Clinical Utility of the MDASI-BT in Patients with Brain Metastases

Transcription

1 Vol. 37 No. 3 March 2009 Journal of Pain and Symptom Management 331 Original Article Clinical Utility of the MDASI-BT in Patients with Brain Metastases Terri S. Armstrong, PhD, Ibrahima Gning, DrPH, Tito R. Mendoza, PhD, Jeffrey S. Weinberg, MD, Mark R. Gilbert, MD, Melissa L. Tortorice, BS, and Charles S. Cleeland, PhD Department of Integrative Nursing Care (T.S.A.), The University of Texas Health Science Center School of Nursing; and the Departments of Symptom Research (I.G., T.R.M., C.S.C.), Neuro-Oncology (T.S.A., M.R.G., M.L.T.), and Neuro-Surgery ( J.S.W.), The University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA Abstract Symptom occurrence has been shown to predict treatment course and survival in cancer patients. The M. D. Anderson Symptom Inventory-Brain Tumor Module (MDASI-BT) was recently validated as a tool for primary brain tumor patient self-report of symptoms. This study evaluated the reliability and validity of the MDASI-BT in patients with brain metastases. Data collection included demographic and clinical factors, and the MDASI-BT (0e10 scale). Construct validity was assessed using confirmatory factor analysis, and known-group validity was evaluated by detecting group differences due to disease severity and treatment approach. For reliability, Cronbach s alpha values were computed for each subscale. A sample of 124 patients participated, of which 53.2% were women. Participants were primarily white (79.8%) and married (78.2%), and a variety of solid tumor malignancies were represented. Factor analysis revealed six underlying constructs, including affective symptoms, cognitive dysfunction, focal neurologic deficits, constitutional and gastrointestinal symptoms, and interference with life. The solution with these factors explained 68.4% of the variance. Mean symptom scores were 1.2 and 2.6, and mean interference scores were 1.8 and 4.3 for patients with good and poor Karnofsky scores, respectively (P < 0.001). These subscales were also sensitive to opioid analgesic use, with group differences of 1.5 and 2.2 (P < 0.001). Cronbach s alpha was 0.9 for each of the two subscales. Fatigue, sleep disturbance, drowsiness, distress, and dry mouth were the most severe symptoms. The MDASI-BT demonstrated validity and reliability in brain metastases patients and can be used to identify and monitor symptom occurrence in relation to treatment course and survival. J Pain Symptom Manage 2009;37:331e340. Ó 2009 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. This work was supported in part by the Oncology Nursing Society. Address correspondence to: Terri S. Armstrong, PhD, Department of Integrative Nursing Care, The University of Texas Health Science Center School of Ó 2009 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved. Nursing, 6901 Bertner Avenue, Room 791, Houston, TX 77030, USA. Terri.S.Armstrong@ uth.tmc.edu Accepted for publication: February 19, /09/$esee front matter doi: /j.jpainsymman

2 332 Armstrong et al. Vol. 37 No. 3 March 2009 Key Words Symptoms, assessment, validation, brain, central nervous system, neurologic, carcinoma, neoplasm Introduction Metastatic brain tumors are those that originate in other parts of the body and spread to the brain. It is estimated that 20%e40% of patients with cancer will develop brain metastases. 1,2 An estimated annual incidence of 98,000e170,000 persons per year is diagnosed with a metastatic brain tumor. 3 Patients diagnosed with these tumors frequently suffer from neurologic symptoms, such as seizures, poor cognition, and muscle weakness in addition to the symptoms associated with their systemic disease. 4,5 The occurrence of symptoms has been shown to impact a variety of health outcomes in patients with systemic cancer, including adjustment to illness, functional health status, disease progression, and survival. Symptoms are now recognized to occur in clusters and may be multiplicative in nature, supporting the concurrent measurement of multiple symptoms. 6,7 The use of instruments designed to allow patients to self-report the occurrence and severity of multiple symptoms has been shown to significantly and independently predict changes in patient functioning, treatment failures, and poor therapeutic outcomes. 8,9 Patients with tumors involving the central nervous system (CNS) are a unique group of cancer patients due to the neurologic nature of symptoms that occur. The impact of the tumor on neurologic function is determined by the number and location of metastases within the brain, as well as the associated impact on intracranial pressure. 