67057619

7/29/2019 67057619

1/11

The effectiveness of behavioural therapy for the treatmentof depression in older adults: a meta-analysis

Zara Samad1, Stephen Brealey2 and Simon Gilbody2

1Humber Mental Health Trust, Trust Headquarters, Willerby Hill, Willerby, East Yorkshire, UK2Department of Health Sciences, University of York, York, UKCorrespondence to: S. Brealey, E-mail: [email protected]

Objective: To systematically review the effectiveness of behavioural therapy in depressed older adults.

Methods: Electronic databases were searched to July 2009. Reference lists of systematic reviews andidentified studies from the search strategy were also screened. Randomised controlled trials (RCTs) ofbehavioural therapy compared with waiting list controls or other psychotherapies in older adults (aged!

55 years) with clinical depression were included. One author independently identified studies forinclusion. Two authors extracted data and assessed the included studies for risk of bias. Estimates ofdepression were combined using a random effects model and the I2 statistic to examine heterogeneity.

Results: Four RCTs were included in the meta-analysis. For post-treatment self-rated depressionsymptoms, behavioural therapy was not significantly more effective than a waiting list control[standardised mean difference (SMD) of 0.52, 95% confidence interval (CI) 1.35 to 0.30,

p 0.21, n 117], cognitive therapy (SMD of 0.23, 95% CI 0.24 to 0.70, p 0.33, n 134) or briefpsychodynamic therapy (SMD of0.37, 95% CI 0.84 to 0.11, p 0.13, n 69). For post-treatmentclinician-rated depression, behavioural therapy was not significantly more effective than cognitivetherapy or brief psychodynamic therapy but was significantly more effective than a waiting list control(weighted mean difference (WMD) of5.68, 95% CI 7.71 to 3.66, p

7/29/2019 67057619

2/11

and Older People Forum, 2008), it is likely that theproblem of depression in older adults will increase. Interms of prevalence, it is estimated that 25% of peopleaged over 65 have symptoms of clinical depression(Department of Health, 2009). If left untreated,

depression in older adults is associated with increasedmortality and a range of negative outcomes includingpoor quality of life, difficulty with social functioningand an increased risk of suicide (Blazer, 2003; Unutzer,2007).

Depressed individuals engage in fewer pleasurableand more aversive activities resulting in less positivereinforcement from interactions with the environment(MacPhillamy and Lewinsohn, 1974). This can leadto self-critical cognitions (Beck et al., 1979), whichcan lower activity engagement further and thereforenegatively reinforce the cognitions. In depressed olderadults, the reduced activity and self-critical cognitions

can arise for reasons including: threats to competency,health or independence and role changes due tospousal bereavement (Fiske et al., 2009). Behaviouraltherapy addresses the negative cognitions andemotions associated with depression in an indirectway (Hopko et al., 2003). It has been described asbeing different from traditional cognitive approachesin that it seeks to help patients modify theirenvironment, not their thinking. It is negative lifecircumstances and patients difficulty in changing thesecircumstances that may lead to passivity in depressedpatients (Jacobson et al., 2001). Treatment originally

involved scheduling an increase in pleasant activitiesand positive interactions with the environment(Lewinsohn and Graf, 1973). Recent developmentsinclude positive activation which places an increasedemphasis on reducing negative reinforcement charac-terised by avoidance behaviour (Jacobson et al., 2001).Other developments include the positive model whichinvolves: baseline assessment of activity; identifyingbehavioural goals within a number of life areas (e.g.hobbies, relationships, employment etc.); listingthese goals in a hierarchy of easiest to most difficultand then planning implementation of these goals viaweekly diary sheets (Hopko et al., 2003). Behavioural

therapy may be delivered via a range of formats (e.g.face to face sessions, telephone support, workbooks)and can be supported by a range of professionals (e.g.psychology graduates, mental health nurses) whoprovide various levels of support.

For use in older adults, behavioural therapypotentially has several advantages. In comparison topharmacological interventions, it can be considered asafer option as anti-depressants are likely to result innegative reactions with other medications (De Leo and

Dieksta, 1990), with older patients on average 3.6 timesmore prescriptions than younger adults (those aged1559 years) (Wong et al., 2004). Fears about theside effects of antidepressant medication in olderadults often result in the prescription of medication at

sub-therapeutic levels (Laidlaw et al., 2008). Further-more, prescribing medication does not address thechanges in behaviour that are needed to overcomedepression which is important given the complexbeliefs and attitudes that have been found in olderadults towards mental illness (Quinn et al., 2009). Asbehavioural therapy is a relatively simple and briefintervention which requires less intensive professionaltraining and support compared to other psychologicalinterventions it could be more cost-effective (Centrefor Economic Performances Mental Health PolicyGroup, 2006; Ekers et al., 2007). In addition, as it is abriefer, simpler intervention it might benefit older

adults with cognitive impairment or a low educationallevel (Porter et al., 2004) who find cognitiverestructuring, as used in cognitive behavioural therapy(CBT), more challenging than younger adults (Hertzogand Hultsch, 2000).

