Psychosomatics 42:241-246, June 2001
© 2001 The Academy of Psychosomatic Medicine
A New Psychosocial Screening Instrument for Use With Cancer Patients
James Zabora, Sc.D.,
Karlynn BrintzenhofeSzoc, D.S.W.,
Paul Jacobsen, Ph.D.,
Barbara Curbow, Ph.D.,
Steven Piantadosi, M.D., Ph.D.,
Craig Hooker, B.S.,
Albert Owens, M.D., and
Leonard Derogatis, Ph.D.
Received April 28, 2000; revised November 27, 2000; accepted November 29, 2000. From The Johns Hopkins Oncology Center, Baltimore, MD; the H. Lee Moffitt Cancer Center, Tampa, Florida; and the University of Maryland, Baltimore, MD. Address reprint requests to Dr. Zabora, The Johns Hopkins Oncology Center, 600 North Wolfe Street, Baltimore, MD 21287.
|
ABSTRACT
|
The authors performed a principal components factor analysis on the 18-item Brief Symptom Inventory (BSI-18), a new brief screening inventory. The factor analysis, in which four factors were specified, is consistent with findings in a previous community sample. The study sample consisted of 1,543 cancer patients who completed the full BSI as part of their entry into care at a regional cancer center. The reliability of the BSI-18 was determined based on the calculation of the internal consistency, mean item scores, and correlations with the total score of the BSI. In addition, sensitivity and specificity was calculated to determine the ability of the BSI-18 to discriminate positive and negative cases. The BSI-18 is a shortened version of the BSI that can serve as a brief psychological screening instrument. The BSI-18 can be incorporated into outpatient clinics to prospectively and rapidly identify cancer patients with elevated levels of distress who are in need of clinical interventions. Early identification of distress with appropriate interventions can reduce distress, enhance quality of life, and decrease health care costs.
Key Words: Syndromes Secondary to General Medical Disorders • Screening Instrument
|
INTRODUCTION
|
Often the distress associated with the diagnosis of cancer may not become manifest to the health care team until the patient reaches an observable crisis event.1 Frequently, referrals to psychosocial providers occur when the patient is severely depressed or anxious, is suicidal, or is experiencing significant conflicts with the family.2 Although psychosocial interventions at acute points in a patient's care have been shown to be effective, the question remains whether an early psychological screening program would enable health care providers to identify patients at risk for distress associated with a cancer diagnosis.3,4 If so, clinical interventions could be initiated during the early phase of care rather than at crisis events.5 Because psychological distress may actually increase health care costs and prolong medical treatments,6 early identification and intervention may also produce a significant financial benefit. The costs associated with early identification and intervention may be significantly less than the costs of adverse medical events and unnecessary hospitalizations related to elevated distress.
Several prevalence studies7–9 suggest that one of every three newly diagnosed cancer patients will experience significantly higher levels of distress and may benefit from social work, psychological, or psychiatric intervention.10 Table 1 details the three major studies related to psychological distress among cancer patients. Given a relatively consistent prevalence of distress at approximately 30%, use of structured interviews or assessments in large-volume clinics may require a significant amount of staff time with a low positive case yield. Self-report measures for the purpose of screening offer a valuable alternative for cancer centers that see a high number of cancer patients per year and wish to screen for psychological distress.11
Given these initial issues, screening should not be confused with assessment. Screening is a rapid method to prospectively identify patients who may potentially experience significant difficulty in their attempts to cope and adapt to their diagnoses and treatments. Screening is a predictive model. Assessment seeks to accomplish a series of tasks in the early phases of a relationship with a patient. These tasks include an estimate of the severity of the patient's distress, definition of the initial course of action, development of a dynamic understanding of the patient, the establishment of a diagnosis, and the first step in the development of a therapeutic relationship.12
Any method of psychosocial screening must be brief and pragmatic. Although improvements in screening techniques continue to be developed, further research is necessary to identify brief, simple, and accurate tools to accomplish this vital task. Through screening, the early incorporation of psychosocial and behavioral interventions in cancer treatment may be more readily accepted by patients and less stigmatizing.13 In addition, these interventions complement cancer therapies and may enhance medical outcomes while reducing the overall costs of health care.6
|
METHODS
|
Sample
Since 1988, staff in a comprehensive cancer center have experimented with psychosocial screening in an effort to develop a prospective model of psychosocial care with early identification of risk linked to appropriate interventions. Since the inception of these efforts, approximately 10,000 psychological profiles have been amassed through the use of the Brief Symptom Inventory (BSI).14 For this project, cases beginning in January 1995 were selected in succession until an acceptable sample size was constructed in order to test the ability of the BSI-18, a recently abbreviated version of the scale,15 to successfully identify distressed cancer patients. An analysis was conducted to calculate the alpha for the BSI-18 in an oncology population. After this analysis, the BSI-18 was subjected to a principal components factor analysis. This analysis was conducted by specifying four factors consistent with the recent findings of Derogatis,15 in an effort to further support the structural hypothesis of the BSI-18 in a cancer population.
