Response to Burnout as a State: Random-Intercept Cross-Lagged Relationship Between Exhaustion and Disengagement in a 10-Day Study [Letter].

Dear editor

We thank Basinska and Gruszczynska1 for their interesting and sophisticated analyses of relationships between exhaustion and disengagement as assessed by the Oldenburg Burnout Inventory (OLBI)2 over 10 consecutive workdays. In a study not cited in their article, we recently examined the relationship between exhaustion and disengagement in an academic anesthesia department.3 Our study involved a one-time administration of the OLBI. We noted that the mean for exhaustion exceeded that for disengagement, ie, burnout symptoms were reflected more in exhaustion than disengagement. The means ± standard deviations were 2.52 ± 0.51 for exhaustion, 2.27± 0.52 for disengagement, and 0.25± 0.42 for their difference, t(129) = 6.68 by paired t-test, effect size (Cohen’s d) = 0.59. Moreover, higher exhaustion means were generally associated with higher disengagement means, r = 0.66 by Pearson correlation coefficient. For comparison with our and other studies (which commonly involve one-time administrations), it would be informative if Basinska and Gruszczynska would report a paired t-test and r for their first day of ratings, indicating the significances of the difference between exhaustion and disengagement and of the correlation between these two components of burnout. This information may be available or derivable from the results they report, but is not obvious (to us, at least). None of the studies in our literature review3– which was based on a PubMed search and, therefore, captured primarily, but not exclusively, health-care related samples of staff and students – tested the statistical significance of the difference between exhaustion and disengagement, as we did. This seemed surprising to us, considering the centrality of the distinction between exhaustion and disengagement to the conceptual model underlying the OLBI.2 Although it is clear that exhaustion and disengagement are intended to be compared in this model,2 as far as we know, the OLBI development process and subsequent studies have not established expected values of exhaustion and disengagement (or their difference), for any large, broadly representative normative population. Our study and literature review3 were intended, in part, to highlight the gap in knowledge concerning differences in exhaustion and disengagement. We think that filling this gap in our study and subsequent studies, such as that of Basinska and Gruszczynska,1 and trying to understand why there has been pronounced variability in past studies with respect to the direction and magnitude of differences between exhaustion and disengagement,3 might enhance the utility of the OLBI compared to other questionnaires for assessing burnout.