The aim of this systematic review of randomisedcontrolled trials (RCTs) was to determine theeffectiveness of behavioural therapy compared toother psychological approaches for the treatment ofdepression in older adults.

Methods

Identification of suitable studies

We searched electronic databases mostly from 1980until July 2009 to coincide with the introduction ofbehavioural therapy into clinical practice: CochraneLibrary (from 1980), Web of Science (from 1980),Medline (from 1950), EMBASE (from 1980), CINAHL(from 1982), AMED (from 1985), PsychINFO (from1987), British Nursing Index (from 1995) and NHSevidence for mental health. MESH headings and freetext along with truncation and wild cards were used

and randomised controlled trial filters. Searches werestructured as four concepts: diagnosis (e.g. depression,depressive disorder), intervention (e.g. activityscheduling), age (e.g. frail older, aging population)and design (e.g. randomised controlled trials).Reference lists were also screened of existing systematicreviews on psychotherapeutic interventions andof identified studies from the search strategy. Tominimise publication bias grey literature was identifiedby searching for unpublished doctoral theses (via

Copyright # 2011 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2011; 26: 12111220.

1212 Z. Samad et al.

7/29/2019 67057619

3/11

http://opensigle.inist.fr/), conference proceedings (viathe Web of Science Conference Proceedings CitationIndex database) and government publications (viathe British Official Publications Current AwarenessService and the Health Management Information

Consortium). The World Health Organisation (WHO)International Clinical Trials registry was also searchedto identify relevant ongoing clinical trials. One author(ZS) screened abstracts and the full text of selectedstudies for eligibility and consulted with a secondauthor (SB) when necessary.

Inclusion criteria

All available RCTs in any language were includedto reduce the potential for publication bias (Khan andKleijnen, 2002). Studies included participants: who

were aged !55 years (although studies that included ayounger population were eligible if it was possibleto extract data for patients aged !55 years); with orwithout physical co-morbidities; and with a diagnosisof clinical depression using structured diagnosticinterviews such as DSM-IV-TR (American PsychiatricAssociation, 2000), clinician-rated scales (e.g. Hamil-ton Depression Rating Scale (HDRS); Hamilton, 1967)or self-report scales (e.g. Beck Depression Inventory(BDI); Beck et al., 1961). Studies were excludedwhich included participants with co-morbid dementiaor severe cognitive impairment or met a diagnosis

of psychosis, bi-polar disorder, substance-misuse ora primary diagnosis for any other mental healthdisorder.

We included RCTs in the behavioural therapygroup which involved an intervention based oneither the basic behavioural principles (e.g. increasingaccess to positive reinforcement) or the more recentdeveloped forms of behavioural therapy (e.g. learningabout the maintenance of depression symptoms andrestarting avoided pleasant/routine/necessary activi-ties). Studies were excluded where behavioural therapywas used in combination with cognitive techniqueswhich directly addressed the negative cognitions

associated with depression. The comparators whichcould be included are described below.

Treatment as usual. This could include a number ofoptions such as usual General Practitioner treatment(e.g. prescribed anti-depressants) or being on a waitinglist for psychological therapy.

CBT. This included interventions that directlymodified cognitions using both cognitive techniques

(e.g. examining evidence for or against dysfunctionalcognitions) and behavioural methods (e.g. behaviouralexperiments to test out predictions based on dysfunc-tional cognitions).

Interpersonal therapy (IPT). Approaches that lookedat links between depressed mood and difficulties in thefollowing areas of conflict: transition, bereavement,interpersonal relationships with significant othersand long term difficulties in forming and maintainingrelationships (Klerman et al., 1984).

Psychodynamic therapy. This concerns approachesused to bring repressed thoughts and feelings intoconsciousness and to develop new ways of toleratingand coping with the emotional pain (Leiper, 2006).

Supportive counselling. This approach focuses on anindividual exploring any problems they may have andto develop ways to resolve them (Rogers, 1961).

Outcome measures

The primary outcome measure was a change indepression symptoms using self-rated (e.g. BDI) orclinician-rated measures (e.g. HDRS) presented ascontinuous data (e.g. means and standard deviations).As psychotherapy trials often present multiple depres-sion measures we gave validated self-report measuresprecedence over clinician-rated measures. As a proxyfor acceptability, dropout rates from treatment wererecorded as dichotomous data.