Measures
Brief Symptom Inventory
The full BSI is a 53-item measure of psychological distress that contains three global scales and nine subscales.16 The BSI is written at a sixth-grade reading level and only requires 5 to 7 minutes to complete. Each item is rated on a 5-point Likert scale from 0 (not at all) to 4 (always). The patient is asked to respond to each item in terms of "how they have been feeling during the past 7 days." Positive cases can be identified by a Global Severity Index (GSI) score of  63 or any two subscales where the T-score is 63.14 The BSI has been used in prevalence studies related to psychological distress9,17 and has been tested for its efficacy as a screening instrument against the omega instruments that are based on the variables previously defined.3 In addition, the BSI possesses characteristics that are more suitable for screening than other instruments, such as the General Hospital Questionnaire or the Hospital Anxiety and Depression Scale.18 Table 2 details the psychometric properties of the BSI.
Each BSI subscale consists of 4–7 items that account for 49 of the overall 53 items. Four additional items (poor appetite, trouble falling asleep, thoughts of death and dying, and feelings of guilt) load on more than one factor but remain separate from the 9 subscales. These 4 configural items have been maintained as part of the BSI due primarily to their clinical predictive significance. For example, "poor appetite" loads on both the depression and somatization factors and can have discriminate significance regarding the nature of depressive symptoms.
Application of a Scale for Use With a Cancer Population
To validate the use of a scale on a specific population (i.e., generalize its valid use in that population), a number of issues must be considered. First, the scale, as a measure of a concept such as distress, should be practical and acceptable to patients, clinicians, and investigators. Practicality and acceptability can be achieved through short completion time, understandable directions and items, and low costs for administration of the scale. Second, the fundamental principles of validity must be established for the population. Finally, scales must minimize measurement error. Scales with measurement error produce the potential for bias and misinterpretation of results.19
Statistical Analyses
As mentioned previously, internal consistency of the BSI-18 in a cancer population was established through the calculation of Cronbach alphas. In addition, the total score on the BSI-18 was correlated with the GSI raw score of the full BSI. We subjected the 18-item BSI to a principal component factor analysis with a varimax rotation while specifying four factors. Values are means±standard deviations for each item calculated. The total variance accounted for by the BSI-18 was also calculated. Finally, the sensitivity and specificity were calculated to determine the capability of the BSI-18 to discriminate positive from negative cases, and the positive predictive value was derived.20
|
RESULTS
|
Sample
Our test sample consisted of 1,543 cancer patients with a mean age of 55.5±14.5 (range=14–90 years). The majority of the sample were men (52.0%). Because consecutive cases were used, over 35 different diagnoses were represented in the total sample. The top ten cancer diagnoses were as follows: breast (19.9%), prostate (13.0%), leukemia (9.8%), lymphoma (9.7%), lung (5.5%), head and neck (4.5%), pancreatic (4.1%), gynecological (3.8%), and brain (2.1%). These cases represent patients seen from 1995 to 1998, and the BSI forms from this time period were computer scanned so that individual item responses could be analyzed. Table 3 presents characteristics of the sample used in the analysis of the BSI-18.
