Original Article. 318 Journal of Pain and Symptom Management Vol. 24 No. 3 September 2002

Transcription

1 318 Journal of Pain and Symptom Management Vol. 24 No. 3 September 2002 Original Article Use of a Depression Screening Tool and a Fluoxetine-Based Algorithm to Improve the Recognition and Treatment of Depression in Cancer Patients: A Demonstration Project Steven D. Passik, PhD, Kenneth L. Kirsh, PhD, Dale Theobald, PhD, MD, Kathleen Donaghy, PhD, Elizabeth Holtsclaw, BA, Sarah Edgerton, MA, and William Dugan, MD Symptom Management and Palliative Care Program (S.D.P., K.L.K.), Markey Cancer Center, University of Kentucky, Lexington, Kentucky; and Community Cancer Care (D.T., K.D., E.H., S.E., W.D.) Indianapolis, Indiana, USA Abstract Helping oncologists to identify and treat depression is an important step in improving the overall care of people with cancer. In previous work performed in our community-based, ambulatory oncology outreach network, we validated a depression screening tool, put into place depression screening programs, and taught oncologists how to follow up on screening with brief, reliable clinical interviews. Subsequently, we provided these oncologists with a fluoxetine-based antidepressant algorithm to follow for the treatment of their depressed patients. In this article, we report on the initial experience identifying and treating 35 ambulatory oncology patients who were screened with the Zung Self-rating Depression Scale (ZSDS). Structured follow-up interviews by their oncologist determined whether the patients qualified for a diagnosis of a major depressive episode. These patients then received 1 of 4 treatments based on the algorithm (no treatment, fluoxetine alone, fluoxetine plus bedtime doxepin, or fluoxetine plus methylphenidate). Patients were matched by their oncologist to a prototype patient for each treatment arm based on their symptomatic presentation (i.e., patients requiring a side effect minimization approach were to be placed on fluoxetine alone; patients who had significant insomnia, weight loss, or neuropathic pain were placed on the fluoxetine plus doxepin regimen; those with prominent fatigue were to receive fluoxetine plus methylphenidate). Patients were followed weekly for one month, and then every two weeks for two more months, with telephone assessments of their depression, associated symptoms and overall quality of life. Results suggested that oncologists most often chose the simplest regimen (fluoxetine alone) but that patients uniformly benefited in terms of improved mood and overall quality of life throughout the 12 weeks of follow-up. Our initial experience suggests that oncologists can be empowered to recognize and treat depression in their patients with a screen- Address reprint requests to: Steven D. Passik, PhD, Symptom Management and Palliative Care Program, University of Kentucky College of Medicine, 800 Rose Street, Lexington, KY , USA. Accepted for publication: May 17, U.S. Cancer Pain Relief Committee, /02/$ see front matter Published by Elsevier, New York, New York PII S (02)

2 Vol. 24 No. 3 September 2002 Treatment of Depression in Cancer Patients 319 ment of depression and other distressing physical and psychological symptoms in our rural, community-based, ambulatory oncology outreach network. 9 Our group began our work in this area by validating the Zung Self-Rating Depression Scale (ZSDS) as a rapid screening tool for the identification of depressed cancer patients. We subsequently taught oncologists how to reliably conduct brief, follow-up clinical interviews to establish a depression diagnosis. 13 We also recognized that many busy oncology clinics did not have the resources to interview large numbers of patients who might screen high on the ZSDS. Therefore, we subsequently described the criterion validity of the tool in order to establish its use as a laboratory value for depression, which might allow treatment to be initiated with a reasonable degree of certainty of the diagnosis, without extensive follow-up. 14 Finally, we are also in the process of describing how single items on the ZSDS for physical and cognitive symptoms could be reliably used to screen for these symptoms in the presence or absence of depression. 15,16 The next logical step in our ongoing effort to improve the identification and treatment of depression in our network was to address the need to remove barriers to intervention. In our initial screening work, we identified more than one-third of our patients as suffering with potentially clinically significant levels of depressive symptoms, but only 7% of them were receiving interventions of any kind. 11 Interventions for depression in cancer patients include psychosocial and behavioral interventions (such as group or individual counseling and psychoeducational programs) and pharmacotherapy. 