Assessment of risk of bias in included studies

Each eligible study was assessed for risk of bias (yes,no or unclear) by two independent authors (ZS andSB) using the Cochrane Collaboration risk of biastool (Higgins and Green, 2008). Disagreements were

resolved through discussion. This assessment addressesmethodological issues such as adequate sequencegeneration in treatment allocation, adequate conceal-ment of treatment allocation, blinding of treatmentallocation, addressing incomplete outcome data, non-selective outcome reporting and other potential threatsto study validity. The extent of agreement between theauthors was expressed as a percentage and Kappastatistic. The implications of risk of bias to the results ofincluded studies are discussed narratively.


Behavioural therapy in older adults 1213

7/29/2019 67057619

4/11

Data extraction

Data were extracted for each eligible study indepen-dently by one author (ZS) and checked by anotherauthor (SB). Characteristics of included studies

were extracted regarding: participants (e.g. age, base-line depression), the intervention and control groups(e.g. mode of delivery, therapist level) and outcomemeasures. If data were missing from individual studiesthe authors were contacted via email.

Data synthesis

Analyses were conducted for self-report measures andclinician-rated measures separately. When individualstudies reported more than one self-rated depressionmeasure, precedence was given to the Geriatric

Depression Scale (GDS) as this measure is validatedspecifically for the older adult population. In theabsence of the GDS, scores from the BDI were used asit is one of the most frequently used instruments indepression research and has widespread acceptability(Rokke et al., 1999). For clinician-rated depression,scores from the HDRS were used in the data analysis.For self-rated depression symptoms the standardisedmean differences (SMD) were calculated across trialsto facilitate analysis of the same outcome usingdifferent scales. We assigned effect sizes accordingto the standard convention where the SMD is small

(00.32), medium (0.330.55) and large (0.56 or more)(Lipsey and Wilson, 1993). For clinician-rated depres-sion symptoms using the HDRS, the weighted meandifference (WMD) was calculated and interpretedwith reference to its effect size. Data for drop outfrom treatment were presented as odds ratios (OR) toindicate the likelihood of these events occurring inthe intervention group compared to the comparisongroups. In anticipation of variation in the delivery oftherapies (e.g. number of sessions, therapy approaches,setting) data were pooled with 95% confidenceintervals (CI) using a random effects model takinginto account both within- and between-study variance

(Sutton et al., 1998).

Exploration of heterogeneity

We measured statistical heterogeneity using the I2

statistic which describes the percentage of variability ineffect size that can be attributed to study heterogeneityrather than due to chance (Higgins and Thompson,2002). Values of the I2 statistic of 25% are considered

to be low, 50% moderate and 75% high (Higgins et al.,2003). Four sources of clinical heterogeneity wereidentified a priori: diagnosis of depression at baseline;type of professional who delivered the behaviouraltherapy intervention; mode via which the treatment

was delivered and presence of physical co-morbidity.The impact of the potential sources of clinical hetero-geneity on the overall treatment effects were explored,when possible, using sensitivity analyses (Higgins andThompson, 2002).

Results

Searching electronic database identified 633 studiesof which 579 were excluded from screening titlesand abstracts as shown in Figure 1. Full copies of54 publications were then retrieved and assessed for

eligibility. From these publications another 4 poten-tially eligible studies were identified from referencelists. Of these 58 studies there were four RCTs whichmet the inclusion criteria (Gallagher and Thompson,1982; Thompson et al., 1987; Scogin et al., 1989; Rokkeet al., 1999). Table 1 summarises the main character-istics of the included studies and shows that whilstthey were all undertaken in the community, includepatients of a similar age and used similar outcomemeasures, there is some variability in the assessment ofbaseline depression, the interventions being compared,and particularly the delivery of the intervention such

as whether this was face-to-face or in the form ofbibliotherapy.

Risk of bias in included studies

For the assessment of risk of bias the percentageagreement between the two independent authors was83% with a Kappa of 0.72 (95% CI 0.470.96). Allincluded studies stated the use of random assignmentin treatment allocation. Figure 2 shows, however,that it was unclear as to whether there was adequateallocation sequence or concealment. In addition, none

of the included studies reported sufficient detailsabout whether treatment allocation was blinded to theoutcome assessors who completed the clinician-ratedmeasures. Incomplete outcome data did not appear tobe a source of bias in one study with missing outcomedata balanced in numbers across intervention groups(Gallagher and Thompson, 1982). Details were insuffi-cient in explaining or dealing with incomplete outcomedata in the other three studies. For selective reporting,three of the studies presented the results of all outcome



7/29/2019 67057619

5/11

measures which they stated to have used (Gallagher andThompson, 1982; Scogin et al., 1989; Rokke et al.,1999), although there was a lack of clarity as to what wasthe primary outcome in two of these studies (Gallagherand Thompson, 1982; Scogin et al., 1989). For other

potential threats to study validity, three of the studieswere potentially at risk of another source of bias. Thisconcerned how patients were approached to take partin the trial (Scogin et al., 1989), how participants in theno-choice condition were yoked to participants in thechoice condition (Rokke et al., 1999), and how patientswho dropped out of treatment were replaced (Gallagherand Thompson, 1982).

