Journal of Affective Disorders 179 (2015) 14–22

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research report

Collaborative care for patients with bipolar disorder: Effects on functioning and quality of life Trijntje Y.G. van der Voort a,b,n, Berno van Meijel a,b,c, Adriaan W. Hoogendoorn a, Peter J.J. Goossens d,e,f, Aartjan T.F. Beekman a, Ralph W. Kupka a,g a

VU University Medical Center, Department of Psychiatry, Amsterdam, The Netherlands Inholland University of Applied Sciences, Department of Health, Sports & Welfare/Cluster Nursing, Research Group Mental Health Nursing, Amsterdam, The Netherlands c Parnassia Psychiatric Institute, Parnassia Academy, The Hague, The Netherlands d Dimence Mental Health, Deventer, The Netherlands e Radboud University Medical Center, Nijmegen, The Netherlands f GGZVS, Institute for the Education of Clinical Nurse Specialists in Mental Health, Utrecht, The Netherlands g Altrecht Mental Health, Utrecht, The Netherlands b

art ic l e i nf o

a b s t r a c t

Article history: Received 16 December 2014 Received in revised form 1 March 2015 Accepted 2 March 2015 Available online 11 March 2015

Background: Functioning and quality of life are impaired in bipolar patients. Methods: Collaborative Care (CC) is a multi-component intervention, provided by a multidisciplinary team, in which a nurse-care manager plays a central role. Effects on functioning and quality of life were tested in a clinical trial. We also investigated the mediating role of depression severity on these outcome variables. Results: Patients randomized to CC showed more improvement in overall functioning compared to patients in the control group who obtained care as usual (CAU), with a small effect size (ES ¼0.3, z¼  2.5, p¼ 0.01). In the domains of autonomy and leisure time, a medium effect was found in favor of CC (autonomy: ES¼ 0.5, z ¼  2.9, p ¼0.004; leisure-time: ES¼0.4, z ¼  2.4, p¼0.02). No differences between conditions were found in the other domains of functioning. Concerning quality of life, patients in CC improved more in the domain physical health (ES ¼0.4, z¼ 2.5, p ¼0.01), if compared to CAU. No differences were found in overall quality of life. Half of the effects on functioning are mediated through the effects of CC on depression severity. Limitations: At baseline, differences on the main outcomes existed between conditions. Two teams stopped participation in the experimental condition after randomization. Sample size was limited. Conclusion: Besides effects on depressive symptoms, CC seems to have direct beneficial effects on both level of functioning and aspects of quality of life. & 2015 Elsevier B.V. All rights reserved.

Keywords: Bipolar disorder Collaborative Care Depression Functioning Quality of life

1. Background Bipolar disorder (BD) is a serious mental illness associated with functional impairment and decreased quality of life. On average, patients suffer from manic or depressive symptoms for 50% of the time despite treatment (Judd et al., 2002; Kupka et al., 2007). In particular sub threshold depressive symptoms, which are present in many patients even during euthymic periods, have been associated with impaired functioning and quality of life (Altshuler et al., 2006;

n Corresponding author at: Inholland University of Applied Sciences, Faculty of Health, Sports and Social Work, de Boelelaan 1109, 1081HV Amsterdam, The Netherlands. Tel.: +31 6 81152652. E-mail addresses: [email protected], [email protected] (T.Y.G. van der Voort).

http://dx.doi.org/10.1016/j.jad.2015.03.005 0165-0327/& 2015 Elsevier B.V. All rights reserved.

Bonnin et al., 2012; Ishak et al., 2012; Michalak et al., 2005; Strejilevich et al., 2013; Vieta et al., 2008). Treatment of patients with bipolar disorder aims at three dimensions of the disorder: mood and related symptoms, psychosocial and cognitive functioning, and overall quality of life. In the current study we broadly define functioning as ‘what a person does or can do’, and quality of life as ‘satisfaction with several life domains’ (World Health Organization, 2001; World Health Organization, 1997). Quality of life may be viewed as the overall concept, encompassing both symptoms and functioning, as well as the patients' perception of this. The question if treatment succeeds in improving these dimensions, and if so, along which pathways, has still to be answered. Expectedly, as symptoms remit, functioning improves as a consequence. We hypothesize several possible relationships between treatment and these three dimensions of bipolar disorder (Fig. 1).

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

Quality of life

Symptoms

Functioning

Collaborative Care

Fig. 1. Possible pathways of treatment effects.

Patients with mental illness and their relatives increasingly ask for treatment that improves not only their symptoms but beyond that improves their functioning and quality of life (Bellack et al., 2007). However, guidelines for bipolar disorder have a considerable focus on pharmacotherapy for symptomatic recovery (National institute for health and clinical excellence, 2014; Yatham et al., 2013). To address a broader range of outcomes, several Collaborative Care (CC) programs with integrated multidisciplinary health care have been developed (Wagner et al., 1996) in which in general nurses play a central role. In a recent review (Beekman et al., 2013) several of these CC-programs were found to be effective in a recent review. To date, three studies have tested the effectiveness of CC in patients with BD with promising results on manic but not depressive symptoms (Bauer et al., 2006). More recently, CC-programs were extended to also address medical problems (Kilbourne et al., 2008) with positive effects on functioning (Bauer et al., 2006) and quality of life (Bauer et al., 2006). We performed a cluster-randomized clinical trial investigating the effectiveness of a CC program for patients with BD (van der Voort et al., 2011). Elsewhere (van der Voort et al., 2015), we reported that patients who received CC experienced less time with depressive symptoms, and less severity of depression, when compared to Care as Usual (CAU). Aims of the current study are (1) to study the effects of CC on functional impairment and quality of life, and (2) to examine to which extend the severity of depressive symptoms mediates functional recovery and quality of life.