Reliability and Validity of the BSI-18
Means±standard deviations for the 18 items and total score for the sample are given in Table 4. In addition, internal consistency reliability was examined for the BSI-18 and the 53-item BSI. Cronbach's alpha was 0.89 for the BSI-18 and 0.95 for the full BSI. The BSI-18 total score was significantly correlated with the GSI of the BSI with r=0.84 (P<0.001). Item-to-total correlations ranged from 0.34 to 0.70.
Principal Components Factor Analysis
A factor analysis of the BSI-18 was conducted on our sample of 1,543 cancer patients, which consisted of a principal components analysis with a varimax rotation. Four factors were identified that had eigenvalues greater than (or very close to) 1.0 and met the scree test for retention. The four factors accounted for 57.8% of the variance in the matrix. The factor pattern matrix from this analysis is given in Table 5.
The results of this analysis of the BSI-18 strongly confirm the hypothesized dimensional structure of the instrument and are very similar to a factor analysis reported by Derogatis15 on a community sample of 1,134 individuals. That analysis also isolated four factors that met retention criteria and accounted for 57.2% of the variance. Much as was with the case of the Derogatis15 analysis with community individuals, even though the BSI-18 was hypothesized to support a three-dimensional structure, the observed four-dimensional solution fits well with the test's hypothesized dimensional composition.
The first factor identified in the present analysis is clearly the anxiety factor, with 5 of the BSI-18 anxiety items demonstrating heavy loadings on this factor. Factor I accounted for approximately 20% of the variance in the matrix. Unexpectedly, the anxiety item, "Feeling so restless I couldn't sit still" did not load on Factor I but instead loaded on the second factor. In addition, the depression item "Feeling lonely" showed a split loading on this factor and the second factor.
Factor II is a very explicit representation of the BSI-18 depression factor. Five of the six hypothesized depression items showed saturated loadings on this factor, which accounted for an additional 17% of the variance. Of the hypothesized depression items, only "Suicidal ideation" did not load on Factor II.
Factor III demonstrates high loadings on all six items of the hypothesized somatization dimension of the BSI-18 and little else. Much as was the case in Derogatis' test15 of the hypothesized structure in a community population, the somatization dimension showed almost perfect empirical confirmation. Factor III accounted for 14% of the variance.
The fourth factor isolated in the analysis fell just short (0.98) of the eigenvalue retention criteria of 1.0, although it did pass the scree test. The final factor revealed only one item ("Suicidal ideation") as correlating strongly with the dimension and accounted for almost 7% of the variance. Interestingly, although loadings fell slightly below the cutoff for reporting (i.e., 0.40), the two items of "Feelings of terror and panic" (0.38) and "Suddenly scared for no reason" (0.35) both showed mild correlations with this dimension. This finding is analagous to the fourth factor reported in the community population,19 which combined items reflecting panic and suicidal thoughts. Suicidal ruminations are a unique aspect of psychopathology that recent research (e.g., Vollrath et al.21) has demonstrated are often associated with panic states, just as they are known to accompany serious depression. Confronted with a diagnosis of cancer, and all the potential hardships that such a diagnosis implies, it is understandable that vulnerable individuals may experience panic, and with it, some cognitive ruminations of an ultimate means of escape.
The current analysis provides strong confirmation of the hypothesized dimensional structure of the BSI-18 in an oncology population. In doing so, the results not only complement prior research with the instrument in a community population, but these results demonstrate a substantial level of construct generalizability.22
Scoring the BSI-18
Standardized scores were developed in psychometrics primarily to facilitate comparisons of the standing or performance of an individual on some attribute of interest (e.g., depression, intelligence, well-being, etc.) to a relevant reference group (e.g., newly diagnosed cancer patients). In a clinical application, such "norms" enable us to better judge how distressed or ill a patient is, as well as better appreciate the magnitude and significance of any change observed over the course of treatment.