Effectiveness of behavioural therapy compared with

waiting list control

Three RCTs were included in this comparison (Thomp-son et al., 1987; Scogin et al., 1989; Rokke et al., 1999).Data were used from the behavioural therapy group attreatment completion and from the delayed treatmentgroup at the end of the waiting period. For post-treatment self-rated depression symptoms the analysisused the GDS (Scogin et al., 1989; Rokke et al., 1999)and the BDI (Thompson et al., 1987). Figure 3 presents apooled SMD of0.52 (95% CI1.35 to 0.30) suggestinga medium effect in symptom level scores of depression

favouring the behavioural therapy group but whichwas not statistically significant (p 0.21). There wassignificant heterogeneity (x2 8.94, df 2, p 0.01)with the I2 statistic suggesting that 78% of variation inthe effect size was due to between study heterogeneity.

For post-treatment clinician rated depression, HDRSscores were used in all studies. The pooled WMD wasestimated to be 5.68 (95% CI 7.71 to 3.66)demonstrating a highly significant difference insymptom-level scores of depression favouring thebehavioural therapy group (p< 0.001). For thisanalysis statistical heterogeneity was not significant(x2 0.11, df 2, p 0.94, I2 0%). There wereinsufficient studies to explore the impact of oura priori sources of clinical heterogeneity.

In terms of acceptability of behavioural therapy,Thompson et al. (1987) found that 16% (4 of 25) ofpatients in the behavioural therapy group dropped-out

of treatment; Scogin et al. (1989) found 17.3% (4 outof 23) of patients dropped out; and Rokke et al. (1999)found 53% (9 of 17) of patients dropped out.


cognitive therapy

All four RCTs contributed to this comparison(Gallagher and Thompson, 1982; Thompson et al.,

Figure 1 Flow of studies through the systematic review.



7/29/2019 67057619

6/11

1987; Scogin et al., 1989; Rokke et al., 1999). Theanalysis used data from the immediate treatmentcompletion stage which compares to a follow-up from1 to 3 months. For post-treatment self-rated depres-

sion symptoms, three studies used the GDS (Thomp-son et al., 1987; Scogin et al., 1989; Rokke et al., 1999)and the other study used the BDI (Gallagher andThompson, 1982). Figure 4 presents a pooled SMD of

Table 1 Main characteristics of included studies

First namedauthor (year)

SettingMean age (SD)Sex (% female)

Baselinedepressiondiagnosis

Interventions(number of patients)

Mode of deliveryTherapist level

Session number(duration)

Outcome measures

Gallagher(1982)

Community67.8 (6.1)

77%

ResearchDiagnostic

Behavi our al therapy (10) Face-to-Face Self rated: BDI andZung depressionscale

Criteria forMajor

Cognitive therapy (10) Doctoral levelpsychologists

Clinician rated: HDRS

DepressiveDisorder

Brief psychotherapy (10) 16 (90min) over12 weeks

Thompson(1987)

Community67.1 (5.8)

67%

ResearchDiagnostic

Behavi our al therapy (25) Face-to-Face Self rated: BDI, GDSand BSI (depression)

Criteria forMajor

Cognitive therapy (27) Doctoral levelpsychologists


DepressiveDisorder

Brief psycho-dynamictherapy (24)

1620 (twice a weekfirst 4 weeks and

once a week thereafter)6 week delayed treatment

control (19)Scogin(1989)

Community68.3 (6.8)

85%

HDRS!10 Behavioural bibliotherapy (23) Work book and weeklytelephone support

Self rated: GDS

Cognitive bibliotherapy (22) Researchers Clinician rated: HDRS4 week delayed treatment

control to bibliotherapy (22)

Rokke(1999)

Community66.3 (5.3)

38%

HDRS!10,BDI!10

and GDS!11

Choice of behavioural therapyor cognitive therapy (15)

Face-to-Face Self-rated: BDIand GDS

No choice of behaviouraltherapy or cognitive

therapy (20)

Doctoral/masters levelpsychologist or graduate

students in clinicalpsychology/counselling


10 week delayed treatmentcontrol (29)

10 (1 h) weekly

HDRS, Hamilton Depression Rating Scale; BDI, Beck Depression Inventory; GDS, Geriatric Depression Scale; BSI (Depression), Depression subscale of the

Brief Symptom Inventory.

Figure 2 Summary of risk of bias assessment in included studies.



7/29/2019 67057619

7/11

0.23 (95% CI0.24 to 0.70) suggesting a small effect inthe difference in symptom level scores of depressionfavouring the cognitive therapy group but which wasnot statistically significant (p 0.33). There was nosignificant heterogeneity between studies (x2 5.24,df 3, p 0.15, I2 43%).

