Mental Health & Prevention (2013) 1, 19–25

2212-6570/$ - see frohttp://dx.doi.org/1

nCorresponding auE-mail address: O

(O. Van den Bergh).1Currently at Thom

Effectiveness of a six session stress reductionprogram for groups

Tom Van Daelea,1, Chantal Van Audenhovea,Debora Vansteenwegenb, Dirk Hermansc, Omer Van den Berghd,n

aPolicy Research Centre Welfare, Health and Family & Research Group on Health Psychology, University ofLeuven, BelgiumbCentre for the Psychology of Learning and Experimental Psychopathology & ISW Limits, University ofLeuven, BelgiumcCentre for the Psychology of Learning and Experimental Psychopathology, University of Leuven, BelgiumdResearch Group on Health Psychology, University of Leuven, Tiensestraat 102, 3000 Leuven, Belgium

Received 5 March 2013; received in revised form 5 June 2013; accepted 7 October 2013Available online 17 October 2013

KEYWORDSStress reduction;Psychoeducation;Cognitive-behavioral;Intervention;Matched control;Design

nt matter & 20130.1016/j.mhp.201

thor.mer.VandenBergh

as More Universit

AbstractThis study set out to determine the effectiveness of a 6-week cognitive-behavioral stressreduction course for groups. Two groups (intervention group N=47; matched control groupN=47) completed questionnaires on stress, depression, anxiety, worrying, and stress manage-ment skills pre and post-intervention, at 6 months and at 1 year follow-up. Results showeddecline for all symptoms in the intervention group (linear trends pso.05), whereas stressmanagement skills remained stable. Clinically significant and reliable change for almost 30% ofparticipants confirmed these findings. No such change was found for the control group. Overall,the data showed small but reliable, long-lasting effects.& 2013 Elsevier GmbH. All rights reserved.

Introduction

Lazarus and Folkman (1984) defined psychological stress as“… a particular relationship between the person and theenvironment that is appraised by the person as taxing orexceeding his or her resources and endangering his or herwell-being.” Almost three decades later, chronic stress isconsidered a major burden in modern society, compromising

Elsevier GmbH. All rights reserve3.10.001

@ppw.kuleuven.be

y College, Antwerp, Belgium

both physical and mental health (American PsychologicalAssociation). High levels of self-perceived stress are, forexample, closely related to several adverse health condi-tions like metabolic syndrome (Chandola, Brunner, &Marmot, 2006) and coronary heart disease (Jood, Redfors,Rosengren, Blomstrand, & Jern, 2009; Rosengren et al.,2004; Xu, Zhao, Guo, Yanhong, & Gao, 2009). There is also aclear link between high levels of stress and the subsequentonset of mental disorders such as depression (van Praag,2004; Wang, 2005).

Considering the scope of the burden of stress, no healthservice will ever be able to provide adequate treatment forall, even in more affluent countries (van't Veer-Tazelaaret al., 2009). This emphasizes the need for large scale

d.

T. Van Daele et al.20

prevention, for example by reducing stress in the generalpopulation. In mental healthcare, prevention can be situ-ated within a stepped-care approach. This represents anattempt to maximize the efficiency of resource allocation intherapy: low threshold and low cost interventions areoffered first, and more intensive and costly interventionsare reserved for those who are not sufficiently helped by theinitial intervention (Haaga, 2000). A recent meta-analysisincluding a variety of programs confirmed that the averageparticipant of a stress reduction program obtains a signifi-cant reduction of perceived stress. When long-term changesare considered, however, results are less clear. The limitednumber of studies that include follow-up for up to 6 months orless find mixed results (Van Daele, Hermans, Van Audenhove,& Van den Bergh, 2012).