5 Symptoms affect patients functional abilities, social interactions, and emotional well-being, and thus their quality of life. In addition to the neurologic symptoms associated with the lesion in the brain, patients who have metastatic brain tumors often experience symptoms related to the primary site of cancer and other metastatic sites. 4 In general, symptoms in patients with brain metastases include signs of increased intracranial pressure or focal irritation. The most common presenting symptom is headache. 10e12 Focal symptoms are dependent on tumor location and are the same as any lesion occurring in a particular area of the brain. In patients with supratentorial tumors, cognitive deficits have been reported to occur frequently. These can include problems with memory (particularly short-term memory), abstract reasoning, and comprehension. Other common symptoms include hemiparesis, or weakness on one side of the body, aphasia, and hemisensory loss. 10,12 Ataxia, manifest as uncoordinated movements, is the most common symptom if the tumor is in the posterior fossa, a result of impairment of cerebellar function or cerebellar pathways. The M. D. Anderson Symptom Inventory Several instruments have been developed to allow patients to self-report symptoms. One such instrument, the M. D. Anderson Symptom Inventory (MDASI), also known as MDA- SI-Core, uses a 0e10 numerical rating scale to rate 19 items measuring patients health status within the last 24 hours prior to evaluation. 13 It includes 13 symptoms commonly reported by cancer patients, and six items that measure how much symptoms interfered with patient daily life (interference items). The tool has established reliability (Cronbach s alpha 0.85e0.94) and validity in persons with cancer. This 11-point scale (0e10) can be used to assess symptom severity both in the clinic as well as be administered remotely through an Interactive Voice Response system. Each of the 13 symptoms is rated from 0 to 10 to indicate the presence and severity of symptoms, with 0 being not present and 10 being as bad as you can imagine. The six interference items include general activity, mood, work (includes both work outside the home and housework), relations with other people, walking, and enjoyment of life. They also are measured on a 0e10 scale, with 0 being did not interfere and 10 being interfered completely. The mean of the 13 core symptoms constitutes the summary measure for

3 Vol. 37 No. 3 March 2009 MDASI-BT and Metastases 333 the symptom subscale, while the average of the interference items is used as a measure of overall symptom burden. The M. D. Anderson Symptom Inventory- Brain Tumor Module (MDASI-BT) was recently developed and validated in patients with primary tumors involving the CNS. This instrument was built on the MDASI-Core, and developed to allow patients to self-report symptoms related specifically to the tumor in the brain and the symptoms associated with therapy. In addition to the 13 symptoms and six interference items of the MDASI-Core, the MDASI-BT includes nine symptoms common to patients with brain tumors. This module includes a total of 22 symptoms; all have demonstrated good reliability and validity for use in patients with primary brain tumors. 14,15 Given the high incidence of brain metastases in patients with systemic cancer, we performed this study to validate the MDASI-BT for use in this patient population. Additionally, this study sought to determine the severity and prevalence of symptoms and to evaluate the psychometric properties of the MDASI-BT in patients with metastatic brain tumors. Methods Sample Institutional Review Board approval was obtained prior to the initiation of the study. Patients were considered eligible for the study if they were 18 years of age or older, had the diagnosis of a metastatic brain tumor, and were able to speak and read English. Patients with metastases from a variety of cancers were purposively included in this study to assure broad applicability to the brain metastases population. They were excluded if they presented evidence of an expressive or receptive aphasia, or gross cognitive dysfunction limiting memory or ability to complete a self-report questionnaire. This was evaluated by a review of the medical record, and consultation with the clinician of record before approaching the patient about participation. Patients presenting to the outpatient clinic or inpatient unit were screened for inclusion in this study. A total