Use of benzodiazepine drugs and perceived job stress in a cohort of working men and women in Belgium. Results from the BELSTRESS-study

Abstract

The aim of the Belstress Study was to see whether use of benzodiazepines is associated with perceived job stress as measured by Karasek's job-strain model. This model has as its central tenet that the most adverse health outcomes are to be expected in high strain jobs characterized by high job demands and low job control. An extension of the model states that the most noxious combination is high job demands, low control and low social support at work. Sample subjects were recruited from 25 Belgian companies between 1994 and 1998, and cover a wide range of occupations. A 5.6% of 16,094 men and 9.3% of 5012 women aged 35–59 years report use of benzodiazepines during the last month. A clear association is displayed between self-reported use of benzodiazepines and a high strain job compared to a low strain job (men: OR=1.93, 99% CI=1.4–2.6; women: OR=1.66, 99% CI=1.0–2.7), after adjustment is made for socio-demographic confounders (age, level of education, occupational group, employment sector, living situation). The independent association with quartile level of job demands is a striking feature (men: OR of highest quartile compared to lowest quartile group=1.91, 99% CI=1.4–2.6; women: OR=1.99, 99% CI=1.3–3.1). In men, an inverse association with quartile level of job control is observed (OR= 0.65, 99% CI=0.5–0.9) whereas in women a clear tendency in that direction is displayed (OR=0.62, 99% CI=0.4–1.1). The association with low social support is less clear; an independent association between use of benzodiazepines and iso-strain was observed particularly in men.

Introduction

Over the last decade, a number of studies have focused on the association between psychotropic drug use and work related stress (Appelberg, Romanov, Honkasalo, & Koskenvuo, 1993; Belin & Montastruc, 1990; Estryn-Behar et al., 1990; Jacquinet-Salord, Lang, Fouriaud, Nicaulet, & Bingham, 1993; Kristensen, 1991; Moisan et al., 1999; Pelfrene et al., 2002; Schifferdecker, Schmidt, Loevenich, & Krahl, 1996; Watts & Short, 1990). Unfortunately, results did not always prove to be consistent, a fact that has been attributed to differences in the applied methodology in general and to the way job stressors were measured in particular (Moisan et al., 1999).

Another reason for the differences in results might be that the outcome of psychotropic drugs is too vague a category, covering both drugs like benzodiazepines that are often prescribed for anxiety symptoms, nervousness and sleep problems (Olfson & Pincus, 1994; Simon, Von Korff, Barlow, Pabiniak, & Wagner, 1996; Simon & Von Korff, 1997) which indeed may directly connect with stressful living conditions, and other drugs like antidepressants or neuroleptics which are usually recommended in relation to major psychiatric disorders (Bogaert & Reuse, 1997). A more focused definition of the outcome would add to a better understanding of the subject.

Our focusing on benzodiazepines is also motivated by the fact that these drugs are reportedly often used chronically without adequate professional surveillance (Barbui, Gregis, & Zappa, 1998). As such, the extensive use of benzodiazepines has become a major concern for healthcare providers and policy makers (Foy, 1993; Tu, Mamdani, Hux, & Tu, 2001). This is particularly the case for Belgium, where elevated levels of chronic usage have repeatedly been observed (Balter, Manheimer, Mellinger, & Uhlenhuth, 1984; Belgisch Centrum voor Farmacotherapeutische Informatie, 2002; Habraken, Blondeel, & Soenen, 1995). Habraken et al. (1995) for example, reported that almost 19% of their sample in the province of East-Flanders used benzodiazepines in the year before the interview, with 9% of the sample classified as chronic users. Generally, the concern is about adverse health effects of overuse and misuse, of which addiction, psychological and physical dependency, effect-tolerance, rebound-insomnia, memory disorder and impairment of other cognitive functions are noted (Ashton (1994); Bogaert & Reuse, 1997; Gudex, 1991; Hay, 2000; Kan, Breteler, & Zitman, 1997).

Few studies have reported a link between the use of benzodiazepines and stressful living conditions and results are difficult to compare. For example, in one study among a Viennese population (Austria) aged 15 years and over, the respondents’ “subjective feeling of being under stress” significantly influenced benzodiazepine use (Pakesch et al., 1993). Likewise, another study reported that following surgery residency training, Nigerian doctors considered their work to be stressful, which went along with some abuse of tobacco, alcohol, benzodiazepines and barbiturates (Adebamowo, Ezeome, Ajuwon, & Adekunle, 1998).

To compare results however, some unequivocal definition of work-related stress is mandatory. Beyond doubt, a leading theoretical framework in research on the relationship between work-related stress and health outcomes is the Job Demand-Control (DC) model, proposed by Robert Karasek in the late 1970s (Bonn & Bonn, 2000; Karasek, 1979; Karasek & Theorell, 1990). Basically, the model's strain hypothesis postulates that job strain and concomitant adverse health effects do not result from a single aspect of the work environment but from the joint effects of both high job demands and low job control. Job demands refer to the perception of the workload, the pressure of time and the complexity of tasks. Job control is related to the potential control workers have over their tasks and behaviour during a working day. It is also called decision latitude and is understood to be determined either by the worker's authority to make decisions on the job or by the breadth of skills used by the worker.

On the basis of empirical research, especially in the field of cardiovascular diseases, the DC-model has been extended with the dimension workplace social support (Johnson & Hall, 1988; Johnson, Hall, & Theorell, 1989; Johnson & Hall, 1994). Social support at work refers to “overall levels of social interaction available on the job from both co-workers and supervisors” (Karasek & Theorell, 1990, p. 69). In this extended, so-called Job Demand-Control-Support (DCS) model, it is hypothesized that the most unfavourable health outcomes are to be expected in workers perceiving high job demands, low job control and low worksite social support. This is referred to as the iso-strain hypothesis, emphasizing “isolated, high strain” conditions.

In this paper, our aim is to study the association between job strain or iso-strain and the use of benzodiazepine drugs in a large cohort of workers covering a broad range of occupations. More specifically, we evaluate which of the job stress dimensions of the extended DCS-model is most strongly related to the usage of this particular and well-defined group of psychotropic drugs. Also, we will test whether the associations are stronger in women than in men because a much higher prevalence of use of benzodiazepine drugs has consistently been observed in women. (Broom, 1995; Cafferata, Kasper, & Bernstein, 1983; Groenewegen, Leufkens, Spreeuwenberg, & Worm, 1999; Pakesch et al., 1993; Rosser, 1989).