See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/48872638

The Irrational Beliefs Inventory (IBI): Development and psychometric

evaluation

Article  in  European Journal of Psychological Assessment · January 1994

DOI: 10.1037/t15478-000 · Source: OAI

CITATIONS READS

72 1,950

4 authors, including:

Robbert Sanderman Paul Emmelkamp

University of Groningen University of Amsterdam
488 PUBLICATIONS   20,118 CITATIONS    614 PUBLICATIONS   21,905 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Virtual reality View project

Blended Smoking Cessation Treatment View project

All content following this page was uploaded by Robbert Sanderman on 14 June 2016.

The user has requested enhancement of the downloaded file.

Diabetes Care Volume 37, September 2014 2427

K. Annika Tovote,1 Joke Fleer,1

Individual Mindfulness-Based Evelien Snippe,1 Anita C.T.M. Peeters,2

CLIN CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL
Paul M.G. Emmelkamp,3,4
Cognitive Therapy and Cognitive Robbert Sanderman,1,5 Thera P. Links,6
and Maya J. Schroevers1
Behavior Therapy for Treating
Depressive Symptoms in Patients
With Diabetes: Results of a
Randomized Controlled Trial
Diabetes Care 2014;37:2427–2434 | DOI: 10.2337/dc13-2918

OBJECTIVE
Depression is a common comorbidity of diabetes, undesirably affecting patients’
physical and mental functioning. Psychological interventions are effective treat-
ments for depression in the general population as well as in patients with a chronic
1
disease. The aim of this study was to assess the efficacy of individual mindfulness- Department of Health Psychology, Univer-
sity Medical Center Groningen, University of
based cognitive therapy (MBCT) and individual cognitive behavior therapy (CBT) in Groningen, Groningen, the Netherlands
comparison with a waiting-list control condition for treating depressive symptoms 2
Department of Internal Medicine, Hospital
in adults with type 1 or type 2 diabetes. Rivierenland Tiel, Tiel, the Netherlands
3
Department of Clinical Psychology, University of
RESEARCH DESIGN AND METHODS Amsterdam, Amsterdam, the Netherlands
4
The Center for Social and Humanities Research,
In this randomized controlled trial, 94 outpatients with diabetes and comorbid King Abdulaziz University, Jeddah, Saudi Arabia
depressive symptoms (i.e., Beck Depression Inventory-II [BDI-II] ‡14) were ran- 5
Department of Psychology, Health and Tech-
domized to MBCT (n = 31), CBT (n = 32), or waiting list (n = 31). All participants nology, University of Twente, Enschede, the
completed written questionnaires and interviews at pre- and postmeasurement Netherlands
6
Department of Endocrinology, University Medical
(3 months later). Primary outcome measure was severity of depressive symptoms Center Groningen, University of Groningen,
(BDI-II and Toronto Hamilton Depression Rating Scale). Anxiety (Generalized Anx- Groningen, the Netherlands
iety Disorder 7), well-being (Well-Being Index), diabetes-related distress (Problem Corresponding author: Maya J. Schroevers,
Areas In Diabetes), and HbA1c levels were assessed as secondary outcomes. [email protected].
Received 13 December 2013 and accepted 27
RESULTS April 2014.
Results showed that participants receiving MBCT and CBT reported significantly Clinical trial reg. no. NCT01630512, clinicaltrials
greater reductions in depressive symptoms compared with patients in the waiting- .gov.
list control condition (respectively, P = 0.004 and P < 0.001; d = 0.80 and 1.00; This article contains Supplementary Data online
clinically relevant improvement 26% and 29% vs. 4%). Both interventions also had at http://care.diabetesjournals.org/lookup/
suppl/doi:10.2337/dc13-2918/-/DC1.
significant positive effects on anxiety, well-being, and diabetes-related distress. No
A slide set summarizing this article is available
significant effect was found on HbA1c values. online.
CONCLUSIONS © 2014 by the American Diabetes Association.
Readers may use this article as long as the work
Both individual MBCT and CBT are effective in improving a range of psychological is properly cited, the use is educational and not
symptoms in individuals with type 1 and type 2 diabetes. for profit, and the work is not altered.
2428 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37, September 2014

