Comparison of four measurements of sickness absence in workers in the Veneto Region: interpretation and applicability. Data from Italian National Social Security Institute, 1997-2002

Abstract BACKGROUND: In 2005, 52% of workers in the Veneto Region reported one or more sick leave spells in the previous 12 months, compared with 16% reported in the European Survey on Working Condition in 2000 (ESWC), although health conditions were better in the Veneto Region than in the ESWC. OBJECTIVES: Since the above findings were based on answers to a questionnaire, we used an objective source of information in order to further investigate the problem. METHODS: The source of data consisted of the sick leave records of workers in the Veneto Region, 1997-2002, which were obtained from the Italian National Social Security Institute (INPS). Four measurements were used: (1) frequency = number of sick-leave spells during the study period/number of insured persons at risk for sickness absence during the study period; (2) cumulative incidence = number of persons with at least one new sick-leave spell during the study period/number of insured persons at risk for sickness absence during the study period; (3) sick-leave rate = number of sick-leave days during the study period/number of insured persons at risk for sickness absence during the study period; (4) duration of absence = number of sick-leave days in new sick leave spells during the study period/number of new sick-leave spells during the study period. Cumulative incidence could only be analyzed by statistical methods (binomial regression). RESULTS: The first three measurements gave overlapping results. The values were highest in industry, lowest in agriculture, and intermediate in crafts and services; they were highest in middle-aged subjects and lowest in younger and older individuals; the time trend increased up to 1999 and then decreased slowly. Cumulative incidence was 52.0%, 38.0%, 36.4% and 22.9%, in industry, services, crafts, and agriculture, respectively, and 42.8% in the total population. The results of binomial regression confirmed the above pattern, as well as an excess in women, in white collar workers, and in various occupational categories which, however, were probably not due to occupational exposures since in each occupational category the risk was always higher in industry than in crafts. By contrast, the sick-leave rate was higher in crafts and agriculture and lower in services and industry; it increased exponentially with age and did not show real differences between males and females. CONCLUSIONS: The choice of measurements influences the results; duration of absence could be used to detect areas of suspected work-related diseases, while cumulative incidence might be more helpful to detect areas of suspected absenteeism/presenteeism.