Preventing occupational stress in healthcare workers

Preventing occupational stress in healthcare workers (Review)

Marine A, Ruotsalainen J, Serra C, Verbeek J

This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library2006, Issue 4

http://www.thecochranelibrary.com

1Preventing occupational stress in healthcare workers (Review)

Copyright © 2006 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

http://www.thecochranelibrary.com

T A B L E O F C O N T E N T S

1ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3CRITERIA FOR CONSIDERING STUDIES FOR THIS REVIEW . . . . . . . . . . . . . . . . . .

3SEARCH METHODS FOR IDENTIFICATION OF STUDIES . . . . . . . . . . . . . . . . . . .

4METHODS OF THE REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5DESCRIPTION OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7METHODOLOGICAL QUALITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10NOTES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10POTENTIAL CONFLICT OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . .

10ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14Characteristics of included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19Characteristics of excluded studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19ADDITIONAL TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19Table 01. Search strategy (MEDLINE) . . . . . . . . . . . . . . . . . . . . . . . . . . .

21Table 02. Study quality and intervention type . . . . . . . . . . . . . . . . . . . . . . . . .

22Table 03. Levels of evidence: legend . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22Table 04. Levels of evidence: Person-directed interventions: Stress . . . . . . . . . . . . . . . . . .

22Table 05. Levels of evidence: Person-directed interventions: Maslach Burnout Inventory . . . . . . . . . .

23Table 06. Levels of evidence: Person-directed interventions: State Trait Anxiety Inventory . . . . . . . . . .

23Table 07. Levels of evidence: Person-directed interventions: general symptoms . . . . . . . . . . . . . .

23Table 08. Levels of evidence: Work-directed interventions: Stress . . . . . . . . . . . . . . . . . .

23Table 09. Levels of evidence: Work-directed interventions: Maslach Burnout Inventory . . . . . . . . . . .

23Table 10. Levels of evidence: Work-directed interventions: general symptoms . . . . . . . . . . . . . .

24ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

24Comparison 01. Person-directed intervention vs no intervention (Short Term) . . . . . . . . . . . . . .

24Comparison 02. Person-directed intervention vs no intervention (Medium and Long Term) . . . . . . . . .

24Comparison 03. Work-directed intervention vs no intervention (Short Term) . . . . . . . . . . . . . .

24Comparison 04. Work-directed intervention vs no intervention (Medium Term) . . . . . . . . . . . . .

24COVER SHEET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26GRAPHS AND OTHER TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26Analysis 01.01. Comparison 01 Person-directed intervention vs no intervention (Short Term), Outcome 01 Beck &

Srivastava Stress Scale and Perceived Stress Scale . . . . . . . . . . . . . . . . . . . . . .

26Analysis 01.02. Comparison 01 Person-directed intervention vs no intervention (Short Term), Outcome 02 Maslach

Burnout Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27Analysis 01.03. Comparison 01 Person-directed intervention vs no intervention (Short Term), Outcome 03 State Trait

Anxiety Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28Analysis 01.04. Comparison 01 Person-directed intervention vs no intervention (Short Term), Outcome 04 General

Health Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28Analysis 02.01. Comparison 02 Person-directed intervention vs no intervention (Medium and Long Term), Outcome

01 Maslach Burnout Inventory (at 2 years) . . . . . . . . . . . . . . . . . . . . . . . .


02 Stait Trait Anxiety Inventory (Jones at 3 months and Yung at 1 month) . . . . . . . . . . . . .

iPreventing occupational stress in healthcare workers (Review)



03 Perceived Stress Scale (at 1 month) . . . . . . . . . . . . . . . . . . . . . . . . .


04 General Health Questionnaire (at 1 month) . . . . . . . . . . . . . . . . . . . . . .

30Analysis 03.01. Comparison 03 Work-directed intervention vs no intervention (Short Term), Outcome 01 Maslach

Burnout Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31Analysis 03.02. Comparison 03 Work-directed intervention vs no intervention (Short Term), Outcome 02 Nursing

Stress Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31Analysis 03.03. Comparison 03 Work-directed intervention vs no intervention (Short Term), Outcome 03 General

symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31Analysis 04.01. Comparison 04 Work-directed intervention vs no intervention (Medium Term), Outcome 01 Nursing

Stress Scale (at 6 months) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

iiPreventing occupational stress in healthcare workers (Review)


Preventing occupational stress in healthcare workers (Review)

Marine A, Ruotsalainen J, Serra C, Verbeek J

Status: New

This record should be cited as:

Marine A, Ruotsalainen J, Serra C, Verbeek J. Preventing occupational stress in healthcare workers. Cochrane Database of SystematicReviews 2006, Issue 4. Art. No.: CD002892. DOI: 10.1002/14651858.CD002892.pub2.

This version first published online: 18 October 2006 in Issue 4, 2006.

Date of most recent substantive amendment: 21 August 2006

A B S T R A C T

Background

Healthcare workers can suffer from occupational stress which may lead to serious mental and physical health problems.

Objectives

To evaluate the effectiveness of work and person-directed interventions in preventing stress at work in healthcare workers.

Search strategy

We searched the Cochrane Depression Anxiety and Neurosis Group trials Specialised Register, MEDLINE, PsychInfo and Cochrane

Occupational Health Field database.

Selection criteria

Randomised controlled clinical trials (RCT) of interventions aimed at preventing psychological stress in healthcare workers. For work-

directed interventions interrupted time series and prospective cohort were also eligible.

Data collection and analysis

Two authors independently extracted data and assessed trial quality. Meta-analysis and qualitative synthesis were performed where

appropriate.

Main results

We identified 14 RCTs, three cluster-randomised trials and two crossover trials, including a total of 1,564 participants in intervention

groups and 1,248 controls. Two trials were of high quality.

Interventions were grouped into 1) person-directed: cognitive-behavioural, relaxation, music-making, therapeutic massage and mul-

ticomponent; and 2) work-directed: attitude change and communication, support from colleagues and participatory problem solving

and decision-making, and changes in work organisation.

There is limited evidence that person-directed interventions can reduce stress (standardised mean difference or SMD -0.85; 95%CI

-1.21, -0.49); burnout: Emotional Exhaustion (weighted mean difference or WMD -5.82; 95%CI -11.02, -0.63) and lack of Personal

Accomplishment (WMD -3.61; 95%CI -4.65, -2.58); and anxiety: state anxiety (WMD -9.42; 95%CI -16.92, -1.93) and trait anxiety

(WMD -6.91; 95%CI -12.80, -1.01). One trial showed that stress remained low a month after intervention (WMD -6.10; 95%CI

-8.44, -3.76). Another trial showed a reduction in Emotional Exhaustion (Mean Difference or MD -2.69; 95%CI -4.20, -1.17) and

in lack of Personal Accomplishment (MD -2.41; 95%CI -3.83, -0.99) maintained up to two years when the intervention was boosted

with refresher sessions. Two studies showed a reduction that was maintained up to a month in state anxiety (WMD -8.31; 95%CI

-11.49, -5.13) and trait anxiety (WMD -4.09; 95%CI -7.60, -0.58).

There is limited evidence that work-directed interventions can reduce stress symptoms (Mean Difference or MD -0.34; 95% CI -0.62,

-0.06); Depersonalization (MD -1.14; 95% CI -2.18, -0.10), and general symptoms (MD -2.90; 95% CI -5.16, -0.64). One study

showed that the difference in stress symptom level was nonsignificant at six months (MD -0.19; 95%CI -0.49, 0.11).



Authors’ conclusions

Limited evidence is available for the effectiveness of interventions to reduce stress levels in healthcare workers. Larger and better quality

trials are needed.

P L A I N L A N G U A G E S U M M A R Y

Preventing occupational stress in healthcare workers

Healthcare workers suffer from work-related or occupational stress often resulting from high expectations coupled with insufficient

time, skills and/or social support at work. This can lead to severe distress, burnout or physical illness, and finally to a decrease in quality

of life and service provision. The costs of stress and burnout are high due to increased absenteeism and turnover.

We conducted a systematic search of the literature on preventing stress or burnout in healthcare workers. We then appraised the quality

of the studies found and combined their results.

Person-directed interventions that include a cognitive-behavioural approach (e.g. coping skills training), combined with relaxation

techniques or not, can be effective in reducing burnout, anxiety, stress and general symptoms in healthcare workers when compared to

no intervention. Work-directed interventions that include communication or nursing delivery change can also be effective in reducing

burnout, stress and general symptoms in healthcare workers when compared to no intervention. At best, the results of stress or burnout

reducing interventions may still be apparent from six months to two years after the end of the interventions.

Most of the studies are small and of poor quality, and it is not clear how large a change in a stress or burnout score is meaningful.

B A C K G R O U N D

Healthcare workers can suffer from work-related or occupational

stress as a result of organisational factors and an imbalance of de-

mands, skills and/or social support at work. Under some condi-

tions this can lead to severe distress, burnout or psychosomatic

diseases, and the resulting deterioration in quality of life and ser-

vice provision (Weinberg 2000). Burnout has been defined as “a

persistent, negative, work-related state of mind in ’normal’ indi-

viduals that is primarily characterized by exhaustion, which is ac-

companied by distress, a sense of reduced effectiveness, decreased

motivation, and the development of dysfunctional attitudes and

behaviours at work. This psychological condition develops grad-

ually but may remain unnoticed for a long time for the individual

involved. It results from a misfit between intentions and reality at

the job. Often burnout is self-perpetuating because of inadequate

coping strategies that are associated with the syndrome” (Schaufeli

2003). Burnout and occupational stress are both considered to

result from exposure to stressors at work, the effects of which are

mediated by personal coping or the ability to deal with environ-

mental stressors at a personal level (Cooper 2001; Schaufeli 2003).

Burnout is considered here as a specific form of psychological stress

and not as a clinical diagnosis (Firth-Cozens 1999). The economic

impact of such conditions is high, as can be inferred from data on

absenteeism and turnover (Jacobson 1996; Raiger 2005). More

than 10% of total claims for occupational diseases are attributed to

stress at work (Williamson 1994). There are a variety of stress fac-

tors in the workplace of healthcare workers that have been shown

to increase the risk of distress and burnout such as increasing

workload, emotional response to contact with suffering and dying

patients, and organisational problems and conflicts (Payne 1987;

McNeely 2005). In addition, many studies have shown that levels

of dissatisfaction, distress and burnout at work are quite high in

healthcare workers (Deckard 1994; Karasek 1992; Maslach 1982;

Ramirez 1996; Raiger 2005).

In order to lessen the impact of different psychosocial risk fac-

tors, programs and intervention activities have been designed with

many focussing on healthcare workers. These interventions are

either directed at risk factors in the working environment, or at

individual persons - such as increasing the individual’s ability to

cope with stress - or at both the working environment and the

individual person.

An extensive number of reviews have been published on the effec-

tiveness of interventions to prevent or treat stress (DeFrank 1987;

Murphy 1995; van der Hek 1997; van der Klink 2001). However,

apart from Mimura 2003, there are no other reviews that focus

specifically on interventions directed at stress prevention in health-

care workers. According to Michie 2003, the levels of psycholog-

ical ill health may be higher in healthcare than in non-healthcare

workers. Therefore, because the characteristics of interventions

designed for healthcare workers may be different from those of

other occupations, the aim of this review is to determine the effec-

tiveness of interventions to reduce stress specifically in healthcare

workers.



O B J E C T I V E S

1. To evaluate the effectiveness of work- and person-directed in-

terventions in preventing stress in healthcare workers.

2. To compare the effectiveness of different kinds of interventions

in preventing stress in healthcare workers.

C R I T E R I A F O R C O N S I D E R I N G

S T U D I E S F O R T H I S R E V I E W

Types of studies

For person-directed interventions we considered only randomised

controlled trials.

