Articles. Funding Cancer Research UK and Chief Scientist Office of the Scottish Government.

Transcription

1 Integrated collaborative care for comorbid major depression in patients with cancer (SMaRT Oncology-2): a multicentre randomised controlled effectiveness trial Michael Sharpe*, Jane Walker*, Christian Holm Hansen, Paul Martin, Stefan Symeonides, Charlie Gourley, Lucy Wall, David Weller, Gordon Murray, for the SMaRT (Symptom Management Research Trials) Oncology-2 Team Summary Background Medical conditions are often complicated by major depression, with consequent additional impairment of quality of life. We aimed to compare the effectiveness of an integrated treatment programme for major depression in patients with cancer (depression care for people with cancer) with usual care. Methods SMaRT Oncology-2 is a parallel-, multicentre, randomised controlled effectiveness trial. We enrolled outpatients with major depression from three cancer centres and their associated clinics in Scotland, UK. Participants were randomly assigned in a 1:1 ratio to the depression care for people with cancer intervention or usual care, with stratification (by trial centre) and minimisation (by age, primary cancer, and sex) with allocation concealment. Depression care for people with cancer is a manualised, multicomponent collaborative care treatment that is delivered systematically by a team of cancer nurses and psychiatrists in collaboration with primary care physicians. Usual care is provided by primary care physicians. Outcome data were collected up until 48 weeks. The primary outcome was treatment response ( 5% reduction in Symptom Checklist Depression Scale [SCL-2] score, range 4) at 24 weeks. Trial statisticians and data collection staff were masked to treatment allocation, but participants could not be masked to the allocations. Analyses were by intention to treat. This trial is registered with Current Controlled Trials, number ISRCTN Findings 5 participants were enrolled between May 12, 28, and May 13, 211; 253 were randomly allocated to depression care for people with cancer and 247 to usual care. 143 (62%) of 231 participants in the depression care for people with cancer and 4 (17%) of 231 in the usual care responded to treatment: absolute difference 45% (95% CI 37 53), adjusted odds ratio 8 5 (95% CI ), p< 1. Compared with patients in the usual care, participants allocated to the depression care for people with cancer programme also had less depression, anxiety, pain, and fatigue; and better functioning, health, quality of life, and perceived quality of depression care at all timepoints (all p< 5). During the study, 34 cancer-related deaths occurred (19 in the depression care for people with cancer, 15 in the usual care ), one patient in the depression care for people with cancer was admitted to a psychiatric ward, and one patient in this attempted suicide. None of these events were judged to be related to the trial treatments or procedures. Interpretation Our findings suggest that depression care for people with cancer is an effective treatment for major depression in patients with cancer. It offers a model for the treatment of depression comorbid with other medical conditions. Funding Cancer Research UK and Chief Scientist Office of the Scottish Government. Introduction Major depression is a leading cause of disability and, when comorbid with a chronic disease, is associated with reduced quality of life and increased health-care costs. 1 3 However, the treatment of comorbid depression is often inadequate. 4 6 Cancer is becoming a chronic disease for a rapidly increasing number of people. 7 Major depression affects roughly 1% of patients with cancer and is associated with worse anxiety, pain, fatigue, and functioning; suicidal thoughts; and poor adherence to anticancer treatments. 6,8 13 The need for better management of comorbid depression that is integrated into patients cancer care has been emphasised in several recent reports However, at present we do not have good evidence for how best to treat major depression in patients with cancer and how to integrate this depression treatment into their cancer care To address this problem, we developed a multicomponent integrated treatment programme, called depression care for people with cancer. 2 This manualised treatment programme is based on the collaborative care model in which a psychiatrist and a care manager collaborate with the patient s primary care physician to provide systematic, proactive treatment and follow-up. 21 In depression care for people with cancer, the collaborative care model is expanded to include integration with the patient s specialist medical care. In a previous single-centre efficacy Lancet 214; 384: Published Online August 28, S (14) See Articles Lancet Psychiatry 214; published online Aug 28 S (14)7313-X See Articles Lancet Oncology 214; published online Aug 28 S (14) See Comment page 176 *Joint first authors Psychological Medicine Research, University of Oxford Department of Psychiatry, Warneford Hospital, Oxford, UK (Prof M Sharpe MD, J Walker PhD); MRC Tropical Epidemiology Group, London School of Hygiene and Tropical Medicine, London, UK (C Holm Hansen PhD); Psychological Medicine Research, University of Edinburgh Department of Psychiatry, Edinburgh, UK (P Martin MSc); University of Edinburgh Cancer Research UK Centre, Western General Hospital, Edinburgh, UK (S Symeonides PhD, Prof C Gourley PhD, L Wall MD); and University of Edinburgh Centre for Population Health Sciences, Teviot Place, Edinburgh, UK (Prof D Weller PhD, Prof G Murray PhD) Correspondence to: Prof Michael Sharpe, Psychological Medicine Research, University of Oxford Department of Psychiatry, Warneford Hospital, Oxford OX3 7JX, UK michael.sharpe@psych.ox.ac.uk See Online for podcast interview with Michael Sharpe and Jane Walker Vol 384 September 2,

