Individual Mindfulness-BasedCognitive Therapy and CognitiveBehavior Therapy for TreatingDepressive Symptoms in PatientsWith Diabetes Results of aRandomized Controlled TrialDiabetes Care 2014372427ndash2434 | DOI 102337dc13-2918

OBJECTIVE

Depression is a common comorbidity of diabetes undesirably affecting patientsrsquophysical and mental functioning Psychological interventions are effective treat-ments for depression in the general population aswell as in patientswith a chronicdisease The aim of this study was to assess the efficacy of individual mindfulness-based cognitive therapy (MBCT) and individual cognitive behavior therapy (CBT) incomparisonwith awaiting-list control condition for treating depressive symptomsin adults with type 1 or type 2 diabetes

RESEARCH DESIGN AND METHODS

In this randomized controlled trial 94 outpatients with diabetes and comorbiddepressive symptoms (ie Beck Depression Inventory-II [BDI-II] Dagger14) were ran-domized to MBCT (n = 31) CBT (n = 32) or waiting list (n = 31) All participantscompleted written questionnaires and interviews at pre- and postmeasurement(3 months later) Primary outcomemeasure was severity of depressive symptoms(BDI-II and Toronto Hamilton Depression Rating Scale) Anxiety (Generalized Anx-iety Disorder 7) well-being (Well-Being Index) diabetes-related distress (ProblemAreas In Diabetes) and HbA1c levels were assessed as secondary outcomes

RESULTS

Results showed that participants receiving MBCT and CBT reported significantlygreater reductions in depressive symptoms compared with patients in the waiting-list control condition (respectively P = 0004 and P lt 0001 d = 080 and 100clinically relevant improvement 26 and 29 vs 4) Both interventions also hadsignificant positive effects on anxiety well-being and diabetes-related distress Nosignificant effect was found on HbA1c values

CONCLUSIONS

Both individual MBCT and CBT are effective in improving a range of psychologicalsymptoms in individuals with type 1 and type 2 diabetes

1Department of Health Psychology Univer-sity Medical Center Groningen University ofGroningen Groningen the Netherlands2Department of Internal Medicine HospitalRivierenland Tiel Tiel the Netherlands3Department of Clinical Psychology University ofAmsterdam Amsterdam the Netherlands4The Center for Social and Humanities ResearchKing Abdulaziz University Jeddah Saudi Arabia5Department of Psychology Health and Tech-nology University of Twente Enschede theNetherlands6Departmentof Endocrinology UniversityMedicalCenter Groningen University of GroningenGroningen the Netherlands

Corresponding author Maya J Schroeversmjschroeversumcgnl

Received 13 December 2013 and accepted 27April 2014

Clinical trial reg no NCT01630512 clinicaltrialsgov

This article contains Supplementary Data onlineat httpcarediabetesjournalsorglookupsuppldoi102337dc13-2918-DC1

A slide set summarizing this article is availableonline

copy 2014 by the American Diabetes AssociationReaders may use this article as long as the workis properly cited the use is educational and notfor profit and the work is not altered

K Annika Tovote1 Joke Fleer1

Evelien Snippe1 Anita CTM Peeters2

Paul MG Emmelkamp34

Robbert Sanderman15 Thera P Links6

and Maya J Schroevers1

Diabetes Care Volume 37 September 2014 2427

CLIN

CAREED

UCATIO

NN

UTR

ITIONPSYC

HOSO

CIAL

Depression is a common and persistentcondition in patients with diabetes withmajor depression diagnosed in 11 ofall patients and depressive symptomsin 31 (1) Alongside its contributionto decreased quality of life coexistingdepression in diabetes may represent agreat burden for patientsrsquo health andthe health care system Depression hasbeen related to lower adherenceto medication dietary and exercise rec-ommendations poorer glycemic con-trol and increased health care costs(23) The high comorbidity of depres-sion in diabetes and the potential nega-tive health consequences warrant theidentification of effective treatments toimprove patient functioning Both antide-pressant medication and psychologicaltreatment have been found effective fortreating depression (4) yet the latteris preferred by the majority of diabeticpatients (5)One potential effective psychological

treatment consists of mindfulness-based cognitive therapy (MBCT) In thelast decade application of MBCT for thetreatment of a wide variety of psycho-logical disorders including depressivesymptoms has grown exponentiallyMBCT focuses on cultivating mindful-ness which can be defined as beingaware of the present moment by meansof paying attention on purpose andwithout judgment (6) Several meta-analyses have demonstrated thatMBCT results in reduction of depressivesymptoms and increases in well-beingin a variety of populations (7ndash9) How-ever little is known about the applicabil-ity and effectiveness of MBCT in patientswith diabetes So far only one random-ized controlled trial (RCT) has investigatedthe effects of MBCT in patients with di-abetes demonstrating a reduction of de-pressive symptoms and anxiety as well asan increase in quality of life (10)Usually MBCT is delivered in a group

format Yet not all participants experi-ence this as beneficial (11) and a groupof patients prefers individual treatmentto group MBCT (12) This is in line with astudy on psychological treatment pref-erences in general demonstrating that70 of people preferred individualtreatment above group treatment (13)This motivated us to investigate theeffectiveness of an individual MBCTprogram In a pilot RCT we foundthat patients in the individual MBCT

condition showed greater reductionsin depressive symptoms and diabetes-related distress compared with a waiting-list condition (14) These positive resultswarrant further investigation of the effi-cacy of individual MBCT for patients withdiabetes which is the focus of the currenttrial

Recent reviews and meta-analysesstrongly recommend comparing MBCTnot only to a passive control group butalso to an active evidence-based inter-vention (1516) In the treatment of de-pressive symptoms the most frequentlyapplied and evidence-based treatmentis cognitive behavior therapy (CBT)(17) CBT has been related to significantimprovements in psychological symp-toms in patients with a diversity ofsomatic diseases especially when deliv-ered individually (18) Regarding CBT inpatients with diabetes and comorbid de-pression five RCTs have investigatedand demonstrated its efficacy (19ndash23)Thus CBT is characterized as the goldstandard against which to assess the ef-ficacy of a relatively new and promisingintervention like MBCT