Depression is a common and persistent condition showed greater reductions superior over the other. The secondary
condition in patients with diabetes, with in depressive symptoms and diabetes- objective was to investigate the effects
major depression diagnosed in 11% of related distress compared with a waiting- of MBCT and CBT in improving anxiety,
all patients and depressive symptoms list condition (14). These positive results well-being, and diabetes-related dis-
in 31% (1). Alongside its contribution warrant further investigation of the effi- tress. In addition, we explored the ef-
to decreased quality of life, coexisting cacy of individual MBCT for patients with fects of MBCT and CBT on glycemic
depression in diabetes may represent a diabetes, which is the focus of the current control, as indicated by HbA1c values.
great burden for patients’ health and trial. When proven efficacious, individual
the health care system. Depression has Recent reviews and meta-analyses MBCT can be established as a sound al-
been related to lower adherence strongly recommend comparing MBCT ternative to CBT for treating depressive
to medication, dietary and exercise rec- not only to a passive control group but symptoms in patients with diabetes and
ommendations, poorer glycemic con- also to an active evidence-based inter- thereby improving quality of psycho-
trol, and increased health care costs vention (15,16). In the treatment of de- logical care. This availability of distinct
(2,3). The high comorbidity of depres- pressive symptoms, the most frequently evidence-based effective interventions
sion in diabetes and the potential nega- applied and evidence-based treatment is particularly important given the finding
tive health consequences warrant the is cognitive behavior therapy (CBT) that preferences and attitudes toward
identification of effective treatments to (17). CBT has been related to significant treatment can influence treatment out-
improve patient functioning. Both antide- improvements in psychological symp- come (26).
pressant medication and psychological toms in patients with a diversity of
treatment have been found effective for somatic diseases, especially when deliv- RESEARCH DESIGN AND METHODS
treating depression (4), yet the latter ered individually (18). Regarding CBT in Study Design
is preferred by the majority of diabetic patients with diabetes and comorbid de- The Mood Enhancement Therapy Inter-
patients (5). pression, five RCTs have investigated vention Study is a multicenter RCT with
One potential effective psychological and demonstrated its efficacy (19–23). three conditions, namely MBCT, CBT,
treatment consists of mindfulness- Thus, CBT is characterized as the gold and a waiting-list control condition.
based cognitive therapy (MBCT). In the standard against which to assess the ef- We chose the latter control condition
last decade, application of MBCT for the ficacy of a relatively new and promising rather than treatment as usual for ethi-
treatment of a wide variety of psycho- intervention like MBCT. cal reasons, as all participants had ele-
logical disorders, including depressive As MBCT combines mindfulness with vated levels of depressive symptoms at
symptoms, has grown exponentially. elements from CBT, MBCT and CBT can randomization. The study protocol re-
MBCT focuses on cultivating mindful- be regarded as related therapies. Yet, ceived ethical approval from the Medi-
ness, which can be defined as being the treatment components and overall cal Ethical Committee of the University
aware of the present moment by means aim of the interventions are distinct. Medical Center Groningen and was con-
of paying attention on purpose and MBCT mainly involves practicing medi- ducted in accordance with the principles
without judgment (6). Several meta- tation and yoga exercises to increase of the Declaration of Helsinki (version
analyses have demonstrated that awareness and acceptance of dysfunc- 2008) and the Medical Research Involv-
MBCT results in reduction of depressive tional thoughts and accompanying neg- ing Human Subjects Act. A detailed
symptoms and increases in well-being ative emotions (6). CBT encourages description of the design has been pub-
in a variety of populations (7–9). How- patients to maintain and increase the lished elsewhere (27).
ever, little is known about the applicabil- frequency of pleasant activities and to
Participants
ity and effectiveness of MBCT in patients lower negative mood by changing the
Eligible participants were patients with
with diabetes. So far, only one random- content of dysfunctional thoughts into
type 1 or 2 diabetes diagnosed at least 3
ized controlled trial (RCT) has investigated more helpful thoughts (24). To date,
months prior to inclusion, aged between
the effects of MBCT in patients with di- only one small RCT directly compared
18 and 70 years, and having symptoms
abetes, demonstrating a reduction of de- group MBCT to group CBT in people
of depression as indicated by a Beck De-
pressive symptoms and anxiety as well as with depression, demonstrating that
pression Inventory-II (BDI-II) score of
an increase in quality of life (10). both interventions were equally effica-
$14. Exclusion criteria were not being
Usually, MBCT is delivered in a group cious (25). No RCT of CBT and MBCT
able to read and write Dutch, preg-
format. Yet, not all participants experi- has been conducted in patients with
nancy, severe psychiatric comorbidity,
ence this as beneficial (11), and a group diabetes.
acute suicidal ideations, receiving an al-
of patients prefers individual treatment The purpose of the current Mood En-
ternative psychological treatment dur-
to group MBCT (12). This is in line with a hancement Therapy Intervention Study
ing or ,2 months prior to starting
study on psychological treatment pref- was to examine the effectiveness of in-
participation in the study, and unstable
erences in general, demonstrating that dividual MBCT and CBT for depressive
treatment with an antidepressant in the
70% of people preferred individual symptoms in patients with diabetes in
last 2 months prior to inclusion in the
treatment above group treatment (13). comparison with a waiting-list control
study.
This motivated us to investigate the condition. We hypothesized that both
effectiveness of an individual MBCT MBCT and CBT were more effective Procedure
program. In a pilot RCT, we found than a waiting-list control condition, Patients were recruited from June 2011
that patients in the individual MBCT with neither MBCT nor CBT being to February 2013 at four hospitals
care.diabetesjournals.org Tovote and Associates 2429

