THREE AUTHORS REPLY

We thank Drs. Smith and Beaton (1) for their interests in our paper (2) and their views on the issues related to measuring changes in perceived job strain. Their Figure 1 shows different scenarios of changes in job-strain ratio. They questioned whether the small change in individual D who passed the threshold could be considered a meaningful change and suggested that changes within the variance of scores may not be real changes. It should be noted that the science of epidemiology is not to make inference about specific individuals but to generalize from research findings to populations (3). Our study showed that, in general, participants who reported changes from low to high perceived job strain were more likely to have developed major depression; those who changed from high to low perceived job strain had a similar risk of major depression as the reference group. At the individual level, these may not apply to some participants. The use of instruments is always associated with variability in clinical practices and scientific research. However, this does not prevent scientists from making frequent measurements and comparing changes in scores. With respect to whether a change within the variance of the scores should be considered as a real change, this is debatable, as we know that the magnitude of variance is largely related to sample size. The variance of scores in one study may not be applicable to another. Drs. Smith and Beaton agreed that our results indicated that “moving across a job-strain threshold of 1 is related to depression onset” (1, p. 132). They challenged that our results “cannot be extended to infer that change in job strain affects depression” (1, p. 132). As an additional analysis in our paper, we investigated the relation between the changes in the values of job strain ratios between 1994–1995 and 2000–2001 and the risk of major depression (2, p. 1087). The changes in the values of job strain ratios were analyzed as a continuous variable (the threshold of 1 was not involved). We found that positive changes in the values, that is, more job strain, were positively associated with the risk of major depression (odds ratio = 2.03, 95% confidence interval: 1.02, 4.05), which is consistent with our results. Drs. Smith and Beaton encouraged us to examine if the effects on depression were confounded by changes in occupation. We agree that such changes can be an important confounder. We are not sure whether individuals would frequently change their occupations, as this may be restrained by the education and training they received. However, it is very possible that people may leave one employer for another, or within one corporation, moving from one division to a different one. As we said in the paper, the National Population Health Survey as a general health survey did not collect detailed information about the work environment, and “it was not clear why job strain ratio changed over the course of the follow-up period and how other changes in workplaces affected the risk of major depressive episode” (2, p. 1090). It may be that changing occupations/employers is associated with greater job strain. Given that the point of the paper is to look at the effects of perceived job strain changes on depression, it really doesn't matter why there is increased job strain. What matters is that the patterns in the data indicate that there is an association with job strain and depression. Drs. Smith and Beaton said that we “used a relatively meaningful cutpoint, that is, where job demands exceeded job control, an improvement on the use of an arbitrary cutpoint based on the median score” (1, p. 132). This is not correct. As one can see from the Statistics Canada documentation (4), the job-strain ratio is calculated as the score of psychological demands/(score of decision authority + score of skill discretion). Score is reversed for certain questions. No arbitrary cutpoint based on median score was involved in the calculation. Conflict of interest: none declared.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.