2. Design We carried out a controlled cluster-randomized trial with a follow-up period of one year in which Collaborative Care (CC) was compared with Care as Usual (CAU). We included 16 mental health outpatient clinics in the Netherlands. Measurements were obtained at baseline, six and twelve months. The primary outcome measures were time spent with depressive or manic symptoms, and severity of symptoms, and these results are reported elsewhere (van der Voort et al., 2015). In the current report, effects on functioning and quality of life will be reported. 2.1. Randomization and inclusion Clustered randomization was performed on the level of outpatient teams that treated at least 20 patients with bipolar disorder and were willing to participate. To reach comparable numbers of respondents in both conditions, teams were matched pairwise on the number of participating nurses. We used an Internet Random Generator, to assign the two teams within every pair randomly to either the experimental or the control condition. The random assignment was performed blindly by the second author (BM). Next, a nurse or a psychiatrist in each team made a

15

list of patients who fulfilled the inclusion criteria and therefore could be invited to participate in our study (see below). If the patient agreed, the researcher contacted the patient to provide detailed oral information about the study. If patients were willing to participate, they received additional written information, and were asked to sign an informed consent form. The study protocol was approved by the Medical Ethical Committee of the VU University Medical Center. 2.2. Patients Patients aged 18–65 years with a diagnosis of bipolar disorder (BD-I; BD-II, BD-NOS) according to DSM- IV-TR (American Psychiatric Association, 2000) were included. Diagnoses were derived from the medical records, and confirmed by the treating psychiatrist, via the Questionnaire of Bipolar Illness (Leverich et al., 2001). Patients in a severe manic or depressive episode at time of inclusion were excluded. We also excluded patients who were sufficiently stable to function well with only low-intensity treatment, given the fact that the program would be too intensive for them. Based on these considerations we applied the following exclusion criteria: (i) severe or very severe depression or mania, with a score of six or seven on the Clinical Global ImpressionBipolar Disorder (CGI-BP (Spearing et al., 1997); (ii) a stable course of illness over the past year, allowing low intensity of treatment with a maximum of four consultations with the psychiatrist or nurse per year; (iii) insufficient command of the Dutch language; (iv) not able or willing to give informed consent. 2.3. Blinding At the time their informed consent was asked, patients were aware of the condition which their treatment team was assigned to. Given the nature of the intervention, nurses and psychiatrist could not be blinded for the condition. 2.4. Collaborative care Our CC-program (van der Voort et al., 2011) consisted of the following elements. All decisions concerning treatment were made in a Collaborative Care team, including the patient, a relative of the patient, the nurse and the psychiatrist. The CC-team would be extended with other professionals if required for optimum treatment. The team met at least three times a year. Coordination of care was provided by the nurse in his/her role as care manager. Care needs were systematically assessed, using the Camberwell Assessment of Needs (Phelan et al., 1995), that formed the basis for an individualized treatment plan. This plan was formulated as a contract, in which goals and treatment activities were recorded, and signed by all team members. All activities and results of treatment were monitored and evaluated in the CC-team. Patients were encouraged to chart their mood by means of the Life Chart Method (Denicoff et al., 2000). A relapse prevention plan was constructed, containing a description of early warning signs of relapse, and early interventions (Goossens et al., 2010). Psycho education was provided in groups to patients and their relatives (Honig et al., 1997). Nurses performed Problem Solving Treatment (PST). PST is a comprehensive therapy, in which patients learn that effectively solving problems in daily life may result in enhanced mood (Mynors Wallis et al., 2000). Pharmacotherapy and somatic care were provided as usual. 2.5. Care as usual In the Netherlands, quality of CAU in mental health care is relatively high. However, the extent to which available

16

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

interventions are being delivered differs among teams is unknown, and no reliable data are available about content and quality of usual care. We assumed that CAU shows considerable variation between treatment teams. Therefore, it was decided to invite all teams willing to participate in the trial. 2.6. Procedures The investigators developed a manual-based CC-training program in collaboration with an expert panel consisting of expert nurses, psychiatrists, a patient and a family member. Fifteen nurses in the experimental condition were trained during three days. The training was aimed at enhancing knowledge about the interventions to be delivered, as well as at offering skills training to perform the interventions adequately. Psychiatrists received a short overview of the intervention program, both in writing and orally. CC was active during one year. During the execution of the intervention program, the nurses were coached by the primary investigator (TV), and the PST-trainer provided supervision on PST to all participating nurses. Nurses in the CAU condition received no form of training, coaching, or supervision. 2.7. Measures Outcomes were measured at baseline (T0) and after six (T6) and twelve months (T12). At baseline, demographics, illness history, diagnosis, illness characteristics, and current treatment were recorded by both the patient and the treating psychiatrist or nurse, with the Questionnaire for Bipolar Illness (QBP-NL, patientversion and clinician version). Functioning was measured with the Functioning Assessment Short Test (FAST-NL-P), which has shown to have good psychometric properties (Rosa et al., 2007). The FAST was designed to assess impairments in functioning as experienced by patients with mental health problems, in particular patients with BD. It is a short instrument, patient-rated, and scores are rated on a four points Likert scale. Higher scores indicate more functional impairments. It comprises 24 items, and covers 6 areas of functioning: autonomy, occupational functioning, cognitive functioning, financial issues, interpersonal relationships and leisure time. Autonomy refers to the patient's capacity to do things alone and make his/her own decisions. Occupational functioning involves the capacity to keep a job, efficiency of executing tasks required, working in the field in which the patient was educated and earning a salary conform this job. Cognitive functioning is related to the ability to concentrate, to perform simple calculations, to solve problems, to learn and to remember new information. Financial issues refer to the capacity to manage finances in a balanced way. Interpersonal relationships involve relations with friends and family, involvement in social activities, sexual relationships, and the ability to defend one's own interests. Leisure time refers to the capability of performing physical activities (sport, exercise) and maintaining hobbies (Rosa et al., 2007). Quality of life was assessed with the World Health Organization Quality of Life Questionnaire, abbreviated version (WHO-QoLbref), of which psychometric properties have been thoroughly established (Skevington et al., 2004; Trompenaars et al., 2005). The WHO-QoL-bref is self-rated and is scored on a five point scale, higher scores indicating higher quality of life. The questionnaire consists of two overall scores (general quality of life and health related quality of life) and scores on four domains: physical, psychological, social relationships, and environment. Physical health refers to: activities of daily living; dependence on medicinal substances and medical aids; energy and fatigue; mobility; pain and discomfort; sleep and rest; and work capacity. Psychological