17 The literature on the utility of these interventions, alone or in combination, is sparse considering their widespread use. Published works are generally very supportive of psychosocial interventions However, given the realities of community-based oncology care in ruand-intervene approach. Such an approach may benefit patients, and, if kept simple, can be incorporated into day-to-day care of people with cancer. J Pain Symptom Manage 2002; 24: U.S. Cancer Pain Relief Committee, Key Words Cancer, depression, fluoxetine Introduction Cancer patients have a high rate of psychiatric comorbidity. Approximately half of all patients diagnosed with cancer manifest levels of distress that qualify for formal psychiatric diagnoses. 1 The majority of these disorders (fully 60% of the diagnoses) are adjustment disorders with mixed depressed and anxious features. The next most common diagnosis is major depression. The prevalence of major depression in people with cancer is estimated between 3 50% depending on site, stage, and assessment methods, 2 5 the estimated prevalence overall is approximately 20%. 6 Depressive disorders are associated with diminished adaptation to life with the disease and diminished overall quality of life. Thus, the identification and treatment of depression in ambulatory oncology settings is an important step in improving the overall care of people with cancer. The identification of depression in cancer patients is a significant clinical and methodological challenge in community-based oncology. Time constraints, stoicism on the part of patients, and lack of familiarity with proper depression assessment on the part of staff combine to lead to a high rate of under-recognition, and, thereby, undertreatment, of depression in cancer patients. In the absence of screening, both medical oncologists and oncology nurses have been found to dramatically underestimate the severity of depressive symptoms in their patients with cancer. Their assessments tend to be overly focused on manifest depressed mood, rather than the more reliable and stable cognitive/ideational symptoms of depression. 7,8 In such circumstances, staff can thus be expected to miss the most depressed patients, especially if they do not display outward signs of depressed mood at the time of the assessment. Over the past five years, we have taken several steps to improve the recognition and treat-

3 320 Passik et al. Vol. 24 No. 3 September 2002 ral Indiana, characterized as it is by limited psychiatric resources, psychosocial interventions requiring specialized staff had limited applicability to our setting. Although the literature contains very few well-controlled studies of antidepressant medication, the few studies that have been published (along with extensive anecdotal and clinical experience) suggest that they may have utility in the treatment of depression in cancer patients. 21,22 Perhaps the most rigorous study in the literature is one supporting the use of fluoxetine and desipramine in depressed patients with cancer. 21 Thus, we decided to create a fluoxetine-based algorithm to help our oncologists choose an antidepressant(s) and to examine the impact of this intervention on depression, associated symptoms and overall quality of life. Although not a typical trial of antidepressants in cancer patients, this approach allowed us to examine the impact of the various treatments and combinations of treatments on a series of patients versus their own baselines. This approach, characterized by systematic follow-up of outcomes, was designed to provide some basic data that could be subsequently used to implement more sophisticated studies of antidepressant efficacy. Some studies have recently been conducted in academic settings 23 and suggest positive outcomes, such as decreased depression and improved quality of life, for patients treated with fluoxetine. We hoped to shed some light on the impact of a program that increases staff awareness of depression, enhances assessment and monitoring of patients, and promotes the use of antidepressants (though the impact of the various elements of the program would remain uncertain at this point in time). The algorithm employed in this study was an attempt to embody a basic aspect of selecting an antidepressant drug for medically ill patients, namely the side effect minimization versus side effect mobilization approach. 