As specified a priori, to explore the influence of aformal depression of diagnosis at baseline (e.g. Majordepressive disorder diagnosis) the two studies which

did not require a formal depression diagnosis (Scoginet al., 1989; Rokke et al., 1999) were removed fromthe meta-analysis. The pooled SMD was then 0.02(95% CI 0.44 to 0.48), which was not statisticallysignificant (p 0.93). To explore the influence of thetype of professional used to deliver the intervention(e.g. mental health vs. non mental health professional),data from one study was removed from the analysis asresearchers delivered the intervention (Scogin et al.,1989). The pooled SMD was then 0.00 (95% CI0.40 to 0.41), which was not statistically significant(p 1.00). To explore the influence of the mode ofdelivery of an intervention meant that the same study

should be removed from the analysis as bibliotherapywas used to deliver the intervention whereas theother studies used face-to-face sessions. This analysisproduced the same results as for removing the non-mental health professional. There was insufficientdata to allow us to explore the effect of physical co-morbidities on self-rated depression symptoms.

For post-treatment clinician-rated depression,HDRS scores were calculated for all four RCTs.The pooled WMD was estimated to be 0.05 (95% CI

2.10 to 2.00) in favour of the behavioural therapygroup which was not statistically significant (p 0.96).There was no significant heterogeneity between studies(x2 3.45, df 3, p 0.33, I2 21%). The influenceof sources of clinical heterogeneity on treatment effectas described above was again repeated. When formaldepression was diagnosed at baseline the pooledWMD was 1.25 (95% CI 4.30 to 1.79) in favourof behavioural therapy which was not statistically

significant (p 0.42). The overall effect remained infavour of behavioural therapy when the interventionwas delivered by a mental health professional or face-to-face with a pooled WMD of1.40 (95% CI 3.84to 1.03), which remained non-significant (p 0.26).It was not possible to explore the effect of physical co-morbidities on clinician-rated depression symptoms.

The number of participants dropping out fromtreatment between baseline and treatment completionwas reported in all four included RCTs. The pooleddropout was greater for cognitive therapy with an oddsratio of 2.04 (95% CI 0.87 to 4.78), which was notstatistically significant (p 0.10).


brief psycho-dynamic therapy

Two studies were included in this comparison usingdata at the end of treatment which is comparable with a3-month follow-up (Gallagher and Thompson, 1982;Thompson et al., 1987). For post-treatment self-rateddepression symptoms, the analysis used the GDS

Figure 3 Self-rated effectiveness of behavioural therapy compared with waiting list control at treatment completion.

Figure 4 Self-rated effectiveness of behavioural therapy compared with cognitive therapy at post-treatment (or 13 month follow-up).



7/29/2019 67057619

8/11

(Thompson et al., 1987) and BDI (Gallagher andThompson, 1982). Figure 5 presents a pooled SMD of0.37 (95% CI 0.84 to 0.11) suggesting a moderateeffect in favour of behavioural therapy which wasnot statistically significant (p 0.13). The evidence forstatistical heterogeneity was non-significant (x2 0.00,df 1, p 0.96, I2 0%). For post-treatment clinicianrated depression, HDRS scores were used for bothstudies. The pooled WMD was estimated to be

1.56 (95% CI4.64 to 1.52) in favour of behaviouraltherapy which was not statistically significant(p 0.32). For this analysis statistical heterogeneitywas non-significant (x2 0.24, df 1, p 0.62, I2 0).There were insufficient studies to explore the impact ofour a priori sources of clinical heterogeneity.

For dropout rates, the pooled odds ratio was 1.50(95% CI 0.326.96) with patients more likely todropout of the behavioural therapy intervention whichwas not statistically significant (p 0.61).

Discussion

The main findings from this review show thatbehavioural therapy for older people is significantlymore effective than waiting list control when measuredby clinician-rated depression using the HRDS, butnot significantly different when measured by patientself-report. It is unclear as to whether clinicianswho administered the HRDS were blind to treatmentallocation (Schulz et al., 1995), which could haveinfluenced its completion and thus the estimatedeffectiveness of behavioural therapy. We also foundthat behavioural therapy is not significantly different in

effectiveness compared with cognitive therapy or briefpsycho-dynamic therapy whether using self-reportedmeasures or clinician-rated assessment.