primarily in the northern part of the Beck et al. (24). The main components This semistructured clinical interview
Netherlands (see Supplementary Data of CBT were behavioral activation and was administered by trained psycholo-
for a list of participating investigators). cognitive restructuring. A description gists at pre- and postmeasurement. At
Recruitment took place through stan- of the protocols can be found in Supple- premeasurement, the assessors of the
dard screening at outpatient clinics, mentary Table 1. HAM-D7 were blinded to the treatment
referral by physician, and self-referral. To assess adherence, all treatment condition. However, at postmeasure-
When patients had elevated levels of sessions of patients who provided ment, the HAM-D7 was administered
depressive symptoms, they were invited consent for this were videotaped. In together with an evaluation of the
for an intake, during which they were addition, patients were asked to report treatment for individuals randomized
screened again and assessed for eligibil- their daily homework practice on to MBCT or CBT, and therefore, the as-
ity. Patients who fulfilled our criteria weekly evaluation forms. Based on rat- sessors were not blinded. The HAM-D7
and gave written informed consent for ings of the videotaped sessions by consists of seven items about depressed
participation were included in the two independent observers, we found mood, feelings of guilt, and anxiety dur-
study. Patients in the MBCT and CBT that therapists’ adherence to the treat- ing the last week. The items are scored
conditions were assessed before ran- ment manuals was sufficiently good on a five-point scale, ranging from 0–4
domization and start of treatment (pre- (85% in MBCT and 83% in CBT). Also, (except for one item, which ranges from
measurement) and immediately after patients’ homework compliance was 0–2). A sum score of $4 represents mild
ending of treatment (postmeasure- sufficient (61% in MBCT and 79% depression, a score between 12 and 20
ment; on average, 3 months after the in CBT). moderate depression, and a score .20
first assessment). Patients assigned represents severe depression. Reliability
Waiting-List Condition
to the waiting-list control condition in the current study was acceptable (a =
Participants in the waiting-list condition
undertook a baseline assessment (pre- 0.65).
received no psychological intervention
measurement) and an assessment at
for 3 months.
the end of the 3-month waiting period Secondary Outcome Measures
(postmeasurement). Assessments
The Well-Being Index (WHO-5) was used
to assess emotional well-being (31). This
Descriptive Measures
Randomization self-report instrument consists of five
Data on the following demographic
Computerized randomization was car- items that are scored on a six-point scale
variables were collected through self-
ried out stratified by sex, use of antide- from 0 (“not present”) to 5 (“constantly
report questionnaires: age, sex, educa-
pressant medication, and baseline BDI-II present”). The items are about positive
tion, marital status, and occupation.
score. Before randomization, patients mood, vitality, and general interest in
Disease-specific characteristics were
were blinded for the treatment condi- relation to the last 2 weeks. The total
retrieved from patients’ records, namely
tion. Accordingly, patients did not re- sum score is converted to a score be-
time since diagnosis, type of diabetes,
ceive any specific information about tween 0 and 100, with a score #50 in-
treatment regimen, comorbidities, com-
the type of intervention or the waiting- dicating poor well-being. In this study,
plications, and BMI. For 14 patients, we
list condition. They were only told that the scale’s reliability was good (a =
could not access the medical records,
they were to be randomized to a psy- 0.82).
and thus this information was retrieved
chological treatment that focuses on Anxiety was assessed by means of the
from the questionnaire.
reducing depression and that treatment Generalized Anxiety Disorder 7 (GAD-
was to start within 3 months after Primary Outcome Measure 7), a seven-item self-report instrument
randomization. The primary outcome measure, severity (32). Respondents are asked to report
of depressive symptoms, was assessed the frequency with which they experi-
Interventions with the BDI-II (29). The BDI-II is a 21-item ence worrying and feeling restless,
Intervention Conditions self-report questionnaire, scored on a annoyed, or afraid during the last 2
Both MBCT and CBT are protocolized in- four-point scale ranging from 0 (“not at weeks. Each item is scored 0 (“not at
terventions aimed at reducing depres- all”) to 3 (“most of the time”). It measures all”) to 3 (“nearly every day”). A total
sive symptoms. The treatments were symptoms of depression such as sadness, sum score of $5 indicates mild
delivered individually in eight weekly loss of interest, and hopelessness during anxiety, a score of 11–15 moderate
sessions of 45–60 min. Patients were the last 2 weeks. A score from 14–19 anxiety, and a score of .15 indicates
also instructed to do daily homework indicates mild depression, a score from severe anxiety. Cronbach a in this study
for 30 min. Both interventions were de- 20–28 moderate depression, and a score was good (a = 0.88).
livered by trained therapists who re- $29 indicates severe depression. The re- The Problem Areas In Diabetes (PAID)
ceived supervision every 3 weeks liability of the BDI-II was good in the cur- was used to measure diabetes-related
throughout the intervention period. rent study (a = 0.84). distress (33,34). The PAID consists of
MBCT was based on the protocol as de- In addition to the self-report depres- 20 items, which are rated on a five-point
veloped by Segal et al. (28). The central sion measurement, and in order to as- scale. The scoring ranges from 0 (“not a
components of MBCT were formal med- sess depressive symptoms in a more problem”) to 4 (“serious problem”). The
itation, yoga exercises, and informal objective manner, symptoms were also items cover various common negative
daily mindfulness practices. CBT was measured using the Toronto Hamilton emotions related to living with and man-
based on the protocol developed by Depression Rating Scale (HAM-D7) (30). aging diabetes. The sum of all items is
2430 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37, September 2014

