authors’ view of the critical factors for the implementation process (better risk assessment, new action plans, and better documentation), “IC status qualitative measure”.
The survey consisted of 45 items including questions on the size of the enterprise, sector (private or public), HES activity level prior to IC implementation (HES activities accomplished in last 3 years such as: assessed psychosocial work environment factors, risk assessment analyses, having worked out action plans to improve work environment, etc.), HES result level prior to IC introduction (absenteeism and lost time accident levels in 1990), HES competence (having a worker HES representative and an occupational health service), and time pressure (an ordinal variable with 4 points related to increased productivity and reduced staff). Although no information on the validity or reliability of these measures was provided, the interview had been tested in previous research with 500 enterprises that had started implementation of IC before 1992. 19
In the Saksvik and Nytro (1996) study, information on accidents and absenteeism was obtained from respondents during interviews and was, in most cases, from archival data. Ordinal variables with three levels (higher – stable – lower) were developed for both the rate of absenteeism and rate of loss time accidents (one day of ill health after the accident) based on reported levels in 1990 compared to 1991-1992 (up to first half of 1992). Multiple regression models for absenteeism and accident development from 1990 to 1992 were developed. They found that all three measures of IC implementation (status) performed well in the model for absenteeism and were all statistically significant; the strongest IC status measure was the qualitative variable (p<0.01, beta 0.09). In the accident model, none of the IC status measures were statistically significant. R-squared values for the two final models were 0.05 and 0.12, respectively.
The primary findings from the Saksvik and Nytro (1996) study were that the reported decrease in absenteeism was, as hypothesized, related to IC status but that there was no statistically significant relationship between IC status and change in accident rates. The authors pointed out that the finding for absenteeism may have been an artifact as the IC regulations may favour enterprises with a tradition for systematic long-range planning in other parts of their business management (thus, this may lead to confounding or effect modification). Additionally, the authors point out that their analyses were conducted at the enterprise level and that part of the variance in the dependent variables can be attributed to individual factors or interactional factors (subject-enterprise). Although the direction of change in accident rates was from high to low, the relationship with IC status was not significant. It is possible that the period of observation was too short to
19 Firms in the offshore oil industry and the onshore aluminum industry were required to establish IC program before 1990.
Institute for Work & Health