Norms for the BSI-18 have been developed by the author of the BSI-1815 for the three primary dimension scores (somatization, depression, and anxiety) and the Global Severity Index (GSI) from approximately 1,500 cancer patients in the current sample. In addition, community norms are also available.19 In recognition of the consistent observation that manifestations of symptomatic distress differ substantially across men and women,14 all norms are gender keyed.
As was the case with both the SCL-90-R23 and the BSI,14 parent instruments to the BSI-18, the metric selected to serve as the standardized score for the BSI-18 is the area T-score. The area T-score is characterized like all T-scores by a distribution with a mean of 50 and a standard deviation of 10 but possesses substantial advantages over linear standardized scores. Because area transformations are normalizing transformations, the area T-scores carry with them meaningful and interpretable percentile equivalents. An area T-score of 60 always places the respondent in the 84th percentile, while a score of 70 is equivalent to the 98th percentile. This is true of linear standardized scores only when the underlying distribution is normal. A detailed discussion of characteristics of various standardized scores is given elsewhere.24
To identify appropriate cutoff scores for the BSI-18, the cases were separated by gender. In examining the distribution of the scores and what is known about the prevalence of distress using the 53-item BSI with this population,17 the 25th percentile was used to determine the cutoff point for caseness. For men, the 25th percentile fell at a score of 10, and for women, it fell at 13. To further substantiate the norms for the BSI-18, sensitivity and specificity were calculated to determine how well the BSI-18 identified positive cases using a score of 10 for men and 13 for women. These results are detailed in Table 6. Finally, given a sensitivity of 91.2% and a specificity of 92.6%, the positive predictive value (PPV) for the BSI-18 was also calculated (Sens/1-Spec), which produced a PPV of 12.32. Any PPV >10 indicates that a test has a strong likelihood for positive case identification.
|
DISCUSSION
|
These results furnish significant support for the use of the BSI-18, a new psychosocial screening instrument, with cancer patients. Principal components factor analysis showed confirmation for the hypothesized dimensional composition. The significant correlations between the BSI-18 total score and the GSI score of the 53-item BSI substantiates the continuity of the BSI-18 in identifying the psychological distress construct.
Significance of the Findings
Given that 30% of all cancer patients experience elevated levels of psychological distress, early identification of vulnerable patients is essential to their comprehensive management. Undetected and untreated distress creates the potential for patients to medicalize psychological symptoms and greatly complicates treatment of the primary neoplastic disorder. Because the focal point of care is the tumor site, patients are often reluctant to consistently reveal their distress to members of the health care team. Patients may perceive that a diversion from the tumor to focus on an emotional response may jeopardize the outcome of their cancer therapy. Given the high patient volumes in many cancer centers, little time, if any, is available for the health care team to assess the emotional concerns of every patient. Further, even when concerns are detected, the patient has often reached a crisis state, and the increased distress has become apparent to members of the team. At this point, referrals for psychosocial care occur when the patient is agitated, acutely depressed, or suicidal.2 Although clinical interventions can be effective at these crisis points, the major question is, "Could these patients be identified as being high risk early in the course of their cancer treatment?" thus allowing earlier initiation of preventive interventions.
These findings are an important step in the development of a prospective psychosocial care delivery system for cancer patients. The BSI-18 is a brief tool designed to identify high-risk patients in the actual clinical setting where cancer therapies are delivered. Given the brevity and simple scoring, the BSI-18 can be easily incorporated into the clinic registration areas and yields an outcome that allows support staff to initiate a referral for psychosocial care. These steps can be achieved within the first or second outpatient visit, and psychosocial services can be incorporated as a component of comprehensive cancer care. The reliability and validity of this instrument support the use of the BSI-18 as a psychological screening instrument for use with cancer patients. Future prospective studies will determine the test-retest reliability and more specific predictive validities of the BSI-18 for this patient population.
