Psychosomatics 41:121-127, April 2000
© 2000 The Academy of Psychosomatic Medicine
Factor Analysis of the Zung Self-Rating Depression Scale in a Large Ambulatory Oncology Sample
Steven D. Passik, Ph.D.,
Jeremy C. Lundberg, M.S.W., L.S.W.,
Barry Rosenfeld, Ph.D.,
Kenneth L. Kirsh, M.S.,
Kathleen Donaghy, Ph.D.,
Dale Theobald, M.D., Ph.D.,
Elizabeth Lundberg, B.A., and
William Dugan, M.D.
Received April 15, 1999; revised August 16, 1999; accepted September 7, 1999. From Oncology Symptom Control and Research, Community Cancer Care, Indianapolis, IN; Memorial Sloan-Kettering Cancer Center, New York, NY; Indiana University and Purdue University, Indianapolis; Indiana Oncology and Hematology Consultants, Community Cancer Care, Inc., Indianapolis, IN. Address correspondence and reprint requests to Dr. Passik, Community Cancer Care, Inc., Oncology Symptom Control Research, 115 W. 19th Street, Indianapolis, IN 46202.
|
ABSTRACT
|
Screening cancer patients for depression with self-report inventories presents clinical and methodological challenges. Many investigators separate "somatic" from "cognitive" symptoms when adapting such measures to oncology settings. However, this practice has rarely been empirically validated through factor-analytic studies. The following study describes a factor analysis of the Zung Self-Rating Depression Scale (ZSDS) from a large ambulatory sample of cancer patients (N=1,109). A four-factor solution emerged, consisting of a cognitive symptom factor, a manifest depressed mood factor, and two somatic factors (eating and non–eating related). These factors accounted for 20% (cognitive), 13% (mood), 8% (non-eating), and 7% (eating) of the variance on the Zung, respectively. The authors discuss the implications of these results as they pertain to screening cancer patients for depression.
Key Words: Depression • Zung Depression Scale • Cancer
|
INTRODUCTION
|
The identification of depression in cancer patients is a significant clinical and methodological challenge. Studies have estimated that between 15% and 53% of cancer patients qualify for a diagnosis of a depressive disorder.1–6 The challenge in identifying depression in cancer patients is due, in part, to the confounding of the neurovegetative or "somatic" symptoms of depression (e.g., insomnia, weight loss, fatigue) by symptoms of cancer or side effects of cancer treatment. Thus, investigators and clinicians typically have adopted strategies for either separating the cognitive and somatic symptoms or developing substitute criteria for the somatic symptoms. In research settings, modified diagnostic criteria have been employed to address this issue.3 Where the use of self-report questionnaires for screening depression is concerned, some investigators have eliminated somatic symptoms entirely in the development of depression questionnaires intended for medically ill patients (e.g., The Hospital Anxiety and Depression Scale, HADS).7 Another alternative has been to form cognitive and somatic "subscales" within traditional measures of depression based on theoretical distinctions and clinical observation, though rarely on the basis of empirical justification. More specifically, the use of factor analytic techniques to assess both the degree to which various somatic and cognitive symptoms co-occur as well as the significance of different clusters of symptoms in assessing depression, has been noticeably absent from the research literature.
Our research group is not immune to this criticism, having recently published the results of a study in which depression was assessed by eliminating the somatic items from a self-report measure of depression.2 The Zung Self-Rating Depression Scale (ZSDS), a 20-item self-report measure of depressive symptoms, was administered to a large sample of oncology patients (N=1,109) receiving treatment at 25 ambulatory oncology clinics throughout Indiana. Responses to the entire scale and the utility of a modified version of the scale composed of the 11 cognitive scale items (eliminating the 9 somatic items) were examined. This latter measure, which we termed the Brief Zung Self-Rating Depression Scale (BZSDS), was highly correlated with the original measure (r=0.91) and yielded similar patterns of correlations with measures of disease status (e.g., advanced stage, degree of physical disability, primary tumor type) and clinical treatment (e.g., patients receiving antidepressant medications had higher scores on the ZSDS and BZSDS). This abbreviated version of the ZSDS identified somewhat fewer patients as having clinically significant levels of depressive symptoms compared to the full 20-item scale (31.1% and 35.9%, respectively). Despite the intuitive appeal of both this method to derive an abbreviated depression scale as well as the apparent utility of such a measure, the lack of empirical evidence for a two-factor breakdown of the ZSDS (and other scales in cancer patients and other medically ill populations) is troubling.
