean Employment conditions, work, health and quality of life survey (ENETS). Type of contract was the index test (temporary employment being the exposure category) and the Chilean version of the multidimensional Employment Precariousness Scale (EPRES-Ch), the reference standard. We examined alternative cut-off scores for the EPRES. Results Sensitivity was low, 29.1% (95% CI: 25.0%–33.4%) and specificity high, 92.3% (95% CI: 89.1%–94.6%). There was a 4% (95% CI: 2.9%–5.9%) of false positive results and 33% (95% CI: 29.6%–36.3%) of false negative results. Conclusions Low sensitivity, irrespective of good specificity, results in non-differential exposure misclassification, with a large proportion of workers in precarious jobs classified as non-precarious irrespective of the cut-off score applied to the Employment Precariousness Scale. This study does not imply that contract type is unimportant. Most temporary jobs are precarious and the intensity of precariousness is higher among them than among permanent jobs. However, used as a monitoring instrument, temporary employment will underestimate the magnitude of the problem, and epidemiologic studies affected by exposure misclassification will produce results that underestimate its association with workers’ health. For properly informing policy, it is critical for both research and monitoring that multidimensional measures of precarious employment be implemented as well.