The current study therefore aims at consolidating theevidence base for stress reduction programs, both in the shortand long term. In the present case, we are interested in howthe intervention performs in the real-life context of commu-nities, resembling common practice. This provides a moreaccurate view of intervention effectiveness in everyday life.The intervention itself is a stress reduction program, developedwithin the cognitive-behavioral therapy (CBT) tradition as anadaptation of a program by White (2000) that was originallydeveloped to reduce anxiety. It is being offered to large groupsof self-registering community dwellers. Since they self-register,participants may have various initial complaints and motiva-tions constituting a heterogeneous group of participants with‘typical’ elevated stress symptoms, but also participants withlow levels of stress whose main interest is to learn more aboutstress and how it may affect them. Whereas White's course wasmore focused on curing participants with elevated complaintlevels, the current course has therefore more characteristics ofa selective preventive intervention.

The goal of the program is to reduce stress by altering therelationship between the person and the environment. Morespecifically, stress reduction is intended to occur through twomain routes. One focuses on strengthening the participants'resources through developing social and self-managementskills. The other attempts to change cognitive representa-tions through targeting negative appraisals and unhelpfulperseverative thinking, such as worrying and ruminatingwhich may mediate the relationship between stressors andpsychopathology (Brosschot, Gerin, & Thayer, 2006). Becausethe program aims to initiate a learning process, the reductionof stress-related symptoms is expected to occur graduallyand to continue in the months following the intervention.

Changes were assessed through self-report questionnaires.Stress scores were considered as the primary outcomemeasure, depression and anxiety as secondary outcomemeasures, and reduction in worrying and increase in stressmanagement skills as the means for stress reduction. We useda pre–post matched control design with two follow-upmoments, one after 6 and one after 12 months. Becauseparticipants needed time to process all the information andpractice the skills taught during the course, it was hypothe-sized that in the months following the intervention, a steady,gradual decline in worrying and a gradual increase in stressmanagement skills would be accompanied by a decline instress and depression and anxiety. The strongest effect isexpected to occur for those participants who present them-selves with higher levels of initial symptoms.

Method

Recruitment and screening

In order to participate, respondents had to reside in one ofthree regions in Flanders (Belgium). In each region, localorganizations were contacted to help distribute informationleaflets through their own networks and communicationchannels, including general practitioners, (sports) clubs,libraries and local press. Exclusion criteria were definedand potential participants who met at least one of thesewere informed that the current intervention might notcompletely suit their needs and that additional professionalhelp might be necessary. Subsequently, they could decide tocontinue following the course or not, but they were alwaysadvised to contact the local centre for ambulatory mentalhealthcare. The centers were informed about these poten-tial contacts and agreed to give these requests priority.If participants continued to follow the course, they wereremoved from the study sample. The exclusion criteriawere the answers on (1) question 15 of the Web ScreeningQuestionnaire (Donker, Straten, van Marks, & Cuijpers,2009) indicating suicidal tendencies (Answering ‘I woulddo it given the opportunity’ on the question whether theidea of harming yourself or taking your life, recentlycame into their mind), (2) the General Anxiety DisorderQuestionnaire-7 (Spitzer, Kroenke, Williams, & Löwe, 2006)showing they suffered from a severe generalized anxietydisorder (15+ on a 21 point scale), (3) three questions ofthe Alcohol Use Disorders Screening Test (Saunders,Aasland, Babor, de la Puente, & Grant, 1993) pointing toproblematic substance abuse (which could lead to alcoholinduced violence, endangering fellow participants). Duringthe course, participants could also be excluded if theteacher-therapist noticed signs of psychotic disorders orsevere deviant behavior.

To study long-term effects, the original goal was torandomly allocate participants to a stress managementcourse or to a 1-year non-intervention control group. This,however, raised practical and ethical concerns in localpartners endangering course implementation: local partnerswere reluctant to advertise the study when half of theparticipants would be denied treatment for 12 months orwould receive some kind of placebo treatment. A matchingprocedure was therefore used to collect control datainstead of using randomized non- or pseudo-interventioncontrols. In the matching procedure, a large sample wasrecruited from the general population through local news-papers, answering an advertisement to participate in aquestionnaire study concerning their general well-being.Subsequently, a selected number of them were matchedone-on-one to the course participants according to prede-termined criteria: stress scores, depression and anxiety, aswell as age, socioeconomic status and gender. Participantsin this control group were not aware of the intervention andhad not expressed an explicit desire to participate in thestress course. This design proved to be acceptable for localpartners. It was subsequently also approved by the ethicscommittee of the Faculty of Psychology and EducationalStudies of the University of Leuven. Controls received €10per data collection wave for participating.