2 See Online for appendix 558 declined to participate 26 not interested 52 did not feel depressed 12 uncontactable 18 did not want help for depression 23 too busy 22 felt physically too unwell 26 preferred other type of help 37 did not want to attend appointments 253 allocated to depression care for people with cancer 22 received 4 sessions 23 received 1 3 sessions 4 died 19 declined further sessions 1 received no sessions 1 died 9 declined sessions Data completeness At 12 weeks: 235 obtained, 8 died, 1 missed At 24 weeks: 231 obtained, 9 died, 13 missed At 36 weeks: 224 obtained, 14 died, 15 missed At 48 weeks: 223 obtained, 19 died, 11 missed Figure 1: Trial profile *Regularly seeing a mental health specialist. trial (SMaRT Oncology-1) we obtained proof-of-principle of this approach to treating depression in patients with cancer. 22 We subsequently developed the depression care for people with cancer programme, including rigorous training and quality assurance procedures to make it suitable for widespread implementation. In this multicentre effectiveness trial (SMaRT Oncology-2), we aimed to establish whether or not depression care for people with cancer is better than usual care in achieving a clinically useful improvement in depression for patients with cancer who have comorbid major depression and a survival prognosis of at least 1 year. Methods Study design and participants We did a parallel, two-, multicentre randomised controlled trial in three cancer centres in Scotland, UK 1428 patients assessed for eligibility 5 patients randomised 37 not eligible 3 unable to provide consent 9 no cancer diagnosis 21 prognosis <12 months 11 depressed for <4 weeks 6 cognitive/communication impairment 33 unable to attend sessions 62 chronic depression 58 psychiatric disorder requiring alternative treatment 1 medical disorder requiring alternative treatment 5 other clinical reason 8 needed urgent psychiatric care 6 cerebral metastases 48 receiving active psychiatric care* 247 allocated to usual care Data completeness At 12 weeks: 241 obtained, 4 died, 2 missed At 24 weeks: 231 obtained, 8 died, 8 missed At 36 weeks: 226 obtained, 13 died, 8 missed At 48 weeks: 224 obtained, 15 died, 8 missed (Glasgow, Edinburgh, and Dundee) and their associated clinics. The trial protocol, including an outline of the statistical analysis plan, has previously been published. 23 We enrolled adults (aged 18 years) with a diagnosis of cancer, a good cancer prognosis (predicted survival 12 months estimated by their cancer specialist) and major depression (Diagnostic and Statistical Manual of Mental Disorders, 4th Edition [DSM-IV] criteria using the inclusive approach to diagnosis) 24,25 of at least 4 weeks duration. We excluded patients if they were unable to participate in depression care for people with cancer (those with substantial cognitive or communication difficulties, or who could not attend regular sessions) or if the depression care for people with cancer programme was inappropriate to their needs (those with continuous depression for 2 years, a psychiatric or medical condition requiring alternative treatment, known cerebral metastases, or those already regularly seeing a mental health specialist). Potential participants were identified by a clinical service offering depression screening to all patients attending selected National Health Service (NHS) cancer clinics in Scotland, UK. All patients who were diagnosed with probable major depression by this screening service (according to the major depression section of the Structured Clinical Interview for DSM-IV [SCID]) 26 were offered referral to the trial. Research clinicians (nurses and psychiatrists) obtained written informed consent from those patients who agreed to this referral, assessed whether they were eligible for trial participation (including reassessment of depression), and informed eligible patients of their treatment allocation after randomisation. The trial was approved by the Scotland A Research Ethics Committee (8/MRE/23) and trial conduct was overseen by trial steering and data monitoring committees (appendix p 5). All participants provided written informed consent. Randomisation and masking Research clinicians collected baseline data immediately before enrolment and entered these data in a secure web-based randomisation database implemented by a trials unit. A database software algorithm allocated participants to either depression care for people with cancer or usual care in a 1:1 ratio using a combination of stratification (by trial centre) and minimisation (by age, primary cancer, and sex). Minimisation variables were balanced within each trial centre using a random element in the algorithm to ensure allocation concealment. Trial statisticians and staff who collected outcome data were masked to allocated interventions; however, participants could not be masked because of the nature of depression care for people with cancer. Procedures In the usual care, the participant s primary care physician and oncologist were informed about the major depression diagnosis and asked to treat their 11 Vol 384 September 2, 214

