Table 1

Findings and appraisals of studies using stress scale

	Taylor (1991), USA	Kwandt (1992), USA	Razavi et al (1993), Belgium	Tsai and Crockett (1993), Taiwan	Carson et al (1999), UK	Lee and Crockett (1994), Taiwan	Proctor et al (1998), UK
Symbols: •, Seems no problem; T, treatment group; (E), environmental management; P, placebo group; (P), personnel support; C, control group.
Design features
Type of study	RCT	RCT	RCT	RCT	RCT	RCT	PCS
Population	102 nurses	31 psychiatric nurses	72 nurses	134 nurses	53 mental health nurses	57 nurses	84 care assistants
	Ta35, Tb34, C33	Ta11, Tb10, P10	T36, C36	T ?, P?	T27, P26	T29, P28	T42, C42
				Power analysis
Focus of intervention	Ta: exercise	Ta: humour	Education + role-playing (P)	Education + relaxation (P)	Social support education (P)	Cognitive techniques (P)	Education + training (E)
	Tb: music (P)	Tb: relaxation (P)
Duration	6 week (360h)	1 session (3h)	8 weeks (24h)	5 weeks (13.5h)	5 weeks (10h)	2 weeks (12h)	6 months regularly
Reliability and validity of scale	Previous R&V data reported	Previous R&V mentioned	Previous R stated	Previous V data reported	Previous R&V mentioned	Present R data reported	R&V not mentioned
			Modified by authors
Authors’ conclusion	Effective	Not successful but can reduce stress	Effective	Effective	No significant advantage	Effective	Not clear (further research needed)

Validity of results
Randomisation	Strategy not clear	Strategy not clear	•	Questionable	•	Not reported	N/A
Sample attrition	9, characteristics considered	11, not considered in analysis	4, not considered in analysis	3, not considered in analysis	11, not considered in analysis	3, not considered in analysis	14, not considered in analysis
Blinding	Not given	Researcher blind	Not given	Not given	Not given	•	Not given
Similarity at start	Not considered	Stress levels not reported	Stress levels not reported	Not clear	•	•	•
Confounding factors	Characteristics of control not considered	Equal treatment and effects of placebo not considered	Equal treatment aside from intervention not reported	Effects of placebo on stress not considered	Effects of placebo on stress not considered	Placebo intervention might worsen stress	Managerial climate change during the study

Authenticity of differences
Main outcome (scale)	Ta=−4.1 Tb=−1.5	Ta=128.9 Tb=137.9	Not reported	T=−0.7 C=−0.2	T=−3.0 P=−11.5	T=−8.1 P=−2.0	Larger increase in C than T
	C=+1.2	C=125.6
Statistical test	Post hoc χ²	ANCOVA	Fisher	ANCOVA	Wald test	ANOVA	Mann-Whitney
	Ta-Tb: not significant	Not significant	Not significant	Significant	Significant	Significant	Not significant on all subscales
	Ta-C Tb-C: significant
Way of comparison	•	•	Within group analysis	Measurement timing not proper	•	•	Total score not analysed
Clinical significance	Not clear	Described but not discussed	Not reported	Not clear	Not clear	Not clear	Not clear
Conclusion of appraisals	Potentially effective	Evaluation impossible	Evaluation impossible	Potentially effective	Questioned but possibly effective	Effective but questionable	No evidence of effectiveness but possibly effective