To examine sources of heterogeneity and to improveour understanding of what factors might influence theeffectiveness of behavioural therapy in older adults itwas possible to explore this for the four RCTs includedin the meta-analyses comparing behavioural therapywith cognitive therapy. We found that when twostudies were excluded (Scogin et al., 1989; Rokke et al.,

1999), which did not include a formal depression ofdiagnosis at baseline this reduced the effectiveness ofcognitive therapy compared to behavioural therapy inself-reported depression. Furthermore, when a studywas excluded (Scogin et al., 1989) because a researcherdelivered the intervention, which was not face-to-facebut in the form of work books and telephone support(i.e. bibliotherapy), then there was no clinical orstatistically significant difference in self-reported

depression at all. Therefore, it appears that whodelivers the intervention and how it is delivered such aswhether bibliotherapy is used or not is important fordetermining the effectiveness of behavioural therapy.We could not explore whether the presence of physicalco-morbidities influenced the effectiveness of inter-ventions as three studies did not provide sufficientdetail to judge whether they included participants withphysical co-morbidities (Gallagher and Thompson,1982; Thompson et al., 1987; Scogin et al., 1989) andone study excluded those patients with physical co-morbidities (Rokke et al., 1999). Because we had

relatively few studies we were unable to more formallyexplore sources of heterogeneity using techniques suchas meta-regression.

For this review a systematic approach was usedto identify the literature and there was independentdata extraction and assessment of study quality bytwo authors and used meta-analytical techniques tocombine the results of studies to increase power andto obtain a combined estimate of effect. It was notmeaningful to explore publication bias statisticallywith only four included RCTs and expect its presence isunlikely with the thorough searches undertaken ofgrey literature. There are, however, limitations to the

evidence presented. First, for the included RCTs it wasunclear as to how randomisation was implementedand whether there was blinded assessment of outcomein clinician-rated measures which are importantsources of bias (Schulz et al., 1995). None of thestudies were designed to have sufficient power todetect statistically significant differences and only hadsmall sample sizes. It was also only possible to combinestudies with short-term follow-up of around 13months. Second, with regards the intervention itself, it

Figure 5 Self-rated effectiveness of behavioural therapy compared with brief psycho-dynamic therapy at post-treatment (or 3 month follow-up).



7/29/2019 67057619

9/11

is not clear as to whether it was adapted to the deliveryof behavioural therapy for older adults. Nor were anydetails provided about the efforts to remove treatmentbarriers typically faced by older adults. This becameapparent as reasons for patients dropping out included

transportation problems (Gallagher and Thompson,1982; Thompson et al., 1987), physical illness(Thompson et al., 1987; Rokke et al., 1999) anddifficulty reading bibliotherapy due to visual impair-ment (Scogin et al., 1989). In contrast, other studieshave adapted an existing behavioural treatmentmanual for use in older adults by adding writtenexamples of activities more relevant to the older andusing large print (Martell et al., 2001; Yon and Scogin,2009) and for older adults in assisted living andmedical settings emphasising the need for changes inthe environment (Lichtenberg et al., 1998; Cernin andLichtenberg, 2009). In an attempt to remove treatment

barriers such as access, other studies have usedcommunity based non-mental health professionals(e.g. nurse, social worker, occupational therapist) todeliver the behavioural intervention within partici-pants homes (Unutzer et al., 2002; Ciechanowskiet al., 2004; Quijano et al., 2007; Cernin andLichtenberg, 2009). Third, there are problems withgeneralising results from the meta-analysis as the RCTsincluded older adults with mild to moderate depres-sion, who were predominantly in their 60s and livingindependently in the community. Caution should betaken when generalising results to older adults with

more severe depression and those in residential caresettings. Additionally, three of the included studiesrecruited participants through non-traditional meanssuch as media announcements and community flyers(Thompson et al., 1987; Scogin et al., 1989; Rokkeet al., 1999) and two of the studies did not require aformal diagnosis of major depressive disorder forinclusion of patients into the study (Scogin et al., 1989;Rokke et al., 1999). This is problematic as these studyparticipants are not necessarily representative of thosewho normally access mental health services. Fourth, alsowith respect to the age of participants included in theRCTs and generalisability, our inclusion criteria specifies

that older adults aged !55 years were eligible forinclusion when usually people !65 years of age arerecognised as being older adults. This was done becauseif a threshold of!65 years of age was used as an inclusioncriterion then none of the RCTs would have beeneligible. The mean age of patients included in the eligibleRCTs was between 66 and 68 and therefore most patientsincluded in the meta-analysis were !65 years of age.

In conclusion, this review has shown that beha-vioural therapy is potentially as effective as alternative

psychotherapies. This concurs with evidence from othersystematic reviews of the effectiveness of behaviouraltherapy mainly in adult populations which found that asa treatment for depression it has outcomes comparableto that of the current recommended psychologicalinterventions (Cuijpers et al., 2007; Ekers et al., 2007).These findings are of interest given that behaviouraltherapy is a relatively safe, simple and brief interventionwhich can be delivered by non-mental health pro-fessionals and thus potentially be more cost-effective.However, these findings should be interpreted with

caution as there is insufficient data from the eligibleRCTs to answer with adequate certainty about whetherbehavioural therapy is more or less effective thanalternative psychological approaches. Further research isrecommended for which larger sample sizes are required,with more clarity on trial design and the adaptation ofthe intervention for older adults, longer term follow-upand economic evaluations alongside clinical trials.