21Effectiveness of a six session stress reduction program

Intervention

The intervention was an adaptation of a program called“Stress Control” (White, 2000; ISW Limits, 2006). Teacherswere trained psychologists from local centers for ambula-tory mental healthcare. They led this course, which com-prised six weekly lessons of 2 h during which participantsmostly listen and are not required to interact. In lesson 1,participants were offered general information on stress; itserved as a general course introduction, followed by twohomework assignments. For the first homework assignmentparticipants needed to evaluate their own stress level,reaction patterns, and general well-being. This wasfollowed by determining concrete goals they wanted toreach during the time of the course. A second homeworkassignment implied reading an overview of basic self-helptips and techniques that could help them with short-termstress reduction, including distraction, practicing sports,breathing exercises, and a proper diet. The goals of thesehomework assignments were to familiarize participants withthe course, to help them apply the general information totheir own personal situation, and to create a personalizedframe of reference for the course content. This allowedcourse participants to select specific, relevant techniquespresented in the following lessons. During lesson 2 theyfocused on the effects of stress on the body and controllingbodily sensations. After some theoretical background, parti-cipants learned the techniques of progressive relaxation andbreathing exercises. Furthermore, the importance of activerecreation was emphasized. In lesson 3 cognitive techniqueswere demonstrated and participants focused on becomingaware of fallacies and challenging dysfunctional thoughts. Inlesson 4 techniques were taught from problem-solving andparticipants learned how to confront their fears, end safetybehaviors, and increase their assertiveness. In lesson 5 andthe first part of lesson 6 the knowledge from Lesson 1 and thetechniques learned in the previous lessons were rehearsed.These were subsequently applied those to problems ofanxiety, panic, sleeping disorders and feelings of depression,and tension and burn-out. In the second part of Lesson 6,guidance was provided on how to control future stress.

Measures

The Depression Anxiety Stress Scales-21 (DASS-21) (Lovibond& Lovibond, 1995), Dutch version by de Beurs, Van Dyck,Marquenie, Lange, and Blonk (2001) is a 21-item self-reportquestionnaire measuring stress (7 items, α=.89) in the pastweek. Symptoms of depression (7 items, α=.94), and symp-toms of anxiety (7 items, α=.91) are also considered assecondary outcome measures.

The Penn State Worry Questionnaire (PSWQ) (Meyer,Miller, Metzger, & Borkovec, 1990), Dutch version by vanRijsoort, Vervaeke, and Emmelkamp (1997) is a 16-item self-report questionnaire used to measure worrying on a fivepoint Likert scale ranging from 1 ‘Not at all typical for me’to 5 ‘Very typical of me’. The questionnaire has a highinternal consistency both for normal (α=.90) and clinical(α=.86) populations.

The Coping Strategies Indicator (CSI) (Amirkhan, 1990),Dutch version by Bijttebier and Vertommen (1997) is a

33-item self-report questionnaire that measures three cop-ing styles: problem-solving (11 items, α=.87), social supportseeking (11 items, α=.90), and avoidance (11 items, α=.73).Because the course supports problem-solving and socialsupport seeking and tries to diminish avoidance when facedwith stress, this questionnaire is therefore suited to evaluatethe change in stress management skills. The questionnaireuses a three point Likert scale (3 ‘a lot’, 2 ‘a little’, 3 ‘not atall’). The higher the scores, the more commonly a strategy isused. High scores are therefore considered positive forproblem-solving and social support seeking, and negativefor avoidance.