For work-directed interventions (see types of interventions) it is

usually unfeasible to randomise study participants because the in-

terventions are usually brought about at the environmental or

group level. This can be overcome by using the cluster-randomised

design where work places or departments are randomised to the

intervention or control groups. However, as this is a difficult and

fairly new design, we also considered additional study designs

for inclusion that allow for environmental interventions at the

group level: controlled before and after studies (otherwise known

as prospective cohort studies or quasi-experimental studies) and

interrupted time-series. The interrupted time-series design is de-

fined by the Cochrane Effective Practice and Organisation of Care

review group as a series of at least three outcome measurements

before and at least three after the implementation of an interven-

tion (EPOC 1998).

Types of participants

We included studies in which the interventions were directed at

workers who had not actively sought help for conditions such

as burnout, depression or anxiety disorder. This included inter-

ventions directed at healthcare workers officially employed in any

healthcare setting or at student nurses or physicians otherwise in

training to become a professional who were also doing clinical

work. This excluded studies in which the participants were just

caregivers and were not employed by a healthcare organisation.

Types of intervention

We included studies with any kind of intervention aimed at pre-

venting or reducing stress arising from work. Prevention was de-

fined as an intervention directed at healthy individuals not diag-

nosed with a specific disease.

Interventions were categorised into two groups:

1. Interventions directed at factors at work, or work-directed inter-

ventions containing measures to change the working environment,

work tasks or working methods. For example, interventions aimed

at decreasing workload or changing the organisation of work.

2. Interventions focusing on people, or person-directed interven-

tions, aimed at teaching personal skills, techniques or remedies

to decrease the effects of stressors at the individual level. For ex-

ample, interventions aimed at mitigating the effects of stress by

relaxation techniques or cognitive-behavioural techniques that in-

creased coping skills, without changing the work environment or

work tasks.

For both work- and person-directed interventions we included

studies with both non-intervention controls and with one or more

alternative intervention as a control group.

Types of outcome measures

Primary outcome measures

As primary outcomes we considered all validated self-report ques-

tionnaires measuring stress or burnout. These included mea-

sures such as: Karasek’s job content questionnaire (Karasek 1998),

Maslach Burnout Inventory (MBI) (Maslach 1982) and the Nurs-

ing Stress Scale (Gray-Toft 1981).

Secondary outcome measures

As secondary outcomes we considered all outcome measures of the

detrimental effects of stress or burnout. These included measures

such as:

(a) psychological symptoms: anxiety and depression, such as

the State Trait Anxiety Inventory (Spielberger 1970), General

Health Questionnaire (Goldberg 1991), Beck Depression Inven-

tory (Beck 1961)

(b) physical symptoms and physiological parameters, such as hor-

mone levels such as prolactin, corticosteroids or others.

We considered the effects measured with all the above instruments

on the following time-scales:

(i) Short Term: up to one month

(ii) Medium Term: from one month to a year

(iii) Long Term: over a year

All other outcomes that did not measure stress or its effects on

individuals were deemed beyond the scope of this review and were

thus excluded from analyses. Examples of excluded outcomes are:

coping skills, knowledge or attitude change, work-performance,

patient satisfaction and claims from clients, employee absenteeism

and turnover.

S E A R C H M E T H O D S F O R

I D E N T I F I C A T I O N O F S T U D I E S

See: Cochrane Depression, Anxiety and Neurosis Group methods

used in reviews.

See Collaborative Review Group search strategy for the search

strategy in the CCDAN database.

1. Electronic databases

(a) MEDLINE (1966- May 2005): see strategy in Table 01.



(b) PsycINFO: we used the following strategy up to May 2005:

(((“Nursery-School-Students” in MJ,MN) or (“Nursery-Schools”

in MJ,MN) or (“Nurses-” in MJ,MN) or (“Nursing-” in

MJ,MN) or (“Nursing-Education” in MJ,MN) or (“Nursing-

Homes” in MJ,MN) or (“Nursing-Students” in MJ,MN) or

(“Psychiatric-Nurses” in MJ,MN) or (“Public-Health-Service-

Nurses” in MJ,MN) or (“School-Nurses” in MJ,MN)) or

(“Health-Personnel” in MJ,MN)) and ((“Stress-Management” in

MJ,MN) or (“Occupational-Stress” in MJ,MN)))

and (intervention or effectiveness or evaluation)

(c) Cochrane Collaboration Specialised Registry of Clinical Trials

of the Depression, Anxiety and Neurosis Group (CCDANCTR):

latest search, February 2002, using the following strategy:

Keyword = burnout or “psychological stress”

and

Free-text = healthcare or staff or employ* or job or work* or nurs*

(d) Cochrane Occupational Health Field database

2. Reference lists

Reference lists from articles and reviews were also carefully

reviewed.

3. Handsearch

All issues between January 1987 and March 2000 of Work &

Stress were handsearched.

M E T H O D S O F T H E R E V I E W

Selection of studies

Two reviewers independently selected the studies to include

in the review according to the inclusion criteria. If there was

any disagreement concerning the inclusion of a study, this was

discussed jointly, and where necessary a third reviewer resolved the

disagreement.

Data extraction and management

The extraction of data was carried out using a made-to-measure

data extraction form filled in independently by two reviewers for

each study.

(a) If studies used a cross-over design we used the results from just

after the implementation of the intervention in the intervention

group, compared to the results in the concurrent waiting-list

control group.

(b) If studies used continuous measures and reported only means

and p-values, we calculated standard deviations from p-values in

the article to be used in the meta-analysis, according to the method

stated in the Cochrane Handbook for Reviewers.

(c) If studies compared more than one active intervention we chose

which of the active interventions to enter into the meta-analysis.

Assessment of methodological quality of included studies

Evaluation of the quality of clinical trials included in the review

was done by means of the checklist developed by Downs 1998.

We had clear reasons for using the Downs 1998 checklist. Since

we needed to include non-randomised studies of work-directed

interventions, we wanted to have an instrument that could assess

the quality of both randomised and non-randomised studies.

According to the review by Deeks 2003, the Downs 1998 checklist

was the best choice for this specific purpose. The checklist contains

separate subscales for reporting, external validity, internal validity

and power with a maximum number of points respectively 10,

2, 13 and 5, and a total maximum score of 30. The individual

items of the check list were rated either as yes (=1) or no/

unable to determine (=0). Two evaluators independently evaluated

the quality of the studies. Where necessary, study authors were

contacted for clarification. The scales of internal validity were used

to rate the quality of studies. Studies were considered to have high

internal validity if they scored higher than 75% of the maximum

score of the two scales combined. The other scales of the checklist

were reported separately.

Measures of treatment effect

The results of each trial were plotted as means and standard

deviations (SD) for continuous outcomes. For continuous

measures, in cases where the same instrument was used, preference

was given to analysing the results with weighted mean differences

(WMD) because these results are easier to interpret for clinicians

and other readers. If this was not possible because different

instruments were used, then standardised mean differences (SMD)

or effect sizes were used.

Unit of analysis issues

For studies that employed a cluster-randomised design and that

reported sufficient data to be included in the meta-analysis and

that did not make an allowance for the design effect we calculated

the design effect based on a fairly large assumed intra-cluster

correlation of 0.10. Even though we did not find information

for the intra-cluster correlation for these types of studies we

assumed that 0.10 would be a realistic estimate. We based

this assumption by analogy on studies about implementation

research (Campbell 2001). We followed the methods stated in the

Cochrane Handbook for the calculations.

Dealing with missing data

Where necessary, missing statistics data (means and standard

deviations) were sought from authors. Since all studies used

questionnaires to measure outcomes using continuous data, we

were unable to conduct an intention-to-treat analysis. We refrained

from standard approaches such as last observation carried forward

or imputing baseline outcomes since we did not have access to

raw patient data. Therefore all results were based on available case

analysis.

Assessment of heterogeneity

We tested for statistical heterogeneity by means of the I2 in

the meta-analysis graphs. If this is greater than 50% there is



substantial heterogeneity between studies. When studies were

statistically heterogeneous, a random-effects model was used;

otherwise a fixed-effect model was used. All estimates included a

95% confidence interval (95% CI).

Data synthesis

The choice between qualitative and quantitative pooling was based

first on clinical homogeneity. Clinically homogeneous studies

were defined as those with similar populations, interventions,

and outcomes measured at the same follow-up point. All studies

concerned healthcare workers. Interventions were divided into

person- and work-directed interventions. Outcomes were judged

similar if they used the same instrument such as the Maslach

Burnout Inventory (Maslach 1982). Statistical pooling was then

done for each subscale of the instrument with RevMan 4.2

software. If different instruments were used but the instruments

measured the same concept such as exposure to stressors, the results

were pooled as well.

Analysis of the data

The results were summarised within the two main categories of

interventions: work-directed interventions and person-directed

interventions. We used weighted mean differences (WMD) to

combine the results of studies that used similar questionnaires

to measure the outcome. For similar outcomes that were

measured with different questionnaires we used standardised mean

differences (SMD) to combine the results.

In addition, because only a proportion of the studies reported data

in sufficient detail for meta-analysis, we used a qualitative synthesis

to grade the evidence according to the methods suggested by the

Cochrane Back Review Group (van Tulder 2003). Although levels

of evidence are essentially subjective and arbitrary they are explicit

and reproducible by giving exact meaning to labels like strong

and limited evidence (de Vet 2003). In other words, this method

allows one to make a judgement (level) of the entirety of results

(evidence) presented in the included studies. We adapted the van

Tulder 2003 grading as follows (see also Table 03):

(i) Strong evidence: if there are two or more high quality studies

with similar positive or negative effects that are confirmed by the

quantitative analysis

(ii) Limited evidence: if there is only one high quality study and/or

multiple low quality studies with similar positive or negative effects

or the results of the high quality studies are not confirmed by the

quantitative analysis

(iii) No evidence: if there are no studies available

(iv) Conflicting evidence: if the results of studies point in opposite

directions.

Outcomes are defined as positive if there is a significant positive

effect at the p = 0.05 level, and negative if there is a significant

negative effect at the p = 0.05 level. If there is no significant effect

this will not be considered as a conflicting outcome.

We followed Van Tulder’s advice to explicitly describe conflicting

outcomes and criteria for high quality studies. We simplified

the levels of evidence as originally described by Van Tulder by

combining the levels of moderate and limited evidence into one

level of limited evidence. Moreover, we added the requirement of

confirmation by meta-analysis as an extra criterion to be met to

reach the level of strong evidence.

Publication bias

We planned to assess publication bias by means of funnel plots.

Methods for future updates

We will expand this review to also cover the cost-effectiveness of

interventions. The search strategy will also be updated to better

reflect the variety in healthcare job titles outside nursing. The

present sample is nonetheless the largest and most specific of

systematic reviews so far conducted. Therefore we feel that even

though our search strategy was initially developed with a different

set of inclusion criteria (all hospital staff including e.g. cleaners),

it still provided an effective and comprehensive search of all the

relevant literature.

D E S C R I P T I O N O F S T U D I E S

From the initial references, a set of 51 potentially eligible studies

were identified. These were then scrutinized further, and another

32 studies were excluded leaving a total of 20 studies. However, we

noticed that two studies (Lökk 1997; Lökk 2000) actually reported

separate sets of results of a single study so we included them as one

study. Thus, our final sample consisted of 19 studies (see table of

included studies).

Additional information regarding study details and/or statistical

data was sought from 10 authors and was received from six. Three

of the authors provided statistical data that had not been published

in their articles which enabled two of these three studies to be

entered into meta-analyses.