sample of 120 patients participated in this study. The principal investigator reviewed patient characteristics weekly to ensure a representative sample was recruited. Instruments Three tools were used in this study, including the MDASI-BT, a demographic data tool and a Clinician Checklist. The MDASI-BT was described previously. The clinical assessment tool included information on disease status (whether newly diagnosed or recurrent disease); tumor type; tumor location and number; concurrent medications (i.e., steroids, anticonvulsants, opioid analgesics, antidepressants, anticoagulants); performance status; and treatment status (prior and current therapies). This tool was designed to collect disease and treatment-related data that may impact the symptoms experienced. The demographic data tool was used to describe the sample and to determine characteristics that might influence patient symptom experience. Factors measured include gender, race, age, level of education, marital status, religious background, employment status, and income. This form was adapted from a similar tool used during the initial validation of the MDASI. 13 Each participant was asked to complete the demographic information sheet and the MDA- SI-BT only once, to minimize issues related to attrition and subject burden. In the initial validation study of the MDASI, as well as with the initial validation of the MDASI-BT, patients were reported to be able to complete the instrument in less than 10 minutes. 13,15 The Clinician Checklist was completed by the principal investigator and the clinical team caring for the patient. Statistical Analysis Data were analyzed using SPSS version All statistical tests were two-tailed, conducted at alpha level of 0.05, with adjustment for Type I error. Effect sizes were computed where applicable to assess minimally important differences using standardized mean difference method. Differences resulting in at least half a standard deviation unit were deemed minimally important. Descriptive statistics were computed for demographic and disease characteristics. Psychometric validation of the instrument was performed using construct and known-group validities, as well as reliability analyses.

4 334 Armstrong et al. Vol. 37 No. 3 March 2009 Table 1 Demographic and Disease Characteristics of the Patient Sample (n ¼ 124) Statistics (n ¼ 124) Patient Characteristic n % Age Mean (SD) 56 (11) Median (range) 57 (30e78) Below 60 years years and older Gender Female Male Education Grade 12 and below Grade 13 and higher Ethnicity White non-hispanic Minority Patient type Inpatient Outpatient KPS, Range, 30e100 Poor KPS (30e80) Good KPS (90e100) Primary cancer Breast Genitourinary Lung Melanoma Other (gastrointestinal, gynecological, thyroid) Number of brain metastases Single Multiple Other metastasis sites Bone, spine Bone, other Lymph nodes Soft tissue Viscera Other Area of brain involved Left Right More than one area Supratentorial Infratentorial Treatment type Radiation therapy Radiosurgery Chemotherapy Surgery Bone marrow transplant Hormonal therapy Biotherapy Treatment in last month Radiation therapy Radiosurgery Chemotherapy Surgery (Continued) Table 1 Continued Statistics (n ¼ 124) Patient Characteristic n % BMT Hormonal therapy Biotherapy Opioid analgesic No Yes Descriptive Statistics. Frequency and percents for the demographic and disease characteristics were reported. Symptom occurrence and severity were ascertained by computing means, standard deviation, range, and confidence interval of the mean, for each item on the instrument. Mean symptom severity scores were computed by averaging the ratings of all 22 symptoms, the 13 core symptoms, and the 9 brain tumor symptoms, into three different subscale summary scores. Mean interference subscale was obtained from an average of the six interference items. In addition, for the interference subscale, an average of activity, walking, and work produces a physical component, whereas the average of relations, mood, and enjoy constitutes the affective component. 17 In addition the proportion of patients with at least a moderate score (5e10), as well as severe scores (7e10), was generated. Previous studies have shown that at least for pain and fatigue, patient s functioning is significantly impaired at this level. 18 Construct Validity. Validity refers to the ability of the instrument to measure the phenomena it is supposed to be measuring. 19 Principal axis factoring with oblimin rotation was used to assess construct validity of the MDASI-BT in patients with metastatic brain tumor. Confirmatory factor analysis was performed to obtain a solution identifying constructs in the data, as previously found in the initial validation of the instrument. A factor solution that is clear, interpretable, and makes clinical sense was adopted. Model fit test was conducted according to Harman s criteria [n ( 1/2) ]. 20 Cluster analysis was conducted, attempting to identify clusters among the 22 symptoms. Ward s method was