Course implementation

The intervention was implemented using a framework calledempowerment implementation. This strategy offers interven-tions room to be flexible to local needs, while still maintainingadherence to strict implementation guidelines. More details onthe actual implementation and the underlying framework canbe found in Van Daele, Van Audenhove, Hermans, Van denBergh, and Van den Broucke (advanced online publication).

Statistical analysis

All data were analyzed using SPSS (SPSS 16.0, IBM). Groupby time interaction was evaluated using a generalized linearmodel (GLM) and compared self-reported symptoms andskills at each measurement point of the design. Further-more, when group by time interactions were at leastmarginally significant (po.10), group by trend-over-timeinteractions were also conducted, testing for linear andquadratic trends in the time variable.

Reliable and clinically significant changes were used assecondary measures for course effectiveness. According toJacobson and Truax (1991), the optimal way to determineclinically significant change is according to ‘criterion C’. Thismethod assumes that both the normal and the clinicalpopulation are normally distributed. In order to achieveclinically significant change (CSC), symptoms severity ofparticipants should be closer to the mean score of the normalpopulation than to those of the clinical population. There-fore, a cut-off score is determined. Because such a cut-offscore is arbitrary and small changes may also lead to a changefrom one side of the cut-off score to the other, reliablechange (RC) is used as an additional criterion. RC indicatesthat the change reported is larger than expected by chance.Both types of change are independent of one another, but thechange aimed for is one that is both reliable and clinicallysignificant. For this analysis, DASS-stress-scores were used todetermine treatment effects, because for this questionnaireDutch normative data were available both for the normal(M=8.5, SD=8.0) and clinical population (M=15.8, SD=9.8).

Results

Participants

Questionnaires were administered pre-intervention, post-inter-vention, after 6 months, and 1 year after course completion.

T. Van Daele et al.22

A total of 77 participants completed the first questionnaire.Two participants who had completed the first questionnairedropped out after the first session, one due to lack of interest,the other due to unexpected surgery. Furthermore, based onthe exclusion criteria, one participant with a suicide risk wasinformed that the intervention might not suit her needs and shewas referred to a local centre for ambulatory mental health-care. This participant nevertheless decided to continue follow-ing the course, but was removed from the study sample.Overall, 75 participants from three different locations (N=28,N=34 and N=13) completed the course, of which 47 partici-pants (63% of the completers) filled in the questionnaires at allfour times. Their mean age was 44.1 years (SD=10.1, range 21–63). From a total of 158 controls that completed the ques-tionnaire at pre-intervention, 139 (88%) completed the ques-tionnaires at all four times. From these 139 controls, 47 werematched with the intervention participants within one standarddeviation on depressive, anxiety and stress scores, as well as forage, socioeconomic status and gender. The mean age was 39.2years (SD=13.4, range 20–69). Other socio-demographics forboth groups can be found in Table 1.

Missing data

By the end of the study 28 (37%) of the 75 completers had notreturned one or more questionnaires and were therefore leftout of the final analyses. These non-responders were com-pared to responders on pre-intervention scores for the DASS,PSWQ, and CSI. Only for stress management skills, a signifi-cant difference was found with non-responders having higherproblem-solving scores, CSI-problem-solving: F(1,73)=4.20,p=.04. Other measures did not show a difference betweenboth groups. Overall, these analyses showed little differencebetween both groups, which makes the results outlined belowrelevant for the group of participants as a whole.

Course effectiveness

An overview of the data for all measurements can be foundin Table 2. Analyses apply to all four repeated measures,which were done over the course of 1 year. Detailed results

Table 1 Sociodemographics for intervention group (N=47) and

Marital status Not marriedMarried or living togetherWidowed or divorced

Degree Lower secondary education (or lessHigher secondary educationHigher education

Gender MaleFemale

of all these group by time interactions, group by trend overtime interactions and trend over time interactions can befound in Table 3, together with Cohen's ds for the interven-tion group at post and 1 year follow-up.