Study designs

Of the 19 included studies, 14 were randomised controlled trials,

three were cluster-randomised trials and two were cross-over stud-

ies. All three included cluster-randomized trials had a unit of anal-

ysis error. We calculated the design effect as 2.2 for the Lökk 1997

and Lökk 2000 study (r=2, m=13); 2.73 for the Schrijnemaekers

2003 study (r=16, m=18.3) and 1.72 for the Proctor 1998 study

(r=12, m=8.2). We used the design effect to diminish the actual

number of participants in both intervention and control groups

if quantitative outcomes were used. We were aiming to also in-

clude studies with interrupted time-series and controlled before-

after designs but the final sample contained none of these.

Country and time period

Eleven of the included studies had been conducted before the year

2000 and eight at or after the year 2000. Eight studies had been



carried out in Europe, another eight in North America and three

in Asia.

Type of settings and participants

Altogether 12 of the studies had been run in hospitals and the

remaining seven in varying settings ranging from a secure forensic

mental health unit to residential care homes for the elderly or the

disabled.

Sample sizes

The total number of participants in various intervention groups

was 1564, and in control groups 1248. The number of partici-

pants in the smallest study groups was less than 20 in seven stud-

ies, between 21 and 50 in six studies, between 51 and 100 in four

studies and more than 101 in two studies (Heaney 1995; Schri-

jnemaekers 2003). These two studies were the only ones deemed

a priori to have sufficient power to detect relevant differences.

Interventions

1. Person-directed interventions

Some interventions were quite straightforward but many used dif-

ferent methods.

(a) Cognitive-behavioural training: Lee 1994, Rowe 1999; von

Baeyer 1983; West 1984.

(b) Relaxation training: Tsai 1993; Yung 2004

(c) Music-making: Bittman 2003. The intervention consisted of

recreational music-making where participants engaged in playing

various drum and percussion instruments in response to questions

that had been designed to inspire deep thought and mutual respect.

(d) AMMA touch therapy (a form of therapeutic massage): McEl-

ligott 2003.

(e) Multicomponent interventions: Cohen-Katz 2005 employed

a pre-existing Mindfulness-Based Stress Reduction program con-

taining group sessions of formal didactic instruction and exercises

to enhance communication skills, stress reactivity and self-com-

passion. Ewers 2002 used a regime of training about practical skills

in reducing distress and improving functioning with schizophrenic

patients. In the Lökk study (Lökk 1997 and Lökk 2000) the inter-

vention consisted of cognitive-behavioural training coupled with

group therapy. Jones 2000 and Norvell 1987 both employed in-

terventions combining cognitive-behavioural and relaxation skills

training.

2. Work-directed interventions

Delvaux 2004 and Razavi 1993 used psychological training pro-

grams with theory, role-playing and experiential exchanges in-

tended to improve attitudes, communication skills and occupa-

tional stress. Heaney 1995 ran a training program about mobi-

lizing support from colleagues and about learning participatory

problem solving and decision-making skills. Melchior 1996, Proc-

tor 1998 and Schrijnemaekers 2003 employed interventions that

introduced innovations in nursing delivery via changes in work

organisation, knowledge and skills training and support and ad-

vice from supervisors.

Multiple interventions

Three studies compared more than one active intervention against

no intervention (Rowe 1999; West 1984; Yung 2004). The re-

sults of two studies could be entered into meta-analyses. Rowe

1999 compared an adaptive coping training condition with the

same training coupled with refresher sessions. Yung 2004 com-

pared stretch-release relaxation with cognitive relaxation. With

these two studies a choice had to be made regarding which active

interventions to enter into the meta-analyses. In the case of Rowe

1999, we chose the training with booster sessions condition as it

was the more intensive option of the two interventions, and with

Yung 2004 we chose the cognitive relaxation condition because

we felt that a cognitive intervention for a cognitive problem was

conceptually on firmer ground.

Type of control group (no intervention vs. other active or “pas-

sive” interventions)

Almost all (n=15) of the included studies used a no intervention

or waiting list control group (Bittman 2003; Cohen-Katz 2005;

Delvaux 2004; Ewers 2002; Heaney 1995; Jones 2000; Melchior

1996; Norvell 1987; Proctor 1998; Razavi 1993; Rowe 1999;

Schrijnemaekers 2003; von Baeyer 1983; West 1984; Yung 2004).

Four studies used another, “passive” intervention aiming to control

for the Hawthorne effect etc. (Lee 1994; Lökk 1997 & Lökk 2000;

McElligott 2003; Tsai 1993). Finally, three studies also compared

one or more active interventions (Rowe 1999; West 1984; Yung

2004).

Follow-up

(i)Short Term:

There were six studies in which the last outcome measurement

was either at the end of the intervention (Ewers 2002; McElligott

2003; Norvell 1987; Proctor 1998; von Baeyer 1983) or at less

than a month after intervention (Tsai 1993).

(ii) Medium Term:

In nine studies the last follow-up measurement was between one

and six months after intervention (Heaney 1995; Lee 1994; Razavi

1993; Yung 2004; Bittman 2003; Cohen-Katz 2005; Delvaux

2004; Lökk 1997 & Lökk 2000; West 1984).

(iii) Long Term:

There were also four studies in which the last follow-up measure-

ment was at twelve months after intervention or later (Jones 2000;

Melchior 1996; Rowe 1999; Schrijnemaekers 2003).

Excluded studies

Reasons for excluding the 32 studies were the following (see table

of excluded studies):

(a) No intervention (Bradley 2002; Fernandez 1995; Teasdale

2001).

(b) No intention to prevent or treat stress or burnout (Schinke

1977).

(c) No stress or burnout or health effects outcome (Calomeni

1999; McPhail 1990; Nissen 1997; Taylor 2001; Weir 1997).



(d) Non-randomised study design (Beddoe 2004; Berg 1994;

Bruneau 2004; Bunce 1996; Freedy 1994; Jones 1998; Keyes

1988; Kushnir 1998; McCue 1991; Murphy 1983; Pålsson 1996;

van Dierendonck 1998; Young 2001; Zimber 2001).

(e) Participants were not healthcare workers or nursing students

with clinical duties (Drain 1991; Lloyd 2002; Toivanen 1993).

(f ) Participants were nursing students but did not have clinical

duties (Godbey 1994; Heaman 1995; Hughes 2003; Johansson

1991; Mancini 1983; Sharif 2004).

M E T H O D O L O G I C A L Q U A L I T Y

Allocation concealment and randomisation

Only one study (Jones 2000) reported concealing intervention

versus control group allocation until it was completed and irrevo-

cable. This was also the only study that described an acceptable

method of randomisation, using random number tables. One fur-

ther study (Tsai 1993) reported having used coin tossing as the

method of randomisation whereas the rest did not describe their

choice of randomisation method.

Blinding

There was mention in only one of the studies (Delvaux 2004)

of blinding those rating the outcome measurements i.e. Nursing

Stress Scale questionnaires. In all but one of the included studies

(McElligott 2003), the blinding of participants to the interven-

tions they were receiving would have been impossible due to the

nature and aims of interventions being self-evident.

Attrition

The loss of participants was low throughout, with only two of the

studies (Norvell 1987; Tsai 1993) suffering a loss exceeding 20%

of the initial sample.

Quality assessment with the Downs 1998quality checklist

Two reviewers (JR and JV) independently rated all included studies

with the Downs 1998 quality checklist. The overall Cohen’s Kappa

of agreement for the entire checklist was 0.54 and for items 13-

25 (internal validity) it was 0.62. The two reviewers then rated all

studies a second time but now working together. All disagreements

were solved by discussion and better definition of the criteria used

for assessment.

The results of the second rating are presented broken down into

the five subscales of the checklist:

A: reporting,

B: external validity,

C: internal validity - bias,

D: internal validity - confounding/ selection bias

E: power (see Table 02: Study quality and intervention type).

The overall mean score (± standard deviation) of the 19 included

studies for the checklist was 17.4 ± 3.8 out of a maximum attain-

able score of 30, which means that on average 58% of the max-

imum attainable score was achieved. The minimum score was 9

and the maximum score 23.

A: The mean score of the quality of reporting scale was 6.6 ± 0.4

out of a maximum of 10.

B: For the external validity subscale the mean score was 0.5 ± 0.1

out of a maximum of two.

C+D: For the combined internal validity subscales the mean score

was 7.9 ± 0.3 out of a maximum of 13. There were only two

studies that got more than 75% on the internal validity subscales

and were therefore rated as high quality studies (Delvaux 2004;

Schrijnemaekers 2003).

E: For statistical power the mean score was 2.4 ± 0.4 out of a

maximum of five.

Studies on work-directed interventions scored on average higher

than studies on person-directed interventions on the checklist as

a whole and on all the subscales. This difference was, however,

only significant for the subscale of statistical power. Most person-

directed studies were fairly small in contrast to the work-directed

interventions that had much bigger numbers of participants. The

two studies that were rated as being high quality were both studies

on work-directed interventions.

R E S U L T S

1. Person-directed interventions

1.1 Primary outcome Stress and Burnout symptoms

(a) Stress symptoms

There were two studies (Jones 2000; Lee 1994) that measured

stress directly and were thus included in a meta-analysis. Since

both used different scales (Jones 2000: Beck and Srivastava Stress

Inventory (Beck 1991); Lee 1994; Perceived Stress Scale (Cohen

1983)) we used standardised mean differences (SMD) in the analy-

sis instead of weighted mean differences (WMD). Person-directed

interventions reduced stress significantly when compared to no in-

tervention and when measured by the abovementioned two stress

scales (SMD -0.85; 95% CI -1.21 to -0.49).

In the medium term, stress remained lower in the intervention

group than it did in the control group in one study (Lee 1994),

(mean difference -6.10; 95% CI -8.44 to -3.76).

The levels of evidence table showed a positive effect in four stud-

ies of person-directed interventions on stress and one inconclu-

sive outcome (Rowe 1999) (Table 04). The scales used to measure

stress that produced results which could not be entered into meta-

analysis were: Stress Assessment Inventory (Rowe 1999), Nursing

Stress Checklist (Benoliel 1990) and the Job-Related Tension In-

dex (Kahn 1964).

(b) Burnout symptoms

Altogether three studies (Cohen-Katz 2005; Ewers 2002; Rowe

1999) were included in a meta-analysis of person-directed inter-



ventions intended to reduce burnout symptoms measured at the

short term with the Maslach Burnout Inventory (MBI) (Maslach

1996). The summary effect of two subscales of the Maslach

Burnout Inventory was in favour of the intervention: Emotional

Exhaustion (WMD -5.82; 95% CI: -11.02 to -0.63), and lack

of Personal Accomplishment (WMD -4.89; 95% CI: -8.71 to

-1.07). The Depersonalization subscale confidence interval just

overlapped with zero: (WMD -2.44; 95% CI: -5.06 to 0.17).

There was considerable heterogeneity between the three studies in

the meta-analysis (I2 ranged from 67.6% to 86.7%; p=0.05 and

p=0.0005 respectively), with Rowe 1999 having a smaller decrease

in the MBI scores (respectively -2, -2, -0.6) than the other two

studies.

For the long term, the results of Rowe 1999 showed that when

compared to six weeks of cognitive-behavioural training, having

refresher sessions at 5, 11 and 17 months led to significantly lower

Emotional Exhaustion (mean difference -6.00; 95% CI -8.16 to

-3.84) and lack of Personal Accomplishment (mean difference

-5.82; 95% CI -7.89 to -3.75) at two years.

The levels of evidence table showed a positive effect in two studies

(Ewers 2002; Rowe 1999) with person-directed interventions on

all burnout subscales as measured with the Maslach Burnout In-

ventory (Maslach 1996) and in three other studies (Cohen-Katz

2005; Norvell 1987; West 1984) on one or two subscales (Table

05). For one study (Bittman 2003), the results were inconclusive

for all subscales.

