NTERNATIONAL R EVIEW OF I NDUSTRIAL AND O RGANIZATIONAL P SYCHOLOGY 2005 coronary heart disease, such as hypertension and dyslipidemia, are also

284 I NTERNATIONAL R EVIEW OF I NDUSTRIAL AND O RGANIZATIONAL P SYCHOLOGY 2005 coronary heart disease, such as hypertension and dyslipidemia, are also

known to be associatedwith risk of Type 2 DM (Beckman et al., 2002; Jacobsen, Bonaa, & Njolstad, 2002).

Is burnout causally implicated in the onset of diabetes? Indirect evidence was provided by a study that found a moderate correlation between feelings of excessive tiredness and the insulin resistance syndrome, a risk factor for diabetes, in 90 healthy males (Raikkonen, Keltikangas-Jarvinen, Adlercreutz, & Hautanen, 1996). Another relatedstudy reportedthat, among 292 insulin- treated diabetic employees, both work-related factors and diabetes-related factors were associatedwith excessive fatigue (Weijman et al., 2003). Direct evidence from a recent study on burnout and Type 2 DM risk (Melamed, Shirom, & Froom, 2003) provides initial support for such a possibility. This study was conducted among primarily white-collar Israeli workers. After excluding those who had a history of DM or other chronic diseases, 633 workers were followed up for a period of 3–5 years. During this periodthere were 17 new cases of treatedType 2 DM. Burnout, as measured by the SMBM, was foundto be associatedwith increasedrisk of Type 2 DM (OR ¼ 1.83, 95% CI 1.20–2.77), even after controlling for age, sex, body mass index, smoking, period of follow-up and job category. Thus, this finding suggests that burnout might be a risk factor for Type 2 DM in Israeli workers.

Common infections One study found small but significant associations between the MBI and self-

reportedepisodes of coldor flu, but failedto finda significant correlation with cholesterol ratio (Hendrix, Steel, Leap, & Summers, 1991). In a study conducted during the Gulf War it was found that pre-war burnout (measured by the SMBM) was associatedwith wartime threat appraisal (worry) and upper respiratory infections among 162 Israeli civilians who carriedon with their employment duties throughout the Gulf War (Kushnir & Melamed, 1992). In a prospective 3-year cohort study of 12,140 employees in the Netherlands (Mohren et al., 2003), a version of the MBI was used to predict increased incidence of common infections. In the longitudinal analysis, a significant effect was foundfor gastroenteritis, andonly for the exhaustion subscale of the MBI. Mohren andher colleagues obtainedthese results after controlling for age, gender, educational level, and the presence of long-standing illness.

Male and female infertility There is yet an additional domain of human health that may be influenced by

burnout, that of male andfemale infertility. Regard ing male infertility, research findings supported the hypothesis that stress has a negative

285 impact on semen quality (e.g., Clarke, Klock, Geoghegan, & Travassos,

B URNOUT AND H EALTH R EVIEW

1999; Giblin, Poland, Moghissi, Ager, & Olson, 1988; Harrison, Callan, & Hennessey, 1987). Based on this set of findings, a case control study was recently initiatedto explore the possibility that burnout wouldalso have negative implications for male fertility. The results confirmedthis suspicion. Males with infertility problems (assessedusing the combinedcriteria of the quantity andquality of motility, an index of sperm concentration) were found to have significantly higher burnout scores (on the SMBM) comparedwith controls (Sheiner, Sheiner, Carel, Potashnik, & Shoham-Vardi, 2002). This finding, if replicated, would point to yet another research frontier concerning the relationship of burnout with health impairments.

In yet another study that used the SMBM (Sheiner, Sheiner, Potashnik, Carel, & Shoham-Vardi, 2003) it was found that stress and burnout levels were not significantly different between a group of women attending a fertility clinic due to female infertility (n ¼ 64) anda control group of women attending the clinic due to their partners’ reproductive impairment. This finding may indicate that there are gender-related differences in the way stress andburnout are implicatedin male andfemale reprod uctive problems.

Self-rated health Self-ratedhealth is a simple andvalidproxy measure for health status

(McGee, Liao, Cao, & Cooper, 1999) since it has been foundto consistently predict mortality even after adjustment for physical ill health at the baseline (Idler & Kasl, 1991). It is most commonly measured with a single question asking respondents to assess their general state of health with a few response options like ‘very poor’, ‘poor’, ‘fair’, ‘good’, ‘very good’ with a dose response association commonly demonstrated in predicting all-cause mortality (cf. Kaplan et al., 1996). The VE measure has been linkedto self-reportedill health or disease states. Using a measure of self-reported general health in a two-wave study of healthy males in Sweden, Halford and her colleagues (Halford, Anderzen, & Arnetz, 2003) found it to negatively correlate with VE. Self-ratedill health has also been foundto be closely correlatedwith burnout in other studies (Gorter et al., 2000; Kahill, 1988; Soderfeldt, Soderfeldt, Ohlson, Theorell, & Jones, 2000). In a longitudinal study of staff burnout in a psychiatric hospital, self-reportedfrequency of serious illness shared10% of the variance with the emotional exhaustion scale of the MBI, after controlling for social support andother confounding variables (Corrigan et al., 1994). Similar findings were also reported in respect of the variables in another study by Bhagat, Allie, and Ford (1995). This rather important association between burnout andpoor self-ratedhealth suggests that burnout may reflect impairedphysical health.