As MBCT combines mindfulness withelements from CBT MBCT and CBT canbe regarded as related therapies Yetthe treatment components and overallaim of the interventions are distinctMBCT mainly involves practicing medi-tation and yoga exercises to increaseawareness and acceptance of dysfunc-tional thoughts and accompanying neg-ative emotions (6) CBT encouragespatients to maintain and increase thefrequency of pleasant activities and tolower negative mood by changing thecontent of dysfunctional thoughts intomore helpful thoughts (24) To dateonly one small RCT directly comparedgroup MBCT to group CBT in peoplewith depression demonstrating thatboth interventions were equally effica-cious (25) No RCT of CBT and MBCThas been conducted in patients withdiabetes

The purpose of the current Mood En-hancement Therapy Intervention Studywas to examine the effectiveness of in-dividual MBCT and CBT for depressivesymptoms in patients with diabetes incomparison with a waiting-list controlcondition We hypothesized that bothMBCT and CBT were more effectivethan a waiting-list control conditionwith neither MBCT nor CBT being

superior over the other The secondaryobjective was to investigate the effectsof MBCT and CBT in improving anxietywell-being and diabetes-related dis-tress In addition we explored the ef-fects of MBCT and CBT on glycemiccontrol as indicated by HbA1c valuesWhen proven efficacious individualMBCT can be established as a sound al-ternative to CBT for treating depressivesymptoms in patients with diabetes andthereby improving quality of psycho-logical care This availability of distinctevidence-based effective interventionsis particularly important given the findingthat preferences and attitudes towardtreatment can influence treatment out-come (26)

RESEARCH DESIGN AND METHODS

Study DesignThe Mood Enhancement Therapy Inter-vention Study is a multicenter RCT withthree conditions namely MBCT CBTand a waiting-list control conditionWe chose the latter control conditionrather than treatment as usual for ethi-cal reasons as all participants had ele-vated levels of depressive symptoms atrandomization The study protocol re-ceived ethical approval from the Medi-cal Ethical Committee of the UniversityMedical Center Groningen and was con-ducted in accordancewith the principlesof the Declaration of Helsinki (version2008) and the Medical Research Involv-ing Human Subjects Act A detaileddescription of the design has been pub-lished elsewhere (27)

ParticipantsEligible participants were patients withtype 1 or 2 diabetes diagnosed at least 3months prior to inclusion aged between18 and 70 years and having symptomsof depression as indicated by a Beck De-pression Inventory-II (BDI-II) score of$14 Exclusion criteria were not beingable to read and write Dutch preg-nancy severe psychiatric comorbidityacute suicidal ideations receiving an al-ternative psychological treatment dur-ing or 2 months prior to startingparticipation in the study and unstabletreatment with an antidepressant in thelast 2 months prior to inclusion in thestudy

ProcedurePatients were recruited from June 2011to February 2013 at four hospitals

2428 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37 September 2014

primarily in the northern part of theNetherlands (see Supplementary Datafor a list of participating investigators)Recruitment took place through stan-dard screening at outpatient clinicsreferral by physician and self-referralWhen patients had elevated levels ofdepressive symptoms they were invitedfor an intake during which they werescreened again and assessed for eligibil-ity Patients who fulfilled our criteriaand gave written informed consent forparticipation were included in thestudy Patients in the MBCT and CBTconditions were assessed before ran-domization and start of treatment (pre-measurement) and immediately afterending of treatment (postmeasure-ment on average 3 months after thefirst assessment) Patients assignedto the waiting-list control conditionundertook a baseline assessment (pre-measurement) and an assessment atthe end of the 3-month waiting period(postmeasurement)

RandomizationComputerized randomization was car-ried out stratified by sex use of antide-pressant medication and baseline BDI-IIscore Before randomization patientswere blinded for the treatment condi-tion Accordingly patients did not re-ceive any specific information aboutthe type of intervention or the waiting-list condition They were only told thatthey were to be randomized to a psy-chological treatment that focuses onreducing depression and that treatmentwas to start within 3 months afterrandomization

Interventions

Intervention Conditions

Both MBCT and CBT are protocolized in-terventions aimed at reducing depres-sive symptoms The treatments weredelivered individually in eight weeklysessions of 45ndash60 min Patients werealso instructed to do daily homeworkfor 30 min Both interventions were de-livered by trained therapists who re-ceived supervision every 3 weeksthroughout the intervention periodMBCT was based on the protocol as de-veloped by Segal et al (28) The centralcomponents of MBCT were formal med-itation yoga exercises and informaldaily mindfulness practices CBT wasbased on the protocol developed by

Beck et al (24) The main componentsof CBT were behavioral activation andcognitive restructuring A descriptionof the protocols can be found in Supple-mentary Table 1

To assess adherence all treatmentsessions of patients who providedconsent for this were videotaped Inaddition patients were asked to reporttheir daily homework practice onweekly evaluation forms Based on rat-ings of the videotaped sessions bytwo independent observers we foundthat therapistsrsquo adherence to the treat-ment manuals was sufficiently good(85 in MBCT and 83 in CBT) Alsopatientsrsquo homework compliance wassufficient (61 in MBCT and 79in CBT)

Waiting-List Condition

Participants in the waiting-list conditionreceived no psychological interventionfor 3 months

Assessments

Descriptive Measures

Data on the following demographicvariables were collected through self-report questionnaires age sex educa-tion marital status and occupationDisease-specific characteristics wereretrieved from patientsrsquo records namelytime since diagnosis type of diabetestreatment regimen comorbidities com-plications and BMI For 14 patients wecould not access the medical recordsand thus this information was retrievedfrom the questionnaire

Primary Outcome Measure

The primary outcome measure severityof depressive symptoms was assessedwith the BDI-II (29) The BDI-II is a 21-itemself-report questionnaire scored on afour-point scale ranging from 0 (ldquonot atallrdquo) to 3 (ldquomost of the timerdquo) Itmeasuressymptoms of depression such as sadnessloss of interest and hopelessness duringthe last 2 weeks A score from 14ndash19indicates mild depression a score from20ndash28 moderate depression and a score$29 indicates severe depression The re-liability of the BDI-II was good in the cur-rent study (a = 084)

In addition to the self-report depres-sion measurement and in order to as-sess depressive symptoms in a moreobjective manner symptoms were alsomeasured using the Toronto HamiltonDepression Rating Scale (HAM-D7) (30)