5 Vol. 37 No. 3 March 2009 MDASI-BT and Metastases 335 Table 2 Symptom Severity and Prevalence (n ¼ 124) MDASI-BT Mean SD Range LCL UCL % $ 5 a % $ 7 b Symptoms (ranked) Fatigue e Sleep disturbance e Drowsiness e Distress e Dry mouth e Pain e Lack of appetite e Shortness breath e Irritability e Numbness e Sadness e Difficulty remembering e Change in bowel pattern e Change in vision e Weakness e Change in appearance e Nausea e Difficulty speaking e Difficulty understanding e Difficulty concentrating e Vomiting e Seizures e Interference Items General activity e Mood e Work including housework e Relations with other people e Walking e Enjoyment of life e Subscale Scores Mean severity (22 items) e Mean core (13 items) e Mean brain tumor (9 items) e Mean interference (6 items) e WAW (walk-activity-work) e REM (relate-enjoy-mood) e Factor Scores Factor 1dconstitutional e Factor 2dcognitive e Factor 3dinterference e Factor 4dfocal neurologic deficit e Factor 5dgastrointestinal e Factor 6daffective e LCL ¼ lower 95% confidence limits; UCL ¼ upper 95% confidence limits. a Percent moderate to severe. b Percent severe. used to generate clustering relationships and display dendrograms. 21 Known-Group Validity. Known-group validity was assessed by looking for differences in scores based on disease severity rated using Karnofsky Performance Status (KPS). Patients were divided into two groups based on their KPS score. The group with good performance was defined by scores of 90e100, whereas scores of 80 and below represented the group with poor performance. Independent sample t-test was used to conduct group comparisons. It was anticipated that patients with a lower KPS rating would experience a higher symptom burden. Groups were compared using mean severity, mean core, and mean brain tumor symptom subscales, as well as mean interference subscale. Internal Consistency (Reliability). For reliability, Cronbach s alpha coefficients were computed for the core, brain tumor, and interference

6 336 Armstrong et al. Vol. 37 No. 3 March 2009 Table 3 Factor Analysis Showing Six Constructs of the MDASI-BT and Model Fit (n ¼ 124) Factor Symptoms Fatigue Drowsiness Dry mouth Pain Irritability Shortness of breath Sleep disturbance Change in bowel pattern Difficulty understanding L Difficulty speaking L Change in vision L Difficulty remembering L Difficulty concentrating L Change in appearance L Work including housework L General activity L Mood L Enjoyment of life L Relations with other people L Walking L Weakness Numbness Seizures Nausea Vomiting Appetite Distress L0.540 Sad L0.410 Model fit test (Harman a ) n SD n ( 1/2) Residuals Sample SD ¼ standard deviation. a Harman s criteria: Results are adequate if the SD of the residuals is slightly less than or approximately equal to the reciprocal of the square root of the sample size (Harman, 1976). Boldfaced numbers indicate inclusion in this factor. subscales. In addition, alpha values were computed for each individual item. Cronbach s alpha values of 0.7 were considered adequate. 22 Results A sample of 124 patients participated in this study, of whom 53.2% were women. There were six patients who refused participation in this project, and six patients were found ineligible after initially being approached. Participants were primarily White (79.8%), and married (78.2%), with a variety of solid tumor malignancies represented. Table 1 displays the demographic and clinical characteristics of the sample. Fatigue, sleep disturbance, drowsiness, distress, and dry mouth were rated the most severe symptoms. Forty-one percent of the sample reported fatigue severity of $5, and over 20% reported sleep disturbance, drowsiness, and distress of at least 7 on a 0e10 scale. Nearly all symptoms had variable reporting of the full range of the scale from 0 to 10, except seizures and difficulty remembering (0e9), and difficulty understanding, speaking, and concentrating (0e7). Descriptive statistics for the items and subscales are shown in Table 2. Factor analysis revealed six underlying constructs, including affective symptoms, cognitive dysfunction, focal neurologic deficits, constitutional symptoms, gastrointestinal symptoms, and interference with daily life. The affective construct consisted of distress and sadness. The cognitive construct consisted of difficulty understanding, difficulty speaking, change in vision, difficulty remembering, difficulty concentrating, and change in appearance. The focal neurologic deficit construct consisted of weakness, numbness, and seizures. The