Stress, depression and anxiety. When the group by timeinteraction was evaluated using GLM, no significant effecton stress scores was found. For the secondary outcomemeasures, a significant effect was found for depressivesymptoms, but not for anxiety. Additionally, the group bytrend-over-time interactions did show a clear trend forstress scores with a linear decline in the intervention group,which was absent in the control group. This was also thecase for depressive symptoms, with a linear decline in theintervention group, whereas the control group remainedstable. No such changes were found for anxiety. Further-more, the intervention group could be split into two: onegroup with higher and one group with lower initial symp-toms. For each measure, the split value is determined as thevalue of the intersection point between the distributions ofthe normal and clinical population (Figure 1). When low(N=12) and high (N=35) stressed participants were com-pared using GLM, there was a strong group by time inter-action, F(3, 135)=4.24, p=.007. Similar results were foundfor low (N=22) and high (N=25) depressed participants,F(3,135)=4.70, p=.004, and for low (N=25) and high(N=22) anxious individuals, F(3, 135)=6.05, p=.001. Addi-tional analyses on the group by trend-over-time interactionshowed clear linear trends for stress F(1, 45)=9.61,p=.003, depression, F(1, 45)=11.39, p=.002, and anxiety,F(1, 45)=6.76, p=.013, with participants with high levelsof initial symptoms showing a decline, whereas patientswith low initial levels remained stable or showed a limitedincrease.

Worrying. An analysis of the group by time interactionwith GLM showed a strong effect on worrying. This wasconfirmed with an analysis on the group by trend-over-time interaction which showed a clear decline for theintervention group whereas the control group remainedstable.

Stress management skills. When the group by timeinteraction was evaluated using GLM, a significant effectwas found for problem-solving in which scores for theintervention group increased from pre to post followed by

matched control group (N=47) in percent.

Group

Intervention Matched control

13 4968 3619 15

) 8 626 2866 66

17 1783 83

Table 2 Evolution of intervention group (N=47) and matched control group (N=47).

Questionnaire May 10’ June 10’ December 10’ June 11’M (SD) M (SD) M (SD) M (SD)

Intervention groupDASSStress 17.5 (9.4) 15.2 (8.3) 14.0 (9.0) 13.7 (9.9)Depression 11.2 (9.2) 9.9 (9.1) 9.2 (9.5) 7.7 (8.3)Anxiety 8.5 (8.0) 7.2 (6.1) 7.1 (8.3) 7.0 (8.4)

PSWQ (worrying) 56.5 (12.7) 52.6 (11.7) 51.7 (12.9) 49.2 (10.7)CSI (coping)Problem solving 24.1 (6.0) 26.0 (5.7) 25.8 (4.7) 24.9 (4.8)Social support seeking 23.6 (6.4) 24.9 (4.1) 24.9 (5.3) 24.5 (4.3)Avoidance 18.9 (4.6) 17.4 (4.1) 18.7 (4.2) 20.1 (2.7)

Matched control groupDASSStress 12.6 (7.9) 11.4 (6.4) 11.8 (8.4) 13.4 (7.4)Depression 7.5 (8.5) 5.4 (5.4) 7.5 (8.5) 7.8 (7.3)Anxiety 5.9 (5.5) 5.7 (5.8) 5.4 (6.3) 5.0 (6.3)

PSWQ (worrying) 48.4 (12.2) 49.2 (11.4) 49.1 (12.0) 48.7 (11.9)

CSI (coping)Problem solving 23.5 (4.4) 23.9 (5.5) 21.5 (5.8) 23.6 (4.5)Social support seeking 22.8 (5.8) 23.4 (5.5) 22.7 (5.5) 22.5 (5.9)Avoidance 17.3 (3.9) 16.1 (3.6) 17.4 (4.5) 16.9 (4.2)

Table 3 F-values for group by time interactions, group by trend over time interactions and trend over time interactions andCohen's d for the intervention group at post and 1-year follow-up for each of the subscales.