24,25 This approach to antidepressant selection is based on the idea that for some patients, the optimal intervention for depression is based on utilizing the cleanest drug possible. Such an approach minimizes side effects, as these patients presentation of depression does not necessitate rapid help with symptoms that might respond to the side effects of a dirtier drug. The minimization approach is basically the same as one might take to treating depression in a physically healthy patient. The utility of the minimizing of antidepressant side effects has been best exemplified by the very successful use of the specific serotonin re-uptake inhibitors (SSRIs) in that population. There are cancer patients for whom such an approach is also desired. For example, a depressed woman with early breast cancer is unlikely to have a need for weight gain, sedation, or help with neuropathic pain, and is likely to do best on an SSRI alone. On the other hand, more symptomatic depressed cancer patients may need a drug or combination of drugs with more side effects that can be mobilized to give them quicker relief of distress. Patients with severe insomnia, cachexia or appetite loss, or neuropathic pain are likely to need such treatments and might be expected to receive such benefit from a tricyclic antidepressant (TCA), or an atypical antidepressant such as mirtazapine. 26 Finally, patients with depression and problematic fatigue might also require the addition of a psychostimulant such as methylphenidate, alone or in combination with their antidepressant, to help them with both mood elevation as well as increased energy. As described below, we implemented a system for screening patients, follow-up interviews by oncology staff and then algorithm-based treatment for depression. The study was opened to several sites and the entire oncology group agreed to participate and was trained in the protocol. In this article, we describe and discuss our experience with the first 35 patients to be identified and treated with this algorithm-based approach. Patients were considered eligible for the study if they had either a solid tumor or hematologic malignancy, had the presence of depressive symptoms, were over 17 years of age, were expected to have a survival rate over three months, and were not currently on antidepressant medication. In addition, patients needed to be able to tolerate oral medications and be competent enough to offer full informed consent to take part in the study. Methods Procedures Data collection occurred at eight time periods (See Figure 1 and Table 1). At baseline (time 1), patients were screened with the

4 Vol. 24 No. 3 September 2002 Treatment of Depression in Cancer Patients 321 ZSDS, and those with a raw score above 48 were deemed eligible for a structured clinical interview. After consent was obtained, patients also completed a Numeric Rating Scale (NRS) assessing pain, fatigue, appetite, and insomnia, along with the FACT-G and a demographics sheet. Results of the screening were placed at the front of the patient s chart along with the structured interview questions and the treatment algorithm. The oncologists then assessed patients who had an elevated score with the Mini-International Neuropsychiatric Interview (MINI) and a diagnosis was made of no depression or major depression to determine final eligibility. Based upon prior training, 13 we concluded that these oncologists had been trained to apply the MINI in a reliable fashion to arrive at a diagnosis of depression. Once the patient was assessed, the oncologist also made the final determination regarding which antidepressant (if any) or combination of antidepressants was appropriate for the patient by matching their patients to prototypes supplied for each treatment approach. Time 2, 3, and 4 testing occurred during consecutive weeks and consisted of the ZSDS and NRS. Time 5 8 testing occurred every two weeks thereafter and again consisted of the ZSDS and NRS. In addition, the FACT-G was administered at time 4, 6, and 8. The oncologists were kept apprised of their patients ZSDS scores over the course of the study and attended a lecture recounting the protocol and performance at the completion of the study. Fig. 1. Decision algorithm for placing subjects on the various study arms. Table 1 Timeline of the Study and Assessments Collected at Each Time Point Collection Week ZSDS NRS FACT-G MINI Demos T1 Baseline X X X X X T2 Week 2 X X T3 Week 3 X X T4 Week 4 X X X T5 Week 6 X X T6 Week 8 X X X T7 Week 10 X X T8 Week 12 X X X Measures Demographics. A standard demographic questionnaire was utilized in the study. The demographic questionnaire obtained information on age, sex, race, marital status, education level, and disease- and treatment-related variables. The Mini-International Neuropsychiatric Interview (MINI). The MINI is a brief structured diagnostic interview designed as an alternative to the Semi-structured Clinical Interview for DSM-IV Disorders (SCID) and the Composite International Diagnostic Interview (CIDI). The MINI contains a core set of diagnostic questions and focuses only those time frames that are useful in making decisions in clinical settings. The clinician-rated version assessing major depressive episodes was used in this study. 27 The MINI was chosen because it provides a brief reliable alternative to longer structured interviews and is one that could easily be mastered and employed by busy oncology staff. Zung Self-Rating Depression Scale (ZSDS). The Zung Self-Rating Depression Scale (ZSDS) 10,28 is a