This semistructured clinical interviewwas administered by trained psycholo-gists at pre- and postmeasurement Atpremeasurement the assessors of theHAM-D7 were blinded to the treatmentcondition However at postmeasure-ment the HAM-D7 was administeredtogether with an evaluation of thetreatment for individuals randomizedto MBCT or CBT and therefore the as-sessors were not blinded The HAM-D7consists of seven items about depressedmood feelings of guilt and anxiety dur-ing the last week The items are scoredon a five-point scale ranging from 0ndash4(except for one item which ranges from0ndash2) A sum score of$4 represents milddepression a score between 12 and 20moderate depression and a score 20represents severe depression Reliabilityin the current study was acceptable (a =065)

Secondary Outcome Measures

TheWell-Being Index (WHO-5) was usedto assess emotional well-being (31) Thisself-report instrument consists of fiveitems that are scored on a six-point scalefrom 0 (ldquonot presentrdquo) to 5 (ldquoconstantlypresentrdquo) The items are about positivemood vitality and general interest inrelation to the last 2 weeks The totalsum score is converted to a score be-tween 0 and 100 with a score 50 in-dicating poor well-being In this studythe scalersquos reliability was good (a =082)

Anxiety was assessed by means of theGeneralized Anxiety Disorder 7 (GAD-7) a seven-item self-report instrument(32) Respondents are asked to reportthe frequency with which they experi-ence worrying and feeling restlessannoyed or afraid during the last 2weeks Each item is scored 0 (ldquonot atallrdquo) to 3 (ldquonearly every dayrdquo) A totalsum score of $5 indicates mildanxiety a score of 11ndash15 moderateanxiety and a score of 15 indicatessevere anxiety Cronbach a in this studywas good (a = 088)

The Problem Areas In Diabetes (PAID)was used to measure diabetes-relateddistress (3334) The PAID consists of20 items which are rated on a five-pointscale The scoring ranges from 0 (ldquonot aproblemrdquo) to 4 (ldquoserious problemrdquo) Theitems cover various common negativeemotions related to livingwith andman-aging diabetes The sum of all items is

carediabetesjournalsorg Tovote and Associates 2429

primarily in the northern part of theNetherlands (see Supplementary Datafor a list of participating investigators)Recruitment took place through stan-dard screening at outpatient clinicsreferral by physician and self-referralWhen patients had elevated levels ofdepressive symptoms they were invitedfor an intake during which they werescreened again and assessed for eligibil-ity Patients who fulfilled our criteriaand gave written informed consent forparticipation were included in thestudy Patients in the MBCT and CBTconditions were assessed before ran-domization and start of treatment (pre-measurement) and immediately afterending of treatment (postmeasure-ment on average 3 months after thefirst assessment) Patients assignedto the waiting-list control conditionundertook a baseline assessment (pre-measurement) and an assessment atthe end of the 3-month waiting period(postmeasurement)

RandomizationComputerized randomization was car-ried out stratified by sex use of antide-pressant medication and baseline BDI-IIscore Before randomization patientswere blinded for the treatment condi-tion Accordingly patients did not re-ceive any specific information aboutthe type of intervention or the waiting-list condition They were only told thatthey were to be randomized to a psy-chological treatment that focuses onreducing depression and that treatmentwas to start within 3 months afterrandomization

Interventions

Intervention Conditions

Both MBCT and CBT are protocolized in-terventions aimed at reducing depres-sive symptoms The treatments weredelivered individually in eight weeklysessions of 45ndash60 min Patients werealso instructed to do daily homeworkfor 30 min Both interventions were de-livered by trained therapists who re-ceived supervision every 3 weeksthroughout the intervention periodMBCT was based on the protocol as de-veloped by Segal et al (28) The centralcomponents of MBCT were formal med-itation yoga exercises and informaldaily mindfulness practices CBT wasbased on the protocol developed by

Beck et al (24) The main componentsof CBT were behavioral activation andcognitive restructuring A descriptionof the protocols can be found in Supple-mentary Table 1

To assess adherence all treatmentsessions of patients who providedconsent for this were videotaped Inaddition patients were asked to reporttheir daily homework practice onweekly evaluation forms Based on rat-ings of the videotaped sessions bytwo independent observers we foundthat therapistsrsquo adherence to the treat-ment manuals was sufficiently good(85 in MBCT and 83 in CBT) Alsopatientsrsquo homework compliance wassufficient (61 in MBCT and 79in CBT)

Waiting-List Condition

Participants in the waiting-list conditionreceived no psychological interventionfor 3 months

Assessments

Descriptive Measures

Data on the following demographicvariables were collected through self-report questionnaires age sex educa-tion marital status and occupationDisease-specific characteristics wereretrieved from patientsrsquo records namelytime since diagnosis type of diabetestreatment regimen comorbidities com-plications and BMI For 14 patients wecould not access the medical recordsand thus this information was retrievedfrom the questionnaire

Primary Outcome Measure

The primary outcome measure severityof depressive symptoms was assessedwith the BDI-II (29) The BDI-II is a 21-itemself-report questionnaire scored on afour-point scale ranging from 0 (ldquonot atallrdquo) to 3 (ldquomost of the timerdquo) Itmeasuressymptoms of depression such as sadnessloss of interest and hopelessness duringthe last 2 weeks A score from 14ndash19indicates mild depression a score from20ndash28 moderate depression and a score$29 indicates severe depression The re-liability of the BDI-II was good in the cur-rent study (a = 084)

In addition to the self-report depres-sion measurement and in order to as-sess depressive symptoms in a moreobjective manner symptoms were alsomeasured using the Toronto HamiltonDepression Rating Scale (HAM-D7) (30)

This semistructured clinical interviewwas administered by trained psycholo-gists at pre- and postmeasurement Atpremeasurement the assessors of theHAM-D7 were blinded to the treatmentcondition However at postmeasure-ment the HAM-D7 was administeredtogether with an evaluation of thetreatment for individuals randomizedto MBCT or CBT and therefore the as-sessors were not blinded The HAM-D7consists of seven items about depressedmood feelings of guilt and anxiety dur-ing the last week The items are scoredon a five-point scale ranging from 0ndash4(except for one item which ranges from0ndash2) A sum score of$4 represents milddepression a score between 12 and 20moderate depression and a score 20represents severe depression Reliabilityin the current study was acceptable (a =065)

