Characteristics of the screened sample
We made a detailed case-note assessment, of 792 (14%) of the 5613 attenders screened. The majority (558 out of 792; 71%) were female with a mean age of 60 years. Breast cancer was the largest group (377 out of 792; 48%), with the remaining being approximately equally divided between prostate, bladder, testes, and colorectal and gynaecological cancers.
Number of cases of MDD identified
is a flow chart showing the number of patients on whom data were recorded at each stage of the screening process. It includes both those screened in the general oncology clinics by a touch-screen computer, and those screened in the breast unit clinics by paper questionnaire. There was a higher completion rate of the HADS screening questionnaire in the oncology clinic (76% compared with 64%), yielding an overall completion rate of 70% (3938 out of 5613).
Number of patients included at each stage of the screening process.
The data from these two outpatient departments were combined. In the combined sample, 891 out of 3938 or 23% (95% confidence intervals 21–24%) of completers scored above the cutoff of 15 or more on the HADS. Although the majority of these high scorers (570 out of 891; 64%) were assessed further by telephone interview, a substantial number of patients (321) were lost to further assessment at this stage. A number of patients (n=79) were ineligible for further assessment, almost all because of poor health or poor prognosis. For a further 224 patients, interviews were not achievable. The main reasons were failure to obtain permission from the oncologist to contact the patient, administrative reasons in 87 and refusal of permission in 27. Further losses were due to failure to achieve contact with the patient (n=96) and patient discharge from the clinic (n=14).
A comparison of the sample of patients on whom telephone interview was achieved with those on whom it was not (see
) showed small but statistically significant differences in age and gender. The interviewed sample was slightly older and more likely to be that of a female. The mean total HADs score of the two groups was similar.
Differences between those patients interviewed and those who were not interviewed
Major depressive disorder was diagnosed in 196 out of 570 (34%; 95% confidence interval 31–38%) of the high-scoring patients on whom SCID interviews were achieved. This represents 196 out of 891 (22%) of all the high scorers on the HADS if those not interviewed are included. In order to compensate for cases missed because of failure to obtain an interview and given that the mean HADs score of those interviewed was comparable to those not interviewed, we have assumed that a similar percentage of the high scorers not interviewed would have MDD. This is an additional 109 cases (34% of 321). We can then estimate the total number of cases of MDD in all high scorers, had we been able to interview them, as 305. This gives an overall estimated prevalence of MDD in those patients who completed the HADS screen of 305 out of 3938 (7.8%; 95% confidence intervals 6.9–8.6%). However, this estimated prevalence in the population of oncology attenders does not take into account those patients who refused to complete the HADS at initial contact (1675 out of 5613; 30%). Therefore, the ‘true’ prevalence rate may be higher than found in this sample, as there was an impression in the screening clinics that depressed individuals were more likely to refuse invitations to participate in the screening procedure.
Description of patients with MDD
Of the 196 patients identified as having MDD, seven refused to provide further information. Of the remainder, 37 were considered inappropriate for treatment by a cancer centre depression service: 20 patients had chronic MDD (depression with onset at least 1 year prior to their first cancer diagnosis). Most of these patients had received or were receiving specialist psychiatric treatment. A further 19 patients were excluded for other reasons listed in . The final sample of cases of MDD considered suitable for treatment in a cancer centre service and on whom assessments were competed was therefore 150. The characteristics of these 150 patients are shown in
Demographic and medical characteristics of 150 patients with comorbid MDD.
The majority were females with breast cancer. Most had inactive disease, with only a minority currently undergoing cytotoxic treatment (other than taking anticancer medications such as tamoxifen) at the time of screening. The median interval since cancer diagnosis was 3 years, with a mode of 1 year. For a small number of patients, 10 or more years had elapsed since their original cancer diagnosis.
Nature of MDD
The median duration of the patient's current depressive episode was 6 months. For half of the sample, this was their first episode of depression (see
Depression characteristics of the study sample of 150 patients with comorbid MDD
The treatment the patients with MDD reported having received for their current episode of depression is shown in . Overall, only 23 out of 150 (15%) of the patients were judged to have received or were currently receiving potentially effective evidence-based treatment for MDD. Likely reasons for the failure to deliver effective treatment were identified at every stage of their management and are shown in .
Flow chart documenting self-report of management of depression.
First, only half (74 out of 150; 49%) of the patients said that they had even spoken with their GP or oncologist about feeling depressed. Second, less than a third (43 out of 150; 29%) reported being offered antidepressant drugs, and less than half of these were taking an adequate dose (as defined in methods) at the time of screening. Finally, only a small minority of patients (11 out of 150; 7%) had been referred to specialist mental health services (psychology or psychiatry) and only (7 out of 150; 5%) had received any formal psychological treatment for depression. These findings indicate a potentially substantial unmet treatment need.