1.2 Secondary outcomes: Anxiety and General Health com-

plaints

(a) Anxiety

Three studies (Jones 2000; von Baeyer 1983; Yung 2004) were

included in a meta-analysis of person-directed interventions with

the State-Trait Anxiety Inventory (Spielberger 1970) as the out-

come. The data of two studies could not be entered into the analy-

sis because Rowe 1999 did not report the results of the State-Trait

Anxiety Inventory measurements and in the case of West 1984

the results were not in a useable format. Based on Jones 2000, von

Baeyer 1983, Yung 2004, person-directed interventions signifi-

cantly reduced both state anxiety (WMD -9.42; 95% CI -16.92

to -1.93) as well as trait anxiety (WMD -6.91; 95% CI -12.80

to -1.01) when compared to no intervention and when measured

by the State-Trait Anxiety Inventory. There was considerable het-

erogeneity (I2 = 83.1%; p=0.003) with the study of Yung 2004

having a smaller decrease on both scales (respectively -5 and -2)

than the other two studies in the meta-analysis.

In the medium term, based on the studies of Jones 2000 and Yung

2004 state-anxiety and trait-anxiety scores (WMD -8.31; 95% CI

-11.49 to -5.13) and trait anxiety scores (WMD -4.09; 95% CI

-7.60 to -0.58) remain lower in the intervention group than in the

control group.

The levels of evidence table showed a positive effect in three in-

cluded studies (Jones 2000; von Baeyer 1983; West 1984) of per-

son-directed interventions on both state and trait anxiety and one

study (Yung 2004) was positive in the state anxiety domain and

inconclusive in the trait anxiety domain (Table 06).

(b) General symptoms

There were two studies (Jones 2000; Yung 2004) that employed

the General Health Questionnaire (Goldberg 1991) and could

therefore be included in a meta-analysis. In the meta-analysis of

these studies person-directed interventions did not reduce general

symptoms significantly more than no intervention (WMD -11.87;

95% CI -27.24 to 3.49).

Three other studies used general health symptoms as an outcome

but their results could not be used for meta-analysis. They also did

not contribute to the levels of evidence because of inconclusive re-

sults (Cohen-Katz 2005; Norvell 1987; Tsai 1993). In the levels of

evidence tables the same studies as in the meta-analysis contribute

positive results based on the authors conclusions (Table 07).

In the medium term, based on the results of Yung 2004 cogni-

tive relaxation training maintains the reduction in General Health

Questionnaire scores for at least one month whereas stretch-re-

lease relaxation training does not (mean difference -7.10; 95% CI

-10.58 to -3.62).

2. Work-directed interventions

Three of the five studies on work-directed interventions reported

results that could be quantified. The other two studies only re-

ported the results of multivariate analysis.

2.1 Primary outcome: Stress and Burnout symptoms

(a) Stress symptoms

The results of Delvaux 2004 show that psychological training on

attitudes, communication skills and occupational stress alleviate

stress (mean difference -0.34; 95% CI -0.62 to -0.06) when com-

pared to no intervention and when measured with the Nursing

Stress Scale (Gray-Toft 1981). This difference became non-signifi-

cant at six-months (mean difference -0.19; 95% CI -0.49 to 0.11)

The addition of the study by Proctor 1998 in the levels of evidence

tables did not change this result because it was inconclusive (Table

08).

(b) Burnout symptoms

According to the results of Melchior 1996, support and advice

given by nurse managers or quality care coordinators alleviate

symptoms on one of the subscales of the Maslach Burnout Inven-

tory: Depersonalization (mean difference -1.14; 95% CI -2.18 to

-0.10) when compared to no intervention. The other two sub-

scales of the Maslach Burnout Inventory, Emotional Exhaustion

and lack of Personal Accomplishment, were non-significant.

The addition of the study by Schrijnemaekers 2003 in the levels

of evidence table showed positive study results for the Maslach



Burnout Inventory subscale lack of Personal Accomplishment,

whereas the Melchior 1996 study was inconclusive (Table 09).

2.2 Secondary outcome: Anxiety and General Health com-

plaints

Anxiety was not used as an outcome measure in any of the studies

on work-directed interventions.

Proctor 1998 used a combination of training of knowledge and

skills and individual programme planning to decrease general

symptoms (mean difference -2.90; 95% CI -5.16 to -0.64) when

compared to no intervention and when measured with the Gen-

eral Health Questionnaire (Goldberg 1991).

Adding the study by Heaney 1995, which measured general health

complaints with the Symptom Checklist-90-Revised (Derogatis

1976), into the levels of evidence tables did not change the results

(Table 10).

Publication bias

We planned to assed publication bias using funnel plots, but no

more than three studies could have been included, making it diffi-

cult draw conclusions. However, as most studies had only a small

sample size and all reported positive outcomes, it is conceivable

that there may have been publication bias. If there are more studies

included in future updates, we will assess the extent of publication

bias by means of funnel plots.

D I S C U S S I O N

The meta-analyses and the qualitative analyses in this review show

that there is limited evidence that person-directed interventions

in healthcare workers effectively reduce the levels of burnout, anx-

iety, stress and general symptoms. With regard to work-directed

interventions, we also found limited evidence that work-directed

interventions in healthcare workers effectively reduce the levels of

stress, general symptoms and burnout on the Depersonalization

and lack of Personal Accomplishment subscales of the Maslach

Burnout Inventory (Maslach 1996).

The programmes for managing stress at work vary greatly with

respect to objectives, types of interventions, duration and the pro-

fessional group at which they are aimed. It therefore makes sense

to try to classify the different types of interventions in order to

compare them in terms of efficacy. Person-directed interventions

are the most common interventions to have been evaluated.

We did not find it feasible to further subdivide the person-directed

interventions because almost all studies used several different in-

gredients. Most claim that it is possible to change the participants’

cognitions about stressful elements at work. We refrained from a

further sensitivity analysis based on differences in quality because

most studies scored about average on the quality scale. Nor did we

perform a sensitivity analysis based on differences in the content

of the interventions such as relaxation techniques versus cogni-

tive behavioural techniques because most interventions comprised

both elements.

We also examined interventions aimed at changing the organi-

sation of work. Although broad organisational interventions are

generally regarded as being more useful and appropriate in terms

of primary prevention, studies evaluating them are very scarce or

lack a robust evaluative design.This has not been confirmed in

this review since on average studies on work-directed interventions

scored higher on the quality checklist, and the two studies with

the highest quality scores were evaluations of work-directed inter-

ventions. Since there were no direct comparisons between work-

directed and person-directed interventions, we did not attempt

to compare the outcomes of the work-directed with the person-

directed interventions. It is clear however, that the problem of at-

trition is greater in the evaluation of work-directed intervention

studies in which the personal interest of participants is apparently

different.

In the meta-analysis of person-directed interventions there was

considerable heterogeneity with Rowe 1999 showing a smaller

decrease in MBI scores in the short term. The same holds for

the Yung 2004 study that showed less decrease in anxiety scores.

However, in the medium and long term follow-up these differences

disappeared. We could not find a plausible explanation for the

heterogeneity in the short term.

The quality of the evidence that we found was not very high.

Some studies applied rigorous methods but contended with attri-

tion problems. Most RCTs were small and in all but one of the

included RCT studies (Jones 2000) the method of randomisation

was not reported or not valid. It was also difficult to get a good

impression of the concealment of allocation to researchers. We as-

sumed that outcomes that were measured by questionnaire were

reported blind to the researchers but we could not be certain about

this assumption. The reporting of most studies was of low quality

as well which is reflected in the low number of studies that could

be used for the meta-analysis. All in all, according to our classifi-

cation of the studies, only two of 19 studies were high quality in

terms of internal validity.

Even though we found significant results it is not possible to say

how this is related to the clinical relevance of the changes achieved

in the studies. With the Maslach Burnout Inventory (Maslach

1996) for example, there is no generally accepted change that

would be regarded as clinically relevant. The authors of the ques-

tionnaire were contacted but could not help clarify this issue for

us.

Since most studies had only a small sample size and all reported

positive outcomes, it is conceivable that there may have been pub-

lication bias. However, it was not possible to assess the extent to

which publication bias may have been present due to the small



number of studies that reported outcomes that could be used for

statistical analysis.

Comparison with other reviews

When compared to the Mimura 2003 review, the present review

used more rigorous inclusion criteria ,only accepting randomised

controlled trials and randomised crossover trials. We found twice as

many (14) RCTs than Mimura 2003 (7). Where Mimura 2003 re-

frained from drawing clear conclusions based on the studies found,

we could make better inferences. The quality of the evidence that

we found was also better than Michie 2003 found, as their sample

consisted almost entirely of cross-sectional studies. Another review

by van der Klink 2001 involved a meta-analysis of interventions

to prevent or treat stress in all occupations. They synthesized all

study outcomes regardless of how they were measured and used all

possible figures available in primary studies to be included in their

meta-analysis. This made it difficult to decipher the meaning of

their findings. Compared to van der Klink’s review, we managed to

relate our results to concrete outcomes such as the Nursing Stress

Scale (Gray-Toft 1981), the Maslach Burnout Inventory (Maslach

1996) or the State-Trait Anxiety Inventory (Spielberger 1970). In

contrast with our review, van der Klink 2001 concluded that there

was no evidence for work-directed interventions. Due to the differ-

ent methods used it is difficult to explain this contrasting finding.

A recent review by Gilbody 2006 about interventions to improve

morale of staff was restricted to mental healthcare workers only.

It reached more or less the same conclusion as the present review,

even though the number of studies included was fewer and they

used a wider range of outcome measures and study designs. We

were more precise in our outcome measures and were therefore

better able to quantify our results.

A U T H O R S ’ C O N C L U S I O N S

Implications for practice

Limited evidence is available for the effectiveness of person- and

work-directed interventions to reduce stress levels in healthcare

settings. At best, the results of these interventions are still apparent

six months to two years after the end of interventions. There is

evidence from one trial that interventions which contain cognitive

elements yield better results than those with behavioural elements.

Implications for research

For person-directed interventions, larger studies with adequate

randomisation and blinding of allocation for the researchers are

needed to confirm the results of this systematic review. More stud-

ies are needed that contrast various stress or burnout reducing tech-

niques with one another. For work-directed interventions cluster-

randomised studies are feasible, but care should be taken to avoid

attrition. Studies that contrast work-directed interventions with

person-directed interventions will show if one type of intervention

is more effective in reducing stress levels than the other. Studies

are also needed comparing different person- and work-directed

interventions with one another.

N O T E S

The title of this review has changed from ’Effectiveness of occu-

pational stress management programmes’ to ’Preventing occupa-

tional stress in healthcare workers’.

P O T E N T I A L C O N F L I C T O F

I N T E R E S T

None known.

A C K N O W L E D G E M E N T S

Ingalill Rahm Hallberg, Dirk von Dierendonck, Martyn Jones,

Deborah McElligott, Michelle Rowe and Joanne Cohen-Katz for

providing further information about their studies; to Albert West-

ergren for kindly facilitating contact with one of the authors. To

Wilmar Schaufeli for his expertise on burnout. To Joan Martí and

to Hugh McGuire for their support in identifying studies for this

review. To Elinor Thompson for technical editing. We thank the

Cochrane Occupational Health Field for help and enthusiastic in-

volvement in finalizing the review.