Secondary Outcome Measures

TheWell-Being Index (WHO-5) was usedto assess emotional well-being (31) Thisself-report instrument consists of fiveitems that are scored on a six-point scalefrom 0 (ldquonot presentrdquo) to 5 (ldquoconstantlypresentrdquo) The items are about positivemood vitality and general interest inrelation to the last 2 weeks The totalsum score is converted to a score be-tween 0 and 100 with a score 50 in-dicating poor well-being In this studythe scalersquos reliability was good (a =082)

Anxiety was assessed by means of theGeneralized Anxiety Disorder 7 (GAD-7) a seven-item self-report instrument(32) Respondents are asked to reportthe frequency with which they experi-ence worrying and feeling restlessannoyed or afraid during the last 2weeks Each item is scored 0 (ldquonot atallrdquo) to 3 (ldquonearly every dayrdquo) A totalsum score of $5 indicates mildanxiety a score of 11ndash15 moderateanxiety and a score of 15 indicatessevere anxiety Cronbach a in this studywas good (a = 088)

The Problem Areas In Diabetes (PAID)was used to measure diabetes-relateddistress (3334) The PAID consists of20 items which are rated on a five-pointscale The scoring ranges from 0 (ldquonot aproblemrdquo) to 4 (ldquoserious problemrdquo) Theitems cover various common negativeemotions related to livingwith andman-aging diabetes The sum of all items is

carediabetesjournalsorg Tovote and Associates 2429

transformed into a scale from 0ndash100with scores of$40 being used to definepatients at risk for high diabetes-relateddistress The internal consistency wasexcellent in the current study (a = 095)Finally glycemic control indicated

by HbA1c values was retrieved frompatientsrsquo records As premeasurementthe average of all assembled valuesof 0ndash6 months prior to interventionwas used and as postmeasurementwe used the average of all valuesbetween 1 and 6 months after theintervention

Sample Size CalculationThe sample size calculation was basedon expected differences in the level ofposttreatment depressive symptomsbetween the waiting-list control groupand eitherMBCT or CBT Assuming a sta-tistical power of 08 and an a of 005 42participants were required in eachgroup enabling us to detect differenceswith an effect size of 06 (35)

Statistical AnalysesAll analyses were performed based onthe intention-to-treat method Missingvalues were estimated by means ofmultiple imputations using the linear re-gression method We performed sensi-tivity analyses based on participantswith no missing data and datasets with5 and 20 imputations As analysesrevealed a similar pattern of results5 complete datasets were imputed after20 iterations SPSS Statistics 20 (SPSSInc) was used for all analyses and allaccording assumptions were metANOVAs and x2 tests were used to ana-lyze if there were differences at baselinebetween the groups regarding demo-graphic and clinical variables as well asprimary and secondary outcome mea-sures Separate ANCOVAs were per-formed for MBCT and CBT to examinethe effects of the interventions in com-parison with the waiting-list conditionPostmeasurement values of the primaryand secondary outcomes were used asdependent variables condition wasused as factor and premeasurementsof the outcomes were used as covariateBetween-group effect sizes were calcu-lated using Cohen d with values rangingfrom 02 to 05 indicating small effectsvalues from 05 to 08 indicating moder-ate effects and values08 indicatinglarge effects (36) Clinically relevant

improvement was defined as havingimproved and being recovered A post-measurement score below the cutoffof the primary outcome measure (ieBDI-II14) indicated improvement Re-covery was calculated by the ReliableChange Index which refers to the differ-ence between an individualrsquos pre- andpostmeasurement scores divided bythe SE of the difference A score 196indicates recovery (37)

RESULTS

Recruitment and AttritionAs is shown in Fig 1 3145 patients wereroutinely screened at a hospital and14 referred themselves in awarenessof their treating physician Of the2266 patients who completed andreturned the screening questionnaire613 (27) had an elevated score (BDI-II$14) Less than half (n = 255) acceptedthe invitation for a face-to-face intakeAn additional six patients were referredby their physician for an intake Duringthe intake patients were screened againand elaborately assessed for eligibilityAlmost one-third of the patients (n = 78)who received an intake were not eligiblefor the trial and an additional one-third(n = 89) did not agree to participatemostly because they did not feel theneed for treatment Finally 94 patientsgave consent and were randomized31 participants to MBCT 32 partici-pants to CBT and 31 participants tothe waiting-list control condition Inboth MBCT and CBT nine patients didnot finish the intervention (ie receivedless than six sessions) Reasons for drop-out were intervention content related(MBCT n = 4 CBT n = 3) lack of time(MBCT n = 3 CBT n = 2) severe illness(MBCT n = 1 CBT n = 2) improvementof depression after a few sessions(MBCT n = 1 CBT n = 1) and no interestin participating in research anymore(CBT n = 1) Two participants in theMBCT condition and four participantsin both CBT and the waiting-list condi-tion did not fill in the postmeasurementquestionnaire

Baseline CharacteristicsTable 1 provides an overview of thebaseline characteristics of the partici-pants There were no statistically signif-icant differences among the threeconditions regarding the demographicor clinical baseline characteristics as

well as primary and secondary out-comes measures (averages shown inTable 2)

Primary Outcome MeasuresThe mean scores and the outcomes ofthe statistical analyses are presented inTable 2 When comparing MBCT andCBT to the control condition both inter-vention groups had significantly less de-pressive symptoms than the controlgroup at postmeasurement (P = 0004and P 0001 respectively) The effectsizes of the change from pre- to post-measurement between MBCT and CBTversus thewaiting list were large (Cohend = 080 and d = 100 respectively)Given the difference in effect sizesbetween the two interventions wealso compared effects of the MBCT andCBT group directly and found no signifi-cant differences (P = 034 not shown inTable 2) Assessing depressive symptomswith the HAM-D7 revealed similar re-sults both MBCT and CBT had signifi-cantly higher outcome improvementthan the waiting-list condition (P 0001 and P = 0001 respectively) Thebetween-group effect sizes in comparisonwith thewaiting list were large (MBCT d =117 CBT d = 109)