Conflict of interest

None declared.

References

American Psychiatric Association. 2000. Diagnostic and Statistical Manual of MentalDisorders (DSM-IV-TR). American Psychiatric Association: Washington DC.

Beck AT, Rush AJ, Shaw BF, Emery G. 1979. Cognitive Therapy of Depression. GuilfordPress: New York.

Beck AT, Ward C, Mendelson M, Mock J, ErbaughJ. 1961. An inventory for measuringdepression. Arch Gen Psychiatr 4: 561571.

Blazer DG. 2003. Depression in late life: review and commentary. J Gerontol58: 249265.

Key Points

Older adults are particularly vulnerable todepression. Behavioural therapy for the treatmentof depression is a relatively safe, simple and briefintervention which requires less intensive pro-

fessional training and support compared to otherpsychological interventions.

Behavioural therapy in depressed older adults hascomparable effectiveness with alternative psy-chotherapies in terms of patient self-reporteddepression symptoms.

Further research is recommended, however, inparticular larger sample sizes are required, withmore clarity on trial design and adaptation of theintervention for older adults, longer term follow-up, and concomitant economic evaluation.



7/29/2019 67057619

10/11

Carney RM, Freedland KE. 2003. Depression, mortality and medical morbidity inpatients with coronary heart disease. Biol Psychiatr 54: 241247.

Centre for Economic Performances Mental Health Policy Group. 2006. The Depres-sion Report. A New Deal for Depression and Anxiety Disorders . London School ofEconomics: London.

Cernin P, Lichtenberg P. 2009. Behavioural treatment for depressed mood: a pleasantevents intervention for seniors residing in assisted living. Clin Gerontol32: 324331.

Ciechanowski P, Wagner E, Schmaling K, et al. 2004. Community integrated home

based depression treatment in older adults: a randomized controlled trial. JAMA291: 15691577.

Cuijpers P, Van Straten A, Warmerdam L. 2007. Behavioural activation treatments ofdepression: a meta-analysis. Clin Psychol Rev 27: 318326.

De Leo D, Dieksta R. 1990. Depression and Suicide in Later Life. Hogrefre and Huber:Toronto.

Department of Health. 2009. Improving Access to Psychological Therapies (IAPT) OlderPeople: Positive Practice Guide. Department of Health: London.

Djernes JK. 2006. Prevalence and predictors of depression in populations of elderly: areview. Acta Psychiatr Scand 113: 372387.

Ekers D, Richards D, Gilbody S. 2007. A meta-analysis of randomised trials ofbehavioural treatment of depression. Psychol Med 38: 611623.

Fiske A, Wetherell J, Gatz M. 2009. Depression in older adults. Ann Rev Clin Psychol5:363389.

Gallagher D, Thompson W. 1982. Treatment of major depressive disorder in olderadult outpatients with brief psychotherapies. Psychotherapy19: 482490.

Hamilton M. 1967. Development of a rating scale for primary depressive illness. Br JSoc Clin Psychol 6: 278296.

Hertzog C, Hultsch D. 2000. Metacognition in adulthood and old age. In Handbook of

Aging and Cognition, Fergus C, Salthouse TA (eds). Lawrence Erlbaum Associates:Mahwah, NJ; 417466.

Higgins JP, Thompson SG, Deeks JJ, Altman DG. 2003. Measuring inconsistency inmeta-analyses. Br Med J 327: 557560.

Higgins JTP, Thompson SG. 2002. Quantifying heterogeneity in a meta- analysis. StatMed 21: 15391558.

Higgins JTP, Green S. 2008. Cochrane Handbook for Systematic Reviews of Interven-tions. Version 5.0.1. The Cochrane Collaboration.

Hopko DR, Leujuez CW, Ruggiero KJ, Eifert GH. 2003. Contemporary behaviouralactivation treatments for depression: procedures, principles and progress. ClinPsychol Rev 23: 699717.

Jacobson NS, Martell CR, Dimidjian S. 2001. Behavioral activation treatment fordepression: returning to contextual roots. Clin Psychol Sci Prac 8: 255270.

Katon WJ. 2003. Clinical and health services relationships between major depression,depressive symptoms and general medical illness. Biol Psychiatr 54: 216226.

Khan K, Kleijnen J. 2002. Phase 4: selection of studies. In Undertaking SystematicReviews of Research on Effectiveness: CRDs Guidance for those Carrying out orCommissioning Reviews, (2nd edn), Khan K, Ter Riet G, Glanville J, Sowden A,

Kleijnen J (eds). Centre for Reviews and Dissemination: University of York.Klerman GL, Weissman MM, Rounsaville BJ, Chevron ES. 1984. InterpersonalPsychotherapy of Depression. Basic Books: New York.

