Cross-sectional studies | ||||||
Author, year | Sampling frame | Participation | Statistical analysis | Adjusted/controlled for | Exposure | Outcome (p<0.05%) |
Bourbonnais, 200529 | Nurses (n=2002) | 77.1% (76% and 84%) | Binominal regression | Coping strategies, social support, domestic load, type A behaviour, personal and socio-occupational characteristics (age, familial status, seniority in the institution, job status, work shift) | Restructuring | Exposed PR 1.84 (1.56–2.16) of psychological distress vs controls |
Dragano, 200530 | Population (n=22 559) | 61% | Bivariate analyses | Age, east/west residency, education, occupational status, physical demands and occupational hazards, weekly working hours and job insecurity | Downsizing | OR 1.53 (1.30–1.79) for exposed men of increased symptom load. OR 1.71 (1.43–2.06) for exposed women |
Greubel, 201131 | Police employees (n=1523) | 76% | ANOVA | Age, gender and shift work | Relocation, extensive changes: downsizing and job changes | Anxiety ↑ depression ↑ |
Karasek, 199032 | White collar (n=8504) | 87% | Mantel-Haenszel | Age, sex | Company initiated job changes | Depression ↑ |
Pepper, 200333 | State employees (n=5889) | 55% | Hierarchical linear modelling | Nothing | Downsizing rate | Mental health component ↔ Perceived Stress ↔ |
Verhaeghe, 200634 | Hospitals (n=2094) | 51% | Logistic regression | Age, sex | Situational changes in working environment | Distress ↑ |
Longitudinal studies | ||||||
Author, year | Sampling frame | Participation (follow-up time) | Statistical analysis | Adjusted/controlled for | Exposure | Outcome (p<0.05%) |
Dahl, 201135 | Population (n=92 869) | Registry based study (6 years) | Multivariate analysis with logit models | Age, gender, children (3 age groups), marital status, stress of parents and spouse, firm tenure, wage, occupation level, firm size, firm age and industry classification | Organisational change | Stress ↑ |
Ferrie, 19982 | White collar (n=7419) | 73% (∼3 years) | ANOVA and logistic regression | Age and employment grade | Change in job description, increase in workload and pace of work | OR 1.56 (1.30–1.86) of minor psychiatric morbidity for exposed men. For women ↔ |
Hansson, 200836 | Hospital (n=226) | 74% (1 year) | One-way and two-way ANOVA | Nothing | Reorganisation of work systems | Biological stress markers ↔ |
Kivimaki, 200737 | Municipal employees (n=26 682) | Registry based study (7 years) | Negative binominal regression | Sex, age, education, occupational status and local government | Downsizing | RR 1.49 (1.10–2.02) of psychotropic drug use for exposed men. RR 1.12 (1.00–1.27) for exposed women |
Loretto, 20103 | Hospital (n=5385) | Baseline 18.4%, 84.3% retention rate of 1st cohort, 76.7 of 2nd cohort, (∼1 year) | Logistic regression | Past GHQ casesness, personal and biographic factors, objective workplace and job characteristics | Perceived amount of overall change | OR 1.21 (1.06–1.38) of GHQ caseness when exposed to change |
Moore, 200638 | Manufacturing company (n=460) | 62%–74% (∼2.5 years) | ANCOVA | Age, gender, education and marital status | Downsizing | Depression ↔ |
Netterstrom, 201039 | Civil servants (n=685) | 44% (2 years) | Logistic regression | Age, leadership, department and occupation | Merger | Depression ↔ |
Probst 200340 | State agency employees (n=313) | 63% (6 months) | Multivariate ANOVA | Nothing | Restructuring | Mental health index ↔ |
Rohall, 200141 | Military officers (n=1536) | 85% (∼2 years) | ANOVA | Nothing | Downsizing | Anxiety ↑ depression ↑ |
Vaananen 201142 | Forest industry corporation (n=6511/4096) | 82.3% (∼4 years) | Cox proportional hazard models | Sex, age, marital status, occupational status, sense of coherence and job characteristics | Merger | Increased risk of postmerger psychiatric event HR 1.60 (1.19–2.14) |
Woodward 199943 | Hospital employees (n=346) | 47% (∼2 year) | ANOVA | Nothing | Re-engineering | Anxiety ↑ depression ↑ |
Where applicable ratios (HR, OR, RR and PR) and 95% CIs were available in the studies, these are listed.
Otherwise relevant significant results are reported using errors: ↓ = decrease in symptoms; ↑ = increase in symptoms; ↔ no significant change in symptoms.
ANOVA, analysis of covariance; GHQ, General Health Questionnaire; PR, prevalence ratio.