3 patients as they normally would (in the UK, the primary care physician might decide to prescribe antidepressants or to refer the patient to mental health services for assessment or for psychological treatment). The patient was encouraged to consult their primary care physician to obtain treatment. In the depression care for people with cancer, the participant s primary care physician and oncologist were informed of the diagnosis of major depression and that the participant would be seen by a specially trained nurse under the supervision of a psychiatrist. We have described depression care for people with cancer in detail elsewhere. 2 In summary, it is an intensive, manualised, collaborative care-based multicomponent treatment programme specifically designed to be integrated with the patient s cancer treatment. The depression care for people with cancer programme is systematically delivered by a team that comprises cancer nurses (trained to competence in depression care for people with cancer [see appendix]) and supervising consultation-liaison psychiatrists working in collaboration with the patient s oncology team and primary care physician. The nurses establish a therapeutic relationship with the patients, provide information about depression and its treatment, deliver brief evidence-based psychological inter ventions (problem-solving therapy and behavioural activation), 27,28 and monitor patients progress (on the Patient Health Questionnaire [PHQ]-9 depression severity scale). 29 The psychiatrists supervise treatment, aiming to achieve and maintain treatment targets (PHQ-9 <1 and 5% drop from baseline), advise primary care physicians about prescribing anti depressants, and provide direct consult ations to patients who are not improving. The initial treatment phase comprises a maximum of ten sessions with the nurse (at the cancer or primary care clinic, or if necessary by telephone) over a 4-month period. After this initial treatment period, PHQ-9 scores are monitored monthly by telephone (through an automated system supplemented by nurse calls) for a further 8 months; additional sessions with the nurse are provided for patients not meeting treatment targets. All cases are reviewed on a weekly basis in supervision meetings attended by nurses and a psychiatrist. Supervision is informed by an electronic registry that generates graphs of patients PHQ-9 depression scores over time, provides alerts when these scores do not meet predefined targets, and stores digital video recordings of every treatment session for easy review (see appendix p 6). In the trial, a random number generator was used to select a 1% sample of these video recordings, which was independently rated for adherence to the treatment manual (with use of a checklist that specified the behaviours that the nurse was expected to use in each session) and quality of delivery (using a 1 scale from very poor to very good). Outcomes The primary (categorical) trial outcome was treatment response measured at 24 weeks. Treatment response was defined as at least a 5% reduction in depression severity from baseline on the self-rated Symptom Checklist Depression Scale (SCL-2) version B. 3 The SCL-2 has 2 items, each of which is rated from 4 and the patient s overall SCL-2 score at each measurement is an average of these (with higher scores indicating more severe symptoms); this widely used scale is derived from the SCL A 5% reduction in score has been shown to be Depression care for people with cancer (n=253) Usual care (n=247) Mean age (years) 56 6 (1 ) 56 1 (1 2) Age range (years) Age (years) 5 77 (3%) 72 (29%) (38%) 97 (39%) >6 79 (31%) 78 (32%) Sex Women 227 (9%) 222 (9%) Men 26 (1%) 25 (1%) Marital status Spouse/partner 156 (62%) 151 (61%) No spouse/partner 97 (38%) 96 (39%) Employment status Not working 137 (54%) 126 (51%) Working 45 (18%) 52 (21%) Retired 71 (28%) 69 (28%) Duration of current depressive episode*(months) (52%) 123 (5%) (21%) 75 (31%) (27%) 46 (19%) Number of previous depressive episodes None 77 (31%) 87 (35%) One 93 (37%) 13 (42%) More than one 82 (33%) 56 (23%) Antidepressant use at trial entry At any dose 11 (4%) 94 (38%) At minimum effective dose or higher 71 (28%) 64 (26%) Time since most recent cancer diagnosis (months) 14 (7 36) 16 (8 38) Primary cancer Breast 14 (55%) 131 (53%) Gynaecological 57 (23%) 64 (26%) Genitourinary 13 (5%) 14 (6%) Other 43 (17%) 38 (15%) Cancer treatment (preceding 2 months) No active treatment 96 (38%) 87 (35%) Surgery 14 (6%) 1 (4%) Hormone therapy 97 (38%) 94 (38%) Radiotherapy 22 (9%) 16 (7%) Chemotherapy 47 (19%) 53 (22%) Other 21 (8%) 17 (7%) (Table 1 continues on next page) Vol 384 September 2,

4 Depression care for people with cancer (n=253) Usual care (n=247) (Continued from previous page) Cancer status Non-active disease: follow-up/adjuvant treatment 2 (79%) 196 (79%) Active disease: radical treatment 19 (8%) 12 (5%) Active disease: palliative treatment 34 (13%) 39 (16%) Baseline scores SCL-2 depression score ( 4) 2 1 ( 62) 2 11 ( 56) SCL-1 anxiety score ( 4) 1 32 ( 84) 1 35 ( 81) Pain ( 1) 44 5 (34 8) 47 (32 6) Fatigue ( 1) 62 4 (22 4) 62 5 (22 2) Physical functioning ( 1) 58 3 (22 7) 6 6 (22 1) Social functioning ( 1) 45 1 (28 7) 46 5 (28 5) Role functioning ( 1) 43 5 (27 7) 47 (27 7) Overall health ( 1) 43 5 (19 9) 44 5 (19 8) Quality of life ( 1) 4 5 (21 4) 43 (19 ) Perceived quality of depression care Poor 8 (32%) 61 (25%) Fair 83 (33%) 89 (36%) Good 52 (21%) 66 (27%) Very good 21 (8%) 25 (1%) Excellent 15 (6%) 6 (2%) Data are mean (SD), n (%), or median (IQR), unless otherwise indicated. SCL=Symptom Check List. Pain and fatigue (for which higher scores indicate poorer health), and physical functioning, social functioning, and role functioning, overall health, and quality of life (for which higher scores indicate better health) are from the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire. *Depression care for people with cancer n=249, usual care n=244. One significant difference was recorded between the s for duration of current depressive episode (p< 5). Depression care for people with cancer n=252, usual care n=246. Most other cancers were gastrointestinal. Other cancer treatments were monoclonal antibodies, bisphosphonates, and trial drugs. Depression care for people with cancer n=251. Table 1: Baseline characteristics comparable to no longer meeting diagnostic criteria for major depression and was chosen to indicate a clinically useful improvement. 