transformed into a scale from 0–100, improvement was defined as having well as primary and secondary out-
with scores of $40 being used to define improved and being recovered. A post- comes measures (averages shown in
patients at risk for high diabetes-related measurement score below the cutoff Table 2).
distress. The internal consistency was of the primary outcome measure (i.e.,
Primary Outcome Measures
excellent in the current study (a = 0.95). BDI-II ,14) indicated improvement. Re-
The mean scores and the outcomes of
Finally, glycemic control, indicated covery was calculated by the Reliable
the statistical analyses are presented in
by HbA1c values, was retrieved from Change Index, which refers to the differ-
Table 2. When comparing MBCT and
patients’ records. As premeasurement, ence between an individual’s pre- and
CBT to the control condition, both inter-
the average of all assembled values postmeasurement scores, divided by
vention groups had significantly less de-
of 0–6 months prior to intervention the SE of the difference. A score .1.96
pressive symptoms than the control
was used, and as postmeasurement, indicates recovery (37).
group at postmeasurement (P = 0.004
we used the average of all values
and P , 0.001, respectively). The effect
between 1 and 6 months after the RESULTS
sizes of the change from pre- to post-
intervention. Recruitment and Attrition measurement between MBCT and CBT
As is shown in Fig. 1, 3,145 patients were versus the waiting list were large (Cohen
Sample Size Calculation routinely screened at a hospital, and d = 0.80 and d = 1.00, respectively).
The sample size calculation was based 14 referred themselves in awareness Given the difference in effect sizes
on expected differences in the level of of their treating physician. Of the between the two interventions, we
posttreatment depressive symptoms 2,266 patients who completed and also compared effects of the MBCT and
between the waiting-list control group returned the screening questionnaire, CBT group directly and found no signifi-
and either MBCT or CBT. Assuming a sta- 613 (27%) had an elevated score (BDI-II cant differences (P = 0.34; not shown in
tistical power of 0.8 and an a of 0.05, 42 $14). Less than half (n = 255) accepted Table 2). Assessing depressive symptoms
participants were required in each the invitation for a face-to-face intake. with the HAM-D7 revealed similar re-
group, enabling us to detect differences An additional six patients were referred sults: both MBCT and CBT had signifi-
with an effect size of 0.6 (35). by their physician for an intake. During cantly higher outcome improvement
the intake, patients were screened again than the waiting-list condition (P ,
Statistical Analyses and elaborately assessed for eligibility. 0.001 and P = 0.001, respectively). The
All analyses were performed based on Almost one-third of the patients (n = 78) between-group effect sizes in comparison
the intention-to-treat method. Missing who received an intake were not eligible with the waiting list were large (MBCT, d =
values were estimated by means of for the trial, and an additional one-third 1.17; CBT, d = 1.09).
multiple imputations using the linear re- (n = 89) did not agree to participate,
gression method. We performed sensi- mostly because they did not feel the Secondary Outcome Measures
tivity analyses based on participants need for treatment. Finally, 94 patients The results of the secondary outcome
with no missing data and datasets with gave consent and were randomized: measures are also presented in Table 2.
5 and 20 imputations. As analyses 31 participants to MBCT, 32 partici- When comparing MBCT and CBT with
revealed a similar pattern of results, pants to CBT, and 31 participants to the waiting list, individuals in both
5 complete datasets were imputed after the waiting-list control condition. In MBCT and CBT had a larger improve-
20 iterations. SPSS Statistics 20 (SPSS both MBCT and CBT, nine patients did ment in levels of well-being (both P ,
Inc.) was used for all analyses, and all not finish the intervention (i.e., received 0.001), anxiety (P = 0.004 and P = 0.01,
according assumptions were met. less than six sessions). Reasons for drop- respectively), and diabetes-related dis-
ANOVAs and x2 tests were used to ana- out were intervention content related tress (P = 0.02 and P = 0.04, respec-
lyze if there were differences at baseline (MBCT: n = 4; CBT: n = 3), lack of time tively). Between-group effect sizes
between the groups regarding demo- (MBCT: n = 3; CBT: n = 2), severe illness were large for well-being and anxiety
graphic and clinical variables as well as (MBCT: n = 1; CBT: n = 2), improvement (range Cohen d = 0.82–0.97) and mod-
primary and secondary outcome mea- of depression after a few sessions erate for diabetes-related distress (d =
sures. Separate ANCOVAs were per- (MBCT: n = 1; CBT: n = 1), and no interest 0.52 and d = 0.57). HbA1c levels did not
formed for MBCT and CBT to examine in participating in research anymore change after MBCT or CBT (P = 0.92 and
the effects of the interventions in com- (CBT: n = 1). Two participants in the P = 0.72, respectively).
parison with the waiting-list condition. MBCT condition and four participants
Postmeasurement values of the primary in both CBT and the waiting-list condi- Clinically Relevant Improvement
and secondary outcomes were used as tion did not fill in the postmeasurement Clinically relevant improvement was
dependent variables, condition was questionnaire. found in 26% of the participants after
used as factor, and premeasurements MBCT and 29% of the participants after
of the outcomes were used as covariate. Baseline Characteristics CBT versus 4% of the patients in the
Between-group effect sizes were calcu- Table 1 provides an overview of the waiting-list condition. When comparing
lated using Cohen d, with values ranging baseline characteristics of the partici- the percentages in the intervention con-
from 0.2 to 0.5 indicating small effects, pants. There were no statistically signif- ditions to the control condition, the dif-
values from 0.5 to 0.8 indicating moder- icant differences among the three ferences were significant (MBCT vs.
ate effects, and values .0.8 indicating conditions regarding the demographic waiting list: P = 0.02; CBT vs. waiting
large effects (36). Clinically relevant or clinical baseline characteristics as list: P = 0.009).
care.diabetesjournals.org Tovote and Associates 2431