Although the factor structure of the ZSDS has not been examined in medically ill populations, several researchers have factor-analyzed the ZSDS in physically healthy samples. For example, Zung8 conducted a factor analysis of ZSDS scores from healthy individuals age 65 and over. The study revealed a primary factor of "loss of self-esteem." Additional factors were all complex in nature, with a second factor consisting of psychomotor retardation and insomnia and two additional factors with constipation and libido items loading on each, respectively. Not only is this factor structure difficult to interpret, but its applicability to medically ill patients is unknown. Kivela and Pahkala9 examined whether age and gender influenced the factor structure of the Zung. They found significant differences in the factor structure of the ZSDS among older men and women, leading the authors to caution against blanket interpretations of the ZSDS data for patients over age 75.
Because of the lack of clear empirical support for an intuitively appealing two-factor breakdown of depressive symptoms into "cognitive" and "somatic" subscales, we examined the factor structure of the ZSDS in the same large sample of cancer patients originally described by Dugan and colleagues.2 We examine the factor structure of the ZSDS and the composition of these factors, as well as the relationship between ZSDS factors and known correlates of depression.
|
METHODS
|
Procedures
A large consecutive convenience sample of subjects (N=1,109) was drawn from patients who entered one of 25 oncology clinics in urban and rural areas throughout Indiana. All patients who visited one of the clinics during the study interval were screened for possible study enrollment, excluding those attending their first clinic appointment. Patients were eligible for the study unless they were judged by the clinician to be too debilitated to complete the research survey (n=65) or could not read and understand the informed consent (n=36). Forty-seven patients refused to participate. Subjects completed the survey before their scheduled physician appointment. The physician recorded information regarding the subject's disease status, pain score, and performance rating after the appointment.
Measures
Zung Self-Rating Depression Scale
The ZSDS is a 20-item self-report measure of the symptoms of depression. Subjects rate each item according to how they felt during the preceding week. Item responses are ranked from 1 to 4 with higher numbers corresponding to more frequent symptoms (although several items are scored in reverse). The sum of the 20 items produces a raw score that is converted into a percentage of the depression measurable by the scale (termed the "SDS index"). Index scores are then categorized into 4 levels to offer a global clinical impression, as recommended by the instrument developers: I, within normal range, no significant psychopathology (SDS Index: <50); II, presence of minimal-to-mild depression (SDS Index: 50–59); III, presence of moderate-to-marked depression (SDS Index: 60–69); and IV, presence of severe-to-extreme depression (SDS Index:  70). Scores on the Zung do not yield a clinical diagnosis of depression but rather indicate levels of depressive symptoms that may be of clinical significance. Several studies have established the Zung scale as a reliable and valid instrument for measuring depressive symptoms.10–12
ECOG Performance Status
The Eastern Cooperative Oncology Group (ECOG13) grading system was used to rate the subject's ability to carry out activities of daily living. Physicians rate the subject's performance status from 0 to 4: Grade 0, normal activity; Grade 1, restricted in strenuous activity but ambulatory and able to do light work; Grade 2, ambulatory and capable of self-care, not able to work; Grade 3, limited self care; and Grade 4, bed-bound. Reliability of this instrument has been demonstrated in several investigations.14,15
Physicians' Depression and Anxiety Ratings
Physicians were asked to assess the level of depression and anxiety they perceived in their patients using separate 0 to 10 numerical rating scales. The scales were anchored at 0 with the statement "none" and at 10 with "severe."
Sociodemographic and Health Characteristics
Sociodemographic factors, including age, gender, marital status, and living arrangements, along with information about the subject's social support systems and use of counseling services, were elicited via questionnaire. Cancer site and stage-of-disease information were recorded from the subject's medical chart. Disease status was defined as follows: disease-free/remission (no current evidence of cancerous tumors or cells); stable disease (cancerous tumors or cells present but not currently spreading); and active disease/relapse (cancerous tumors or cells currently spreading). Disease status does not define treatment status. (Patients were at a variety of treatment levels at the time of the screening, including patients who had no evidence of current disease but were receiving adjuvant chemotherapy treatment.)