32 Secondary outcomes were: average depression severity (SCL-2 score averaged over 24, 36, and 48 weeks); remission of depression (defined as an SCL-2 score < 75 at all of the 24, 36, and 48 week assessments); 22 and self-rated improvement in depression (five-point scale from much worse to much better). Tertiary outcomes were: anxiety (SCL-1, range 4); 31 pain, fatigue, physical, social, and role func tioning, and overall health and quality of life (on the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire [EORTC-QLQ-C3]); 33 and the patient s perceived quality of depression care (five-point scale from poor to excellent). Outcome data were collected by a dedicated team masked to treatment allocation at 12, 24, 36, and 48 weeks post-randomisation by questionnaires sent in the post, supplemented by telephone calls or home visits when necessary. The team delivering depression care for people with cancer recorded the duration of all treatment sessions, related administrative time, and telephone calls to participants and primary care physicians. Data for other treatments were collected by participant report, supplemented by case note review. Doses of antidepressants were coded according to their minimum effective doses (appendix p 11). 34 We defined serious adverse events as deaths from any cause, admission to a psychiatric ward, or attempted suicide. Statistical analysis We estimated that two treatment s of 25 participants each would give 9% power at the 5% significance level to detect a difference of at least 15 (15%) in the proportion of participants achieving the primary outcome of treatment response. The main analysis was done by CHH (who was masked to treatment allocation and used the SAS software version 9.1.3) at the end of the trial, using the intention-to-treat principle that included all randomised participants with usable outcome data. Effects on treatment response and remission were analysed with logistic regression, self-rated improvement with ordinal regression, and depression severity with analysis of covariance. The stratification and minimisation variables were included as covariates in all analyses and depression severity at outcome was also adjusted for baseline score. Tertiary outcomes were analysed with logistic regression, ordinal regression, or analysis of covariance as appropriate. We did preplanned analyses for five subs (defined by the stratification and minimisation variables plus cancer status at baseline). We undertook sensitivity analyses on the primary outcome to establish the effect of clustering (by the treating nurse in the depression care for people with cancer and by participant in the usual care ), using mixed effects logistic regression and to measure the effect of missing data using multiple imputation and worst-case-scenario assumptions. We calculated the cost of delivering depression care for people with cancer by multiplying the cost of each unit of resource by the amount used. This trial is registered as an international standard randomised controlled trial with Current Controlled Trials, number ISRCTN Role of the funding source The funders had no role in the study design; in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit for publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. Results Between May 12, 28, and May 13, 211, 1428 referrals were made to the trial team by the screening service (figure 1, and appendix p 17). 5 people were eligible and agreeable to participation, 253 of whom were allocated to depression care for people with cancer and 247 to usual care. More than half of participants were women and were Vol 384 September 2, 214

5 6 months 6 12 months Depression care for people with cancer Usual care Depression care for people with cancer Usual care Patient contacts with depression care for people with cancer team Number of treatment sessions with nurse* 1 (4 %) (9 1%) (87 %) Median number of sessions (IQR; range) 9 (7 1; 1) Patient contacts with psychiatrist (in session or by telephone) Patients with 1 contact 9 (35 6%) 36 (15 6%) Median number of contacts (IQR; range) ( 1; 5) ( ; 4) Patient telephone contacts with nurse (additional to follow-up) Patients with 1 contact 218 (86 2%) 162 (7 1%) Median number of contacts (IQR; range) 3 (1 5; 19) 1 (1 3; 1) Antidepressant drug treatment Any antidepressant prescribed 181 (8 8%) 133 (58 1%) 177 (81 6%) 129 (59 4%) Antidepressant prescribed at minimum 169 (75 4%) 11 (48 %) 158 (72 8%) 18 (49 8%) effective dose Dose of antidepressant increased 58 (25 9%) 33 (14 4%) 48 (22 1%) 16 (7 4%) New antidepressant prescribed 136 (6 7%) 73 (31 9%) 43 (19 8%) 25 (11 5%) 2 antidepressants prescribed 5 (22 3%) 23 (1 %) 41 (18 9%) 2 (9 2%) Patient contacts with mental health professionals who were not part of the depression care for people with cancer team Number of consultations Patients with 1 consultation 12 (5 1%) 42 (17 5%) 12 (5 3%) 41 (17 9%) Median number of consultations (IQR; range) ( ; 16) ( ; 15) ( ; 5) ( ; 12) Psychological treatments received Counselling or support 12 (5 1%) 37 (15 4%) 8 (3 5%) 36 (15 7%) Formal psychological treatment 2 ( 8%) 1 ( 4%) 3 (1 3%) Patient contacts with other health professionals Primary care physician consultations Patients with 1 consultation 223 (94 9%) 224 (92 9%) 195 (86 3%) 23 (88 6%) Median number of consultations (IQR; range) 5 (3 8; 32) 5 (3 8; 31) 4 (2 7; 3) 3 (1 7; 34) Consultations with oncology team Patients with 1 consultation 143 (6 9%) 183 (75 9%) 134 (59 %) 146 (63 8%) Median number of consultations (IQR; range) 1 ( 3; 48) 2 (1 3; 42) 1 ( 2; 48) 1 ( 2; 6) Data are n (%) unless otherwise specified. Percentages shown are of number of patients with available data, ranging from 217 to 253 in the depression care for people with cancer and from 217 to 241 in the usual care.