Figure 1—Participant recruitment and flow through the study.

CONCLUSIONS anxiety and diabetes-related distress. No treatments (38). Our results are innova-
This is the first RCT that investigated the effects were found for HbA1c values. tive, as this is the first RCT study on the
effectiveness of individually delivered Given the high prevalence and bur- effectiveness of the individual delivery
MBCT and CBT in reducing depressive den of depressive symptoms in patients of MBCT, with currently only evidence
symptoms in outpatients with type 1 with diabetes, a key finding of this RCT for the effectiveness of group-based
and type 2 diabetes. Concordant with is that patients receiving one of the MBCT. We are aware that our results
our hypothesis, results indicated that psychological interventions reported should be replicated to draw more firm
both MBCT and CBT were effective in re- greater reductions in depressive symp- conclusions. Yet, it is promising, given
ducing depressive symptoms compared toms in comparison with the control the well-known effectiveness of CBT, es-
with a waiting-list control condition, condition. Our findings are in line with pecially when individually delivered
with neither MBCT nor CBT being supe- Manicavasgar et al. (25), demonstrating (18), that individual MBCT was as effec-
rior over the other. MBCT and CBT were effectiveness of group MBCT and CBT in tive. Taking into account the differences
also effective in improving a wider range reducing depression. These and our in treatment focus and components of
of patient-relevant outcomes, including findings show that CBT is effective but MBCT and CBT, our results imply that
increases in well-being and reductions in not superior to some other active two evidence-based distinct types of
2432 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37, September 2014

Table 1—Baseline characteristics

MBCT (n = 31) CBT (n = 32) Waiting list (n = 31) Total (n = 94)*
Age (years), mean (SD) 49.8 (13.3) 54.6 (11.3) 54.7 (10.5) 53.1 (11.8)
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 education/advanced 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) 29.3 (7.6) 31.9 (6.6) 30.6 (8.4) 30.6 (7.6)
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) 17.8 (13.0) 15.0 (11.4) 17.0 (11.4) 16.6 (11.9)
Diabetes complications, n (%)†
One 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 . 0.05 in all cases) on any of the demographics and clinical characteristics. †Included diabetes complications
are: retinopathy, neuropathy, nephropathy, and diabetic foot.

psychological interventions can be investigating either MBCT or CBT, also translate into improved self-care and
offered to patients with diabetes. An showing reductions in diabetes-related subsequent glycemic control (4).
important next step would be to inves- distress (14,19). Taken together, MBCT A methodological challenge in the in-
tigate possible moderators of effective- and CBT not only reduce depressive vestigation of improvements in HbA1c
ness; that is, factors related to the symptoms, but also improve other psy- levels is that the HbA1c level is an aver-
differential effectiveness of MBCT and chological outcomes. age value over the previous 3 months. In
CBT within certain subgroups of patients Explorative analysis showed no signif- this study, HbA1c was only included for
(i.e., for whom is which intervention icant reductions in HbA1c values either exploratory reasons in order to burden
more beneficial?). in MBCT or in CBT. This finding is con- the patients as little as possible. There-
Besides depressive symptoms, we cordant with two previous RCTs on fore, HbA1c values were obtained from
were also interested in a possible wider MBCT (10) and CBT (19) that did not patients’ medical records instead of
effect of MBCT and CBT on other indica- find an effect on glycemic control. A scheduling additional measurements at
tors of functioning. It is clinically rele- recent review and meta-analysis studied designated time points. Consequently,
vant to observe that both MBCT and the impact of psychosocial interventions our HbA1c values are crude indications
CBT significantly increase well-being on both psychological and physical of HbA1c values in the months preceding
and reduce anxiety and diabetes-related health in patients with diabetes (39). and following the two active interven-
distress. These findings are consistent No interventions were identified that tions. Also, as patients in the waiting-
with previous research showing that were effective for both medical and men- list condition received care directly after
psychological interventions focusing on tal outcomes at the same time. Alto- patients in the active conditions had fin-
depressive symptoms can also improve gether, our findings and previous results ished the intervention, it was not possi-
anxiety and quality of life (10,20). Re- suggest that alleviating depressive symp- ble to compare CBT and MBCT with the
sults are also in line with previous toms through psychological interventions control condition regarding changes in
studies in patients with diabetes like MBCT or CBT does not automatically HbA1c values.
care.diabetesjournals.org Tovote and Associates 2433

