Validation and Application of the MD Anderson Symptom Inventory for Traditional Chinese Medicine (MDASI-TCM)

Transcription

1 J Natl Cancer Inst Monogr (2017) 2017(52): lgx010 doi: /jncimonographs/lgx010 Article ARTICLE Validation and Application of the MD Anderson Symptom Inventory for Traditional Chinese Medicine (MDASI-TCM) Zhandong Li, Qiuling Shi, Meng Liu, Liqun Jia, Bin He, Yufei Yang, Jie Liu, Hongsheng Lin, Huei-Kai Lin, Pingping Li, Xin Shelley Wang Affiliations of authors: Department of Integrative Medicine, Beijing Cancer Hospital, Peking University, Beijing, PR China (ZL, PL); Department of Symptom Research, The University of Texas MD Anderson Cancer Center, Houston, TX (QS, HKL, XSW); China-Japan Friendship Hospital, Beijing, PR China (ML, LJ); Xiyuan Hospital, China Academy of Chinese Medical Sciences, Beijing, PR China (BH, YY); Guang anmen Hospital, China Academy of Chinese Medical Sciences, Beijing, PR China (JL, HL) Correspondence to: Xin Shelley Wang, MD, MPH, Department of Symptom Research, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1450, Houston, TX ( or Pingping Li, MD, Department of Integrative Medicine, Beijing Cancer Hospital, Peking University, Beijing, PR China ( Abstract Background: The MD Anderson Symptom Inventory (MDASI) is a brief, yet thorough, patient-reported outcomes measure for assessing the severity of common cancer-related symptoms and their interference with daily functioning. We report the development of an MDASI version tailored for use with Traditional Chinese Medicine in China (the MDASI-TCM). Methods: Chinese-speaking patients with mixed cancer types (n ¼ 317) participated in the study. The development and validation process included four steps: 1) identify candidate TCM-specific items, with input from patients, oncologists, and TCM specialists; 2) eliminate candidate TCM items lacking relevance, based on patient report; 3) psychometrically examine the MDASI-TCM s validity and reliability in cancer patients receiving TCM-based care; and 4) cognitively debrief patients to assess the MDASI-TCM s relevance, understandability, and acceptability. Results: Seven TCM-specific symptom items (sweating, feeling cold, constipation, bitter taste, coughing, palpitations, and heat in palms/soles) were clinically and psychometrically meaningful to add to the core MDASI. Approximately 61% of patients had moderate to severe symptoms (rated 5 on the MDASI-TCM s 0 10 scale). Cronbach a coefficients were.90 for symptom-severity items and.93 for interference items, indicating internal consistency reliability. Known-group validity was substantiated by the MDASI-TCM s detection of differences in symptom severity according to performance status (P <.001) and interference levels by cancer stage (P <.05). Cognitive debriefing indicated that patients found the MDASI-TCM to be an understandable, easy-to-use tool. Conclusions: The Chinese MDASI-TCM is a valid, reliable, and concise measure of symptom severity and interference that can be used to assess Chinese cancer patients and survivors receiving TCM-based care. The physical, mental, and emotional symptoms produced by cancer and its treatment greatly affect a patient s ability to function (1 3). Symptom management is therefore a necessary part of cancer care during and after treatment. On a global basis, clinicians engaged in oncology clinical practice and research are increasingly incorporating integrative medicine techniques, such as those found in Traditional Chinese Medicine (TCM) for example, single or combined herbal medications, acupuncture, and other approaches. In China, TCM has long been an active part of cancer care, across various disease types and either alone or in combination with conventional cancer therapies. Given that symptom reduction can be an important therapeutic goal, patient-reported symptoms could be useful as an outcome measure for assessing patients with cancer who opt for Received: December 8, 2016; Revised: January 24, 2017; Accepted: August 11, 2017 The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oup.com. 48