Statistical Analyses
Because we had no a priori hypotheses regarding the number of factors likely to emerge from the 20-item Zung rating scale in cancer patients, principal-component factor analysis was used to extract the factor solution. The number of factors was determined using a combination of several methods, including a minimum eigenvalue of 1.0, examination of the scree plot, and analysis of the factor loadings and residual variance. Each of these methods supported the application of a four-factor solution. This solution was then subjected to a varimax rotation to minimize the overlap between different factors. Factor scores were calculated as the sum of scores on the items making up each factor, and these factor scores were used in correlational and regression analyses to assess the relationships between various factors and other relevant variables (i.e., physician ratings of depression, pain, and anxiety). All data were analyzed using SPSS 7.5.
|
RESULTS
|
Sample Characteristics
A total of 1,109 subjects completed the questionnaire and were included in the survey. The sample was predominantly female (64.2%), between 50 and 80 years old (75.4%), married (67.6%), and living with their spouse (66.7%). Most subjects described their primary caretaker as either their spouse (48.9%) or other family members (18.7%), although 271 subjects (25.5%) indicated that they had no one whom they considered a primary caretaker, and an additional 41 subjects (3.9%) reported relying on paid professionals or volunteers for care.
Nearly one-half of the sample (46.8%) was rated by their physicians as having no impairment in their physical performance ability (ECOG Performance Status, Grade 0). The most common cancer sites were breast (36.4%), colon or rectal (16.0%), lymphoma (10.0%), and lung (9.9%). About half of the study subjects had early-stage disease (24.6% Stage I, 33.0% Stage II) at diagnosis. Over 60% of the subjects were disease-free or stable. Patients' mean ratings of their pain at its worst in the past week was 3.02 (range: 0–10, SD=2.7).
Factor Solution
Principal-components factor analysis generated a four-factor solution to the 20-item Zung rating scale. Although this solution only accounted for 48% of the variance in Zung rating scale item intercorrelations, both an examination of the scree plot as well as the decrease in eigenvalues of additional factors supported this solution (see Table 1). Moreover, inspection of solutions with fewer factors revealed an unacceptable amount of residual variation, whereas solutions with more factors appeared to contain redundant information (i.e., many items loading onto multiple factors). The four-factor solution, on the other hand, generated relatively distinct factors, each composed of between two and eight items, with no items loading on more than one factor. This solution was subsequently rotated, using a varimax rotation, to further reduce the degree of overlap between scales (see Table 2). Items were considered to load on a factor if the rotated factor loading was at least 0.5. Two items were included in factors for which the loadings were only 0.49 because these two items did not load on any of the other factors and were logically consistent with the derived factors. An additional two items (Item 2, "Morning is when I feel the best" and Item 4, "I have trouble sleeping through the night") did not load on any factor.
The first factor (Cognitive), composed of eight items, appeared to primarily reflect cognitive symptoms of depression such as anhedonia, feelings of worthlessness, and difficulty concentrating/making decisions. This factor, which accounted for 20% of the scale variance after rotation, included the following items: "My mind is as clear as it used to be" (Item 11), "I feel hopeful about the future" (Item 14), "I find it easy to do the things I used to do" (Item 12), "I still enjoy the things I used to do" (Item 20), "My life is pretty full" (Item 18), "I feel that I am useful and needed" (Item 17), "I find it easy to make decisions" (Item 16), and "I enjoy looking at, talking to, and being with attractive men/women" (Item 6). This factor had a high degree of internal consistency within these eight items, with an alpha coefficient of 0.85.
The second factor (Manifest Depressed Mood), composed of five items, appeared to reflect symptoms of manifest depressed mood such as sad or blue mood, irritability, crying, and thoughts of death or suicide. This factor accounted for an additional 13% of the variance after rotation. The following items were included in the second factor: "I feel downhearted and blue" (Item 1), "I have crying spells or feel like it" (Item 3), "I am restless and can't keep still" (Item 13), "I am more irritable than usual" (Item 15), and "I feel that others would be better off if I were dead" (Item 19). Despite being composed of only five items, this scale also revealed a reasonable degree of internal consistency, with an alpha coefficient of 0.67, suggesting that the factor solution resulted in a grouping of items with relatively high intercorrelations.