Table 2—Results for primary and secondary outcomes

Time Treatment†
Premeasurement, Postmeasurement,
Measure Group mean (SD) mean (SD) t P value F P value d
Depression (BDI-II) MBCT 23.6 (7.7) 17.1 (11.9) 4.75 ,0.001 9.71 0.004 0.80 (0.27–1.31)
CBT 25.6 (8.7) 17.4 (11.9) 5.58 ,0.001 15.56 ,0.001 1.00 (0.47–1.51)
WAIT 24.3 (8.0) 23.5 (10.3) 0.65 0.52
Depression (HAM-D7) MBCT 8.9 (3.5) 4.7 (4.3) 6.33 ,0.001 17.52 ,0.001 1.17 (0.61–1.69)
CBT 9.4 (3.8) 4.6 (3.4) 5.55 ,0.001 13.06 ,0.001 1.09 (0.55–1.60)
WAIT 7.5 (2.8) 7.1 (3.7) 0.71 0.49
Well-being (WHO-5) MBCT 32.4 (18.4) 49.5 (21.5) 5.90 ,0.001 17.35 ,0.001 0.92 (0.39–1.43)
CBT 26.8 (17.8) 47.4 (20.2) 6.07 ,0.001 18.95 ,0.001 1.02 (0.48–1.53)
WAIT 27.7 (15.9) 30.9 (15.4) 1.09 0.28
Anxiety (GAD-7) MBCT 12.6 (5.3) 6.9 (4.8) 7.00 ,0.001 9.60 0.004 0.98 (0.44–1.49)
CBT 11.9 (4.9) 6.8 (5.0) 6.61 ,0.001 7.42 0.01 0.82 (0.29–1.32)
WAIT 9.8 (5.0) 8.2 (4.6) 2.37 0.02
Diabetes distress (PAID) MBCT 38.3 (20.9) 32.0 (21.8) 3.08 0.002 5.67 0.02 0.52 (0.01–1.02)
CBT 42.0 (22.3) 34.0 (23.4) 2.87 0.004 5.68 0.04 0.57 (0.06–1.07)
WAIT 35.5 (21.5) 36.0 (21.2) 20.26 0.79
HbA1c*
mmol/mol MBCT 63.4 (9.6) 63.1 (10.8) 0.10 0.92
% 8.0 (0.9) 7.9 (1.0)
mmol/mol CBT 67.1 (15.2) 65.9 (13.0) 0.36 0.72
% 8.3 (1.4) 8.2 (1.2)
WAIT, waiting list. *Not measured in the waiting-list condition. †Comparing both MBCT and CBT with the waiting-list condition.

Limitations to participate in the study because of no 8-week interventions delivered on