5 322 Passik et al. Vol. 24 No. 3 September item self-report measure of the symptoms of depression. Subjects rate each item regarding how they felt during the preceding week using a 4-point Likert scale, with 4 representing the most unfavorable response. The sum of the 20 items, after correcting for the 10 items that are reverse-scored, produces a raw score that is converted into a depression score (termed the SDS index). These index scores are then categorized into 4 levels to offer a global clinical impression: 1 within normal range, no significant psychopathology (SDS Index: below 50); 2 presence of minimal to mild depression (SDS Index: 50-59); 3 presence of moderate to marked depression (SDS Index: 60-69); and 4 presence of severe to extreme depression (SDS Index: 70 and above). Scores are not meant to offer strict diagnostic guidelines but rather denote levels of depressive symptomatology that may be of clinical significance. Overall, the ZSDS has been shown to be relatively valid and to have high internal consistency, exhibiting an alpha coefficient of ,29 Functional Assessment of Cancer Therapy - General (FACT-G). The FACT-G (version 2) is a 33-item self-administered questionnaire covering the quality of life domains of physical, social and family, emotional and functional well-being. Items are rated on a 5-point likert scale, from 0 (not at all) to 4 (very much) for how true each statement has been for the patient in the past seven days. After accounting for reverse-scored items, questions are summed across the four subscales and added for a total score, with higher scores indicative of greater overall quality of life. The instrument has been shown to be easy to use, brief, reliable, and valid. 30,31 The measure has high internal consistency, with coefficient alphas ranging from on the subscales and for the total scale. 32,33 The scale has also shown high test-retest reliability (r 0.87) and concurrent validity, as shown by strong correlations with the Functional Living Index Cancer (r 0.80).34 Currently in the fourth version, the measure has been deemed to be appropriate for use with all cancer patients. Numeric Rating Scales (NRS). The NRS consisted of four questions covering pain, fatigue, appetite, and insomnia. These items were added to the study to determine whether fluoexetine or doxepin would be likely to affect these areas. In the case of doxepin, it was expected that pain would decrease and appetite would be stimulated. Patients were asked to circle the number that best described the symptom at its worst during a one-week period. The items utilize a 0 10 scale, with 0 meaning no presence of the symptom and 10 meaning that the symptom was as bad as could be imagined. This information was used to help guide the oncologist with regard to treatment selection. The most intense symptom was targeted by the antidepressant (i.e., to improve it or not worsen it) as a way of matching patients to treatments. For example, patients with severe insomnia were generally slated for the arm that included doxepin. The Fluoxetine-Based Algorithm. The antidepressant algorithm had four specific treatment arms as described below: Option 0: No treatment. The physician could opt to not treat the patient and have them be subjected to follow-up only. This was the option reserved for situations in which the physician thought the patient s depression was situational and would resolve spontaneously, the physician felt indecisive about the diagnosis, or the patient could not tolerate the medications for depression. Option 1: Fluoxetine alone. This regimen embodied the side effect minimization approach. Patients on this arm received daily fluoxetine 20 mg orally in the morning. Option 2: Fluoxetine plus doxepin. This regimen was meant to embody the side-effect mobilization approach. Based upon clinical experience in this population, doxepin was added in the hope of treating insomnia through its sedative effect as well as possible use for pain reduction and increased appetite. Whereas a newer single agent such as mirtazapine or venlafaxine might be used instead, these agents were not available at the time of the study. Patients treated on this arm of the study received fluoxetine 20 mg orally in the morning and doxepin 25 mg (escalated to 50 mg if the patient still exhibited poor sleep and appetite) orally at bedtime, daily. Option 3: Fluoxetine plus methylphenidate. This regimen was employed for the patient with prominent fatigue. Patients treated in this arm received fluoxetine 20 mg orally in the