*nine patients completed the final depression care for people with cancer session in the second 6 months of trial participation. The same types of antidepressants were prescribed in both s and were: amitriptyline, citalopram, clomipramine, dosulepin, duloxetine, escitalopram, fluoxetine, lofepramine, mirtazapine, nortriptyline, paroxetine, sertraline, trazodone, and venlafaxine. Consultations with the primary care physician and oncology staff might have included treatment for depression but the content of these sessions was not recorded. Psychiatrist, psychologist, counsellor, or psychiatric nurse. Oncologist or cancer nurse. Table 2: Treatments received for depression receiving follow-up or adjuvant cancer treatment. Partic i- pant characteristics at baseline did not differ between the two s, except for a slightly longer duration of current depressive episode in the depression care for people with cancer (table 1 and appendix p 1). Of the participants allocated to the depression care for people with cancer intervention, 96% (243/253) received at least one treatment session and 87% (22/253) at least four sessions (table 2). Most (92%; 153/166) of the video-recorded sessions of depression care for people with cancer reviewed were judged to have included all the specified treatment components; the mean quality rating of treatment delivered in these sessions was high (7 9 out of 1). More than half of the participants in both s were prescribed antidepressants. However, rates of prescription at minimum effective doses were higher in the depression care for people with cancer than in the usual care (appendix p 11) and more antidepressant adjustments were made (new prescriptions, dose Vol 384 September 2,

6 Depression care for people with cancer Usual care Adjusted treatment effect estimate (95% CI)* Unadjusted treatment effect estimate (95% CI) Primary outcome Treatment response (5% reduction on 143 (62%) 4 (17%) 8 5 (5 5 to 13 4) 7 8 (5 to 12 ) < 1 SCL-2) at 24 weeks Secondary outcomes Remission of depression (score < 75 on 72 (33%) 8 (4%) 13 1 (6 1 to 28 2) 13 (6 1 to 27 9) < 1 SCL-2 at each of 24, 36, and 48 weeks) Average depression severity (SCL-2 score 94 ( 67) 1 71 ( 7) 78 ( 9 to 66) 77 ( 9 to 64) < 1 [ 4]) at 24, 36, and 48 weeks Self-rated improvement of depression at 4 1 (2 7 to 6 2)** 4 1 (2 7 to 6 1)** < 1 24 weeks Better/much better 181 (79%) 112 (49%) Has not changed 38 (17%) 76 (33%) Worse/much worse 9 (4%) 39 (17%) Data are n (%) or mean (SD), unless otherwise indicated. Effect estimates are odds ratios (95% CI) from logistic regressions unless otherwise specified. SCL-2=Symptom Check List-2 Depression Subscale. *Adjusted for trial centre, age, cancer type, and sex. Average depression severity was also adjusted for baseline SCL-2 depression score. n=231. n=217. n=218. Mean difference from ANCOVA. n=228 in depression care, n=227 in usual care. **Common odds ratio from an ordinal logistic regression assuming proportional odds. Table 3: Trial outcomes p value increases, prescription of more than one antidepressant drug) in those allocated to depression care for people with cancer than in those allocated to usual care (table 2). Participants allocated to usual care were more likely to see mental health specialists other than those delivering the depression care for people with cancer intervention, but very few patients in either received formal psychological treatment from these professionals. We obtained primary outcome data from 462 (92%) of the 5 participants (462 [96%] of 483 patients who were still alive at 24 weeks [231 in each treatment ]; figure 1). The primary outcome of treatment response was achieved by 62% (143/231) of participants allocated to depression care for people with cancer and 17% (4/231) of those allocated to usual care; the absolute difference in outcome between the treatment s was 45% (95% CI 37 53%) (table 3, figure 2). The odds ratio (OR) for the treatment response in the two s was 8 5 (95% CI ; p< 1) and the number needed to treat was 2 24 (95% CI ). Depression care for people with cancer was also better than usual care for all the secondary and tertiary outcomes. 72 (33%) of 217 participants allocated to the depression care for people with cancer intervention achieved remission of their major depression, compared with only eight (4%) of 218 allocated to usual care (OR 13 1, 95% CI ). Depression scores averaged over the 24, 36, and 48 week timepoints were 78 points lower in the depression care for people with cancer than in the usual care (95% CI 66 9) indicating a standardised mean difference, or effect size, of 1 13 (95% CI ). Self-rated improvement in depression at 24 weeks was much greater with depression care for people with cancer than with usual care (p< 1) (figure 2). Anxiety, pain, fatigue, physical functioning, social functioning, role functioning, overall health, and quality of life were all consistently better with depression care for people with cancer than with usual care at all timepoints (all p< 5) (figure 2, appendix pp 12, 18). The participants ratings of their quality of depression care were notably low at baseline, but rated as excellent or very good by 72% (162/226) of those receiving the depression care for people with cancer programme at 12 weeks compared with only 25% (57/229) of those in the usual care (p< 1). This difference was sustained at 48 weeks (appendix p 2). We analysed the primary and secondary outcomes for differential treatment effects within the subs defined by the stratification and minimisation variables (trial centre, age, primary cancer, and sex) plus follow-up or adjuvant treatment versus active cancer status, and recorded no evidence of interaction effects (appendix p 21). The treatment effect estimates for the primary outcome in the cluster, multiple imputation, and worst-case scenario analyses, were all similar to those obtained from the main analysis. Similarly, the effect estimates on other outcomes were found to be robust with use of different methods for handling missing data (appendix pp 13 16). 34 cancer-related deaths (19 in the depression care for people with cancer, 15 in the usual care ), one admission to a psychiatric ward (a patient in the depression care for people with cancer ), and one attempted suicide (also in the depression care for people with cancer ) occurred during the trial. None of these events were judged to be related to the trial treatments or procedures. The mean additional cost per patient of providing depression care for people with cancer was 613 (95% CI ); this value included the cost of treatment Vol 384 September 2, 214