The remaining two factors, which included three and two items respectively, reflected more typically somatic symptoms and somatic preoccupation associated with depression and may be those most highly confounded by an illness such as cancer. The third factor (Somatic: Non-Eating), accounted for 8% of the variance after rotation. This factor contained the following three items: "I have trouble with constipation" (Item 8), "My heart beats faster than usual" (Item 9), and "I get tired for no reason" (Item 10). The fourth factor (Somatic: Eating), composed of only two items, accounted for an additional 7% of the variance. These two items both reflected appetite difficulties or concerns, "I eat as much as I used to" (Item 5) and "I notice that I am losing weight" (Item 7). Because of the small number of items comprising these factors, the low alpha coefficients observed (0.47 and 0.48, respectively) have relatively little interpretive value.
Correlations Within and Between Factors
Summing the Items for Each Factor-Generated Score
Factor scores were then used in analyses to ascertain the correlations between the factors as well as between individual factor scores and other relevant variables. The first of these questions was relatively straightforward: Although the four factors were significantly correlated (not surprisingly, given the large sample size) with one another, these correlations were all relatively modest, ranging from 0.24 to 0.43. Factor I, reflecting primarily cognitive symptoms of depression, was the factor most highly correlated with ZSDS rating scale total scores (r=0.89, P<0.0001), while the remaining factors were somewhat less highly correlated with ZSDS total scores (Factor II: r=0.66, P<0.0001; Factor III: r=0.60, P<0.0001, Factor IV: r=0.53, P<0.0001). This pattern of correlations, however, may in part reflect the number of items comprising each scale (i.e., since Factor I is made up of eight items, or 40% of the original Zung scale, this factor would be expected to yield a higher correlation than scales made up of five, three, or two items).
Correlations between the ZSDS factor scores and physician ratings of depression made on a 0-to-10 numerical rating scale revealed an interesting pattern of results. The factor most highly correlated with physician ratings of depression (r=0.34, P<0.0001) was the factor composed of items tapping manifest (and often obvious) aspects of depressed or irritable mood, crying, and thoughts of death or suicide. The remaining three factors yielded correlations between 0.20 and 0.24 with physician ratings of depression, although these correlations were statistically significant (P<0.0001) in part because of the large sample size. A similar pattern of results emerged for physician ratings of their patients' anxiety (using a 0-to-10 numerical rating scale).
Patients' functional status also generated an interesting pattern of correlations with the four factor scores. Although ECOG ratings were significantly correlated with all four factors, these correlations were strongest for Factors I and IV (r=0.30, P<0.0001). Factor II scores, on the other hand, were substantially less correlated with the ECOG ratings (r=0.17, P<0.0001), although the correlation was still statistically significant, as was the correlation between ECOG ratings and Factor III scores (r=0.28, P<0.0001). Scores on Factors I and II also differed significantly by age (F=3.21, P<0.012 and F=10.98, P<0.0001, respectively); however, no such differences emerged for Factors III and IV. Post hoc analyses (based on a Bonferroni adjustment) revealed that subjects in the oldest age-group (over 80) obtained significantly higher scores on Factor I compared to subjects between 50 and 69 years old. There were no other significant differences among Factor I scores. Factor II scores, on the other hand, were significantly higher among younger subjects. Subjects less than 50 years old, for example, obtained significantly higher scores on Factor II than did subjects over 60. Likewise, subjects between 50 and 59 years old obtained significantly higher scores on Factor II than did subjects over 70. There were no other significant differences in Factor II scores. There were no significant associations between factor scores and patient gender.
|
DISCUSSION
|
The results of this study indicate that the ZSDS has an understandable and clinically interesting factor structure. In a sense, this study represents the first large-scale statistical demonstration of the common clinical assertion that depressive symptoms in cancer patients can be divided along somatic and cognitive lines. These findings go on to suggest that these broad dimensions might require further refinement. Indeed, the four-factor solution suggests that along the cognitive or nonsomatic dimension, symptoms of manifest depressed mood are (broadly) separate from the ideational- and anhedonia-related symptoms. Along the somatic dimension, eating-related somatic symptoms (weight and appetite loss) seem to be distinguishable from other important somatic symptoms among patients with cancer. The failure of diurnal variation and insomnia to cluster with somatic factors in this analysis suggests that the somatic symptoms of depression may not be particularly salient in cancer patients. These results, as well as the implications of the correlations between the physicians' ratings with the individual factor scores, will be discussed below.