S O U R C E S O F S U P P O R T

External sources of support

• Ministry of Social Affairs and Health FINLAND

Internal sources of support

• Cochrane Occupational Health Field FINLAND

• Finnish Institute of Occupational Health FINLAND

• Corporació Parc Taulí (Sabadell) SPAIN

• Network of Centers for Research on Epidemiology and Public

Health SPAIN



R E F E R E N C E S

References to studies included in this review

Bittman 2003 {published data only}

Bittman B, Bruhn KT, Stevens C, Westengard J, Umbach PO. Recre-

ational music-making: a cost-effective group interdisciplinary strat-

egy for reducing burnout and improving mood states in long-term

care workers. Advances in Mind-Body Medicine 2003;19(3-4):4–15.

Cohen-Katz 2005 {unpublished data only}

Cohen-Katz J, Wiley SD, Capuano T, Baker DM, Kimmel S, Shapiro

S. The effects of mindfulness-based stress reduction on nurse stress

and burnout, Part II: A quantitative and qualitative study. Holistic

Nursing Practice 2005;19(1):26–35.

Delvaux 2004 {published data only}

Delvaux N, Razavi D, Marchal S, Bredart A, Farvacques C, Slach-

muylder JL. Effects of a 105 hours psychological training program on

attitudes, communication skills and occupational stress in oncology:

a randomised study. British Journal of Cancer 2004;90(1):106–14.

Ewers 2002 {published data only}

Ewers P, Bradshaw T, McGovern J, Ewers B. Does training in psy-

chosocial interventions reduce burnout rates in forensic nurses?. Jour-

nal of Advanced Nursing 2002;37(5):470–6.

Heaney 1995 {published data only}

Heaney CA, Price RH, Rafferty J. Increasing coping resources at

work: a field experiment to increase social support, improve work

team functioning, and enhance employee mental health. Journal of

Organizational Behavior 1995;16:335–52.

Jones 2000 {unpublished data only}

Jones MC, Johnston DW. Evaluating the impact of a worksite stress

management programme for distressed students: a randomised con-

trolled trial. Psychology and Health 2000;15:689–706.

Lee 1994 {published data only}

Lee S, Crockett MS. Effect of assertiveness training on levels of stress

and assertiveness experienced by nurses in Taiwan, Republic of China.

Issues in Mental Health Nursing 1994;15(4):419–32.

Lökk 1997 {published data only}

Lökk J, Arnetz B. Psychophysiological concomitants of organiza-

tional change in health care personnel: effects of a controlled inter-

vention study. Psychotherapy and Psychosomatics 1997;66(2):74–7.

Lökk 2000 {published data only}

Lökk J, Arnetz B. Impact of management Change and an Interven-

tion Program on Health Care Personnel. Psychotherapy and Psychoso-

matics 2000;69:79–85.

McElligott 2003 {unpublished data only}

McElligott D, Holz MB, Carollo L, Somerville S, Baggett M,

Kuzniewski S, et al. A pilot feasibility study of the effects of touch

therapy on nurses. Journal of the New York State Nurses Association

2003;34(1):16–24.

Melchior 1996 {published data only}

Melchior ME, Philipsen H, bu-Saad HH, Halfens RJ, van de Berg

AA, Gassman P. The effectiveness of primary nursing on burnout

among psychiatric nurses in long-stay settings. Journal of Advanced

Nursing 1996;24(4):694–702.

Norvell 1987 {published data only}

Norvell N, Belles D, Brody S, Freund A. Worksite Stress Management

for Medical Care Personnel: Results from a Pilot Program. Journal

for Specialists in Group Work 1987;57:118–26.

Proctor 1998 {published data only}

Proctor R, Stratton-Powell H, Tarrier N, Burns A. The impact of

training and support on stress among care staffin nursing and res-

idential homes for the elderly. Journal of Mental Health 1998;7(1):

59–71.

Razavi 1993

Razavi D, Delvaux N, Marchal S, Bredart A, Farvacques C, Paes-

mans M. The effects of a 24-h psychological training program on at-

titudes, communication skills and occupational stress in oncology: a

randomised study. European Journal of Cancer 1993;29A(13):1858–

63.

Rowe 1999 {published data only}

Rowe MM. Teaching health-care providers coping: results of a two-

year study. Journal of Behavioral Medicine 1999;22(5):511–27.

Schrijnemaekers 2003 {published data only}

Schrijnemaekers VJ, Van Rossum E, Candel MJ, Frederiks CM, De-

rix MM, Sielhorst H, et al. Effects of emotion-oriented care on work-

related outcomes of professional caregivers in homes for elderly per-

sons. Journals of Gerontology Series B-Psychological Sciences & Social

Sciences 2003;58(1):50–7.

Tsai 1993 {published data only}

Tsai SL, Crockett MS. Effects of relaxation training, combining im-

agery, and meditation on the stress level of Chinese nurses working

in modern hospitals in Taiwan. Issues in Mental Health Nursing 1993;

14(1):51–66.

von Baeyer 1983 {published data only}

von Baeyer C, Krause L. Effectiveness of stress management training

for nurses working in a burn treatment unit. International Journal of

Psychiatry in Medicine 1983;13(2):113–26.

West 1984 {published data only}

West DJ Jr, Horan JJ, Games PA. Component Analysis of Occupa-

tional Stress Inoculation Apllied to Registered Nurses in an Acute

Care Hospital Setting. Journal of Counseling Psychology 1984;31(2):

209–18.

Yung 2004 {published data only}

Yung PM, Fung MY, Chan TM, Lau BW. Relaxation training meth-

ods for nurse managers in Hong Kong: a controlled study. Interna-

tional Journal of Mental Health Nursing 2004;13(4):255–61.

References to studies excluded from this reviewBeddoe 2004

Beddoe AE, Murphy SO. Does mindfulness decrease stress and foster

empathy among nursing students?. The Journal of Nursing Education

2004;43(7):305–12.

Berg 1994

Berg A, Hansson UW, Hallberg IR. Nurses creativity, tedium and

burnout during 1 year of clinical supervision and implementation

of individually planned nursing care: comparisons between a ward

for severely demented patients and a similar control ward. Journal of

Advanced Nursing 1994;20(4):742–9.



Bradley 2002

Bradley JR, Cartwright S. Social support, job stress, health, and job

satisfaction among nurses in the United Kingdom. International Jour-

nal of Stress Management 2002;9(3):163–81.

Bruneau 2004

Bruneau BM, Ellison GT. Palliative care stress in a UK community

hospital: evaluation of a stress-reduction programme. International

Journal of Palliative Nursing 2004;10(6):296–304.

Bunce 1996

Bunce D, West MA. Stress management and innovation interven-

tions at work. Human Relations 1996;49(2):209–32.

Calomeni 1999

Calomeni CA, Solberg LI, Conn SA. Nurses on quality improvement

teams: how do they benefit?. Journal of Nursing Care Quality 1999;

13(5):75–90.

Drain 1991

Drain CB, Dirks KR, Jenkins OC, Campbell JK, Finstuen K, John-

son GR. Comparison of two instructional methods on the levels of

physiologic and psychologic stress as measured by blood cortisol and

the State-Trait Anxiety Inventory. Nurse Anesthesia 1991;2(4):172–

83.

Fernandez 1995

Fernandez San Martin MI, Villagrasa F Jr, Fe GM, Vazquez GJ, Cruz

CE, guirre Trigo MV, et al. The study of occupational satisfaction

and its determining factors among health professionals working in

one of the areas in Madrid. Revista Espanola de Salud Publica 1995;

69(6):487–97.

Freedy 1994

Freedy JR, Hobfoll SE. Stress inoculation for reduction of burnout: a

conservation of resources approach. Anxiety, Stress, and Coping 1994;

6:311–25.

Godbey 1994

Godbey KL, Courage MM. Stress-management program: interven-

tion in nursing student performance anxiety. Archives of Psychiatric

Nursing 1994;8(3):190–9.

Heaman 1995

Heaman D. The quieting response (QR): a modality for reduction

of psychophysiologic stress in nursing students. Journal of Nursing

Education 1995;34(1):5–10.

Hughes 2003

Hughes LC, Romick P, Sandor MK, Phillips CA, Glaister J, Levy K, et

al. Evaluation of an informal peer group experience on baccalaureate

nursing students’ emotional well-being and professional socialization.

Journal of Professional Nursing 2003;19(1):38–48.

Johansson 1991

Johansson N. Effectiveness of a stress management program in reduc-

ing anxiety and depression in nursing students. Journal of American

College Health 1991;40(3):125–9.

Jones 1998

Jones JW, Barge BN, Steffy BD, Fay LM, Kunz LK, Wuebker LJ.

Stress and medical malpractice: organizational risc assessment and

intervention. Journal of Applied Psychology 1998;7:727–35.

Keyes 1988

Keyes JB, Dean SF. Stress inoculation training for direct contact

staff working with mentally retarded persons. Behavioral Residential

Treatment 1988;3:315–23.

Kushnir 1998

Kushnir T, Malkinson R, Ribak J. Rational thinking and stress man-

agement in health workers: a psychoeducational program. Interna-

tional Journal of Stress Management 1998;5(3):169–78.

Lloyd 2002

Lloyd C, King R, Chenoweth L. Social work, stress and burnout: A

review. Journal of Mental Health 2002;11(3):255–65.

Mancini 1983

Mancini J, Lavecchia C, Clegg R. Graduate nursing students and

stress. Journal of Nursing Education 1983;22(8):329–34.

McCue 1991

McCue JD, Sachs CL. A stress management workshop improves res-

ident’s coping skills. Archives of Internal Medicine 1991;151:2273–7.

McPhail 1990

McPhail A, Pikula H, Roberts J, Browne G, Harper D. Primary

nursing. A randomized crossover trial. Western Journal of Nursing

Research 1990;12(2):188–97.

Murphy 1983

Murphy LR. A comparision of relaxation methods for reducing stress

in nursing personnel. Human Factors 1983;25:413–40.

Nissen 1997

Nissen JM, Boumans NP, Landeweerd JA. Primary nursing and qual-

ity of care: a Dutch study. International Journal of Nursing Studies

1997;34(2):93–102.

Pålsson 1996

Palsson MB, Hallberg IR, Norberg A, Björvell H. Burnout, empathy

and sense of coherence among Swedish district nurses before and

after systematic clinical supervision. Scandinavian Journal of Caring

Sciences 1996;10(1):19–26.

Schinke 1977

Schinke SP, Wong SE. Evaluation of staff training in group homes

for retarded persons. American Journal of Mental Deficiency 1977;82

(2):130–6.

Sharif 2004

Sharif F, Armitage P. The effect of psychological and educational

counselling in reducing anxiety in nursing students. Journal of Psy-

chiatric and Mental Health Nursing 2004;11(4):386–92.

Taylor 2001

Taylor B, Lo R. The effects of Healing Touch on the coping ability,

self esteem and general health of undergraduate nursing students.

Complementary Therapies in Nursing & Midwifery 2001;7(1):34–42.

Teasdale 2001

Teasdale K, Brocklehurst N, Thom N. Clinical supervision and sup-

port for nurses: an evaluation study. Nursing and Health Care Man-

agement 2001;33(2):216–24.

Toivanen 1993

Toivanen H, Helin P, Hänninen O. Impact of regular relaxation

training and psychosocial working factors on neck-shoulder tension

and absenteeism in hospital cleaners. Journal of Occupational Medicine

1993;35(11):1123–30.



van Dierendonck 1998

van Dierendonck D, Schaufeli W, Buunk BP. The evaluation of an

individual burnout intervention program: the role of inequity and

social support. Journal of Applied Psychology 1998;83(3):392–407.

Weir 1997

Weir R, Stewart L, Browne G, Roberts J, Gafni A, Easton S, Seymour

L. The efficacy and effectiveness of process consultation in improving

staff morale and absenteeism. Medical Care 1997;35(4):334–53.

Young 2001

Young LE, Bruce A, Turner L, Linden W. Evaluation of mindfulness-

based stress reduction intervention. The Canadian Nurse 2001;97(6):

23–6.