health involves: bodily image and appearance; negative feelings; positive feelings; self-esteem; spirituality, religion, personal beliefs; thinking, learning, memory and concentration. Social relationship refers to personal relationships; social support; and sexual activity. Environment addresses: financial resources; freedom, physical safety and security; health and social care (accessibility and quality); home environment; opportunities for acquiring new information and skills; participation in and opportunities for recreation or leisure activities; physical environment (pollution, noise, traffic, climate); and transport. 3. Statistical analysis Baseline characteristics of the experimental and control groups concerning demographic and clinical variables were described. T-tests for independent groups were used for continuous variables, and χ2 statistics for categorical variables to compare conditions at baseline. These analyses were also performed to compare dropouts and completers on baseline characteristics. Firstly, observed means and standard deviations were calculated for functioning and quality of life, in respondents in both conditions at the three measurements. Data were analyzed according to the intention to treat-principle. Differences in outcome between CC and CAU were evaluated by means of linear mixed models, as this method is appropriate to test differences in time between groups, for modeling nesting of patients within teams and for dealing with missing observations due to drop out. To test for a difference in treatment effect over time, a groupntime-interaction term was entered into the model. Next, effect sizes (ESs) between groups were calculated, based on the standardized estimated differences between T6 and T0, and between T12 and T0, between groups, using pooled pretest SDs (Morris, 2008). To test for mediation by depression severity scores, we carried out mediation analyses (Baron and Kenny, 1986). The analyses were extended using multilevel analyses that take the nesting of measurements (of patients within teams) into account. We tested if the nesting of measurements in our study yielded different results, by testing a model with against a model without taking nesting into account using a likelihood ratio test. Furthermore, we analyzed completers only and compared the results with intention-to-treat analyses. All statistical tests were two-tailed, and alpha was set at 0.05. 4. Sample size calculation The sample size calculation concerned the comparison of the outcomes between the two research conditions at T12 with those at T0. At the time we planned our study, there were no studies sufficiently comparable to ours to estimate an effect size. We assumed the effect size Cohen's d ¼0.5, because we consider this a clinically relevant effect. With an α of 0.05 (two-tailed), and a power (1  β) of 0.80, the required sample size was 2  63 patients. In cases of cluster randomization the rule of thumb is to add 25% to this amount, bringing the total to 2  79. Taking into account an expected drop out of 30%, a sample of 103 patients in each group was deemed necessary. 5. Results 5.1. Flow of participants and sample characteristics One hundred and thirty-eight patients were included in the analysis. Fig. 2 depicts the flow of participants during the process. Informed consent was obtained from 71 participants in the

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

Allocation

Enrollment

Teams assessed for eligibility n=47

17

Teams not participating n=29 Reasons: Too time consuming (n=6) Other priorities (n=17) Not enough BD patients (n=6)

Teams randomized n=18

Teams allocated to CC (n=9)

Teams allocated to CAU (n=9)

Patients assessed for eligibility (n=155) Refused participation: - too intensive (n=5) - not interested (n=7) - too time consuming (n=3) - too burdening (n=14) - no reaction (n=1) - unknown (n=7) Excluded: - Did not meet inclusion criteria (n=21) - referred to other team (n=3) - 2 teams stopped participation in study, due to organizational reasons (n=38, of whom 15 consented already)

Patients assessed for eligibility (n=143) Refused participation:

- too intensive (n=5) - not interested (n=14) - too time consuming (n=7) - too burdening (n=14) - no reaction (n=15) - unknown (n=2) Excluded: -did not meet inclusion criteria (n=4)

Informed consent (n=82) Initiated allocated intervention (n=82)

Informed consent (n=71) Initiated allocated intervention (n=56)

Follow up

T0 n=138

Lost to follow in study (n=11)

Lost to follow up in study (n=10)

Continued allocated intervention (n=43)

Continued allocated intervention (n=78)

-discontinued CC, because of financial reasons (n=2 ), or CC does not fit personal expectations (n=11) - however completed T12: n=3

-discontinued intervention, because referred to other team (n=2), or financial reasons (n=2) - however completed T12: n=2

Completed assessments at T12 n=45(80.4%)

Completed assessments at T12 n=72 (87.8%)

Analysis

T12 n=117 (84,8%)

Analyzed ITT (n=56)

Analyzed ITT (n=82)

If a questionnaire is missing on all measurements, this respondent is excluded from these particular analyses

If a questionnaire is missing on all measurements, this respondent is excluded from these particular analyses

Fig. 2. CONSORT diagram showing the flow of participants through each stage of the trial.

intervention group, and from 82 participants in the control group. However, 2 teams withdrew from the experimental arm of the study, due to organizational complications, which resulted in 15 patients leaving the experimental group. Finally, 56 patients started the intervention in the CC arm and 82 in the control arm. In total 21 participants were lost to follow-up (CC: n ¼11; controls: n ¼10). Baseline characteristics of patients who dropped out of the study were compared with those who continued participation. Patients who dropped out had significantly more often a family history of bipolar disorder. To determine whether drop-out of the CC-program was predicted by illness characteristics at baseline we conducted a logistic regression. Stopping with CC was predicted only by having suffered from more manic symptoms in the six months preceding baseline. The clustering,

i.e., the nesting of patients within teams, could be ignored, since the χ2 statistic of the likelihood ratio testing the model with against the model without the nesting of patients within teams was non-significant, neither on functioning nor on quality of life. Therefore we will only report results without taking this type of nesting into account. Results from per protocol analyses did not differ from Intention-to-Treat (ITT) analyses, for both functioning and quality of life. Thus, we will report ITT analyses only. Not all assessments were fully completed by all respondents, therefore in the final analyses sample sizes may differ per questionnaire. At T12, data of 117 patients (84.8%) were included in the analyses (CC: 80.4%; controls: 87.8%). Sample characteristics are depicted in Table 1. At baseline, for some variables significant differences appeared to exist between

18

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

Table 1 Sample characteristics.