7 Vol. 37 No. 3 March 2009 MDASI-BT and Metastases 337 Fig. 1. Cluster analysis of the MDASI-BT symptoms with dendrogram (n ¼ 124). Distances between symptom scores as rated by patients were calculated using squared Euclidian distances to form clusters of symptoms. Graphical displays with dendrograms showing cohesion between clusters were done using Ward s method. Shaded areas represent the commonly seen patterns of gastrointestinal and affective clusters as described in other symptom validation studies. constitutional construct consisted of fatigue, drowsiness, dry mouth, pain, irritability, shortness of breath, and change in bowel pattern. The gastrointestinal construct consisted of nausea and vomiting. Lastly, the interference with life construct consisted of the interference items, general activity, mood, work, walking, relations with other people, and enjoyment of life. The solution with these factors explained 68.4% of the variance and satisfied Harman s criteria for model fit. Table 3 shows a representation of the different constructs found during factor analysis, whereas Fig. 1 depicts the symptom clustering. Known-group validity was established for the MDASI-BT using KPS. Mean symptom scores were 1.2 and 2.6, and mean symptom interference was 1.8 and 4.3 for patients with good (90e100) and poor (80 and below) KPS, respectively. Fig. 2 represents a graphical visualization of the individual symptoms by KPS performance status. These scales also were sensitive to opioid analgesic use, with group differences of 1.5 and 2.2 (P < 0.001), and whether the patient was an inpatient or outpatient, with group differences of 1.24 and 3.04 (P < 0.001), respectively. The results of the group comparisons are indicated in Table 4. Reliability of the instrument as a whole and for the subscales and constructs was evaluated by calculation of Cronbach s alpha, with eigenvalues >1. Cronbach s alpha values for all scales were between 0.74 and 0.94 except the

8 338 Armstrong et al. Vol. 37 No. 3 March neurologic construct, which was This is most likely a reflection of the low reporting of seizures in the group as a whole. Results of the reliability analysis for the MDASI-BT subscales and factors are displayed with corresponding alpha values (Table 5). Discussion The MDASI-BT was initially developed and validated in patients with primary brain tumors. Patients with brain metastases have neurologic symptoms in addition to those symptoms associated with the primary cancer and treatment. This study evaluated the utility Poor KPS Pain Fatigue Nausea Sleep Distress Shortbreath Remember Appetite Drowsy Drymouth Sad Vomiting Numbness Weakness Difficulty Understanding Difficulty Speaking Seizures Difficulty Concentrating Good KPS Change in Vision Change in Appearance Change in Bowel Pattern Fig. 2. Symptom severity by good (90e100) versus poor (30e80) KPS (n ¼ 124). Note: Mean differences are not significant for nausea, vomiting, seizures, and change in bowel patterns. Table 4 Group Mean Comparison by KPS (n ¼ 124) Irritability of the MDASI-BT for use in this patient population, and demonstrated good reliability, and construct and known-group validities. In addition, the instrument demonstrated sensitivity to the use of opioid analgesics, and inpatient versus outpatient status. The most severe symptoms were constitutional, such as fatigue and difficulty sleeping. In the primary brain tumor patient population, the most severe symptoms were the same. However, all 22 symptoms on the instrument had variable reporting in both study populations. For the current study, between 5% and 20% of patients reported symptom severity of greater than 5 for the neurologic symptoms that were included in the instrument. Because patients MDASI-A Karnofsky Groups n Mean SD Difference P-value Severity Subscale Good (90e100) <0.001 Poor (30e80) Core Subscale Good (90e100) <0.001 Poor (30e80) Brain Tumor Subscale Good (90e100) <0.001 Poor (30e80) Interference Subscale Good (90e100) <0.001 Poor (30e80) WAW Good (90e100) <0.001 Poor (30e80) REM Good (90e100) <0.001 Poor (30e80) DIFF ¼ mean difference; WAW ¼ mean of walking-activity-work; REM ¼ mean of relations-enjoy-mood.