Questionnaires Analyses

Group by Trend over time Cohen's d

Time Trend over time Intervention Control At post At follow-up

DASSa,b,e

Stress 2.1 5.3n 7.6n o1 .14 .40Depression 2.7n 5.4n 7.1n o1 .18 .18Anxiety o1 o1 .26 .39

PSWQ (worrying)a,b,e 5.6nnn 13.2nnn 21.1nnn o1 .32 .62

CSI (coping)c,d

Problem solving 2.95n o1 .41 .20Social support seeking o 1 1.2 .24 .16Avoidance 2.3 3.1 .34 .32

nnpo.01.npo.05.nnnpo.001.adf=276.bdf=92.cdf=264.ddf=88.edf=46.

23Effectiveness of a six session stress reduction program

a limited decline at follow-up whereas those of the controlgroup remained relatively stable. No such differences werefound for social support or avoidance. Analyses on the group

by trend-over-time interaction for stress management how-ever, were not significant for problem solving, social supportseeking, nor avoidance.

Depression Anxiety Stress Scales

Scor

eStress Depression Anxiety

Higher baselineLower baseline

Figure 1 Comparison of DASS-scores between participantswith a low- and high-level baseline of complaints.

T. Van Daele et al.24

Clinical significance

Clinical significance was evaluated using reliable and clini-cally significant change. DASS-stress-scores were used todetermine treatment effects. The cut-off score for clinicallysignificant change was determined at 11.7, which indicatedthat anyone above this score was considered within therange of the clinical population and anyone below this scorewas within the range of the normal population. Further-more, in order for a change to be considered reliable – notattributable to chance – it had to be larger than 6.3.Initially, 85% of all participants reported stress symptomswithin the range of the clinical population. At post-inter-vention, 13% of these participants showed both a reliableand clinically significant change. After 6 months this numberhad increased to 23% and by the 1 year follow-up, 28% of theparticipants had undergone a reliable and clinically signifi-cant change. Because clinically significant change impliedthat participants had to be within the range of the clinicalpopulation, it is not surprising that 92% of the participantsthat underwent a clinically significant and reliable changewere part of the group with higher initial symptoms men-tioned earlier.

Discussion

This study aimed to investigate whether a 6-week CBT stressreduction course for groups was effective in the long term inreducing stress and depression and anxiety, in decreasingworrying, and in increasing stress management skills. Theresults of the trend analyses indicated a linear decline ofstress and depression in the intervention group from pre-intervention to post-intervention and further on throughfollow-up. These effects were not observed in the non-intervention group. There was also a strong linear decline in

worrying, but only little change in stress management skills.Furthermore, measures of clinical significance showed thatalmost 30% of all course participants experienced a clini-cally significant and reliable change in the year followingthe course. Finally, the strongest effect occurred for thoseparticipants who presented themselves with higher levels ofinitial symptoms.

The overall modest mean effects seem to be due to thehigh amount of variation on all measures between partici-pants. Evidence for this explanation can be found in thecomparison of stress, anxiety and depression betweenparticipants with low and high initial symptoms, in whichparticipants with high initial levels show a much strongercontinuous and gradual decline of symptoms, whereasparticipants with low initial symptoms remained stable.This is in line with other stress management programs thatintend to realize long-term changes and which find effectsup to 4 years after the intervention (Rowe, 2000, 2006). Thecourse therefore appears to initiate a long-term processcharacterized by gradual declines in self-reported symp-toms. However, these declines were not strong enough tofind significant effects for all symptoms at all times using aGLM. Furthermore, participants who present themselveswith higher initial symptoms might have chosen to followthe intervention at a time of high life stress, which thenwould have subsided naturally, anyway, even without parti-cipating. Such competing factors and life experiences thatmight affect these aspects in a person's life can never beruled out completely.

