Friday, July 11, 2008
Lancet this week presents a randomized controlled trial of a nurse-delivered educational intervention for depression in cancer patients. This was a single-center Scottish study involving 200 outpatients (mean age 56 years; expected prognosis had to be greater than 6 months for inclusion; a plurality of subjects had breast cancer) who were randomized to usual care or usual care plus the intervention. Outpatients at this center were screened with the Hospital Anxiety and Depression Scale & if this was positive were diagnosed with depression using DSM-IV criteria (using the DSM structured clinical interview materials). The intervention was delivered by 3 study nurses, none of whom had training in psychology, and involved up to 10 one-on-one sessions over 3 months (the mean was 7 sessions). The intervention is described here:
The content of the intervention, Depression Care for People with Cancer, comprisedThe primary outcome was a depression score at 3 months (they used the SCL-20 depression scale; grossly it's a 0-4 composite score from a 20-item scale with 4 being the worst and 0 being zero depressive symptoms). Depression scores decreased for both groups at 3 months (from a mean of 2.3 to 1.2 in the treatment group and 1.55 in the control group - this was statistically signficant). Many econdary outcomes were improved in the treatment group as well (68% of intervention group patients no longer met DSMIV criteria for major depressive disorder vs. 45% of control patients at 3 months, etc.).
education about depression and its treatment (including antidepressant
medication); problem-solving treatment to teach the patients coping strategies designed to overcome feelings of helplessness; and communication about management of major depressive disorder with each patient's oncologist and primary-care doctor.
This was an educational intervention, but they also followed who actually were started on or titrated to 'therapeutic doses' of antidepressant drug therapy during the trial (about 20% of patients were on them at baseline): 69% in the treatment group vs. 42% in the control group. (There's a web-only table defining what they meant by therapeutic doses - it's pretty reasonable although it's unclear how they actually derived these doses: 20mg citalopram, 75 mg venlafaxine, etc. - nothing surprising.)
One might note that these rates of 'therapeutic' drug doses are identical to the remission rate as defined by the DSM structured clinical interview criteria and wonder if the effect of this intervention was really pharmacologic (via encouraging the patients to talk with their docs about depression and be more open to pharmacologic therapy). To be clear, that was, in fact the point of the educational intervention and it appears it worked, albeit only modestly better than usual care.
There are three major problems with interpreting these sorts of trials. First, it's always impossible to get a good sense of what the actual intervention was (what it entailed, could this be something that's feasible at my center, etc.). Second, for these complex non-pharmacologic interventions one always wonders if there was some simple aspect of the intervention which actually lead to its salutary effect (i.e. something which would take 5 minutes once not 45 minutes 7 times) - there's no good way of knowing without more trials etc. etc. That is - despite its welcome findings as a clinician there's an off-putting-ness (forgive me for that hyphenated monstrosity) about it - 'sounds nice but I'm not convinced this whole thing is worth it and it seems too complex to really consider something like this). Third, as this is a constant issue with palliative-care-relevant (it strikes again) research, it necessarily involves scales and indices and numbers and as much as I'd like to know what the patient-relevant difference between 1.2 and 1.55 on the SCL-20 at 3 months is I don't. There is no way for the average clinician to appreciate what that means in our patients' lives (like we could, for instance, by knowing that the number needed to treat to gain a 50% reduction in pain at one month is 3 for drug X - and even some would argue we don't know what that means...).
Anyway, despite the above, it's great to see such well-designed research into treating depression in the setting of cancer, and I hope there's more to come. My own gloss on this is that it likely worked because it improved patients' 'access' to antidepressants; drugs which many patients, even those suffering greatly with depression, are reluctant to try.
STRONG, V., WATERS, R., HIBBERD, C., MURRAY, G., WALL, L., WALKER, J., MCHUGH, G., WALKER, A., SHARPE, M. (2008). Management of depression for people with cancer (SMaRT oncology 1): a randomised trial. The Lancet, 372(9632), 40-48. DOI: 10.1016/S0140-6736(08)60991-5