9 Vol. 37 No. 3 March 2009 MDASI-BT and Metastases 339 Items Table 5 Reliability AnalysisdCronbach s Alpha for the Subscales and Factors (n ¼ 124) Cronbach s n (Items) Alpha Subscales MDASI-BT (28 items) Core symptom items Brain tumor symptom items Symptom interference items Factors Factor 1dgeneral symptoms Factor 2dcognitive Factor 3dinterference Factor 4dneurologic Factor 5dgastrointestinal Factor 6daffective were evaluated at one time point, it will be important to conduct longitudinal studies to evaluate the trajectory of the symptoms over time and in relation to disease and treatment status in individual patients. We currently assume that certain symptoms, such as fatigue and nausea, may be worse during therapy, whereas other symptoms, such as unilateral weakness or seizures, may be worse as the tumor progresses in patients with both metastatic and primary tumors. Once these symptom trajectories and clusters have been identified in a standardized manner using validated instruments such as the MDASI-BT, interventions to reduce symptom severity and prevalence can then be designed and evaluated. Several symptoms included in this instrument were not reported as severe for a majority of patients. As a result, their significance for this patient population may be questioned. However, the mean of the core and brain tumor items as well as all factor constructs, except the gastrointestinal factor, correlated with worse performance status and patient status (inpatient versus outpatient). This was also true in the primary tumor population. In addition, all constructs and mean scores correlated with opioid analgesic use in the metastatic tumor population, underscoring the association of pain with the occurrence of other symptoms in patients with cancer. This highlights the importance of the concordance of symptoms and the importance of evaluating multiple symptoms for impact on patient status, and not focusing on a single symptom item. Seizures were not frequently reported in this sample, and severity scores alone might indicate this item as a candidate for removal from the instrument. However, the clinical significance of seizures in patients with brain metastases warrants its inclusion in the instrument. Seizures are estimated to occur in 30% of patients with brain metastases at some point in the disease trajectory. As these patients were evaluated at one point in time with a recall period set to the last 24 hours, the full impact of seizures in this patient population may not be apparent in this initial utility study. Further use of the instrument and evaluation of the incidence and severity of seizures is needed and should be closely monitored in further use of the instrument. This study provided initial support for the utility of this instrument in the evaluation of patients with brain metastases from a variety of cancers. This broad applicability is important, as most clinical studies of therapeutic approaches often include patients with metastases from a variety of cancers. It is an easy-to-complete instrument that can be used in clinical care and evaluation of therapeutic approaches. The use of such an instrument is especially important to measure the full impact of therapies on the status of patients. Further validation of the instrument, including repeated measures over time to further evaluate instrument sensitivity to treatment status, and evaluation of the reliability of caregiver rating will need to be completed to broaden the validity and use of the instrument in the future. Acknowledgments This study would not have been possible without the collaboration of patients and their families and caregivers, as well as the health care personnel who cared for them. The authors would like to thank the Oncology Nursing Society for their support of this project. References 1. Cairncross JG, Kim JH, Posner JB. Radiation therapy for brain metastases. Ann Neurol 1980; 7(6):529e541. Available from nih.gov/entrez/query.fcgi?cmd¼retrieve&;db¼pub Med&dopt¼Citation&list_uids¼ Accessed July 10, Posner JB. Neurologic complications of cancer. Philadelphia, PA: Davis, 1995.