Secondary Outcome MeasuresThe results of the secondary outcomemeasures are also presented in Table 2When comparing MBCT and CBT withthe waiting list individuals in bothMBCT and CBT had a larger improve-ment in levels of well-being (both P 0001) anxiety (P = 0004 and P = 001respectively) and diabetes-related dis-tress (P = 002 and P = 004 respec-tively) Between-group effect sizeswere large for well-being and anxiety(range Cohen d = 082ndash097) and mod-erate for diabetes-related distress (d =052 and d = 057) HbA1c levels did notchange after MBCT or CBT (P = 092 andP = 072 respectively)

Clinically Relevant ImprovementClinically relevant improvement wasfound in 26 of the participants afterMBCT and 29 of the participants afterCBT versus 4 of the patients in thewaiting-list condition When comparingthe percentages in the intervention con-ditions to the control condition the dif-ferences were significant (MBCT vswaiting list P = 002 CBT vs waitinglist P = 0009)

2430 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37 September 2014

CONCLUSIONS

This is the first RCT that investigated theeffectiveness of individually deliveredMBCT and CBT in reducing depressivesymptoms in outpatients with type 1and type 2 diabetes Concordant withour hypothesis results indicated thatboth MBCT and CBT were effective in re-ducing depressive symptoms comparedwith a waiting-list control conditionwith neither MBCT nor CBT being supe-rior over the other MBCT and CBT werealso effective in improving a wider rangeof patient-relevant outcomes includingincreases in well-being and reductions in

anxiety and diabetes-related distress Noeffects were found for HbA1c values

Given the high prevalence and bur-den of depressive symptoms in patientswith diabetes a key finding of this RCTis that patients receiving one of thepsychological interventions reportedgreater reductions in depressive symp-toms in comparison with the controlcondition Our findings are in line withManicavasgar et al (25) demonstratingeffectiveness of group MBCT and CBT inreducing depression These and ourfindings show that CBT is effective butnot superior to some other active

treatments (38) Our results are innova-tive as this is the first RCT study on theeffectiveness of the individual deliveryof MBCT with currently only evidencefor the effectiveness of group-basedMBCT We are aware that our resultsshould be replicated to draw more firmconclusions Yet it is promising giventhe well-known effectiveness of CBT es-pecially when individually delivered(18) that individual MBCT was as effec-tive Taking into account the differencesin treatment focus and components ofMBCT and CBT our results imply thattwo evidence-based distinct types of

Figure 1mdashParticipant recruitment and flow through the study

carediabetesjournalsorg Tovote and Associates 2431

CONCLUSIONS

This is the first RCT that investigated theeffectiveness of individually deliveredMBCT and CBT in reducing depressivesymptoms in outpatients with type 1and type 2 diabetes Concordant withour hypothesis results indicated thatboth MBCT and CBT were effective in re-ducing depressive symptoms comparedwith a waiting-list control conditionwith neither MBCT nor CBT being supe-rior over the other MBCT and CBT werealso effective in improving a wider rangeof patient-relevant outcomes includingincreases in well-being and reductions in

anxiety and diabetes-related distress Noeffects were found for HbA1c values

Given the high prevalence and bur-den of depressive symptoms in patientswith diabetes a key finding of this RCTis that patients receiving one of thepsychological interventions reportedgreater reductions in depressive symp-toms in comparison with the controlcondition Our findings are in line withManicavasgar et al (25) demonstratingeffectiveness of group MBCT and CBT inreducing depression These and ourfindings show that CBT is effective butnot superior to some other active

treatments (38) Our results are innova-tive as this is the first RCT study on theeffectiveness of the individual deliveryof MBCT with currently only evidencefor the effectiveness of group-basedMBCT We are aware that our resultsshould be replicated to draw more firmconclusions Yet it is promising giventhe well-known effectiveness of CBT es-pecially when individually delivered(18) that individual MBCT was as effec-tive Taking into account the differencesin treatment focus and components ofMBCT and CBT our results imply thattwo evidence-based distinct types of

Figure 1mdashParticipant recruitment and flow through the study

carediabetesjournalsorg Tovote and Associates 2431

psychological interventions can beoffered to patients with diabetes Animportant next step would be to inves-tigate possible moderators of effective-ness that is factors related to thedifferential effectiveness of MBCT andCBTwithin certain subgroups of patients(ie for whom is which interventionmore beneficial)Besides depressive symptoms we

were also interested in a possible widereffect of MBCT and CBT on other indica-tors of functioning It is clinically rele-vant to observe that both MBCT andCBT significantly increase well-beingand reduce anxiety and diabetes-relateddistress These findings are consistentwith previous research showing thatpsychological interventions focusing ondepressive symptoms can also improveanxiety and quality of life (1020) Re-sults are also in line with previousstudies in patients with diabetes

investigating either MBCT or CBT alsoshowing reductions in diabetes-relateddistress (1419) Taken together MBCTand CBT not only reduce depressivesymptoms but also improve other psy-chological outcomes

Explorative analysis showed no signif-icant reductions in HbA1c values eitherin MBCT or in CBT This finding is con-cordant with two previous RCTs onMBCT (10) and CBT (19) that did notfind an effect on glycemic control Arecent review andmeta-analysis studiedthe impact of psychosocial interventionson both psychological and physicalhealth in patients with diabetes (39)No interventions were identified thatwere effective for bothmedical andmen-tal outcomes at the same time Alto-gether our findings and previous resultssuggest that alleviating depressive symp-toms through psychological interventionslike MBCT or CBT does not automatically

translate into improved self-care andsubsequent glycemic control (4)

A methodological challenge in the in-vestigation of improvements in HbA1c

levels is that the HbA1c level is an aver-age value over the previous 3months Inthis study HbA1c was only included forexploratory reasons in order to burdenthe patients as little as possible There-fore HbA1c values were obtained frompatientsrsquo medical records instead ofscheduling additional measurements atdesignated time points Consequentlyour HbA1c values are crude indicationsof HbA1c values in themonths precedingand following the two active interven-tions Also as patients in the waiting-list condition received care directly afterpatients in the active conditions had fin-ished the intervention it was not possi-ble to compare CBT and MBCT with thecontrol condition regarding changes inHbA1c values

Table 1mdashBaseline characteristics

MBCT (n = 31) CBT (n = 32) Waiting list (n = 31) Total (n = 94)

Age (years) mean (SD) 498 (133) 546 (113) 547 (105) 531 (118)

Sex n ()Male 17 (55) 16 (50) 15 (48) 48 (51)Female 14 (45) 16 (50) 16 (52) 46 (49)