Socio demographic characteristics Age, mean (sd) Gender female, n (%) Partner, yes, n (%) Total years of education, mean (sd) Clinical characteristics Diagnosis, n (%) Bipolar disorder I Bipolar disorder II Bipolar disorder NOS Nr of (hypo-) manic episodes in lifetime, n (%) 0 2–4 5–10 10–20 More than 20 Nr of depressive episodes in lifetime, n (%) 0 2–4 5–10 10–20 More than 20 Age of onset, mean (sd) Duration of illness, mean (sd) Family history, n (%) Depression Bipolar disorder Suicide or suicide attempts Alcohol use disorder Recent course of illness, at baseline Most recent episode in last year, n (%) (hypo-) manic Depressive No episode Nr of months with depressive symptoms in past 6 months, LCM, mean (sd) Nr of months with manic symptoms in past six months, LCM, mean(sd) Current severity of depression in past week, QIDS, mean(sd) Current severity of mania in past week, ASRM, mean (sd) Functioning, total scale, mean (sd) Health related quality of life, mean (sd)

Control condition N ¼ 82a

Collaborative care N ¼ 56a

p

44.7 49 45 16.9

46.8 39 36 14.2

0.3 0.1 0.2 o 0.001

(11) (59) (55) (3.3)

( 9.8) (70) (67) (3.5)

0.1 49 (61) 28 (35) 4 (5)

39 (72) 11 (20) 2 (4)

2 26 21 12 15

3 20 11 5 11

0.7 (3) (34) (28) (16) (20)

(6) (40) (22) (10) (22) 0.4

2 19 17 15 19 23.9 20.5

(3) (26) (24) (21) (26) (10.0) (11.0)

48 21 12 26

(59) (26) (15) (32)

21 42 15 2.3 1.0 8.1 1.8 23 3.1

(27) (54) (19) (2.2) (1.5) (5.1) (2.4) (13.5) (1.0)

3 13 10 4 14 23.5 23.0 31 21 10 17

(7) (30) (23) (9) (32) (11.6) (12.8)

0.8 0.2

(55) (38) (18) (30)

0.7 0.2 0.6 1.0 0.8

11 29 10 3.2 1.0 10.5 2.3 31.0 2.6

(22) (58) (20) (2) (1.5) (5.5) (3.8) (15.6) (1.0)

0.02 1.0 0.01 0.4 0.002 0.02

Socio demographic characteristics, total number of manic episodes, total number of depressive episodes, type of last episode last year, age of onset, duration of illness, family history: retrieved by selfrating, on Questionnaire Bipolar Illness (QBP-NL 2.0). Diagnosis: retrieved by clinician version QBP-NL 2.0. LCM: Life Chart Method. QIDS: Quick Inventory of Depressive Symptoms. ASRM: Altman Selfrating Mania Scale. a

Number of respondents differs slightly among measurement.

conditions. Concerning illness characteristics, patients randomized to CC reported a higher number of months with depressive symptoms during the six months prior to baseline than patients in the control group (means (SD): CC 3.2 (2.1); controls 2.3 (2.2), p ¼0.02). Patients in the experimental condition had a higher severity of depressive symptoms, in the week preceding baseline (mean (SD): CC 10.5 (5.5); controls 8.1 (5.1), p ¼0.01). Furthermore, patients randomized to CC on average had a lower educational level compared to patients in the control condition (mean (SD): CC 14.2 (3.5) controls 16.9 (3.3), p o0.01). Patients in the CC condition experienced more functional impairments at baseline than patients in the control condition (mean FAST-score (SD) controls 23.0 (13.5); CC 31.0 (15.6), p ¼0.002). Patients in the control condition reported at baseline a better quality of life concerning health related quality of life (mean health related QOL (SD): controls 3.1 (1.0); CC 2.6 (1.0), p ¼0.02). 5.2. Functioning After twelve months, overall functioning improved more in patients in the CC group than in participants who received CAU (Table 2), with a small effect size (T0–T12: ES ¼0.3, z ¼  2.5, p ¼0.001). After six months, patients in the CC condition improved

significantly more concerning autonomy than patients in the control condition, and this effect was even stronger after twelve months, with a medium effect size (T0–T12: ES¼0.5, z ¼ 2.9, p¼ 0.004). Patients in the CC group improved more in the domain of leisure time after six and twelve months (T0–T12: ES¼0.4, z¼  2.4, p¼ 0.002), compared to patients who received CAU. Furthermore, after a year, participants in the CC group showed a decrease of impairments of cognitive functioning, compared to participants in the control group (T0–T12: ES¼0.4, z ¼  2.8, p¼ 0.005). In the domains work functioning, financial issues and social life, no differences were found between conditions. 5.3. Quality of life No differences were found between conditions for overall quality of life (T0–T12: ES¼ 0.2, z ¼0.9, p ¼0.4) or for health related quality of life, (T0–T12: ES¼ 0.2, z¼ 1.2, p ¼0.4), as depicted in Table 3. Concerning the domain of physical health, quality of life improved significantly more between baseline and twelve months in patients in the CC group, compared to patients in the control condition (T0–T12: ES ¼0.4, z ¼2.5, p¼ 0.01). No effects were found in the domains of psychological health, social life, and environment.