7 A B 4 Mean score Participants (%) Treatment response at 24 weeks DCPC Usual care Participants (%) Self-rated improvement in depression at 24 weeks DCPC Usual care Much better Better Has not changed Worse Much worse Depression Anxiety Pain Usual care 4 1 DCPC Perceived quality of depression care at 24 weeks DCPC Usual care Excellent Very good Good Fair Poor 1 Fatigue Physical functioning Social functioning 1 1 Mean score 1 Role functioning Overall health Quality of life 1 1 Mean score BL BL BL Time (weeks) Time (weeks) Time (weeks) Figure 2: Trial outcomes (A) Treatment response (5% reduction in Symptom Checklist-2 depression score), self-rated improvement of depression, and perceived quality of depression care at 24 weeks. (B) Mean scores over time at baseline and at 12, 24, 36, and 48 weeks on all continuous scales plotted separately for the DCPC and the usual care. Error bars represent 95% CIs. See also appendix p 12. DCPC=depression care for people with cancer. BL=baseline. sessions ( 429), telephone contacts about depression care ( 48), and treatment supervision ( 136). Full details of treatment costs are described in appendix pp 8 9. Discussion The main finding of SMaRT Oncology-2 is that a much greater number of patients had a clinically useful improvement in depression (treatment response) with depression care for people with cancer than with usual care. The recorded difference in treatment response between the s was substantial (62% in the depression care for people with cancer vs 17% in the usual care ). Depression care for people with cancer also produced a much greater improvement in other important cancer-related symptoms of anxiety, pain, and fatigue, and in functional ability, overall Vol 384 September 2,

8 Panel: Research in context Systematic review We did a systematic review of randomised controlled trials to establish which, if any, treatments had been shown to be effective for patients with diagnoses of both major depression and cancer. 18 We identified relevant published research articles by systematically searching the following electronic databases: Medline ( ), Embase Classic and Embase ( ), PsycINFO ( ), and the Cochrane Central Register of Controlled Trials (CENTRAL). Searches were run for the combination of cancer and depression using both standardised subject terms and free text terms, including synonyms and alternative spellings, and we used the Cochrane Collaboration recommended terms for randomised controlled trials. We found only a small amount of evidence from small trials in patients with a diagnosis of major depression and suggestive evidence from trials of patients with a range of depression severity. Taken together, the available data suggested that antidepressant drugs combined with psychological treatments might be effective, but did not provide definitive evidence for a particular approach. Interpretation SMaRT Oncology-2 compared a multicomponent collaborative care-based treatment programme that was integrated with cancer care (depression care for people with cancer) with usual care. It provides new and robust evidence for the effectiveness of the depression care for people with cancer programme in the treatment of comorbid major depression in patients with cancer and good prognosis. Depression care for people with cancer also improved anxiety, pain and fatigue, functioning, quality of life, and experience of care. The additional cost of delivering this care programme was modest. The trial findings emphasise the potential value of integrated, intensive, and systematically delivered approaches to the management of major depression comorbid with chronic medical conditions. quality of life, and perceived quality of care than did usual care. Furthermore, these relative benefits of depression care for people with cancer all persisted to the final 48-week follow-up. Apart from our own efficacy trial, we are not aware of any other previous studies of a systematic integrated approach to the treatment of comorbid major depression in patients attending cancer services (panel). 18 The most relevant previous studies are two trials of less intensive treatments (for patients with a range of depression severity rather than those with a diagnosis of major depression) and a sub analysis of a primary care-based trial of treatment for depression in elderly people The recorded treatment effect of depression care for people with cancer in SMaRT Oncology-2 (standardised mean difference 1 1) is substantially greater than that reported by any of these previous trials. It is also much greater than the average effect of similar (collaborative care-type) treatments for depression in general (standardised mean difference.34) and in depression comorbid with chronic diseases (standardised mean difference 43). 38,39 The treatment effect is also larger than the one we recorded in our efficacy trial. 