Education n ()Lower level vocational school 8 (26) 10 (31) 5 (16) 23 (25)Secondary educationadvanced level vocational school 14 (45) 15 (47) 18 (58) 47 (50)Higher or university education 9 (29) 7 (22) 8 (26) 24 (25)

Employment n ()Employed 16 (52) 15 (47) 21 (68) 52 (55)Not employed 15 (48) 17 (53) 10 (32) 42 (45)

Relationship status n ()In a relationship 24 (77) 22 (69) 21 (68) 67 (71)Not in a relationship 7 (23) 10 (31) 10 (32) 27 (29)

BMI mean (SD) 293 (76) 319 (66) 306 (84) 306 (76)

Type of diabetes n ()Type 1 15 (48) 11 (34) 11 (36) 37 (39)Type 2 16 (52) 21 (66) 20 (65) 57 (61)

Diabetes treatment n ()Oral medication 4 (13) 4 (12) 4 (13) 12 (13)Oral medication and insulin 10 (32) 14 (44) 11 (36) 35 (37)Insulin 17 (55) 14 (44) 16 (51) 47 (50)

Time since diagnosis (years) mean (SD) 178 (130) 150 (114) 170 (114) 166 (119)

Diabetes complications n ()daggerOne or more complications 9 (29) 13 (40) 9 (29) 31 (33)No complications 22 (71) 19 (60) 22 (71) 63 (77)

Comorbidity n ()One or more comorbidities 14 (45) 18 (56) 18 (58) 50 (53)No comorbidity 17 (55) 14 (44) 13 (42) 44 (47)

Antidepressant use at trial entry n ()Usage 2 (7) 5 (16) 3 (10) 10 (11)No usage 29 (93) 27 (84) 28 (90) 84 (89)

Groups did not significantly differ (P 005 in all cases) on any of the demographics and clinical characteristics daggerIncluded diabetes complicationsare retinopathy neuropathy nephropathy and diabetic foot

2432 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37 September 2014

LimitationsAlthough we carefully designed ourstudy several limitations to this studyneed to be acknowledged First wewere not able to reach a fully poweredsample of at least 42 participants percondition as patient recruitment tookmore time than originally planned Yetfor the actual sample size the powerwas still 68 Second although the ma-jority of the patients were recruited as aconsecutive sample (ie screening n =79) a small group of participants wasrecruited as a convenience samplebased on (self) referral (n = 15) Theformer may not be representativeof treatment-seeking or clinicallyreferred patients while the latter sam-ple may suffer from selection biashereby reducing generalizability of theresults Third attrition rates in bothMBCT and CBT were high as only70 of the randomized participantscompleted treatment These attritionrates are consistent with previous stud-ies targeting distressed patients withdiabetes (1019) Screening a consecu-tive sample may have accountedfor the dropout rate as the majorityof participants did not seek treat-ment themselves but instead were ap-proached and offered treatmentFourth a substantial group of patientswho met the inclusion criteria refused

to participate in the study because of noneed for psychological treatment Ourinclusion rate is comparable to otherstudies using a consecutive samplemethod (1040) As patients wereblinded to the content of the treatmentwe do not assume that refusal to partic-ipation was content related Fifth thefact that assessors of depressive symp-toms with the clinical interview at post-measurementwere not blindedmayhavebiased their ratings Finally as all partic-ipants had elevated levels of depressivesymptoms at randomization for ethicalreasons we included a waiting-list con-trol condition rather than treatmentas usual As patients in the controlcondition received one of the interven-tions after the 3-month waiting periodlong-term effects of CBT andMBCT com-pared with the control condition couldnot be assessed

ConclusionThis is the first RCT examining the ef-fectiveness of individual MBCT and in-dividual CBT in reducing depressivesymptoms in patients with diabetesResults clearly suggest that MBCT aswell as CBT are effective interventionsin treating depressive symptoms inpatients with diabetes Given theireffectiveness and the fact that both in-terventions are short structured

8-week interventions delivered onan individual basis they could be im-plemented in optimizing psychologicalcare for depressed patients withdiabetes

Acknowledgments The authors thank all ofthe patients who participated in the study thepsychologists who delivered the MBCT and CBTsessions the secretaries and research assistantsof the University Medical Center Groningen theMartini Hospital Groningen the Medical CenterLeeuwarden and the Hospital Rivierenland Tielfor the effortsFunding This study was financed by theUniversity of GroningenDuality of Interest No potential conflicts ofinterest relevant to this article were reportedAuthor Contributions KAT constructed thedesign of the study researched data and wrotethe manuscript JF and MJS constructed thedesign of the study and revised the manuscriptES constructed the design of the study re-searched data and reviewed the manuscriptACTMP reviewed the manuscript PMGEand RS constructed the design of the studyand reviewed the manuscript TPL participatedin the design of the study and reviewed themanuscript KAT is the guarantor of this workand as such had full access to all the data in thestudy and takes responsibility for the integrity ofthe data and the accuracy of the data analysis

References1 Anderson RJ Freedland KE Clouse RELustman PJ The prevalence of comorbid depres-sion in adults with diabetes a meta-analysis Di-abetes Care 2001241069ndash1078

Table 2mdashResults for primary and secondary outcomes

Measure GroupPremeasurement

mean (SD)Postmeasurement

mean (SD)

Time Treatmentdagger

t P value F P value d

Depression (BDI-II) MBCT 236 (77) 171 (119) 475 0001 971 0004 080 (027ndash131)CBT 256 (87) 174 (119) 558 0001 1556 0001 100 (047ndash151)WAIT 243 (80) 235 (103) 065 052

Depression (HAM-D7) MBCT 89 (35) 47 (43) 633 0001 1752 0001 117 (061ndash169)CBT 94 (38) 46 (34) 555 0001 1306 0001 109 (055ndash160)WAIT 75 (28) 71 (37) 071 049

Well-being (WHO-5) MBCT 324 (184) 495 (215) 590 0001 1735 0001 092 (039ndash143)CBT 268 (178) 474 (202) 607 0001 1895 0001 102 (048ndash153)WAIT 277 (159) 309 (154) 109 028

Anxiety (GAD-7) MBCT 126 (53) 69 (48) 700 0001 960 0004 098 (044ndash149)CBT 119 (49) 68 (50) 661 0001 742 001 082 (029ndash132)WAIT 98 (50) 82 (46) 237 002