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

19

Table 2 Mean scores and standard deviations of domains and total scale of functioning at baseline and two follow-up measurements, with parameter estimates b, test statistics, p-values and derived effect sizes for between condition and time interaction terms of the Linear Mixed Models. FAST

Autonomy T0 T6 T12 Work functioning T0 T6 T12 Cognitive functioning T0 T6 T12 Financial issues T0 T6 T12 Leisure time T0 T6 T12 Social relationships T0 T6 T12 Total scale T0 T6 T12

Controls

CC

n ¼77a

n¼ 54a

Mean (SD)

Mean (SD)

1.7 (2.1) 2.0 (2.3) 2.1 (2.4)

Interaction term

b

z-statistic

p-value

ES

3.2 (2.9) 2.4 (2.7) 2.4 (2.4)

Condition*T6 Condition*T12

 1.0  1.1

 2.7  2.9

0.008 0.004

0.4 0.5

7.9 (5.0) 7.3 (5.0) 8.6 (5.7)

9.2 (5.7) 8.5 (5.6) 9.0 (5.3)

Condition*T6 Condition*T12

0.0  0.7

0.07  0.9

0.90 0.40

0.1 0.2

5.6 (3.9) 5.5 (3.6) 6.0 (3.7)

8.1 (4.0) 7.0 (4.3) 7.0 (4.5)

Condition*T6 Condition*T12

 0.9  1.5

 1.6  2.8

0.10 0.05

0.2 0.4

1.0 (1.4) 1.1 (1.6) 1.2 (1.7)

1.8 (1.9) 1.7 (1.8) 2.0 (2.1)

Condition*T6 Condition*T12

 0.4  0.1

 1.8  0.5

0.08 0.60

0.2 0.1

1.6 (1.6) 1.7 (1.6) 2.0 (1.8)

2.6 (2.0) 2.0 (2.0) 2.3 (1.9)

Condition*T6 Condition*T12

 0.7  0.7

 2.4  2.4

0.02 0.02

0.4 0.4

5.3 (4.2) 5.0 (3.8) 5.1 (4.0)

6.7 (4.5) 5.7 (4.2) 6.0 (3.9)

Condition*T6 Condition*T12

 0.6  0.7

 1.0  1.1

0.30 0.30

0.1 0.2

23.0 (13.5) 22.6 (13.4) 24.8 (15.5)

31.0 (15.6) 27.0 (16.6) 28.4 (15.3)

Condition*T6 Condition*T12

 3.6  4.8

 1.9  2.5

0.06 0.01

0.2 0.3

The regression parameter Condition*T6 concerns the course of functioning between baseline and 6 months, and reflects to what extent patients in the CC-condition improved more (mean score) than patients in the CAU condition. FAST: Functioning Assessment Short Test. CC: Collaborative Care. SD: Standard deviation. ES: (between condition) effect size derived from estimated means. T0 ¼baseline assessment, T6 ¼ six months assessment, T12¼ twelve months assessment. a

Sample sizes might differ slightly between assessments.

5.4. Mediation through effects on depressive symptoms Prior to any test of mediation, two conditions need to have been met (Baron and Kenny, 1986): (i) the treatment has an effect on the outcome under study, and (ii) the presumed mediating variable predicts the outcome measure. In our case, both conditions were met. The next step in the mediation analyses showed that the effect of CC on functioning at T12 was reduced when depression severity was added to the analysis (b¼ 2.8, SE ¼1.9, t¼  1.4, p¼ 0.15), and that depression severity retained a significant effect on functioning at T12 (b¼ 1.1, SE ¼0.2, t¼5.6, p o0.001). Since the uncorrected effect of the CC condition was  5.5 points on the overall function scale, and the corrected effect (for depression severity) was  2.8 points, we conclude that depression severity accounts for 50% of the effect of CC on overall functioning. Concerning overall quality of life, analyses showed no mediation through depressive symptoms.

6. Discussion We investigated the effect of CC on functioning and quality of life in patients with bipolar disorder, in a cluster-randomized controlled trial. We found that patients who received CC during a year improved significantly more with regard to overall functioning, as well as in the domains autonomy, cognition, and leisure time, when compared to patients treated in Care as Usual, with small to medium effect sizes. No differences were found in the domains work functioning, financial issues, and social life. No differences were found on overall quality of life and health related

quality of life, nor in psychological health, social life, and environment. However, patients randomized to CC showed more improvement in the domain physical health, with a medium effect size. Improvement of depression severity accounted for 50% of the effect on overall functioning, suggesting that CC has a direct effect on functioning next to its indirect effect on functioning via improvement of depressive symptoms (Fig. 1). Our findings confirm findings of other studies, in that depressive symptoms are strongly associated with impaired functioning and quality of life (Altshuler et al., 2006; Bonnin et al., 2012; Goldberg and Harrow, 2011; Goossens et al., 2008; Judd et al., 2005; Rosa et al., 2012). In our study 50% of the effect of CC on overall functioning is explained by reducing depressive symptoms. Studies reporting effects of CC on functioning and quality of life in patients with bipolar disorder are still scarce. In the study of Bauer et al. (2006), improvements were found on functioning as well as quality of life. Since no improvements were found on depressive symptoms, this suggests that in this study the improvement of functioning and quality of life was not mediated by improvement of depression. In a pilot study of a CC program in patients with cardiovascular risks and BD (Kilbourne et al., 2012), a very small but significant effect was found on functioning (ES¼ 0.2). Another CC program, extended to improve medical problems in patients with BD, showed an effect on quality of life, with a small effect size (Kilbourne et al., 2008). We included Problem Solving Treatment in our CC program, specifically aimed at improving depressive symptoms. PST may at least in part account for the effect we found on depressive symptoms, and via this effect, also on functioning. The rationale of PST is that by increasing problem-solving skills, the patients'

20

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

Table 3 Mean scores and standard deviations of domains and total scale of quality of life at baseline and two follow-up measurements with parameter estimates b, test statistics, p-values and derived effect sizes for between condition and time interaction terms of the Linear Mixed Models. Quality of life

Physical health T0 T6 T12 Psychological health T0 T6 T12 Social life T0 T6 T12 Environment T0 T6 T12 Overall QoL T0 T6 T12 Health related QoL T0 T6 T12

Controls

CC

n¼ 77a

n ¼54a

Mean (SD)