Zimber 2001

Zimber A, Rudolf A, Teufel S. A training program to reduce distress

among geriatric caregivers [Arbeitsbelastungen in der Altenpflege re-

duzieren: Ein Trainingsprogramm für Mitarbeiter und Leitugskräfte].

Zeitschrift fur Gerontologie und Geriatrie 2001;34(5):401–7.

References to studies awaiting assessment

Carson 1999

Carson J, Cavagin J, Bunclark J. Effective communication in men-

tal health nurses: did social support save the psychiatric nurse?. NT

Research 1999;4:31–42.

Additional references

Beck 1961

Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory

for measuring depression. Archives of General Psychiatry 1961;4:561–

71.

Beck 1991

Beck DL, Srivastava R. Perceived level and sources of stress in bac-

calaureate nursing students. Journal of Nursing Education 1991;30

(3):127–33.

Benoliel 1990

Benoliel JQ, McCorkle R, Georgiadou F, Denton T, Spitzer A. Mea-

surement of stress in clinical nursing. Cancer Nursing 1990;13(4):

221–28.

Campbell 2001

Campbell MK, Mollison J, Grimshaw JM. Cluster trials in imple-

mentation research: estimation of intra-cluster correlation and sam-

ple size. Statistics in Medicine 2001;20:391–9.

Cohen 1982

Cohen S, Hoberman HM. Positive events and social supports as

buffers of life change stress. Journal of Applied Social Psychology 1982;

13:99–125.

Cohen 1983

Cohen S, Kamarck T, Mermelstein R. A global measure of perceived

stress. Journal of Health and Social Behavior 1983;24(4):385–96.

Cooper 1988

Cooper CL, Sloane SJ, Williams S. The Occupational Stress Indicator.

Windsor, Berkshire: NFER-Nelson, 1988.

Cooper 2001

Cooper CL, Dewe P, O’Driscoll MP. Organizational stress. London:

SAGE, 2001.

de Vet 2003

de Vet HCW, van Tulder MW, Bouter LM. Levels of evidence: Intel-

lectual aid or absolute judgement?. Journal of Clinical Epidemiology

2003;56:917–8.

Deckard 1994

Deckard G, Meterko M, Field D. Physician burnout: an examination

of personal, professional, and organizational relationships. Medical

Care 1994;32(7):745–54.

Deeks 2003

Deeks JJ, Dinnes J, D’Amico R, Sowden AJ, Sakarovitch C, Song F,

et al. Evaluating non-randomised intervention studies. Health Tech-

nology Assessment 2003;7(27):1–173.

DeFrank 1987

DeFrank RS, Cooper CL. Worksite stress management interventions:

their effectiveness and conceptualisation. Journal of Managerial Psy-

chology 1987;2:4–10.

Derogatis 1976

Derogatis L, Rikels K, Rock A. The SCL-90 and MMPI: A step in

the validation of a new self-report scale. British Journal of Psychiatry

1976;128:280–9.

Derogatis 1983

Derogatis LR, Melisaratos N. The Brief Symptom Inventory: an in-

troductory report. Psycholical Medicine 1983;13(3):595–605.

Downs 1998

Downs SH, Black N. The feasibility of creating a checklist for the

assessment of the methodological quality both of randomised and

non-randomised studies of health care interventions. Journal of Epi-

demiology and Community Health 1998;52(6):377–84.

Dunnette 1990

Dunnette MD, Hough LM. Handbook of industrial and organiza-

tional psychology. 2nd Edition. Palo Alto, CA: Consulting Psycholo-

gists Press, 1990.

EPOC 1998

Cochrane Effective Practice and Organisation of Care Group. In-

cluding Interrupted Time Series (ITS) Designs in a EPOC Review.

http://www.epoc.uottawa.ca/methods.htm 1998:1–5.

Firth 1989

Firth H, Britton P. Burnout, absence and turnover amongst British

nursing staff. Journal of Occupational Psychology 1989;62(1):55–9.

Firth-Cozens 1999

Firth-Cozens J, Payne R. Stress in health professionals. Chichester:

Wiley, 1999.

Gilbody 2006

Gilbody S, Cahill J, Barkham M, Richards D, Bee P, Glanville J.

Can we improve the morale of staff working in psychiatric units? A

systematic review. Journal of Mental Health 2006;15(1):7–17.

Goldberg 1991

Goldberg D, Williams P. A User’s Guide to the General Health Ques-

tionnaire. London: NFER–Nelson, 1991.

Gray-Toft 1981

Gray-Toft P, Anderson JG. The Nursing Stress Scale: Development

of an instrument. Journal of Psychopathology and Behavioral Assessment

1981;3(1):11–23.



Jacobson 1996

Jacobson BH, Aldana SG, Goetzel RZ, Vardell KD, Adams TB,

Piedras RJ. The relationship between perceived stress and self-re-

ported illness-related absenteeism. American Journal of Health Pro-

motion 1996;11(1):54–61.

Kahn 1964

Kahn R, Wolfe D, Quinn R, Snoek JD, Rosenthal R. Organizational

Stress: Studies in role conflict and role ambiguity. New York, NY: Wiley,

1964.

Karasek 1992

Karasek R. Stress prevention through work reorganization: a sum-

mary of 19 international case studies. Conditions of Work Digest 1992;

11:23–42.

Karasek 1998

Karasek R, Brisson C, Kawakami N, Houtman I, Bongers P, Amick

B. The Job Content Questionnaire (JCQ): an instrument for inter-

nationally comparative assessments of psychosocial job characteris-

tics. Journal of Occupational Health Psychology 1998;3(4):322–55.

Maslach 1982

Maslach C. Burnout: the cost of caring. New York, NY: Prentice-Hall,

1982.

Maslach 1996

Maslach C, Jackson SE, Leiter SP. Maslach Burnout Inventory manual.

3rd Edition. Palo Alto, CA: Consulting Psychologists Press, 1996.

McNeely 2005

McNeely E. The consequences of job stress for nurses’ health: Time

for a check-up. NursING Outlook 2005;53:291–9.

Michie 2003

Michie S, Williams S. Reducing work related psychological ill health

and sickness absence: a systematic literature review. Occupational and

Environmental Medicine 2003;60(1):3–9.

Mimura 2003

Mimura C, Griffiths P. The effectiveness of current approaches to

workplace stress management in the nursing profession: an evidence

based literature review. Occupational and Environmental Medicine

2003;60(1):10–5.

Murphy 1995

Murphy S. Stress management in work settings: a critical review of

the health effects. American Journal of Health Promotion 1995;11(2):

112–35.

Payne 1987

Payne RL, Firth-Cozens J. Stress in health professionals. Chichester:

Wiley, 1987.

Raiger 2005

Raiger J. Applying a Cultural Lens to the Concept of Burnout. Journal

of Transcultural Nursing 2005;16(1):71–6.

Ramirez 1996

Ramirez AJ, Graham J, Richards MA, Cull A, Gregory WM. Mental

health of hospital consultants: the effects of stress and satisfaction at

work. Lancet 1996;347(9003):724–8.

Schaufeli 2003

Schaufeli WB, Buunk BP. Burnout: An Overview of 25 Years of Re-

search and Theorizing. In: SchabracqMJ, WinnubstJAM, CooperCL

editor(s). The handbook of Work and Health Psychology. 2nd Edition.

Chichester: Wiley, 2003:383–424.

Spielberger 1970

Spielberger CD, Gorsuch RL, Lushene RE. Manual for the State-

Trait Anxiety Inventory (Self-Evaluation Questionnaire). Palo Alto, CA:

Mind Garden, 1970.

van der Hek 1997

van der Hek H, Plomp HN. Occupational stress management pro-

grammes: a practical overview of published effect studies. Occupa-

tional Medicine [London] 1997;47(3):133–41.

van der Klink 2001

van der Klink JJ, Blonk RW, Schene AH, van Dijk FJ. The benefits

of interventions for work-related stress. American Journal of Public

Health 2001;91(2):270–6.

van Tulder 2003

van Tulder M, Furlan A, Bombardier C, Bouter L. Updated method

guidelines for systematic reviews in the cochrane collaboration back

review group. Spine 2003;28(12):1290–9.

Weinberg 2000

Weinberg A, Creed F. Stress and psychiatric disorder in healthcare

professionals and hospital staff. Lancet 2000;355(9203):533–7.

Williamson 1994

Williamson AM. Managing Stress in the Workplace .2. the Scien-

tific Basis (Knowledge-Base) for the Guide. International Journal of

Industrial Ergonomics 1994;14(1-2):171–96.

T A B L E S

Characteristics of included studies

Study Bittman 2003

Methods Cross-over study

Participants 112 randomly selected staff of a retirement community in the USA

Interventions 1) Recreational music making

2) No intervention -control



Characteristics of included studies (Continued )

Outcomes Maslach Burnout Inventory, Profile of Mood States

Notes Person-directed

Allocation concealment D – Not used

Study Cohen-Katz 2005

Methods RCT

Participants 25 nurses, pastoral care, respiratory therapy and social work personnel in the USA

Interventions 1) Mindfulness-based stress reduction programme


Outcomes Maslach Burnout Inventory, Brief Symptom Inventory



Study Delvaux 2004

Methods RCT

Participants 115 oncology nurses in Belgium

Interventions 1) Psychological Training Program


Outcomes Nursing Stress Scale

Notes Organisational


Study Ewers 2002

Methods RCT

Participants 20 forensic mental health nurses in the UK

Interventions 1) Psychosocial Intervention Training


Outcomes Maslach Burnout Inventory



Study Heaney 1995

Methods RCT

Participants 1375 direct care staff and home managers in the USA

Interventions 1) Caregiver support program


Outcomes Social support, Organisational climate, SCL-90-R, Confidence in coping ability



Study Jones 2000

Methods RCT

Participants 79 student nurses reporting significant levels of affective distress at screen in the UK

Interventions 1) multi-modal stress management





Outcomes Derogatis Stress profile, B&S Stress Inventory, GHQ, STAI, BDI, ways of coping, absenteeism


Allocation concealment A – Adequate

Study Lee 1994

Methods RCT

Participants 60 hospital nurses suffering from either: insomnia, headache or gastrointestinal discomfort in Taiwan

Interventions 1) assertiveness training

2) traditional inservice program about computer applications in nursing

Outcomes Perceived Stress Scale, Rathus Assertiveness Schedule



Study Lökk 1997

Methods Cluster-randomised trial

Participants 26 health care personnel in a geriatric hospital in Sweden

Interventions 1) support and advice from a psychologist

2) passive attendance by psychologist

Outcomes Blood pressure and pulse rate, prolactin, cortisol, DHEA, estradiol



Study Lökk 2000


Participants 26 health care personnel in a geriatric hospital in Sweden

Interventions 1) support and advice from a psychologist

2) passive attendance by psychologist

Outcomes A structured made-to-measure stress questionnaire



Study McElligott 2003

Methods RCT

Participants 20 nurses working at a tertiary care center in the USA

Interventions 1) AMMA therapy

2) a standardised touch therapy protocol

Outcomes VAS of anxiety, blood pressure, heart rate, pulse oximetry, skin temperature



Study Melchior 1996

Methods RCT




Participants 161 psychiatric nurses in long-stay settings in the Netherlands

Interventions 1) support and advice given by nurse managers or quality care coordinators





Study Norvell 1987

Methods RCT

Participants 12 respiratory therapists in the USA

Interventions 1) stress management program


Outcomes MBI, C-H Inventory of Phys Symptoms, The Hassles Scale, The Uplifts Scale



Study Proctor 1998


Participants 98 care staff in residential nursing homes in the UK

Interventions 1) developing knowledge and skills and individual programme planning


Outcomes The Occupational Stress Indicator, GHQ



Study Razavi 1993

Methods RCT

Participants 72 oncology nurses in Belgium and France

Interventions 1) a 24-h Psychological Training Program


Outcomes The Nursing Stress Scale



Study Rowe 1999

Methods RCT

Participants 126 health-care professionals from the USA

Interventions 1) Stress management/ adaptive coping training, with refresher sessions








Study Schrijnemaekers 2003


Participants 300 professional caregivers in homes for elderly persons in the Netherlands

Interventions 1) Emotion-oriented care training, clinical lessons and supervision meetings


Outcomes Maslach Burnout Inventory, Job satisfaction



Study Tsai 1993

Methods RCT

Participants 137 nurses in Taiwan

Interventions 1) Training about stress at work, relaxation, breathing, imagery and meditation

2) traditional in-service education about theory analysis

Outcomes Nurse Stress Checklist, Chinese General Health Questionnaire



Study West 1984

Methods RCT

Participants 60 acute care hospital nurses in the USA

Interventions 1) Stress Inoculation (SI) training


Outcomes MBI, Job-related tension, Life satisfaction STAI, RAS, Systolic and diastolic blood pressure



Study Yung 2004

Methods RCT

Participants 65 nurse managers in China

Interventions 1) Cognitive relaxation

2) Stretch-release relaxation


Outcomes C-STAI, C-GHQ



Study von Baeyer 1983

Methods Cross-over study

Participants 14 nurses in a burn treatment unit in Canada

Interventions 1) Cognitive-behavioural stress management training


Outcomes STAI





Characteristics of excluded studies

Study Reason for exclusion

Beddoe 2004 The study does not employ a control group.