10 340 Armstrong et al. Vol. 37 No. 3 March Ries LAG, Melbert D, Krapcho M, et alin: SEER Cancer Statistics Review, 1975e2004, Vol Bethesda, MD: National Cancer Institute, Armstrong TS, Gilbert MR. Metastatic brain tumors: diagnosis, treatment, and nursing interventions. Clin J Oncol Nurs 2000;4(5):217e225. Available from trez/query.fcgi?cmd¼retrieve&;db¼pubmed& dopt¼citation&list_uids¼ Accessed July 10, Rabbitt JE, Page MS. Selected complications in neuro-oncology patients. Semin Oncol Nurs 1998; 14(1):53e60. Available from nih.gov/entrez/query.fcgi?cmd¼retrieve&;db¼ PubMed&dopt¼Citation&list_uids¼ Accessed July 10, Dodd M, Janson S, Facione N, et al. Advancing the science of symptom management. J Adv Nurs 2001;33(5):668e Lenz ER, Pugh LC, Milligan RA, et al. The middle-range theory of unpleasant symptoms: an update. Adv Nurs Sci 1997;19(3):14e Clark JA, Talcott JA. Symptom indexes to assess outcomes of treatment for early prostate cancer. Med Care 2001;39(10):1118e Sarna L. Effectiveness of structured nursing assessment of symptom distress in advanced lung cancer. Oncol Nurs Forum 1998;25(6):1041e Wright DC, Delaney TF, Buckner JC. Treatment of metastatic cancer. In: Devita VT, Hellman S, Rosenberg SA, eds. Cancer: Principles and practice of oncology, 4th ed. Philadelphia, PA: Lippincott, 1993: 2170e Hoskin PJ, Crow J, Ford HT. The influence of extent and local management of the outcome of radiotherapy for brain metastases. Int J Radiat Oncol Biol Phys 1990;19:111e Posner JB. Management of brain metastases. Rev Neurol (Paris) 1992;148:477e Cleeland CS, Mendoza TR, Wang XS, et al. Assessing symptom distress in cancer patients: the M. D. Anderson Symptom Inventory. Cancer 2000; 89(7):1634e Armstrong TS, Cohen MZ, Eriksen L, Cleeland C. Content validity of self-report measurement instruments: an illustration from the development of the Brain Tumor Module of the M.D. Anderson Symptom Inventory. Oncol Nurs Forum 2005; 32(3):669e676. Available from nlm.nih.gov/entrez/query.fcgi?cmd¼retrieve&; db¼pubmed&dopt¼citation&list_uids¼ Accessed July 10, Armstrong TS, Mendoza T, Gning I, et al. Validation of the M.D. Anderson Symptom Inventory Brain Tumor Module (MDASI-BT). J Neurooncol 2006; 80(1):27e35. Available from nih.gov/entrez/query.fcgi?cmd¼retrieve&;db¼pub Med&dopt¼Citation&list_uids¼ Accessed July 10, SPSS. Statistical package for the social sciences. Chicago, IL: SPSS, Inc., Cleeland CS, Nakamura Y, Mendoza TR, et al. Dimensions of the impact of cancer pain in a four country sample: new information from multidimensional scaling. Pain 1996;67(2e3):267e273. Available from query.fcgi?cmd¼retrieve&;db¼pubmed&dopt¼ Citation&list_uids¼ Accessed July 10, Mendoza TR, Wang XS, Cleeland CS, et al. The rapid assessment of fatigue severity in cancer patients: use of the Brief Fatigue Inventory. Cancer 1999;85(5):1186e Lynn MR. Determination and quantification of content validity. Nurse Res 1985;35(6):382e Harman HH. Modern factor analysis. Chicago, IL: University of Chicago Press, Ward JH Jr. Hierarchical grouping to optimize an objective function. J Am Stat Assoc 1963; 58(301):236e Nunnally JC, Bernstein IH. Psychometric theory, 3rd ed. New York: McGraw-Hill, 1994.