Mean (SD)

Interaction term

60.9 (15.1) 58.6 (17.1) 57.7 (17.6)

54.4 (16.2) 53.7 (17.8) 56.5 (18.0)

Condition*T6 Condition*T12

2.3 5.8

52.8 (17.0) 53.6 (17.9) 52.2 (18.4)

47.0 (17.0) 48.2 (17.3) 50.1 (17.5)

Condition*T6 Condition*T12

56.1 (21.3) 56.7 (20.6) 57.8 (19.6)

52.9 (19.5) 55.6 (20.4) 54.7 (16.8)

69.8 (16.9) 72.5 (16.4) 71.6 (15.8)

b

z-statistic

p-value

ES

1.0 2.5

0.3 0.01

0.2 0.4

1.5 4.6

0.6 1.8

0.5 0.07

0.1 0.3

Condition*T6 Condition*T12

3.3 3.3

1.0 0.6

0.3 0.6

0.2 0.1

67.1 (15.1) 67.0 (14.0) 68.9 (14.5)

Condition*T6 Condition*T12

 1.8 1.3

 0.8 0.5

0.4 0.6

0.1 0.1

3.4 (0.9) 3.5 (1.0) 3.3 (0.9)

3.3 (1.0) 3.3 (1.0) 3.4 (0.8)

Condition*T6 Condition*T12

0.0 0.2

0.2 0.9

0.9 0.4

0.1 0.2

3.1 (1.0) 3.0 (1.0) 3.0 (0.9)

2.6 (1.0) 2.9 (1.0) 2.8 (1.1)

Condition*T6 Condition*T12

0.3 0.2

1.4 1.2

0.2 0.2

0.3 0.2

The regression parameter Condition*T6 concerns the course of quality of life between baseline and 6 months, and reflects to what extent patients in the CC-condition improved more (mean score) than patients in the CAU condition. CC: Collaborative Care. SD: Standard deviation. ES: (between condition) effect size derived from estimated means. QoL: Quality of life. T0¼ baseline assessment, T6 ¼six months assessment, T12¼twelve months assessment. a

Sample sizes might differ slightly between assessments.

understanding of the relationship between everyday problems and mood increases, resulting in regaining control over his/her own life, by setting goals and actively working on achieving these goals. It is possible that PST did affect functional impairments not only via depression, but in a direct way as well. PST may have resulted in improving the ability to take decisions (autonomy), improving the ability of solving problems (cognitive functioning), and improving the capacity of performing activities in sport or hobbies, stimulated by the plans that resulted from PST-sessions (leisure time). Moreover, the fact that CC was organized as a protocolized and systematic treatment method, may have contributed to these domains of functioning as well, encouraging patients to become actively involved in treatment, and thereby improving their self-management skills. In quality of life an improvement was found only on the domain of physical health. This domain includes, among others, activities of daily living, energy, and work capacity. The way CC was organized, and PST in particular, may have contributed to this effect, encouraging patients to take more initiative in their lives. Surprisingly, overall quality of life did not improve. Since the WHO-Qol-bref has proven to be sensitive for change (van de Willige et al., 2005), one would expect it to change along with symptoms and functioning. Quality of life is a complex concept, and its association with remission of symptoms and functioning remains to be unraveled (Rubio et al., 2013). Our study has several strong features. First, since only few studies reported on the effect of CC for patients with BD on functioning and quality of life, our study adds in an important way to the limited knowledge. This is also true for insights about the underlying mechanisms of CC. We showed a mediating effect of 50 percent of depressive symptoms on improvement in functioning, and in addition explained possible mechanisms of a direct influence of the program on functioning via activating patients to

self-manage their problems. Second, our study was a pragmatic trial carried out within everyday treatment conditions which contributes to the generalizability of the results. Third, so far, no other CC program for bipolar disorder included PST, which in our study may have added to the improvement of depressive symptoms (Mynors Wallis et al., 2000). Our study has some limitations. First, at baseline, significant differences were found between treatment conditions. This may be explained by the way patients were invited for the study. Patients randomized to the CC condition received detailed information about the intervention. Patients who considered themselves as rather stable may have declined to participate, expecting the intervention to be too intensive for them, given the limited severity of symptoms and problems. Patients in the control condition did not receive this kind of detailed information, since treatment they would receive was not about to change. If less instable patients did not participate in the experimental condition, this may have led to the differences between groups. If and how these baseline differences influenced the results of our study is not straightforward. One possible hypothesis is that patients who are more severely ill at baseline are expected to improve more. Another possible hypothesis however, is that patients who are more severely ill at baseline are expected to remain more severely ill. These are contradicting views, however, both could be true. Moreover, one should not only take the p-value in consideration, in deciding if an intervention is clinically effective or not. Effect sizes are more meaningful outcomes for clinicians. Therefore, we decided to report the uncorrected results in the main article. However, in post-hoc analysis we adjusted for the differences at baseline (Twisk, 2013). The adjusted effects on overall functioning and on the domain leisure time were no longer significant. For the domain of functioning autonomy, a trend for significance remained after adjustment, with a medium effect size (ES ¼0.5,

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

z ¼  1.8, p ¼0.08). Concerning quality of life, on physical health after adjustment a trend for significance was found, with a small effect size (ES¼ 0.3, z¼  1.8, p¼ 0.07). A second limitation is the limited sample size, which may have reduced statistical power. The statistical power of the study was further reduced by the fact that two teams withdraw of the experimental arm of the study after randomization, caused by organizational problems unrelated to the study. A third limitation is the fact that blinding was not performed in the analyses, due to pragmatic reasons. In conclusion, CC is a promising intervention, which may help patients to improve functioning and physical health (e.g. activities of daily living, energy, and work capacity), which is an important aspect of quality of life, in patients with bipolar disorder. Reducing depression severity accounts for half of the improvement in functioning. Moreover, collaboration being systematically performed, as well as PST, may have a direct effect on functioning.