Psycho-oncologists have long argued that the neurovegetative symptoms of depression were separable from the cognitive/ideational symptoms, with the latter being more reliable for diagnostic purposes. The factor structure of the ZSDS reflects this conceptualization. Factor I of the ZSDS, which accounted for 20% of the variance in subjects' responses, is a clearly demarcated cognitive factor composed of concentration problems, outlook, and enjoyment in activities. Factor II, which accounted for 13% of the variance, is composed of items assessing symptoms of manifest depressed mood (i.e., crying, irritability, blue mood, and suicidal ideation). We have suggested elsewhere, based on other analyses, that the former symptoms may be more relevant to diagnosing depression in cancer patients despite the apparent tendency for oncologists to place more importance on the latter symptoms. Moreover, symptoms such as depressed mood, sadness, and irritability, which may be especially prominent when patients are assessed in the doctor's office before a visit, may be more variable and therefore less indicative of a depressive illness. Even suicidal ideation, while often perceived as the penultimate indicator of depression, may reflect a more transient mood symptom in patients with cancer. Although perhaps counterintuitive, our clinical experience has revealed that many patients express such thoughts as an indication of frustration or to "blow off steam" rather than as a genuine desire to die.
The hypothesis that Factor I symptoms are more salient markers of depression, whereas Factor II symptoms are more transient, is supported by the pattern of observed correlations between factor scores and medical/demographic variables. For example, while ratings of functional ability (ECOG scores) should be significantly correlated with severity of depressive symptoms, this correlation was substantially lower for Factor II compared to the other three factors. Similarly, while age was not correlated with scores on Factors I, III, or IV, there was an inverse correlation between age and Factor II scores. This finding indicates that younger subjects were relatively more likely to endorse the manifest symptoms of depressed mood compared to older subjects, a finding that could reflect generational differences in emotional expression or expectations regarding their physical functioning (i.e., with older subjects being less upset over perceived functional disability).
The presence of separate factors corresponding to eating-related and non–eating-related somatic symptoms suggests that appetite disturbances related to the administration of cancer therapies and/or depression may require separation from the other somatic symptoms. Fatigue is the only item on Factor III that is typically considered to be a symptom of depression (constipation and racing heartbeat were likely included in the ZSDS originally to assess somatic preoccupations in healthy subjects). Moreover, all the symptoms included on Factor III could reflect symptoms or side effects from cancer treatment (e.g., treatment with opioids for pain). Furthermore, it is of interest that items assessing diurnal variation and insomnia did not cluster with these somatic symptoms. Whether this reflects actual differences in depressive syndromes compared to those seen in healthy depressed patients or other treatment-related variables is unclear.
The correlations between physicians' ratings and items on the ZSDS have been discussed elsewhere, but warrant reiteration here.1,2 The tendency of physicians to assess depression based on manifest depressed mood and not on the more clinically significant cognitive factors reflects the need for further training and education among oncologists if they are able to accurately assess for mood disturbances in their patients. Ongoing discussions of issues such as physician-assisted suicide presume that physicians are capable of assessing depression in their patients, yet these data suggest that such assumptions may be unjustified. The potential for "missed" diagnoses of depression among patients who display primarily cognitive symptoms certainly exists, and the importance of these symptoms cannot be overemphasized.
Limitations
The results of this study should be interpreted with a degree of caution. Although the sample in this survey was large, it was a community-based, largely rural sample that may differ from other oncology populations (i.e., tertiary care; inner-city urban). Furthermore, before firm conclusions can be drawn, additional confirmatory factor analyses should be performed with a new sample of patients. Finally, the very small number of items loading on Factors III and IV may reflect limited stability in those factors. Whether these factors are meaningful warrants further investigation.
|
CONCLUSION
|
The ZSDS has an interesting and meaningful factor structure in cancer patients. Depressive symptoms do indeed cluster into cognitive and somatic dimensions, although these dimensions may require further refinement. Given the somewhat complex relationships between physical and psychological symptoms among patients with cancer, careful attention to the precise symptoms experienced and the patterns of symptoms reported may have important ramifications for clinical practice as well as further research.