Laidlaw L,Davidson K, TonerH, et al. 2008. A randomised controlled trial of cognitivebehaviour therapy vs. treatment as usual in the treatment of mild to moderate latelife depression. Int J Geriatr Psychiatr 23: 843850.

Leiper R. 2006. Psychodynamic formulation: a prince betrayed and disinherited. InFormulation in Psychology and Psychotherapy: Making Sense of Peoples Problems,Johnstone L, Dallos R (eds). Routledge: East Sussex.

Lewinsohn P, Graf M. 1973. Pleasant activities and depression. J Consult Clin Psychol41: 261268.

Li C, Ford E, Strine T, Mokdad A. 2008. Prevalence of depression amongst U.S. adultswith diabetes: findings from the 2006 behavioral risk factor survellience system.Diabetes Care 31: 105107.

Lichtenberg P, Kimbarrow M, Wall J, Roth R, MacNeil S. 1998. Depression in GeriatricMedical and Nursing Home Patients. A Treatment Manual. Wayne State UniversityPress: Detroit.

Lipsey M, Wilson D. 1993. The efficacy of psychological, educational and behaviouraltreatment. Am Psychol 48: 11811209.

MacPhillamy DJ, Lewinsohn PM. 1974. Depression as a function of levels of desiredand obtained pleasure. J Abnorm Psychol 83: 651657.

Martell C, Addis M, Jacobson N. 2001. Depression in Context: Strategies for GuidedAction. Norton: New York.

Mental Health and Older People Forum. 2008. A collective responsibility to act now onageing and mental health: a consensus statement. Accessed online on 21/08/09 from:www.mentalhealthequalities.org.uk/silo/files/consensus-statement august.pdf

Penninx BWJH, Beekman ATF, Honig A, et al. 2001. Depression and cardiacmortality. Results from a community-based longitudinal study. Arch Gen Psychiatr58: 221227.

Porter J, Spates R, Smitham S. 2004. Behavioural activation group therapy inpublic mental health settings: a pilot investigation. Prof Psychol: Res Pract 35:297301.

Quijano L, Stanley M, Petersen N, et al. 2007. Healthy IDEAS: a depression inter-vention delivered by community based case managers serving older adults. J ApplGerontol 26: 139156.

Quinn KM, Laidlaw K, Murray LK. 2009. Older peoples attitudes to mental illness.Clin Psychol Psychother 16: 3345.

Rogers C. 1961. On Becoming a Person. Houghton Mifflin: Boston, MA.

Rokke P, Tomhave J, Jocic Z. 1999. The role of client choice and target selectionin self-management therapy for depression in older adults. Psychol Aging14:155169.

Schulz KF, Chalmers I, Hayes RJ, Altman DG. 1995. Empirical evidence of bias.Dimensions of methodological quality associated with estimates of treatmenteffects in controlled trials. JAMA 273(5): 408412.

Scogin F, Jamison C, Gochneaur K. 1989. Comparative efficacy of cognitive andbehavioural bibliotherapy for mildly and moderately depressed older adults. JConsult Clin Psychol 57: 403407.

Sutton A, Abrams K, Jones D. 1998. Systematic reviews of trials and other studies.Health Technol Assess 2 (19).

Thomas C, Morris S. 2003. Cost of depression among adults in England in 2000. Br JPsychiatry183: 514519.

Thompson L, Gallagher D, Breckenridge J. 1987. Comparative effectiveness ofpsychotherapies for depressed elders. J Consult Clin Psychol 55: 385390.

Unutzer J. 2007. Late-life depression. N Engl J Med 357: 22692276.Unutzer J, Katon W, Callahan C, et al. 2002. Collaborative care management of late-

life depression in primary care setting: a randomised controlled trial. JAMA 288:

28362845.Von Korff M, Katon W, Lin EH, et al. 2005. Potentially modifiable factors associatedwith disability among people with diabetes. Psychosom Med 67: 233240.

Wong I, Campion P, Coulton S, et al. 2004. Pharmaceutical care for elderly patientsshared between community pharmacists and general practitioners. RESPECT(Randomised Evaluation of Shared Prescribing for Elderly people in the Com-munity over Time). BMC Health Serv Res 4(1): 11.

World Health Organisation. 2001. The World Health Report 2001: Mental Health: NewUnderstanding, New Hope. World Health Organisation: Geneva.

Yon A, Scogin F. 2009. Behavioural activation as a treatment for geriatric depression.Clin Gerontol 32: 91103.



7/29/2019 67057619

11/11

Copyright of International Journal of Geriatric Psychiatry is the property of John Wiley & Sons, Inc. and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.

Documents

67057619