Authors' contributions TV drafted this paper and it was modified by all other authors. AB, RK, BM, PG en TV contributed to the design of the study protocol. TV, RK, PG and BM designed the intervention protocol, which was discussed and enhanced with the expert panel. TV executed the research. AH contributed to the analyses. All authors contributed to and approved the final manuscript.

Role of funding sources Funding sources were VU University, Department of Psychiatry, Amsterdam, The Netherlands. Inholland, University of applied Sciences, Department of Health, Sports & Welfare/Cluster Nursing, Research Group Mental Health Nursing, Amsterdam, The Netherlands Dimence Mental Health, Deventer, The Netherlands.

Conflict of interest No conflict declared.

Acknowledgments We have no acknowledgments to make.

References Altshuler, L.L., Post, R.M., Black, D.O., Keck, P.E.J., Nolen, W.A., Frye, M.A., Suppes, T., Grunze, H., Kupka, R.W., Leverich, G.S., McElroy, S.L., Walden, J., Mintz, J., 2006. Subsyndromal depressive symptoms are associated with functional impairment in patients with bipolar disorder: results of a large, multisite study. J. Clin. Psychiatry 67, 1551–1560. American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders, third edition American Psychiatric Publishing Incorporated, Text Revision (DSM-IV-TR). Baron, R.M., Kenny, D.A., 1986. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J. Personal. Soc. Psychol. 51, 1173–1182. Bauer, M.S., McBride, L., Williford, W.O., Glick, H., Kinosian, B., Altshuler, L., Beresford, T., Kilbourne, A.M., Sajatovic, M., 2006. Collaborative care for bipolar disorder: part II. Impact on clinical outcome, function, and costs. Psychiatr. Serv. 57, 937–945. Beekman, A.T., van der Feltz-Cornelis, C., van Marwijk, H.W., 2013. Enhanced care for depression. Curr. Opin. Psychiatry 26, 7–12. Bellack, A.S., Green, M.F., Cook, J.A., Fenton, W., Harvey, P.D., Heaton, R.K., Laughren, T., Leon, A.C., Mayo, D.J., Patrick, D.L., Patterson, T.L., Rose, A., Stover, E., Wykes, T., 2007. Assessment of community functioning in people with schizophrenia and other severe mental illnesses: a white paper based on an NIMH-sponsored workshop. Schizophr. Bull. 33, 805–822.