22 This larger effect represents the further development of depression care for people with cancer and the even poorer outcome with usual care for the more representative sample of patients recruited into SMaRT Oncology-2. This striking effectiveness of depression care for people with cancer is probably attributable to the fact that it was intensive (all patients were offered both actively managed antidepressant drug treatment and psychological therapy, and most treatment sessions were face-to-face), systematically implemented (nurses were trained for 3 months until they had proven their competence and were regularly supervised during treatment by video recordings of sessions and patient outcomes were monitored regularly), and integrated with the patients cancer and primary care to promote acceptability and adherence. We were surprised by how poor the outcome with usual care was, especially since usual care was enhanced by informing patients, oncologists, and primary care physicians of the patients diagnosis of major depression. Although the frequency of prescription of antidepressant medication in the usual care was higher than that reported in a previous observational study, 5 notably this prescribing was not managed actively (by changing drugs, adjusting dose, or combining drugs, depending on the patient s response). Few patients in the usual care were seen by mental health services and very few received a formal psychological treatment, despite these options being available to patients at all trial centres. Other trials have recorded similarly poor outcomes with usual care for patients with major depression comorbid with medical conditions. 4 Therefore, the frequent clinical practice of managing depression by merely communicating the diagnosis to the patient s primary care physician (who has the option of referral to mental health services) cannot be relied on to produce good outcomes. The strengths of our trial include recruitment by systematic screening of cancer clinics that served a geographically defined population of about 4 million people, addressing depression at the crucial time after completion of initial cancer treatment, 16 good participation in the depression care for people with cancer programme, and negligible missing outcome data. However, our trial also had several limitations. First, the sample was mainly women receiving follow-up or adjuvant treatment for breast and gynaecological cancers, which potentially limits the generalisability of our findings. However, in our planned sub analyses, we recorded no association between treatment effect and patients sex, primary cancer, or cancer status. Furthermore, a trial of a similar treatment programme in patients with a poor prognosis cancer (lung cancer) showed a similar benefit. 41 Second, the trial was done in one particular health-care system, although the organisation of cancer care is similar to that in most developed countries with both primary and specialist health-care systems. Third, we could not mask participants and their primary care physicians to treatment allocation, although outcome data collectors Vol 384 September 2, 214

9 and stat isticians were masked. Fourth, as with other studies of multicomponent treatments, we are not able to establish whether one component of the depression care for people with cancer programme was more important than others. Finally, we only followed patients to 1 year and consequently do not know their longer-term outcome. The findings of this trial add to the accumulating evidence for the effectiveness of collaborative care approaches to the treatment of depression comorbid with medical conditions. 42 They also provide new evidence that large and sustained treatment effects can be achieved if depression treatment is integrated with medical care, intensive, and systematically delivered by a well-trained and supervised team. Furthermore, the additional cost of delivering depression care for people with cancer was quite modest, especially in the context of cancer treatment. The most important implication of this study is that paying the same systematic attention to the management of comorbid major depression as we presently do to the associated medical condition achieves much better outcomes for patients. Contributors MS, JW, and GM conceived and designed the study. JW led the study conduct and CHH led the study analysis under the supervision of MS and GM. PM was responsible for the management of all study data. SS, CG, and LW advised about study design, conduct, and data analysis and interpretation. MS and JW wrote the draft with critical revision from all other authors. All authors have seen and approved the final version. Declaration of interests We declare no competing interests. Acknowledgments The study was jointly sponsored by University of Edinburgh and NHS Lothian. The study was funded by Cancer Research UK as part of a programme grant (grant number C5547/A7375) with additional funding from the Chief Scientist Office of the Scottish Government. Recruitment and follow-up was also supported by the Scottish Mental Health Research Network, funded by NHS Research Scotland. JW is supported by Sir Michael Sobell House Hospice, Oxford, and the NIHR Collaboration for Leadership in Applied Health Research and Care Oxford at Oxford Health NHS Foundation Trust (the views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health). We thank the trial participants, their families and carers, and staff of the cancer clinics, wards, hospices and general practices of NHS Lothian, NHS Greater Glasgow and Clyde, NHS Forth Valley, NHS Fife, NHS Lanarkshire, and NHS Tayside. References 1 Vos T, Flaxman AD, Naghavi M, et al. Years lived with disability (YLDs) for 116 sequelae of 289 diseases and injuries : a systematic analysis for the Global Burden of Disease Study 21. Lancet 212; 38: Moussavi S, Chatterji S, Verdes E, Tandon A, P atel V, Ustun B. Depression, chronic diseases, and decrements in health: results from the World Health Surveys. Lancet 27; 37: Egede LE. Major depression in individuals with chronic medical disorders: prevalence, correlates and association with health resource utilization, lost productivity and functional disability. Gen Hosp Psychiatry 27; 29: Young AS, Klap R, Sherbourne CD, Wells KB. The quality of care for depressive and anxiety disorders in the United States. Arch Gen Psychiatry 21; 58: Sharpe M, Strong V, Allen