6 Vol. 24 No. 3 September 2002 Treatment of Depression in Cancer Patients 323 morning, methylphenidate 5 mg in the morning, and methylphenidate 5 mg with lunch, daily. Based upon the results of the MINI, NRS scores, and the prototypical patients listed for their benefit, oncologists decided on which arm to place the patient. It should be noted that the oncologists were at liberty to reduce or discontinue medications if they were deemed to be causing adverse side effects, such as agitation. However, no patients had their medications reduced or discontinued. Statistical Analyses A series of descriptive statistics were performed. In addition, a series of repeated measures ANOVAs were conducted to compare the differences in scale scores across time for individual treatment arms. Results The study was conducted on the first 35 patients to be enrolled on the algorithm. Once identified and enrolled, all patients completed the study. The average age of the sample was years (SD 12.14). There were 24 women (68.6%) and 11 men (31.4%) (See Table 2). The majority were high school graduates (42.9%, n 15), followed by those with some high school experience (22.9%, n 8) or some exposure to college (17.1%, n 6). The sample was primarily Caucasian (88.9%, n 31) and most were either married (51.4%, n 18) or widowed (31.4%, n 11). Tumor type varied but the majority had breast (42.9%, n 15), lung (11.4%, n 4), or prostate (11.4%, n 4) cancer. The majority of patients were placed on the fluoxetine alone arm (option 1) (n 16, 45.7%) followed by fluoxetine and doxepin (option 2) (n 8, 22.9%), fluoxetine and methylphenidate (option 3) (n 6, 17.1%), and no treatment (option 0) (n 5, 14.3%). Figures 2 and 3 show the breakdown of scores on the ZSDS and FACT-G across the study. As can be seen, mean scores on the ZSDS exhibit a reduction in depressive symptoms from baseline through week 12 for subjects on all treatment arms. In addition, mean quality-of-life scores on the FACT-G showed improvement from baseline to week 12 for subjects from all treatment arms. Although the overall sample size was too small for definitive analysis, an exploratory series of repeated measures ANOVAs were performed utilizing simple contrasts to compare FACT-G and ZSDS scores at study end to baseline for each of the treatment arms. Compared to baseline, FACT- G scores within each arm were significantly improved at study end for option 0 (F 54.40, P 0.005), option 1 (F 15.42, P 0.002), and option 2 (F 36.70, P 0.002). ZSDS scores at study end were only significantly improved for option 1 (F 73.94, P 0.001) and option 2 (F 55.94, P 0.005). Figures 4, 5, 6, and 7 show the mean responses of the patient s on the NRS items, both globally and across treatment arms. Mean scores for the pain and insomnia items improved from baseline to week 12 for options 0, 1, and 2, with only a slight increase for subjects on option 3. Similarly, patients across all options showed improvement in fatigue levels from baseline to week 12. Finally, responses on the insomnia item show that subjects on options 1, 2, and 3 all improved, while option 0 subjects had a slight increase in difficulty with insomnia. A series of repeated measures ANOVAs were performed utilizing simple contrasts to compare the four NRS item scores at study end to baseline for each of the treatment arms. Compared to baseline, there were no significant improvements in the pain NRS item for any of the treatment arms. For the fatigue NRS item, only option 1 was significantly improved at study end compared to baseline (F 6.24, P 0.05). Similarly, option 1 was the only treatment arm to show improvement compared to baseline for both the insomnia NRS item (F 5.91, P 0.05) and the decreased appetite NRS item (F 5.97, P 0.05). These findings are likely due to the larger n in that treatment arm. Discussion In this article, we have described our initial experience with a program of screening for depression, follow-up with oncologist-administered clinical interviews, and interventions guided by a fluoxetine-based antidepressant algorithm. Our initial experiences suggest that such a screen and intervene approach can be valuable for teaching oncologists to recognize depressive symptoms in their patients and to improve the treatment of these problems. In addition, the