Psychosocial screening not only provides the opportunity to prospectively link patients to specific psychosocial services, but use of a standardized screening instrument such as the BSI-18 can also serve as a baseline measure. Consequently, if patients with a high level of distress are directly referred to a counseling service, a standardized measure such as the BSI can be administered upon completion of the intervention. In effect, the screening instrument serves as the pretest measure triaging to the intervention and as a posttest measure to determine any change in the psychological status. For psychosocial programs to survive in a managed care environment, psychosocial services must also position themselves to compete for mental health capitated contracts.
|
REFERENCES
|
-
Weisman AD, Worden JW: The existential plight in cancer: significance of the first 100 days. Int J Psychiatry Med 1977; 7:1–15
-
Rainey LC, Wellisch DK, Fawzy FI: Training health care professionals in psychosocial aspects of cancer: a continuing education model. Journal of Psychosocial Oncology 1983; 1:41–60
-
Weisman AD, Worden JW, Sobel HJ:1980 Psychosocial screening and interventions with cancer patients: a research report. Boston, MA, Harvard Medical School, 1980
-
Zabora JR, Smith-Wilson R, Fetting JH, et al: An efficient method for the psychosocial screening of cancer patients. Psychosomatics 1990; 31:192–196[Abstract/Free Full Text]
-
Zabora JR, Loscalzo MJ: Comprehensive psychological programs: a prospective model of care. Oncology Issues 1996; 11:14–18
-
Allison TG, Williams DE, Miller TD, et al: Medical and economic costs of psychological distress in patients with coronary artery disease. Mayo Clin Proc 1995; 70:734–742[Abstract]
-
Derogatis LR, Morrow GR, Fetting JH, et al: The prevalence of psychiatric disorders among cancer patients. JAMA 1983; 249:751–757[Abstract/Free Full Text]
-
Farber JM, Weinerman BH, Kuypers JA: Psychosocial distress of oncology patients. Journal of Psychosocial Oncology 1984; 2:109–118
-
Stefanek ME, Derogatis LR, Shaw A: Psychosocial distress among oncology outpatients. Psychosomatics 1987; 28:530–539[Abstract/Free Full Text]
-
Zabora JR, Smith ED, Loscalzo MJ: Psychosocial rehabilitation, in Clinical Oncology, 2nd Edition, edited by Abeloff MD, et al. New York, Churchill Livingstone, 2000, pp 2845–2865
-
Derogatis LR: Self-report measures of distress, in Handbook of Stress. New York, Free Press, 1982
-
Tomb DA: Psychiatry for the House Officer, 3rd Edition. Baltimore, MD, Williams & Wilkins, 1988
-
Fawzy FI, Fawzy N, Arndt LA, et al: Critical review of psychosocial interventions in cancer care. Arch Gen Psychiatry 1995; 52:100–113[Abstract/Free Full Text]
-
Derogatis LR: The Brief Symptom Inventory (BSI): Administration, Scoring and Procedures Manual, 3rd Edition. Minneapolis, MN, National Computer Systems, 1993
-
Derogatis LR: BSI-18: Administration, Scoring and Procedures Manual. Minneapolis, MN, National Computer Systems, 2000
-
Derogatis LR, Melisaraatos N: The Brief Symptom inventory (BSI): an introductory report. Psychol Med 1983; 13:595–606[Medline]
-
Zabora JR, BrintzenhofeSzoc KM, Smith ED: Prevalence of psychological distress by cancer site. Proceedings of the American Society of Clinical Oncology 1996; 15:507–507
-
Zabora JR: Screening procedures for psychosocial distress, in The Handbook of Psycho-Oncology, edited by Holland JC. New York, Oxford University Press, 1998
-
Blythe BJ: Measurement of Direct Practice. Thousand Oaks, CA, Sage, 1989
-
Dawson-Saunders B, Trapp RG: Basic and Clinical Biostatistics. Norwalk, CT, Appleton and Lange, 1994
-