K, et al. Major depression in outpatients attending a regional cancer centre: screening and unmet treatment needs. Br J Cancer 24; 9: Walker J, Holm Hansen C, Martin P, et al. Prevalence, associations, and adequacy of treatment of major depression in patients with cancer: a cross-sectional analysis of routinely collected clinical data. Lancet Psychiatry 214; published online Aug org/1.116/s (14)7313-x. 7 Department of Health England. Living with and beyond cancer: taking action to improve outcomes. London: Department of Health, Walker J, Holm Hansen C, Martin P, et al. Prevalence of depression in adults with cancer: a systematic review. Ann Oncol 213; 24: Brown LF, Kroenke K, Theobald DE, Wu J, Tu W. The association of depression and anxiety with health-related quality of life in cancer patients with depression and/or pain. Psychooncology 21; 19: O Connor M, Weir J, Butcher I, et al. Pain in patients attending a specialist cancer service: Prevalence and association with emotional distress. J Pain Symptom Manage 212; 43: Storey DJ, Waters RA, Hibberd CJ, et al. Clinically relevant fatigue in cancer outpatients: the Edinburgh Cancer Centre symptom study. Ann Oncol 27; 18: Walker J, Waters RA, Murray G, et al. Better off dead: suicidal thoughts in cancer patients. J Clin Oncol 28; 26: Colleoni M, Mandala M, Peruzzotti G, Robertson C, Bredart A, Goldhirsch A. Depression and degree of acceptance of adjuvant cytotoxic drugs. Lancet 2; 356: NICE. Depression in adults with a chronic physical health problem: treatment and management. London: National Institute for Health and Clinical Excellence, Institute of Medicine. From cancer patient to cancer survivor: lost in transition. Washington, DC: National Academies Press, Institute of Medicine. Cancer care for the whole patient: meeting psychosocial health needs. Washington, DC: National Academies Press, Li M, Fitzgerald P, Rodin G. Evidence-based treatment of depression in patients with cancer. J Clin Oncol 212; 3: Walker J, Sawhney A, Hansen CH, et al. Treatment of depression in adults with cancer: a systematic review of randomized controlled trials. Psychol Med 214; 44: Fann JR, Ell K, Sharpe M. Integrating psychosocial care into cancer services. J Clin Oncol 212; 3: Walker J, Sharpe M. Depression care for people with cancer: a collaborative care intervention. Gen Hosp Psychiatry 29; 31: Katon W. Collaborative depression care models: from development to dissemination. Am J Prev Med 212; 42: Strong V, Waters R, Hibberd C, et al. Management of depression for people with cancer (SMaRT oncology 1): a randomised trial. Lancet 28; 372: Walker J, Cassidy J, Sharpe M. The second Symptom Management Research Trial in Oncology (SMaRT Oncology-2): a randomised trial to determine the effectiveness and cost-effectiveness of adding a complex intervention for major depressive disorder to usual care for cancer patients. Trials 29; 1: American Psychiatric Association. Diagnostic and statistical manual of mental disorders, 4th edn. Washington, DC: American Psychiatric Association, Kathol RG, Mutgi A, Williams J, Clamon G, Noyes R Jr. Diagnosis of major depression in cancer patients according to four sets of criteria. Am J Psychiatry 199; 147: First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSM-IV axis I disorders. New York: Biometrics Research, New York State Psychiatric Institute, Mynors-Wallis. Problem-solving treatment for anxiety and depression: a practical guide. Oxford and New York: Oxford University Press, Hopko DR, Lejuez CW, Ruggiero KJ, Eifert GH. Contemporary behavioral activation treatments for depression: procedures, principles, and progress. Clin Psychol Rev 23; 23: Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 21; 16: Walker J, Sharpe M, Kroenke K, Murray G. The HSCL-2: one questionnaire, two versions. J Psychosom Res 21; 68: Derogatis LR. The SCL-9 manual 1. Scoring, administration and procedures for the SCL-9. Baltimore: Clinical Psychometric Research, Vol 384 September 2,

10 32 O Connor M, Butcher I, Hansen CH, et al. Measuring improvement in depression in cancer patients: a 5% drop on the self-rated SCL-2 compared with a diagnostic interview. Gen Hosp Psychiatry 21; 32: Aaronson N, Cull A, Kaasa S, Sprangers M. The European Organization for Research and Treatment of Cancer (EORTC) modular approach to quality of life assessment in oncology. Int J Mental Health 1994; 23: Taylor D, Paton C, Kapur N. Maudsley hospital prescribing guidelines, 1th edn. London: Informa Healthcare, Kroenke K, Theobald D, Wu J, et al. Effect of telecare management on pain and depression in patients with cancer: a randomized trial. JAMA 21; 34: Ell K, Xie B, Quon B, Quinn DI, Dwight-Johnson M, Lee PJ. Randomized controlled trial of collaborative care management of depression among low-income patients with cancer. J Clin Oncol 28; 26: Fann JR, Fan MY, Unutzer J. Improving primary care for older adults with cancer and depression. J Gen Intern Med 29; 24 (suppl 2): S Archer J, Bower P, Gilbody S, et al. Collaborative care for depression and anxiety problems. Cochrane Database Syst Rev 212; 1: CD Ekers D, Murphy R, Archer J, Ebenezer C, Kemp D, Gilbody S. Nurse-delivered collaboratice care for depression and long-term physical conditions: a systematic review and meta-analysis. J Aff ect Disord 213; 149: Unutzer J, Katon W, Callahan CM, et al. Collaborative care management of late-life depression in the primary care setting: a randomized controlled trial. JAMA 22; 288: Walker J, Holm Hansen C, Martin P, et al. Integrated collaborative care for major depression comorbid with a poor prognosis cancer (SMaRT Oncology-3): a multicentre randomised controlled trial in patients with lung cancer. Lancet Oncol 214; published online Aug Katon WJ, Lin EH, Von Korff M, et al. Collaborative care for patients with depression and chronic illnesses. N Engl J Med 21; 363: Vol 384 September 2, 214