Berg 1994 The study is non-randomised by design.

Bradley 2002 The study does not employ an intervention. It is an aetiological study.

Bruneau 2004 The study is non-randomised by design.

Bunce 1996 The study is non-randomised by design.

Calomeni 1999 The study does not have a stress outcome.

Drain 1991 The participants of the study are not health care personnel or nursing students with clinical duties.

Fernandez 1995 The study does not employ an intervention.

Freedy 1994 The study is non-randomised by design.

Godbey 1994 The participants are nursing students but do not have clinical duties.

Heaman 1995 The participants are nursing students but do not have clinical duties.

Hughes 2003 The participants are nursing students but do not have clinical duties.

Johansson 1991 The participants are nursing students but do not have clinical duties.

Jones 1998 The study is non-randomised by design.

Keyes 1988 The study is non-randomised by design.

Kushnir 1998 The study is non-randomised by design.

Lloyd 2002 The participants of the study are not health care personnel or nursing students with clinical duties.

Mancini 1983 The participants are nursing students but do not have clinical duties.

McCue 1991 The study is non-randomised by design.

McPhail 1990 The study does not have a stress outcome.

Murphy 1983 The study is non-randomised by design.

Nissen 1997 The study does not have a stress outcome.

Pålsson 1996 The study is non-randomised by design.

Schinke 1977 The study does not intend to prevent or treat stress or burnout.

Sharif 2004 The participants are nursing students but do not have clinical duties.

Taylor 2001 The study does not have a stress outcome.

Teasdale 2001 The study does not employ an intervention.

Toivanen 1993 The participants of the study are not health care personnel or nursing students with clinical duties.

Weir 1997 The study does not have a stress outcome.

Young 2001 The study is non-randomised by design.

Zimber 2001 The study is non-randomised by design.

van Dierendonck 1998 The study is non-randomised by design.

A D D I T I O N A L T A B L E S

Table 01. Search strategy (MEDLINE)

Search strategy



1 exp Burnout, professional

2 burnout.tw.

3 exp job satisfaction

4 job satisfaction.tw.

5 (mental adj25 health adj25 professional$) .ab,ti.

6 (stress adj25 professional$) .ab,ti.

7 (stress adj25 occupational) .ab,ti.

8 randomised controlled trial.pt.

9 controlled clinical trial.pt.

10 randomised controlled trials

11 random allocation

12 double blind method

13 single blind method

14 or/8-13

15 clinical trial.pt.

16 exp Clinical trial

17 (clin$ adj25 trial$) .ti,ab.

18 ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind $ or 12010 mask$)) .ti,ab.

19 19 placebos

20 placebos.ti,ab.

21 random.ti,ab.

22 research design

23 or/15-22

24 comparative study

25 exp “Evaluation and Follow Up”

26 follow up studies

27 prospective studies

28 (control$ or prospectiv$ or volunteer$) .ti,ab.

29 or/24-28

30 14 or 23 or 29

31 animal

32 human

33 31 and 32

34 31 not 33

35 or/1-7

36 exp anxiety/ or exp anxiety disorders

37 anxiety.ti.

38 exp Depression

39 depression.ti.

40 or/35-39

41 exp Health personnel

42 health.mp. and personnel.ti,ab. [mp=title, abstract, registry]

43 Nursing

44 nurse.mp. or nursing.ti,ab. [mp=title, abstract, registry]



Table 01. Search strategy (MEDLINE) (Continued )

Search strategy

45 limit 44 to human

46 or/41-45

47 40 and 46

48 34


50 30


52 47 and 51

53 52 and 8

54 41 or 42

55 54 and 51 and 40

56 health.mp. and personnel.tw. [mp=title, abstract, registry]

57 54 and 56

58 54 or 56

59 58 and 51 and 40

60 (health and related and occupation$) .ab,tw,ti.

61 cleaners.mp. [mp=title, abstract, registry]

62 exp Occupational diseases

63 58 and 62

64 58 or 62

65 from 64 keep 8

Table 02. Study quality and intervention type

Study ID Intervention type Intervention content Reporting External validity Internal validity Power

Bittman Person-directed 8 2 8 4

Cohen-Katz Person-directed 3 0 8 1

Delvaux Organisational communication skills 8 1 10 4

Ewers Person-directed 8 1 9 0

Heaney Organisational participatory problem solving 5 0 6 5

Jones Person-directed 7 0 8 3

Lee Person-directed 8 0 8 3

Lökk Person-directed 7 1 8 1

McElligott Person-directed 5 0 6 0

Melchior Organisational primary nursing 7 1 7 4

Norvell Person-directed 3 0 6 0

Proctor Organisational coping with difficult residents 6 0 7 3

Razavi Organisational communication skills 7 0 8 3

Rowe Person-directed 8 0 9 3



Table 02. Study quality and intervention type (Continued )

Study ID Intervention type Intervention content Reporting External validity Internal validity Power

Schrijnemaekers Organisational emotion-oriented care 7 1 10 5

Tsai Person-directed 5 1 6 4

von-Baeyer Person-directed 8 1 9 0

West Person-directed 6 0 8 1

Yung Person-directed 10 0 9 2

Table 03. Levels of evidence: legend

HQ+ LQ+ HQ- LQ- HQinc LQinc

High quality

positive effect

(strong evidence)

Low quality,

positive effect

(limited evidence)

High quality,

negative effect

(strong evidence)

Low quality,

negative effect

(limited evidence)

High quality,

inconsistent effect

Low quality,

inconsistent effect

Table 04. Levels of evidence: Person-directed interventions: Stress

Stress scale HQ+ LQ+ HQ- LQ- HQinc LQinc

Beck and Srivastava Stress Inventory Jones

Perceived Stress Scale Lee

Stress Assessment Inventory Rowe 1999

Nursing Stress Checklist Tsai

Job-Related Tension Index West

Table 05. Levels of evidence: Person-directed interventions: Maslach Burnout Inventory

MBI subscales HQ+ LQ+ HQ- LQ- HQinc LQinc

Emotional Exhaustion Cohen-Katz

Ewers

Norvell

Rowe

West

Bittman

Depersonalisation Ewers

Rowe

Bittman

Cohen-Katz

Norvell

West

Personal Accomplishment Cohen-Katz

Ewers

Rowe

West

Bittman

Norvell



Table 06. Levels of evidence: Person-directed interventions: State Trait Anxiety Inventory

STAI facets HQ+ LQ+ HQ- LQ- HQinc LQinc

State Jones

von Baeyer

West

Yung

Trait Jones

von Baeyer

West

Yung

Table 07. Levels of evidence: Person-directed interventions: general symptoms

Scale HQ+ LQ+ HQ- LQ- HQinc LQinc

Brief Symptom Inventory Cohen-Katz

GHQ General Health Questionnaire Jones

C-H Inventory of Physical Symptoms Norvell

Chinese GHQ Tsai

Chinese GHQ Yung

Table 08. Levels of evidence: Work-directed interventions: Stress

Stress scale HQ+ LQ+ HQ- LQ- HQinc LQinc

Nursing Stress Scale Delvaux

Occupational Stress Indicator Proctor

Table 09. Levels of evidence: Work-directed interventions: Maslach Burnout Inventory

MBI subscales HQ+ LQ+ HQ- LQ- HQinc LQinc

Emotional Exhaustion Schrijnemaekers Melchior

Depersonalisation Schrijnemaekers Melchior

Personal Accomplishment Schrijnemaekers Melchior

Table 10. Levels of evidence: Work-directed interventions: general symptoms

Scale HQ+ LQ+ HQ- LQ- HQinc LQinc

Symptom Checklist-90-Revised Heaney

GHQ General Health Questionnaire (somatic symptoms) Proctor



A N A L Y S E S

Comparison 01. Person-directed intervention vs no intervention (Short Term)

Outcome titleNo. of

studies

No. of

participants Statistical method Effect size

01 Beck & Srivastava Stress Scale

and Perceived Stress Scale

2 130 Standardised Mean Difference (Fixed) 95% CI -0.85 [-1.21, -0.49]

02 Maslach Burnout Inventory Weighted Mean Difference (Random) 95% CI Subtotals only

03 State Trait Anxiety Inventory Weighted Mean Difference (Random) 95% CI Subtotals only

04 General Health Questionnaire 2 127 Weighted Mean Difference (Random) 95% CI -11.87 [-27.24,

3.49]

Comparison 02. Person-directed intervention vs no intervention (Medium and Long Term)

Outcome titleNo. of

studies

No. of


01 Maslach Burnout Inventory (at

2 years)

Weighted Mean Difference (Random) 95% CI Subtotals only

02 Stait Trait Anxiety Inventory

(Jones at 3 months and Yung at

1 month)

Weighted Mean Difference (Random) 95% CI Subtotals only

03 Perceived Stress Scale (at 1

month)

1 57 Weighted Mean Difference (Fixed) 95% CI -6.10 [-8.44, -3.76]

04 General Health Questionnaire

(at 1 month)

1 48 Weighted Mean Difference (Fixed) 95% CI -7.10 [-10.58, -3.62]

Comparison 03. Work-directed intervention vs no intervention (Short Term)

Outcome titleNo. of

studies

No. of


01 Maslach Burnout Inventory Weighted Mean Difference (Random) 95% CI Subtotals only

02 Nursing Stress Scale 1 115 Weighted Mean Difference (Fixed) 95% CI -0.34 [-0.62, -0.06]

03 General symptoms 1 57 Weighted Mean Difference (Fixed) 95% CI -2.90 [-5.16, -0.64]

Comparison 04. Work-directed intervention vs no intervention (Medium Term)

Outcome titleNo. of

studies

No. of


01 Nursing Stress Scale (at 6

months)

1 111 Weighted Mean Difference (Fixed) 95% CI -0.19 [-0.49, 0.11]

C O V E R S H E E T

Title Preventing occupational stress in healthcare workers

Authors Marine A, Ruotsalainen J, Serra C, Verbeek J

Contribution of author(s) Albert Marine and Consol Serra conceived the review, gathered the data and wrote the first

draft.