Although we carefully designed our need for psychological treatment. Our an individual basis, they could be im-
study, several limitations to this study inclusion rate is comparable to other plemented in optimizing psychological
need to be acknowledged. First, we studies using a consecutive sample care for depressed patients with
were not able to reach a fully powered method (10,40). As patients were diabetes.
sample of at least 42 participants per blinded to the content of the treatment,
condition, as patient recruitment took we do not assume that refusal to partic-
more time than originally planned. Yet, ipation was content related. Fifth, the Acknowledgments. The authors thank all of
for the actual sample size, the power fact that assessors of depressive symp- the patients who participated in the study, the
was still 68%. Second, although the ma- toms with the clinical interview at post- psychologists who delivered the MBCT and CBT
sessions, the secretaries and research assistants
jority of the patients were recruited as a measurement were not blinded may have of the University Medical Center Groningen, the
consecutive sample (i.e., screening; n = biased their ratings. Finally, as all partic- Martini Hospital Groningen, the Medical Center
79), a small group of participants was ipants had elevated levels of depressive Leeuwarden, and the Hospital Rivierenland Tiel
recruited as a convenience sample symptoms at randomization, for ethical for the efforts.
based on (self) referral (n = 15). The reasons, we included a waiting-list con- Funding. This study was financed by the
University of Groningen.
former may not be representative trol condition rather than treatment Duality of Interest. No potential conflicts of
of treatment-seeking or clinically as usual. As patients in the control interest relevant to this article were reported.
referred patients, while the latter sam- condition received one of the interven- Author Contributions. K.A.T. constructed the
ple may suffer from selection bias, tions after the 3-month waiting period, design of the study, researched data, and wrote
hereby reducing generalizability of the long-term effects of CBT and MBCT com- the manuscript. J.F. and M.J.S. constructed the
design of the study and revised the manuscript.
results. Third, attrition rates in both pared with the control condition could E.S. constructed the design of the study, re-
MBCT and CBT were high, as only not be assessed. searched data, and reviewed the manuscript.
;70% of the randomized participants A.C.T.M.P. reviewed the manuscript. P.M.G.E.
completed treatment. These attrition Conclusion and R.S. constructed the design of the study
and reviewed the manuscript. T.P.L. participated
rates are consistent with previous stud- This is the first RCT examining the ef-
in the design of the study and reviewed the
ies targeting distressed patients with fectiveness of individual MBCT and in- manuscript. K.A.T. is the guarantor of this work
diabetes (10,19). Screening a consecu- dividual CBT in reducing depressive and, as such, had full access to all the data in the
tive sample may have accounted symptoms in patients with diabetes. study and takes responsibility for the integrity of
for the dropout rate, as the majority Results clearly suggest that MBCT as the data and the accuracy of the data analysis.
of participants did not seek treat- well as CBT are effective interventions
References
ment themselves, but instead were ap- in treating depressive symptoms in
1. Anderson RJ, Freedland KE, Clouse RE,
proached and offered treatment. patients with diabetes. Given their Lustman PJ. The prevalence of comorbid depres-
Fourth, a substantial group of patients effectiveness and the fact that both in- sion in adults with diabetes: a meta-analysis. Di-
who met the inclusion criteria refused terventions are short, structured abetes Care 2001;24:1069–1078
2434 MBCT and CBT for Depression in Diabetes Diabetes Care Volume 37, September 2014