Promising however is the fact that the reduction insymptoms was also accompanied by a strong decline inworrying, one of the two mechanisms deliberately targetedby the intervention. The other mechanism focused on theimprovement in stress management techniques of partici-pants, which showed little difference. The absence of aneffect could be because the CSI might not be that wellsuited as a questionnaire to measure general changes instress management skills. Since participants were asked toreport how they managed a specific problem they encoun-tered in the month preceding the time of completing thequestionnaire, it might be that stress management in highlyspecific situations was reported, as opposed to the moregeneral stress management skills that were intended. Analternative explanation is that the primary focus of theintervention was on psychoeducation and less on the actualskill training: participants were expected to practice athome in real life situations. Since this was not a controlledenvironment, feedback could not be delivered immediatelyand teachers had little control over whether course parti-cipants actively used the acquired stress managementtechniques in their home situations, an intervention aspectthat should be improved in future versions. This would implythat it might be the reduction in worrying that is primarilyresponsible for the decline in symptoms. Because thecurrent design does not allow making causal inferences, afocus for future research would therefore be to determinethe mechanisms through which the symptom reduction isaccomplished and potentially improve the skill trainingcomponent in order to make it (more) effective.

Finally, the present study did not make use of randomallocation of participants, because looking into the long-term effectiveness of the intervention would have required

25Effectiveness of a six session stress reduction program

withholding (wait list) control participants from followingthe course for over a year, which was not acceptable for theorganizations promoting the course locally. In addition,random allocation to an intervention or control group isdifficult for any study planning to do a long-term follow-up.A matched control design therefore turned out to be themost practical and ethical solution. While in a randomizedcontrolled trial (RCT) initial conditions in both groups areequalized, including the motivation to participate in acourse, the latter aspect was not present in our controlgroup. Neither did they receive some kind of bogus orplacebo treatment. However, we used a matched controldesign in which we tried to control for a wide range ofvariables related to participants' initial level of symptoms,as well as their socio-demographics. Balancing both theneed for long term follow-up and randomized allocation ofparticipants to non-intervention conditions may prove to bea difficult task.

Overall, the main conclusion of this study is that thereare preliminary indications that the intervention has sub-stantial long-term effects on participants' stress and anxietyand depression, especially for those who have higher initialsymptoms and as far as up to 1 year after the intervention.It furthermore also shows that the effectiveness of“interventions in the field” that focus on self-registeringcommunity dwellers should be interpreted carefully. Theheterogeneous groups they attract might not only includepeople who participate in order to obtain an immediatereduction of high symptoms, but also those who have lowlevel symptoms and follow the course in order to preventfuture symptoms. Especially the latter group could beresponsible for an underestimation of intervention effec-tiveness if they are not specifically taken into consideration,as this might cause floor effects.

References

American Psychological Association. Stress in America findings[Electronic version]. Washington, DC: Author.

Amirkhan, J. H. (1990). A factor-analytically derived measure ofcoping: The Coping Strategies Indicator. Journal of Personalityand Social Psychology, 59, 1066–1074.

Bijttebier, P., & Vertommen, H. (1997). Psychometric properties ofthe coping strategy indicator in a Flemish sample. Personalityand Individual Differences, 23, 157–160.

Brosschot, J. F., Gerin, W., & Thayer, J. F. (2006). The perseverativecognition hypothesis: A review of worry, prolonged stress-related physiolocial activation, and health. Journal of Psychoso-matic Research, 60, 113–124.

Chandola, T., Brunner, E., & Marmot, M. (2006). Chronic stress atwork and the metabolic syndrome: Prospective study. BritishMedical Journal, 332, 521–524A.

de Beurs, E., Van Dyck, R., Marquenie, L. A., Lange, A., & Blonk, R.W. B. (2001). De DASS: een vragenlijst voor het meten vandepressie, angst en stress [The DASS: A questionnaire formeasuring depression, anxiety and stress]. Gedragstherapie,34, 35–53.

Donker, T., Straten, A., van Marks, I. M., & Cuijpers, P. (2009).A brief web-based screening questionnaire for common mentaldisorders: Development and validation. Journal of MedicalInternet Research, 11, e19–e35.