7 324 Passik et al. Vol. 24 No. 3 September 2002 Demographics Table 2 Breakdown of Patients Demographics Both Globally and Across Treatment Arms Global n 35 Option 0 n 5 (14.3%) Option 1 n 16 (45.7%) Option 2 n 8 (22.9%) Option 3 n 6 (17.1%) Mean Age (SD) (12.14) (13.26) (14.05) (8.54) (10.09) Sex Male 11 (31.4%) 1 (20.0%) 6 (37.5%) 3 (37.5%) 1 (16.7%) Female 24 (68.6%) 4 (80.0%) 10 (62.5%) 5 (62.5%) 5 (83.3%) Education Level 1 8th grade 2 (5.9%) 1 (20.0%) 5 (31.3%) 1 (14.3%) 1 (16.7%) 8 11th grade 8 (23.5%) 1 (20.0%) NA 1 (14.3%) NA High school grad 15 (44.1%) 1 (20.0%) 9 (56.3%) 2 (28.6%) 3 (50.0%) Some college 6 (17.6%) 1 (20.0%) 2 (12.5%) 1 (14.3%) 2 (33.3%) College degree 2 (5.9%) 1 (20.0%) NA 1 (14.3%) NA Some post-college 1 (2.9%) NA NA 1 (14.3%) NA Marital status Single 2 (5.7%) NA 2 (12.5%) NA NA Married 18 (15.4%) 3 (60.0%) 7 (43.8%) 4 (50.0%) 4 (66.7%) Divorced 4 (11.4%) NA 2 (12.5%) NA 2 (33.3%) Widowed 11 (31.4%) 2 (40.0%) 5 (31.3%) 4 (50.0%) NA Ethnicity African-American 3 (8.8%) NA 1 (6.7%) 1 (12.5%) 1 (16.7%) Caucasian 31 (91.2%) 5 (100.0%) 14 (93.3%) 7 (87.5%) 5 (83.3%) results also offer very tentative evidence that treating patients (in the form of antidepressant medicine accompanied by frequent follow-up phone interviews by staff) led to improvements over baseline in most areas. We are also somewhat encouraged that our efforts to raise awareness around depression is becoming manifest in more patients receiving treatment than was the case when we began our work in At that time, only 7% of all CCC patients were receiving treatment of any kind for depression, while a screening survey at CCC 11 suggested that as many as one-third of the patients treated at the facility were potentially depressed. With the screen and intervene approach in place, 86% (30/35) of the identified patients were treated and there were clinical reasons not to treat the remaining 14% (5/35) with antidepressant therapy. Giving the oncologists a concrete set of responses to a problem like depression is probably a reasonable way to see to it that more patients will be treated. Before discussing the results further, it is important to note that this project is not a clinical trial of the efficacy of antidepressants, but rather, a demonstration project to highlight a particular approach to improving the recognition and treatment of depression in cancer patients. Our hope would be that similar, if not simplified, versions of this approach might be thought to have utility in other settings and that other oncology programs might opt to approach patients distress in ever more systematic ways. The utility of diagnostic algorithms might be developed for other diagnoses that are hard to Fig. 2. Comparison of Zung scores both globally and by treatment arm chosen. Fig. 3. Comparison of FACT-G scores both globally and by treatment arm chosen.

8 Vol. 24 No. 3 September 2002 Treatment of Depression in Cancer Patients 325 Fig. 4. Mean scores for option 0 patients on the NRS items for pain, fatigue, insomnia, and appetite. identify, such as delirium, which is most common in patients with far advanced cancer. Based on the setting (i.e., inpatient versus outpatient), algorithms can be chosen that try to identify the prevailing clinical problems seen most often and may possibly be used for purposes of differentiating between diagnoses. There are several limitations inherent in the methodology employed in the present study with regard to deriving efficacy data on antidepressants. First, because of our attempt to teach our oncologists, the clinical principles of individualizing antidepressant selection, patients were not randomly assigned to treatment groups and therefore the groups were not equivalent across the various treatment arms. Second, all medications were given open-label, again to capture the clinical reality of antidepressant therapy in the real world setting. Thus, patients knew they were on specific drugs with specific potential effects. Thus, no attempt was made to exclude placebo responders. Therefore, the results to be discussed below should be viewed as the outcome of the effects of the program and not of particular antidepressants. Fig. 6. Mean scores for option 2 patients on the NRS items for pain, fatigue, insomnia, and appetite. Overall, all patients benefited, evident from week one, with regard to both improvement in depression scores and overall quality of life. This was evident in decreasing ZSDS scores and improved FACT-G scores. The effect was also present in the no treatment group, suggesting that some depressions in cancer patients are truly transient and can be expected to resolve spontaneously, whereas others are probably responding to the increased dialogue about depression, the follow-up phone calls from the research team, and the attention that this conveys. Another explanation is that patients may have anticipated the hypothesis and responded in a biased way to help the investigators. Upon closer inspection, the patients in option 1 exhibited the greatest amount of overall within-arm improvement, as evidenced by significantly improved scores for quality of life, depression, fatigue, insomnia, and appetite. Also, option 2 patients showed significant within-arm improvement for quality of life and depression. However, caution needs to be used in interpreting these findings due to the small Fig. 5. Mean scores for option 1 patients on the NRS items for pain, fatigue, insomnia, and appetite. Fig. 7. Mean scores for option 3 patients on the NRS items for pain, fatigue, insomnia, and appetite.