Diabetes distress (PAID) MBCT 383 (209) 320 (218) 308 0002 567 002 052 (001ndash102)CBT 420 (223) 340 (234) 287 0004 568 004 057 (006ndash107)WAIT 355 (215) 360 (212) 2026 079

HbA1cmmolmol MBCT 634 (96) 631 (108) 010 092 80 (09) 79 (10)mmolmol CBT 671 (152) 659 (130) 036 072 83 (14) 82 (12)

WAIT waiting list Not measured in the waiting-list condition daggerComparing both MBCT and CBT with the waiting-list condition

carediabetesjournalsorg Tovote and Associates 2433

LimitationsAlthough we carefully designed ourstudy several limitations to this studyneed to be acknowledged First wewere not able to reach a fully poweredsample of at least 42 participants percondition as patient recruitment tookmore time than originally planned Yetfor the actual sample size the powerwas still 68 Second although the ma-jority of the patients were recruited as aconsecutive sample (ie screening n =79) a small group of participants wasrecruited as a convenience samplebased on (self) referral (n = 15) Theformer may not be representativeof treatment-seeking or clinicallyreferred patients while the latter sam-ple may suffer from selection biashereby reducing generalizability of theresults Third attrition rates in bothMBCT and CBT were high as only70 of the randomized participantscompleted treatment These attritionrates are consistent with previous stud-ies targeting distressed patients withdiabetes (1019) Screening a consecu-tive sample may have accountedfor the dropout rate as the majorityof participants did not seek treat-ment themselves but instead were ap-proached and offered treatmentFourth a substantial group of patientswho met the inclusion criteria refused

to participate in the study because of noneed for psychological treatment Ourinclusion rate is comparable to otherstudies using a consecutive samplemethod (1040) As patients wereblinded to the content of the treatmentwe do not assume that refusal to partic-ipation was content related Fifth thefact that assessors of depressive symp-toms with the clinical interview at post-measurementwere not blindedmayhavebiased their ratings Finally as all partic-ipants had elevated levels of depressivesymptoms at randomization for ethicalreasons we included a waiting-list con-trol condition rather than treatmentas usual As patients in the controlcondition received one of the interven-tions after the 3-month waiting periodlong-term effects of CBT andMBCT com-pared with the control condition couldnot be assessed

ConclusionThis is the first RCT examining the ef-fectiveness of individual MBCT and in-dividual CBT in reducing depressivesymptoms in patients with diabetesResults clearly suggest that MBCT aswell as CBT are effective interventionsin treating depressive symptoms inpatients with diabetes Given theireffectiveness and the fact that both in-terventions are short structured

8-week interventions delivered onan individual basis they could be im-plemented in optimizing psychologicalcare for depressed patients withdiabetes

Acknowledgments The authors thank all ofthe patients who participated in the study thepsychologists who delivered the MBCT and CBTsessions the secretaries and research assistantsof the University Medical Center Groningen theMartini Hospital Groningen the Medical CenterLeeuwarden and the Hospital Rivierenland Tielfor the effortsFunding This study was financed by theUniversity of GroningenDuality of Interest No potential conflicts ofinterest relevant to this article were reportedAuthor Contributions KAT constructed thedesign of the study researched data and wrotethe manuscript JF and MJS constructed thedesign of the study and revised the manuscriptES constructed the design of the study re-searched data and reviewed the manuscriptACTMP reviewed the manuscript PMGEand RS constructed the design of the studyand reviewed the manuscript TPL participatedin the design of the study and reviewed themanuscript KAT is the guarantor of this workand as such had full access to all the data in thestudy and takes responsibility for the integrity ofthe data and the accuracy of the data analysis

References1 Anderson RJ Freedland KE Clouse RELustman PJ The prevalence of comorbid depres-sion in adults with diabetes a meta-analysis Di-abetes Care 2001241069ndash1078

Table 2mdashResults for primary and secondary outcomes

Measure GroupPremeasurement

mean (SD)Postmeasurement

mean (SD)

Time Treatmentdagger

t P value F P value d

Depression (BDI-II) MBCT 236 (77) 171 (119) 475 0001 971 0004 080 (027ndash131)CBT 256 (87) 174 (119) 558 0001 1556 0001 100 (047ndash151)WAIT 243 (80) 235 (103) 065 052

Depression (HAM-D7) MBCT 89 (35) 47 (43) 633 0001 1752 0001 117 (061ndash169)CBT 94 (38) 46 (34) 555 0001 1306 0001 109 (055ndash160)WAIT 75 (28) 71 (37) 071 049

Well-being (WHO-5) MBCT 324 (184) 495 (215) 590 0001 1735 0001 092 (039ndash143)CBT 268 (178) 474 (202) 607 0001 1895 0001 102 (048ndash153)WAIT 277 (159) 309 (154) 109 028

Anxiety (GAD-7) MBCT 126 (53) 69 (48) 700 0001 960 0004 098 (044ndash149)CBT 119 (49) 68 (50) 661 0001 742 001 082 (029ndash132)WAIT 98 (50) 82 (46) 237 002

Diabetes distress (PAID) MBCT 383 (209) 320 (218) 308 0002 567 002 052 (001ndash102)CBT 420 (223) 340 (234) 287 0004 568 004 057 (006ndash107)WAIT 355 (215) 360 (212) 2026 079

HbA1cmmolmol MBCT 634 (96) 631 (108) 010 092 80 (09) 79 (10)mmolmol CBT 671 (152) 659 (130) 036 072 83 (14) 82 (12)