Vollrath M, Koch R, Angst J: The Zurich study IX. Panic disorder and sporadic panic: symptoms, diagnosis, prevalence and overlap with depression. Eur Arch Psychiatry Clin Neurosci 1990; 239, 221–230
-
Messick S: Validity of psychological assessment: validation of inferences from persons equals responses and performances as scientific inquiry into score meaning. Am Psychol 1995; 30, 741–749
-
Derogatis LR: Symptom Checklist-90-R (SCL-90-R) Administration, Scoring, and Procedures Manual, 3rd Edition. Minneapolis, MN, National Computer Systems, 1994
-
Tatsuoka MM: Selected Topics in Advanced Statistics: Standardized Scales–Linear and Area Transformations. Champaign, IL, Institute for Personality & Ability Testing, 1969
This article has been cited by other articles:
|
|
|
|
 
K. E. Hoffman, E. P. McCarthy, C. J. Recklitis, and A. K. Ng
Psychological Distress in Long-term Survivors of Adult-Onset Cancer: Results From a National Survey
Arch Intern Med,
July 27, 2009;
169(14):
1274 - 1281.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. J. Fraser, S. Bhatia, K. Ness, A. Carter, L. Francisco, M. Arora, P. Parker, S. Forman, D. Weisdorf, J. G. Gurney, et al.
Impact of chronic graft-versus-host disease on the health status of hematopoietic cell transplantation survivors: a report from the Bone Marrow Transplant Survivor Study
Blood,
October 15, 2006;
108(8):
2867 - 2873.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. R. Schover, R. Jenkins, D. Sui, J. H. Adams, M. S. Marion, and K. E. Jackson
Randomized Trial of Peer Counseling on Reproductive Health in African American Breast Cancer Survivors
J. Clin. Oncol.,
April 1, 2006;
24(10):
1620 - 1626.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Frumovitz, C. C. Sun, L. R. Schover, M. F. Munsell, A. Jhingran, J. T. Wharton, P. Eifel, T. B. Bevers, C. F. Levenback, D. M. Gershenson, et al.
Quality of Life and Sexual Functioning in Cervical Cancer Survivors
J. Clin. Oncol.,
October 20, 2005;
23(30):
7428 - 7436.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. Coultas, J. Frederick, B. Barnett, G. Singh, and P. Wludyka
A Randomized Trial of Two Types of Nurse-Assisted Home Care for Patients With COPD
Chest,
October 1, 2005;
128(4):
2017 - 2024.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Grassi, E. Rossi, S. Sabato, G. Cruciani, and M. Zambelli
Diagnostic Criteria for Psychosomatic Research and Psychosocial Variables in Breast Cancer Patients
Psychosomatics,
December 1, 2004;
45(6):
483 - 491.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. S. Donaldson
Taking Stock of Health-Related Quality-of-Life Measurement in Oncology Practice in the United States
J Natl Cancer Inst Monographs,
October 1, 2004;
2004(33):
155 - 167.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
W. F. Pirl
Evidence Report on the Occurrence, Assessment, and Treatment of Depression in Cancer Patients
J Natl Cancer Inst Monographs,
July 1, 2004;
2004(32):
32 - 39.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. M. Hudson, A. C. Mertens, Y. Yasui, W. Hobbie, H. Chen, J. G. Gurney, M. Yeazel, C. J. Recklitis, N. Marina, L. R. Robison, et al.
Health Status of Adult Long-term Survivors of Childhood Cancer: A Report From the Childhood Cancer Survivor Study
JAMA,
September 24, 2003;
290(12):
1583 - 1592.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. G Glinder and J. B Persons
Review: psychoeducational interventions reduce depressive symptoms in cancer
Evid. Based Ment. Health,
November 1, 2002;
5(4):
105 - 105.
[Full Text]
[PDF]
|
|
|
Get information about faster international access.
Privacy Policy
Copyright © 2001
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