9 326 Passik et al. Vol. 24 No. 3 September 2002 numbers of patients on treatment arms 0 and 3. Thus, it is hard to determine if the depression scores and NRS items would have been significantly improved given a larger sample within those arms. It is interesting to note that we provided the oncologists with an algorithm booklet on the front of the patient s chart. These booklets provided the patient s depression screening results, the follow-up interview questions, and a description of each of the study arms (with color-coded instructions corresponding to each arm for patients). Along with each arm was a prototype patient to whom the oncologists were supposed to match their patient. The oncologists utilized the simplest regimen 43% of the time, eschewing the 2 drug regimens for the most part. If we were to redesign the algorithm, we would likely substitute mirtazapine as the exemplar of the side effect mobilization approach. In an open label trial of this agent recently completed in our network, we found suggestions of benefit for depression, insomnia, appetite loss, and nausea for patients in a six-week trial, with simple bedtime dosing of 30 mg and 15 mg. 26 A single agent approach to side effect mobilization might be preferable to our oncology colleagues. An alternative explanation for the oncologists choices of study arms may be that indeed more patients present to oncologists requiring a side effect minimization rather than mobilization approach. This may be the case in the treatment of ambulatory oncology patients. In any case, such approaches should be kept as elegant and simple as possible so that physicians will be encouraged to utilize them. References 1. Derogatis L, Morrow G, Fetting J. The prevalence of psychiatric disorders among cancer patients. JAMA 1983:249; Fulton C. The prevalence and detection of psychiatric morbidity in patients with metastatic breast cancer. Euro J Cancer Care 1998;7: Hopwood P, Stephens RJ. Depression in patients with lung cancer: prevalence and risk factors derived from quality of life data. J Clin Oncol 2000; 18: Lynch ME. The assessment and prevalence of affective disorders in advanced cancer. J Pall Care 1995;11: Vernon SW, Gritz ER, Peterson SK, et. al. Correlates of psychologic distress in colorectal cancer patients undergoing genetic testing for hereditary colon cancer. Health Psych 1997;16: Bottomley A. Depression in cancer patients: a literature review. Eur J Cancer Care 1998;7: Passik S, Dugan W, McDonald M, et al. Oncologists recognition of depression in their patients with cancer. J Clin Oncol 1998:16; McDonald M, Passik S, Dugan W, et al. Nurses recognition of depression in their patients with cancer. Oncol Nurs Forum 1999:3(26); Passik SD, Theobald DE, Holtsclaw E, et al. The oncology symptom control research (OSCR) program; a psycho-oncology and symptom management program for community-based ambulatory oncology. Symptom Management for the Cancer Patient II Zung W. Depression in the normal aged. Psychosomatics 1967;7: Dugan W, McDonald M, Passik S, et al. Use of the zung self-rating depression scale in cancer patients: feasibility as a screening tool. Psycho-Oncol 1998;7: Passik SD, Lundberg J, Rosenfeld B, et al. Factor analysis of the zung self-rating depression scale in a large ambulatory oncology sample. Psychosomatics 2000;41: Passik SD, Donaghy KB, Theobald D, et al. Oncology staff recognition of depressive symptoms on videotaped interviews of depressed cancer patients: implications for designing a training program. J Pain Symptom Manage 2000;19: Passik