21

Bonnin, C.M., Sanchez-Moreno, J., Martinez-Aran, A., Sole, B., Reinares, M., Rosa, A.R., Goikolea, J.M., Benabarre, A., yuso-Mateos, J.L., Ferrer, M., Vieta, E., Torrent, C., 2012. Subthreshold symptoms in bipolar disorder: impact on neurocognition, quality of life and disability. J. Affect. Disord. 136, 650–659. Denicoff, K.D., Leverich, G.S., Nolen, W.A., Rush, A.J., McElroy, S.L., Keck, P.E., Suppes, T., Altshuler, L.L., Kupka, R., Frye, M.A., Hatef, J., Brotman, M.A., Post, R.M., 2000. Validation of the prospective NIMH-Life-Chart Method (NIMHLCM-p) for longitudinal assessment of bipolar illness. Psychol. Med. 30, 1391–1397. Goldberg, J.F., Harrow, M., 2011. A 15-year prospective follow-up of bipolar affective disorders: comparisons with unipolar nonpsychotic depression. Bipolar Disord. 13, 155–163. Goossens, P.J.J., Hartong, E.G., Knoppert-van der Klein, E.A.M., Achterberg, T.v., 2008. Self-reported psychopathological symptoms and quality of life in outpatients with bipolar disorder. Perspect. Psychiatr. Care 44, 275–284. Goossens, P.J.J., Kupka, R.W., Beentjes, T.A., Achterberg, T.v., 2010. Recognising prodromes of manic or depressive recurrence in outpatients with bipolar disorder: a cross-sectional study. Int. J. Nurs. Stud. 47, 1201–1207. Honig, A., Hofman, A., Rozendaal, N., Dingemans, P., 1997. Psycho-education in bipolar disorder: effect on expressed emotion. Psychiatry Res. 72, 17–22. Ishak, W.W., Brown, K., Aye, S.S., Kahloon, M., Mobaraki, S., Hanna, R., 2012. Healthrelated quality of life in bipolar disorder. Bipolar Disord. 14, 6–18. Judd, L.L., Akiskal, H.S., Schettler, P.J., Endicott, J., Maser, J.D., Solomon, D.A., Leon, A.C., Rice, J.A., Keller, M.B., 2002. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch. Gen. Psychiatry 59, 530–537. Judd, L.L., Akiskal, H.S., Schettler, P.J., Endicott, J., Leon, A.C., Solomon, D.A., Coryell, W., Maser, J.D., Keller, M.B., 2005. Psychosocial disability in the course of bipolar I and II disorders: a prospective, comparative, longitudinal study. Arch. Gen. Psychiatry 62, 1322–1330. Kilbourne, A.M., Goodrich, D.E., Lai, Z., Clogston, J., Waxmonsky, J., Bauer, M.S., 2012. Life Goals Collaborative Care for patients with bipolar disorder and cardiovascular disease risk. Psychiatr. Serv. 63, 1234–1238. Kilbourne, A.M., Post, E.P., Nossek, A., Drill, L., Cooley, S., Bauer, M.S., 2008. Improving medical and psychiatric outcomes among individuals with bipolar disorder: a randomized controlled trial. Psychiatr. Serv. 59, 760–768. Kupka, R.W., Altshuler, L., Nolen, W.A., Suppes, T., Luckenbaugh, D.A., Leverich, G.S., Frye, M.A., Keck Jr., P.E., McElroy, S.L., Grunze, H., Post, R.M., 2007. Three times more days depressed than manic or hypomanic in both bipolar I and bipolar II disorder. Bipolar Disord. 9, 531–535. Leverich, G.S., Nolen, W.A., Rush, A.J., McElroy, S.L., Keck, P.E., Denicoff, K.D., Suppes, T., Altshuler, L.L., Kupka, R., Kramlinger, K.G., Post, R.M., 2001. The Stanley Foundation Bipolar Treatment Outcome Network. I. Longitudinal methodology. J. Affect. Disord. 67, 33–44. Michalak, E.E., Yatham, L.N., Lam, R.W., 2005. Quality of life in bipolar disorder: a review of the literature. Health Qual. Life Outcomes 3, 72. Morris, S.B., 2008. Estimating effect sizes from pretest–posttest–controlgroup designs. Organ. Res. Methods 11, 364–386. Mynors Wallis, L.M., Gath, D.H., Day, A., Baker, F., 2000. Randomised controlled trial of problem solving treatment, antidepressant medication, and combined treatment for major depression in primary care. Gen. Pract. BMJ 320, 26–30. National Institute for Health and Clinical Excellence, 2014. Bipolar Disorder (update): The Management of Bipolar Disorder in Adults, Children and Adolescents in Primary and Secondary Care (Clinical Guideline), 185. National Institute for Health and Clinical Excellence. Phelan, M., Slade, M., Thornicroft, G., Dunn, G., Holloway, F., Wykes, T., Strathdee, G., Loftus, L., McCrone, P., Hayward, P., 1995. The Camberwell Assessment of Need: the validity and reliability of an instrument to assess the needs of people with severe mental illness. Br. J. Psychiatry 167, 589–595. Rosa, A.R., Gonzalez-Ortega, I., Gonzalez-Pinto, A., Echeburua, E., Comes, M., Martinez-Aran, A., Ugarte, A., Fernandez, M., Vieta, E., 2012. One-year psychosocial functioning in patients in the early vs. late stage of bipolar disorder. Acta Psychiatr. Scand. 125, 335–341. Rosa, A.R., Sanchez-Moreno, J., Martinez-Aran, A., Salamero, M., Torrent, C., Reinares, M., Comes, M., Colom, F., Van Riel, W., Ayuso-Mateos, J.L., Kapczinski, F., Vieta, E., 2007. Validity and reliability of the Functioning Assessment Short Test (FAST) in bipolar disorder. Clin. Pract. Epidemiol. Ment. Health 3, 5. Rubio, J.M., Olfson, M., Villegas, L., Perez-Fuentes, G., Wang, S., Blanco, C., 2013. Quality of life following remission of mental disorders: findings from the National Epidemiologic Survey on Alcohol and Related Conditions. J. Clin. Psychiatry 74, e445–e450. Skevington, S.M., Lotfy, M., O’Connell, K.A., 2004. The World Health Organization's WHOQOL-BREF quality of life assessment: psychometric properties and results of the international field trial. A report from the WHOQOL group. Qual. Life Res. 13, 299–310. Spearing, M.K., Post, R.M., Leverich, G.S., Brandt, D., Nolen, W., 1997. Modification of the Clinical Global Impressions (CGI) Scale for use in bipolar illness (BP): the CGI-BP. Psychiatry Res. 73, 159–171. Strejilevich, S.A., Martino, D.J., Murru, A., Teitelbaum, J., Fassi, G., Marengo, E., Igoa, A., Colom, F., 2013. Mood instability and functional recovery in bipolar disorders. Acta Psychiatr. Scand. 128, 194–202. Trompenaars, F.J., Masthoff, E.D., Van Heck, G.L., Hodiamont, P.P., De Vries, J., 2005. Content validity, construct validity, and reliability of the WHOQOL-Bref in a population of Dutch adult psychiatric outpatients. Qual. Life Res. 14, 151–160.

22

T.Y.G. van der Voort et al. / Journal of Affective Disorders 179 (2015) 14–22

Twisk, J.W.R., 2013. Applied Longitudinal Data Analysis for Epidemiology: A Practical Guide. Cambridge University Press, New York. van de Willige, G., Wiersma, D., Nienhuis, F.J., Jenner, J.A., 2005. Changes in quality of life in chronic psychiatric patients: a comparison between EuroQol (EQ-5D) and WHOQoL. Qual. Life Res. 14, 441–451. van der Voort, T.Y.G., van Meijel, B., Goossens, P.J.J., Renes, J., Beekman, A.T.F., Kupka, R.W., 2011. Collaborative care for patients with bipolar disorder: a randomised controlled trial. BMC Psychiatry 11, 133. Vieta, E., Sanchez-Moreno, J., Lahuerta, J., Zaragoza, S., 2008. Subsyndromal depressive symptoms in patients with bipolar and unipolar disorder during clinical remission. J. Affect. Disord. 107, 169–174. van der Voort, T.Y.G., van Meijel, B., Goossens, P.J.J., Hoogendoorn, A.W., Draisma, S., Beekman, A., Kupka, R.W., 2015. Collaborative care for patients with bipolar disorder: randomised controlled trial. Br. J. Psychiatry 2015.

Wagner, E.H., Austin, B.T., Von Korff, M., 1996. Organizing care for patients with chronic illness. Milbank Q. 74, 511–544. World Health Organization. International Classification of Functional Disabilities and Health: ICF. 2001. World Health Organization Division of Mental Health and Prevention of Substance Abuse. WHOQOL Measuring of Quality of Life. 1997. Yatham, L.N., Kennedy, S.H., Parikh, S.V., Schaffer, A., Beaulieu, S., Alda, M., O’Donovan, C., Macqueen, G., McIntyre, R.S., Sharma, V., Ravindran, A., Young, L.T., Milev, R., Bond, D.J., Frey, B.N., Goldstein, B.I., Lafer, B., Birmaher, B., Ha, K., Nolen, W.A., Berk, M., 2013. Canadian Network for Mood and Anxiety Treatments (CANMAT) and International Society for Bipolar Disorders (ISBD) collaborative update of CANMAT guidelines for the management of patients with bipolar disorder: update 2013. Bipolar Disord. 15, 1–44.