Jani Ruotsalainen and Jos Verbeek updated the searches two years later. Jani Ruotsalainen

and Consol Serra re-analysed the new material. Jani Ruotsalainen rewrote the second draft.

Jos Verbeek was responsible for the statistical calculations.

All authors commented on the final draft.



Issue protocol first published 2000/2

Review first published 2006/4

Date of most recent amendment 23 August 2006

Date of most recent

SUBSTANTIVE amendment

21 August 2006

What’s New Information not supplied by author

Date new studies sought but

none found

Information not supplied by author

Date new studies found but not

yet included/excluded

14 August 2006

Date new studies found and

included/excluded

01 May 2005

Date authors’ conclusions

section amended

Information not supplied by author

Contact address Dr Albert Marine

Head of Counselling Programme for Health Professionals

Prevention Service

Corporacio Sanitaria Parc Tauli de Sabadell

Parc Tauli s/n

Sabadell

Catalonia

08208

SPAIN

E-mail: [email protected]

Tel: 34-93-7231010

Fax: 34-93-7233804

DOI 10.1002/14651858.CD002892.pub2

Cochrane Library number CD002892

Editorial group Cochrane Depression, Anxiety and Neurosis Group

Editorial group code HM-DEPRESSN



G R A P H S A N D O T H E R T A B L E S

Analysis 01.01. Comparison 01 Person-directed intervention vs no intervention (Short Term), Outcome 01

Beck & Srivastava Stress Scale and Perceived Stress Scale

Review: Preventing occupational stress in healthcare workers

Comparison: 01 Person-directed intervention vs no intervention (Short Term)

Outcome: 01 Beck % Srivastava Stress Scale and Perceived Stress Scale

Study Intervention Control Standardised Mean Difference (Fixed) Weight Standardised Mean Difference (Fixed)

N Mean(SD) N Mean(SD) 95% CI (%) 95% CI

Jones 2000 39 56.33 (12.74) 34 69.43 (16.40) 55.7 -0.89 [ -1.37, -0.41 ]

Lee 1994 29 22.80 (4.90) 28 26.80 (4.90) 44.3 -0.81 [ -1.35, -0.26 ]

Total (95% CI) 68 62 100.0 -0.85 [ -1.21, -0.49 ]

Test for heterogeneity chi-square=0.05 df=1 p=0.82 I² =0.0%

Test for overall effect z=4.64 p<0.00001

-10.0 -5.0 0 5.0 10.0

Favours intervention Favours control


Maslach Burnout Inventory



Outcome: 02 Maslach Burnout Inventory

Study Intervention Control Weighted Mean Difference (Random) Weight Weighted Mean Difference (Random)


01 Emotional Exhaustion

Cohen-Katz 2005 12 15.00 (10.21) 13 23.31 (9.88) 21.5 -8.31 [ -16.20, -0.42 ]

Ewers 2002 10 10.51 (3.14) 10 18.91 (3.14) 38.1 -8.40 [ -11.15, -5.65 ]

Rowe 1999 42 23.53 (4.09) 42 25.60 (4.60) 40.4 -2.07 [ -3.93, -0.21 ]

Subtotal (95% CI) 64 65 100.0 -5.82 [ -11.02, -0.63 ]


Test for overall effect z=2.20 p=0.03

02 Depersonalization

Cohen-Katz 2005 12 1.58 (1.78) 13 5.54 (6.78) 24.2 -3.96 [ -7.78, -0.14 ]

Ewers 2002 10 2.04 (3.26) 10 5.96 (3.26) 31.4 -3.92 [ -6.78, -1.06 ]

Rowe 1999 42 6.73 (2.84) 42 7.30 (3.41) 44.4 -0.57 [ -1.91, 0.77 ]

Subtotal (95% CI) 64 65 100.0 -2.44 [ -5.06, 0.17 ]



-10.0 -5.0 0 5.0 10.0

Favours intervention Favours control (Continued . . . )



(. . . Continued)



03 Personal Accomplishment (lack of)

Cohen-Katz 2005 12 -42.33 (3.96) 13 -36.92 (5.92) 28.0 -5.41 [ -9.33, -1.49 ]

Ewers 2002 10 -39.64 (2.45) 10 -32.21 (2.45) 35.3 -7.43 [ -9.58, -5.28 ]

Rowe 1999 42 35.33 (3.86) 42 37.38 (4.28) 36.7 -2.05 [ -3.79, -0.31 ]

Subtotal (95% CI) 64 65 100.0 -4.89 [ -8.71, -1.07 ]



-10.0 -5.0 0 5.0 10.0



State Trait Anxiety Inventory



Outcome: 03 State Trait Anxiety Inventory



01 State anxiety

Jones 2000 39 30.77 (7.46) 34 46.35 (12.69) 35.0 -15.58 [ -20.45, -10.71 ]

Yung 2004 18 36.89 (5.75) 30 41.48 (8.16) 36.9 -4.59 [ -8.54, -0.64 ]

von Baeyer 1983 7 27.00 (6.50) 7 35.10 (8.50) 28.1 -8.10 [ -16.03, -0.17 ]

Subtotal (95% CI) 64 71 100.0 -9.42 [ -16.92, -1.93 ]



02 Trait anxiety

Jones 2000 39 37.29 (7.90) 34 45.68 (9.81) 38.3 -8.39 [ -12.52, -4.26 ]

Yung 2004 18 42.06 (6.26) 30 43.90 (7.93) 38.6 -1.84 [ -5.89, 2.21 ]

von Baeyer 1983 7 24.10 (4.80) 7 37.00 (10.80) 23.1 -12.90 [ -21.66, -4.14 ]

Subtotal (95% CI) 64 71 100.0 -6.91 [ -12.80, -1.01 ]



-10.0 -5.0 0 5.0 10.0





General Health Questionnaire



Outcome: 04 General Health Questionnaire



Jones 2000 40 15.00 (8.54) 39 34.73 (14.00) 49.9 -19.73 [ -24.86, -14.60 ]

Yung 2004 18 24.78 (7.03) 30 28.83 (10.35) 50.1 -4.05 [ -8.98, 0.88 ]

Total (95% CI) 58 69 100.0 -11.87 [ -27.24, 3.49 ]

Test for heterogeneity chi-square=18.68 df=1 p=<0.0001 I² =94.6%


-10.0 -5.0 0 5.0 10.0


Analysis 02.01. Comparison 02 Person-directed intervention vs no intervention (Medium and Long Term),

Outcome 01 Maslach Burnout Inventory (at 2 years)


Comparison: 02 Person-directed intervention vs no intervention (Medium and Long Term)

Outcome: 01 Maslach Burnout Inventory (at 2 years)




Rowe 1999 42 19.38 (4.67) 42 25.38 (5.41) 100.0 -6.00 [ -8.16, -3.84 ]

Subtotal (95% CI) 42 42 100.0 -6.00 [ -8.16, -3.84 ]

Test for heterogeneity: not applicable


02 Depersonalisation

Rowe 1999 42 6.87 (2.47) 42 7.49 (3.39) 100.0 -0.62 [ -1.89, 0.65 ]

Subtotal (95% CI) 42 42 100.0 -0.62 [ -1.89, 0.65 ]




Rowe 1999 42 31.21 (4.61) 42 37.03 (5.04) 100.0 -5.82 [ -7.89, -3.75 ]

Subtotal (95% CI) 42 42 100.0 -5.82 [ -7.89, -3.75 ]



-10.0 -5.0 0 5.0 10.0





Outcome 02 Stait Trait Anxiety Inventory (Jones at 3 months and Yung at 1 month)



Outcome: 02 Stait Trait Anxiety Inventory (Jones at 3 months and Yung at 1 month)



01 State anxiety

Jones 2000 39 31.62 (8.48) 34 39.44 (11.34) 46.7 -7.82 [ -12.47, -3.17 ]

Yung 2004 18 32.07 (7.12) 30 40.81 (7.95) 53.3 -8.74 [ -13.09, -4.39 ]

Subtotal (95% CI) 57 64 100.0 -8.31 [ -11.49, -5.13 ]



02 Trait anxiety

Jones 2000 39 36.45 (9.03) 34 42.45 (9.62) 46.7 -6.00 [ -10.30, -1.70 ]

Yung 2004 18 38.07 (6.89) 30 40.48 (6.33) 53.3 -2.41 [ -6.32, 1.50 ]

Subtotal (95% CI) 57 64 100.0 -4.09 [ -7.60, -0.58 ]



-10.0 -5.0 0 5.0 10.0



Outcome 03 Perceived Stress Scale (at 1 month)



Outcome: 03 Perceived Stress Scale (at 1 month)

Study Intervention Control Weighted Mean Difference (Fixed) Weight Weighted Mean Difference (Fixed)


Lee 1994 29 20.90 (4.90) 28 27.00 (4.10) 100.0 -6.10 [ -8.44, -3.76 ]

Total (95% CI) 29 28 100.0 -6.10 [ -8.44, -3.76 ]



-10.0 -5.0 0 5.0 10.0





Outcome 04 General Health Questionnaire (at 1 month)



Outcome: 04 General Health Questionnaire (at 1 month)



Yung 2004 18 20.47 (5.23) 30 27.57 (7.00) 100.0 -7.10 [ -10.58, -3.62 ]

Total (95% CI) 18 30 100.0 -7.10 [ -10.58, -3.62 ]



-10.0 -5.0 0 5.0 10.0


Analysis 03.01. Comparison 03 Work-directed intervention vs no intervention (Short Term), Outcome 01

Maslach Burnout Inventory


Comparison: 03 Work-directed intervention vs no intervention (Short Term)

Outcome: 01 Maslach Burnout Inventory




Melchior 1996 60 14.45 (6.47) 101 15.97 (6.66) 100.0 -1.52 [ -3.61, 0.57 ]

Subtotal (95% CI) 60 101 100.0 -1.52 [ -3.61, 0.57 ]



02 Depersonalisation

Melchior 1996 60 5.61 (2.96) 101 6.75 (3.68) 100.0 -1.14 [ -2.18, -0.10 ]

Subtotal (95% CI) 60 101 100.0 -1.14 [ -2.18, -0.10 ]




Melchior 1996 60 32.28 (3.55) 101 32.21 (3.98) 100.0 0.07 [ -1.12, 1.26 ]

Subtotal (95% CI) 60 101 100.0 0.07 [ -1.12, 1.26 ]



-10.0 -5.0 0 5.0 10.0





Nursing Stress Scale



Outcome: 02 Nursing Stress Scale



Delvaux 2004 57 1.70 (0.76) 58 2.04 (0.78) 100.0 -0.34 [ -0.62, -0.06 ]

Total (95% CI) 57 58 100.0 -0.34 [ -0.62, -0.06 ]



-10.0 -5.0 0 5.0 10.0



General symptoms



Outcome: 03 General symptoms



Proctor 1998 30 5.55 (3.77) 27 8.45 (4.80) 100.0 -2.90 [ -5.16, -0.64 ]

Total (95% CI) 30 27 100.0 -2.90 [ -5.16, -0.64 ]



-10.0 -5.0 0 5.0 10.0


Analysis 04.01. Comparison 04 Work-directed intervention vs no intervention (Medium Term), Outcome 01

Nursing Stress Scale (at 6 months)


Comparison: 04 Work-directed intervention vs no intervention (Medium Term)

Outcome: 01 Nursing Stress Scale (at 6 months)



Delvaux 2004 53 1.62 (0.73) 58 1.81 (0.87) 100.0 -0.19 [ -0.49, 0.11 ]

Total (95% CI) 53 58 100.0 -0.19 [ -0.49, 0.11 ]



-10.0 -5.0 0 5.0 10.0




Documents

Preventing occupational stress in healthcare workers