SD, Kirsh KL, Donaghy KB, et al. An attempt to employ the zung self-rating depression scale as a lab test to trigger follow-up in ambulatory oncology clinics: criterion validation and detection. J Pain Symptom Manage 2001;21: Kirsh KL, Passik SD, Holtsclaw E, et al. I get tired for no reason: a single item screening for cancerrelated fatigue. J Pain Symptom Manage 2001;22(5): Passik SD, Whitcomb LA, Kirsh KL, Theobald DE. An unsuccessful attempt to develop a single item screening for insomnia in cancer patients. J Pain Symptom Manage (in press.) 17. Breitbart W, Passik SD. Psychiatric contributions to palliative care. In: Doyle E, Hanks G, MacDonald N, eds. The oxford textbook of palliative medicine. New York: Oxford University Press, 1993: Fawzy F. Psychosocial interventions for patients with cancer: what works and what doesn t. Eur J Cancer 1999;35: McGrath P. Findings from an educational support course for patients with leukemia. Cancer Pract 1999;7: Spiegel D, Sephton SE, Terr A, et al. Effects of psychosocial treatment in prolonging cancer sur-

10 Vol. 24 No. 3 September 2002 Treatment of Depression in Cancer Patients 327 vival may be mediated by neuroimmune pathways. Ann NY Acad Sci 1998;840: Holland JC, Romano SJ, Heiligstein JH, et al. A controlled trial of fluoxetine and desipramine in depressed women with advanced cancer. Psycho-Oncol 1998;7: Passik SD, Breitbart WS. Depression in patients with pancreatic carcinoma: diagnostic and treatment issues. Cancer 1996;78: Fisch MJ, Lauher PJ, Passik SD, et al. Fluoxetine versus placebo in advanced cancer outpatients: a placebo controlled, double-masked trial of the Hoosier Oncology Group. Proceedings of the 37 th Annual Meeting of ASCO, May 12 15, San Francisco, CA. 383A. 24. Passik SD, McDonald M, Dugan W, et al. Depression in cancer patients recognition and treatment. Medscape Mental Health 1997;2: Theobald D, Passik SD, Donaghy K, et al. An investigation of the analgesic and antiemetic efficacy of mirtazapine (Remeron) in cancer pain patients. Pan American Congress of Psychosocial and Behavioral Oncology, October 20 23, 1999, New York, NY. 26. Theobald DE, Passik SD, Kirsh K, et al. An investigation of the efficacy of mirtazapine in cancer pain patients. 52nd Institute on Psychiatric Services, American Psychiatric Association, October 25 29, Sheehan D, Lecrubier Y, Sheehan K, et al. Comparison of the mini international neuropsychiatric interview (MINI) with the SCID-P and the CIC: a validity study. Psychopharm Bull 1995;31: Zung W. Factors influencing the self-rating depression scale. Arch Gen Psychiatry 1967;16: Tate D, Forchheimer M, Maynard F, et al. Comparing two measures of depression in spinal cord injury. Rehabilitation Psychology 1993;38: Cella D, Tulsky D, Gray G, et al. The functional assessment of cancer therapy scale: development and validation. J Clinical Oncology 1993;11: Winstead-Fry P, Schultz A. Psychometric analysis of the functional assessment of cancer therapy-general (FACT-G) scale in a rural sample. Cancer 1997; 79: Brady M, Cella D, Mo F, et al: Reliability and validity of the functional assessment of cancer therapy breast quality of life instrument. J Clin Oncol 1997; 15: Cella D, Bonomi A, Lloyd S, et al. Reliability and validity of the functional assessment of cancer therapy-lung (FACT-L) quality of life instrument. Lung Cancer 1995;12: Yellen S, Cella D, Webster K, et al. Measuring fatigue and other anemia-related symptoms with the functional assessment of cancer therapy (FACT-G) measurement system. J Pain Symptom Manage 1997; 13:63 74.