2. Egede LE, Ellis C. Diabetes and depression: 15. Galante J, Iribarren SJ, Pearce PF. Effects of cognitive therapy for depressive symptoms in
global perspectives. Diabetes Res Clin Pract mindfulness-based cognitive therapy on mental patients with diabetes: design of a randomized
2010;87:302–312 disorders: A systematic review and meta-analysis controlled trial. BMC Psychol 2013;1:17
3. Katon WJ. The comorbidity of diabetes mel- of randomised controlled trials. J Res Nurs 2013; 28. Segal ZV, Williams MG, Teasdale JD.
litus and depression. Am J Med 2008;121(Suppl. 18:133–155 Mindfulness-Based Cognitive Therapy for De-
2):S8–S15 16. Coelho HF, Canter PH, Ernst E. Mindfulness- pression: A New Approach to Preventing Re-
4. van der Feltz-Cornelis CM, Nuyen J, Stoop C, based cognitive therapy: evaluating current ev- lapse. New York, Guilford Press, 2002
et al. Effect of interventions for major depres- idence and informing future research. J Consult 29. Beck AT, Steer RA, Ball R, Ranieri WF. Com-
sive disorder and significant depressive symp- Clin Psychol 2007;75:1000–1005 parison of Beck Depression Inventories-IA
toms in patients with diabetes mellitus: 17. Butler AC, Chapman JE, Forman EM, Beck and -II in psychiatric outpatients. J Pers Assess
a systematic review and meta-analysis. Gen AT. The empirical status of cognitive-behavioral 1996;67:588–597
Hosp Psychiatry 2010;32:380–395 therapy: a review of meta-analyses. Clin Psychol 30. McIntyre RS, Konarski JZ, Mancini DA, et al.
5. Dwight-Johnson M, Sherbourne CD, Liao D, Rev 2006;26:17–31 Measuring the severity of depression and remis-
Wells KB. Treatment preferences among de- 18. Beltman MW, Voshaar RCO, Speckens AE. sion in primary care: validation of the HAMD-7
pressed primary care patients. J Gen Intern Cognitive-behavioural therapy for depression in scale. CMAJ 2005;173:1327–1334
Med 2000;15:527–534 people with a somatic disease: meta-analysis of 31. Bech P. Measuring the dimension of psy-
6. Kabat-Zinn J. Mindfulness-based interven- randomised controlled trials. Br J Psychiatry chological general well-being by the WHO-5.
tions in context: Past, present, and future. Clin 2010;197:11–19 QOL Newsletter 2004;32:15–16
Psychol Sci Pract 2003;10:144–156 19. van Bastelaar KM, Pouwer F, Cuijpers P, 32. Spitzer RL, Kroenke K, Williams JBW, Löwe
7. Hofmann SG, Sawyer AT, Witt AA, Oh D. The Riper H, Snoek FJ. Web-based depression treat- B. A brief measure for assessing generalized
effect of mindfulness-based therapy on anxiety ment for type 1 and type 2 diabetic patients: anxiety disorder: the GAD-7. Arch Intern Med
and depression: A meta-analytic review. J Con- a randomized, controlled trial. Diabetes Care 2006;166:1092–1097
sult Clin Psychol 2010;78:169–183 2011;34:320–325 33. Polonsky WH, Anderson BJ, Lohrer PA, et al.
8. Khoury B, Lecomte T, Fortin G, et al. 20. Penckofer SM, Ferrans C, Mumby P, et al. A Assessment of diabetes-related distress. Diabe-
Mindfulness-based therapy: a comprehensive psychoeducational intervention (SWEEP) for de- tes Care 1995;18:754–760
meta-analysis. Clin Psychol Rev 2013;33:763– pressed women with diabetes. Ann Behav Med 34. Welch GW, Jacobson AM, Polonsky WH.
771 2012;44:192–206 The Problem Areas in Diabetes Scale. An evalu-
9. Chiesa A, Serretti A. Mindfulness based cog- 21. Lamers F, Jonkers CCM, Bosma H, et al. A ation of its clinical utility. Diabetes Care 1997;
nitive therapy for psychiatric disorders: a sys- minimal psychological intervention in chroni- 20:760–766
tematic review and meta-analysis. Psychiatry cally ill elderly patients with depression: a ran- 35. Cohen J. A power primer. In Methodologi-
Res 2011;187:441–453 domized trial. Psychother Psychosom 2010;79: cal Issues & Strategies in Clinical Research. 3rd
10. van Son J, Nyklı́cek I, Pop VJ, et al. The ef- 217–226 ed. Kazdin AE, Ed. Washington, DC, American
fects of a mindfulness-based intervention on 22. Lustman PJ, Griffith LS, Freedland KE, Kissel Psychological Association, 2003, p. 427-436
emotional distress, quality of life, and HbA(1c) SS, Clouse RE. Cognitive behavior therapy for 36. Cohen J. Statistical Power Analysis for the
in outpatients with diabetes (DiaMind): a ran- depression in type 2 diabetes mellitus. A ran- Behavioral Sciences. New York, L. Erlbaum Asso-
domized controlled trial. Diabetes Care 2013; domized, controlled trial. Ann Intern Med 1998; ciates, 1977
36:823–830 129:613–621 37. Jacobson NS, Truax P. Clinical significance:
11. Griffiths K, Camic PM, Hutton JM. Partici- 23. Safren SA, Gonzalez JS, Wexler DJ, et al. A a statistical approach to defining meaningful
pant experiences of a mindfulness-based cogni- randomized controlled trial of cognitive behav- change in psychotherapy research. J Consult
tive therapy group for cardiac rehabilitation. J ioral therapy for adherence and depression Clin Psychol 1991;59:12–19
Health Psychol 2009;14:675–681 (CBT-AD) in patients with uncontrolled type 2 38. Tolin DF. Is cognitive-behavioral therapy
12. Lau MA, Colley L, Willett BR, Lynd LD. Em- diabetes. Diabetes Care 2014;37:625–633 more effective than other therapies? A meta-
ployee’s preferences for access to mindfulness- 24. Beck AT, Rush AJ, Shaw BF, Emery G. Cog- analytic review. Clin Psychol Rev 2010;30:710–
based cognitive therapy to reduce the risk of nitive Therapy of Depression. New York, Guil- 720
depressive relapsedA discrete choice experi- ford, 1979 39. Harkness E, Macdonald W, Valderas J,
ment. Mindfulness 2012;3:318–326 25. Manicavasgar V, Parker G, Perich T. Coventry P, Gask L, Bower P. Identifying psycho-
13. Lang AJ. Mental health treatment prefer- Mindfulness-based cognitive therapy vs cogni- social interventions that improve both physical
ences of primary care patients. J Behav Med tive behaviour therapy as a treatment for non- and mental health in patients with diabetes:
2005;28:581–586 melancholic depression. J Affect Disord 2011; a systematic review and meta-analysis. Diabe-
14. Schroevers MJ, Tovote KA, Keers JC, Links TP, 130:138–144 tes Care 2010;33:926–930
Sanderman R, Fleer J. Individual mindfulness- 26. Swift JK, Callahan JL, Vollmer BM. Prefer- 40. Strong V, Waters R, Hibberd C, et al. Man-
based cognitive therapy for people with diabetes: ences. J Clin Psychol 2011;67:155–165 agement of depression for people with cancer
A pilot randomized controlled trial. Mindfulness. 27. Tovote KA, Fleer J, Snippe E, et al. Cognitive (SMaRT oncology 1): a randomised trial. Lancet
14 August 2013 [Epub ahead of print] behavioral therapy and mindfulness-based 2008;372:40–48

View publication stats