WAIT waiting list Not measured in the waiting-list condition daggerComparing both MBCT and CBT with the waiting-list condition

carediabetesjournalsorg Tovote and Associates 2433

2 Egede LE Ellis C Diabetes and depressionglobal perspectives Diabetes Res Clin Pract201087302ndash3123 Katon WJ The comorbidity of diabetes mel-litus and depression Am JMed 2008121(Suppl2)S8ndashS154 van der Feltz-Cornelis CM Nuyen J Stoop Cet al Effect of interventions for major depres-sive disorder and significant depressive symp-toms in patients with diabetes mellitusa systematic review and meta-analysis GenHosp Psychiatry 201032380ndash3955 Dwight-Johnson M Sherbourne CD Liao DWells KB Treatment preferences among de-pressed primary care patients J Gen InternMed 200015527ndash5346 Kabat-Zinn J Mindfulness-based interven-tions in context Past present and future ClinPsychol Sci Pract 200310144ndash1567 Hofmann SG Sawyer AT Witt AA Oh D Theeffect of mindfulness-based therapy on anxietyand depression A meta-analytic review J Con-sult Clin Psychol 201078169ndash1838 Khoury B Lecomte T Fortin G et alMindfulness-based therapy a comprehensivemeta-analysis Clin Psychol Rev 201333763ndash7719 Chiesa A Serretti A Mindfulness based cog-nitive therapy for psychiatric disorders a sys-tematic review and meta-analysis PsychiatryRes 2011187441ndash45310 van Son J Nyklıcek I Pop VJ et al The ef-fects of a mindfulness-based intervention onemotional distress quality of life and HbA(1c)in outpatients with diabetes (DiaMind) a ran-domized controlled trial Diabetes Care 201336823ndash83011 Griffiths K Camic PM Hutton JM Partici-pant experiences of a mindfulness-based cogni-tive therapy group for cardiac rehabilitation JHealth Psychol 200914675ndash68112 Lau MA Colley L Willett BR Lynd LD Em-ployeersquos preferences for access to mindfulness-based cognitive therapy to reduce the risk ofdepressive relapsedA discrete choice experi-ment Mindfulness 20123318ndash32613 Lang AJ Mental health treatment prefer-ences of primary care patients J Behav Med200528581ndash58614 Schroevers MJ Tovote KA Keers JC Links TPSanderman R Fleer J Individual mindfulness-based cognitive therapy for people with diabetesA pilot randomized controlled trial Mindfulness14 August 2013 [Epub ahead of print]

15 Galante J Iribarren SJ Pearce PF Effects ofmindfulness-based cognitive therapy on mentaldisorders A systematic review and meta-analysisof randomised controlled trials J Res Nurs 201318133ndash15516 Coelho HF Canter PH Ernst E Mindfulness-based cognitive therapy evaluating current ev-idence and informing future research J ConsultClin Psychol 2007751000ndash100517 Butler AC Chapman JE Forman EM BeckAT The empirical status of cognitive-behavioraltherapy a review of meta-analyses Clin PsycholRev 20062617ndash3118 Beltman MW Voshaar RCO Speckens AECognitive-behavioural therapy for depression inpeople with a somatic disease meta-analysis ofrandomised controlled trials Br J Psychiatry201019711ndash1919 van Bastelaar KM Pouwer F Cuijpers PRiper H Snoek FJ Web-based depression treat-ment for type 1 and type 2 diabetic patientsa randomized controlled trial Diabetes Care201134320ndash32520 Penckofer SM Ferrans C Mumby P et al Apsychoeducational intervention (SWEEP) for de-pressed women with diabetes Ann Behav Med201244192ndash20621 Lamers F Jonkers CCM Bosma H et al Aminimal psychological intervention in chroni-cally ill elderly patients with depression a ran-domized trial Psychother Psychosom 201079217ndash22622 Lustman PJ Griffith LS Freedland KE KisselSS Clouse RE Cognitive behavior therapy fordepression in type 2 diabetes mellitus A ran-domized controlled trial Ann Intern Med 1998129613ndash62123 Safren SA Gonzalez JS Wexler DJ et al Arandomized controlled trial of cognitive behav-ioral therapy for adherence and depression(CBT-AD) in patients with uncontrolled type 2diabetes Diabetes Care 201437625ndash63324 Beck AT Rush AJ Shaw BF Emery G Cog-nitive Therapy of Depression New York Guil-ford 197925 Manicavasgar V Parker G Perich TMindfulness-based cognitive therapy vs cogni-tive behaviour therapy as a treatment for non-melancholic depression J Affect Disord 2011130138ndash14426 Swift JK Callahan JL Vollmer BM Prefer-ences J Clin Psychol 201167155ndash16527 Tovote KA Fleer J Snippe E et al Cognitivebehavioral therapy and mindfulness-based

cognitive therapy for depressive symptoms inpatients with diabetes design of a randomizedcontrolled trial BMC Psychol 201311728 Segal ZV Williams MG Teasdale JDMindfulness-Based Cognitive Therapy for De-pression A New Approach to Preventing Re-lapse New York Guilford Press 200229 Beck AT Steer RA Ball R Ranieri WF Com-parison of Beck Depression Inventories-IAand -II in psychiatric outpatients J Pers Assess199667588ndash59730 McIntyre RS Konarski JZ Mancini DA et alMeasuring the severity of depression and remis-sion in primary care validation of the HAMD-7scale CMAJ 20051731327ndash133431 Bech P Measuring the dimension of psy-chological general well-being by the WHO-5QOL Newsletter 20043215ndash1632 Spitzer RL Kroenke K Williams JBW LoweB A brief measure for assessing generalizedanxiety disorder the GAD-7 Arch Intern Med20061661092ndash109733 PolonskyWH Anderson BJ Lohrer PA et alAssessment of diabetes-related distress Diabe-tes Care 199518754ndash76034 Welch GW Jacobson AM Polonsky WHThe Problem Areas in Diabetes Scale An evalu-ation of its clinical utility Diabetes Care 199720760ndash76635 Cohen J A power primer In Methodologi-cal Issues amp Strategies in Clinical Research 3rded Kazdin AE Ed Washington DC AmericanPsychological Association 2003 p 427-43636 Cohen J Statistical Power Analysis for theBehavioral Sciences New York L Erlbaum Asso-ciates 197737 Jacobson NS Truax P Clinical significancea statistical approach to defining meaningfulchange in psychotherapy research J ConsultClin Psychol 19915912ndash1938 Tolin DF Is cognitive-behavioral therapymore effective than other therapies A meta-analytic review Clin Psychol Rev 201030710ndash72039 Harkness E Macdonald W Valderas JCoventry P Gask L Bower P Identifying psycho-social interventions that improve both physicaland mental health in patients with diabetesa systematic review and meta-analysis Diabe-tes Care 201033926ndash93040 Strong V Waters R Hibberd C et al Man-agement of depression for people with cancer(SMaRT oncology 1) a randomised trial Lancet200837240ndash48

2434 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37 September 2014