Haaga, D. A. F. (2000). Introduction to the special section onstepped care models in psychotherapy. Journal of Consultingand Clinical Psychology, 68, 547–548.

ISW Limits (2006). Stressbeheersing [Stress control]. Leuven:Author.

Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statisticalapproach to defining meaningful change in psychotherapyresearch. Journal of Consulting and Clinical Psychology, 59,12–19.

Jood, K., Redfors, P., Rosengren, A., Blomstrand, C., & Jern, C.(2009). Self-perceived psychological stress and ischemic stroke:A case-control study. BMC Medicine, 7.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping.New York: Springer.

Lovibond, S. H., & Lovibond, P. F. (1995). Manual for the DepressionAnxiety Stress Scales. Sydney, Australia: The Psychology Foun-dation of Australia.

Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990).Development and validation of the Penn State Worry Question-naire. Behavior Research and Therapy, 28, 487–495. http://dx.doi.org/10.1016/0005-7967(90)90135-6.

Rosengren, A., Hawken, S., Ôunpuu, K. S., Zubaid, M., Almahmeed, W.,& Blackett, K. N., et al. (2004).Association of psychosocial riskfactors with risk of acute myocardial infarcation in 11,119 cases and13,648 controls from 52 countries (the INTERHEART study): A case-control study. Lancet, 364, 953–962.

Rowe, M. M. (2000). Skills training in the long-term managementof stress and occupational burnout. Current Psychology, 19,215–228.

Rowe, M. M. (2006). Four-year longitudinal study of behavioralchanges in coping with stress. American Journal of HealthBehavior, 30, 602–612.

Saunders, J. B., Aasland, O. G., Babor, T. F., de la Puente, J. R., &Grant, M. (1993). Development of the Alcohol Use DisordersScreening Test (AUDIT), WHO collaborative project on earlydetection of persons with harmful alcohol consumption. Addic-tion, 88, 791–804.

Spitzer, R., Kroenke, K., Williams, J., & Löwe, B. (2006). The GAD 7.A brief measure for assessing generalised anxiety disorder.Archives Internal Medicine, 166, 1092. (109).

van Praag, H. M. (2004). Can stress cause depression?. Progressin Neuro-Psychopharmacology & Biological Psychiatry, 28,891–907.

van Rijsoort, S. N., Vervaeke, G., & Emmelkamp, P. M. G. (1997).De Penn State Worry Questionnaire en de Worry DomainsQuestionnaire: Eerste resultaten in een normale Nederlandsta-lige populatie [The Penn State Worry Questionnaire and theWorry Domains Questionnaire: First results in a normal Dutch-language population]. Gedragstherapie, 30, 121–128.

Van Daele, T., Van Audenhove, C., Hermans, D., Van den Bergh, O.,& Van den Broucke, S. Empowerment implementation: Enhan-cing fidelity and adaptation in a psychoeducational intervention.Health Promotion International (advance online publication).

Van Daele, T., Hermans, D., Van Audenhove, C., & Van den Bergh,O. (2012). Stress reduction through psychoeducation: A meta-analytic review. Health Education and Behavior, 39, 474–485.

van't Veer-Tazelaar, P., van Marwijk, H. W. J., van Oppen, P.,van Hout, H. P. J., van der Horst, H. E., & Cuijpers, P., et al.(2009).Stepped-care prevention of anxiety and depression inlate life. Archives of General Psychiatry, 66, 297–304.

Wang, J. (2005). Work stress as a risk factor for major depressiveepisode(s). Psychological Medicine, 35, 865–871.

White, J. (2000). Treating anxiety and stress: A group psychoeduca-tional approach using brief CBT. Chichester: Wiley.

Xu, W., Zhao, Y., Guo, L., Yanhong, G., & Gao, W. (2009). Job stressand coronary heart disease: A case-control study using a chinesepopulation. Journal of Occupational Health, 51, 107–113.