|
REFERENCES
|
- Passik SD, McDonald MV, Dugan WJ: Use of an algorithm-based antidepressant treatment in community-based hospital oncology practice: preliminary results. J Pain Symptom Manage 1998; 15:S11
- Dugan WM, McDonald MV, Passik SD, et al: Use of the Zung Self-Rating Depression Scale in cancer patients: feasibility as a screening tool. Psychooncology 1998; 7:483–493[CrossRef][Medline]
- Kathol R, Mutgi A, Williams J: Diagnosis of major diagnosis according to foursets of criteria. Am J Psychiatry 1990; 147:1021–1024
- Plumb B, Holland J: Comparative studies of psychosocial function in patients with advanced cancer I: self-reported depressive symptoms. Psychosom Med 1997; 39:264–276[Abstract/Free Full Text]
- Evans D, McCartney C, Haggerty J: Treatment of depression in cancer patients in association with better life adaptation: a pilot study. Psychosom Med 1988; 50:72–76
- Falcon S, Pinglo F, Hurtado de Mendoza L: Autoassessment Zung Scale (anxiety and depression) use in hospitalized patients for cancer treatment. Proc Am Soc Clin Onc Los Angeles, 1998
- Zigmond AS, Snaith RP: The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983; 67:361–370[Medline]
- Zung W: Factors influencing the Self-Rating Depression Scale. Arch Gen Psychiatry 1967; 16:543–547[Abstract/Free Full Text]
- Kivela S, Pahkala K: Sex and age differences of factor pattern and reliability of the Zung Self-Rating Depression Scale in a Finnish elderly population. Psychol Rep 1986; 59:587–597[Medline]
- Agrell B, Dehlin O: Comparison of six depression rating scales in geriatic stroke patients. Stroke 1989; 20:1190–1194
- Gabrys JB, Peters K: Reliability, discriminant and predictive validity of the Zung Self-Rating Depression Scale. Psych Rep 1985; 57:1091–1096
- Biggs JT, Wilie LT, Ziegler VE: Validity of the Zung Self-Rating Depression Scale. Br J Psychiatry 1978; 132:381–385[Abstract/Free Full Text]
- Eastern Cooperative Oncology Group: Functional Assessment Scale. Clinical Oncology: A Functional Approach. Edited by Rubin P. American Cancer Society, 1983
- Conill C, Verger E, Salamero M: Performance status assessment in cancer patients. Cancer 1990; 65:1864–1866
- Roila F, Lupattelli M, Basurto C: Intra- and interobserver variability in cancer patients' performance status assessed according to Karnofsky and ECOG scales. Ann Oncol 1990; 2:437–439[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
S. F. Garcia, D. Cella, S. B. Clauser, K. E. Flynn, T. Lad, J.-S. Lai, B. B. Reeve, A. Wilder Smith, A. A. Stone, and K. Weinfurt
Standardizing Patient-Reported Outcomes Assessment in Cancer Clinical Trials: A Patient-Reported Outcomes Measurement Information System Initiative
J. Clin. Oncol.,
November 10, 2007;
25(32):
5106 - 5112.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C.P. van Wilgen, P.U. Dijkstra, R.E. Stewart, A.V. Ranchor, and J.L.N. Roodenburg
Measuring Somatic Symptoms With the CES-D to Assess Depression in Cancer Patients After Treatment: Comparison Among Patients With Oral/Oropharyngeal, Gynecological, Colorectal, and Breast Cancer
Psychosomatics,
December 1, 2006;
47(6):
465 - 470.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
Y. Guo, D. L. Musselman, A. K. Manatunga, N. Gilles, K. C. Lawson, M. R. Porter, J. S. McDaniel, and C. B. Nemeroff
The Diagnosis of Major Depression in Patients With Cancer: A Comparative Approach
Psychosomatics,
October 1, 2006;
47(5):
376 - 384.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. C. Trask
Assessment of Depression in Cancer Patients
J Natl Cancer Inst Monographs,
July 1, 2004;
2004(32):
80 - 92.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Lloyd-Williams
Depression--the hidden symptom in advanced cancer
J R Soc Med,
December 1, 2003;
96(12):
577 - 581.
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Akechi, T. Nakano, N. Akizuki, M. Okamura, K. Sakuma, T. Nakanishi, E. Yoshikawa, and Y. Uchitomi
Somatic Symptoms for Diagnosing Major Depression in Cancer Patients
Psychosomatics,
June 1, 2003;
44(3):
244 - 248.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
H. PESSIN, B. ROSENFELD, and W. BREITBART
Assessing Psychological Distress Near the End of Life
American Behavioral Scientist,
November 1, 2002;
46(3):
357 - 372.
[Abstract]
[PDF]
|
|
|
Get information about faster international access.
Privacy Policy
Copyright © 2000
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