2 Z. Li et al. 49 treatment with TCM. Monitoring and evaluating the effectiveness of treatment with TCM agents requires a validated assessment questionnaire with appropriate, relevant symptom items (4 7). Unlike widely adopted tumor outcome evaluation methods, the processes for establishing standardized patient-reported outcome (PRO) measures are generally novel for practitioners and researchers, and utilizing PRO measures in clinical trials (other than for pain assessment) can be unfamiliar and challenging for TCM practitioners (8,9). Although there is a great need to integrate the patient s perspective into routine TCMbased patient care and adverse event reporting, this work is yet to be done. The MD Anderson Symptom Inventory (MDASI) is a wellestablished patient-reported outcome (PRO) tool with satisfactory psychometric proprieties for assessing cancer-related symptom burden, defined as the combined severity and impact of symptoms on a patient s daily functioning (8), irrespective of specific cancer diagnosis or treatment type (1). The core MDASI measures 13 common cancer-related symptoms and six symptom-related interference items at their worst in the last 24 hours, rated on a 0 10 numeric scale. The six interference items reflect the impact of multiple symptoms on the patient s daily functioning. The core MDASI has been psychometrically and linguistically validated in Chinese (10). Assessing symptoms in specific patient populations can accomplished through the use of MDASI modules, which enhance the core MDASI with additional cancer type specific or treatment-specific symptom items (11,12). The number of additional module-specific items is kept small, so that the resulting instrument remains informative but concise. This approach ensures that the MDASI will be easy to use in clinical practice and research and not burdensome when repeated measurement is needed. Because every module includes the core items, module development continually validates the psychometric properties of the core MDASI. MDASI modules are available for various cancer types and treatments and in various languages. The primary objective of the current study was to develop and psychometrically validate a Chinese-language MDASI-TCM module. Our secondary objective was to use the MDASI-TCM to create symptom profiles (symptom prevalence and severity) of critical symptoms that affect daily functioning at a statistically significant level for patients and survivors under TCM care. Methods The multicenter study was approved by the Institutional Review Board of Beijing Cancer Hospital and Institute, Beijing, People s Republic of China. Patients were consecutively recruited from four Beijing institutions. Eligible patients were age 18 years or older, spoke Chinese, had a pathological diagnosis of cancer, were receiving TCM-based care, and could understand the study s intent. They were permitted to be undergoing active cancer treatment or to be under clinical observation. All enrolled patients gave informed consent to participate in the study. Trained study coordinators conducted interviews. Patients completed self-administered questionnaires, including the MDASI-TCM, the Medical Outcomes Survey 36-Item Short-Form Health Survey (SF-36), and a single-item quality-of-life question. Demographic data, current disease variables (including cancer site, cancer stage, presence of metastatic disease), laboratory results, treatment information, comorbidities, and Eastern Cooperative Oncology Group performance status (ECOG PS) were extracted from the medical record by the study coordinators. Development of the MDASI-TCM To develop the MDASI-TCM in accordance with US Food and Drug Administration guidance for establishing the content and validity of a PRO tool (13,14), we 1) derived a list of candidate MDASI-TCM symptom items based on a) expert input by Chinese TCM practitioners from multiple institutions, who suggested common symptoms identified by TCM theory, and b) patient input from a previous study in China that explored multiple TCM-related symptoms (15); we added these to the core MDASI to create a provisional MDASI-TCM; 2) administered the provisional MDASI-TCM to all patients and then refined the candidate TCM symptom list by examining each item for nonrelevance (low severity and low prevalence, defined as > 60% of responses ¼ 0 on the MDASI-TCM s 0 10 scale) and clinical interpretability (applicability for TCM practice, overlap with core items); 3) investigated the resulting MDASI-TCM s psychometric properties (validity, reliability); and 4) conducted cognitive debriefing interviews to characterize patient perception of the completeness and clarity of the final MDASI-TCM (13,16). Statistical Analysis The sample size needed for validating the MDASI-TCM was determined on the basis of the MDASI s ability to distinguish among patients with poor vs good performance status (clinician-rated ECOG PS) (1,17). Patients with good ECOG PS (0 1) would be expected to have lower symptom severity, compared with patients with poor ECOG PS (2). Given that at least 10 patients are needed per item, we aimed for a sample size that would allow detection of a half-standard-deviation difference in symptom severity between the two groups (18,19). We computed a global symptom component score (mean of the 13 core MDASI symptom items plus the TCM-specific candidate items), a score for each of these symptom components separately, and an interference component score (mean of the six interference items). The interference component score was further dichotomized into a subscore for work, activity, and walking (WAW; representing physical functioning) and a subscore for relations with other people, enjoyment of life, and mood (REM; representing mental health/social functioning) (20). Using component scores reduced the number of within-pair comparisons and type I errors. Means of differences, 95% confidence limits (CLs), and statistical significance (P <.05) were tested with independentsample t tests. Validity Convergent validity. We examined the MDASI-TCM s construct validity using hierarchical clustering analysis of the symptom items, visually represented in a dendrogram (21,22). Known-group validity. We examined the MDASI TCM s knowngroup validity by testing its ability to differentiate among patients according to ECOG PS, cancer stage, and current treatment status.

3 50 J Natl Cancer Inst Monogr, 2017, Vol. 2017, No. 52 Criterion validity. We examined the MDASI TCM s correspondence with a generic clinical outcome questionnaire, the SF-36. The SF-36 comprises eight subscales (including a physicalfunctioning subscale and a mental health subscale) that together characterize health-related quality of life; the subscales can be further grouped into a physical health composite score (PCS) and a mental health composite score (MCS) (23). Spearman rank correlation coefficients were calculated to compare the interference component scores or individual items from the MDASI-TCM with the MCS, PCS, and physical functioning and mental health subscales of the SF-36 and the single-item quality-of-life question. Reliability Internal consistency reliability. We examined the MDASI-TCM s internal consistency by calculating Cronbach a coefficients for the symptom severity and symptom interference subscales. Test-retest reliability. All patients completed the MDASI-TCM a second time one week after the initial assessment. Intraclass correlations were computed to demonstrate test-retest reliability. Symptom Profiles The severity and prevalence of the individual MDASI-TCM symptom and interference items were calculated using descriptive analyses. We categorized symptom severity as mild (1 4 on the MDASI-TCM s 0 10 scale), moderate (5 6), or severe (7 10) on the basis of previous research (24 26). We used linear regression analysis to examine symptom impact on daily functioning (the mean interference component score). Exploratory univariate analyses were conducted to identity candidate variables that were examined further in multivariable regression analyses. The multivariable analyses used forward and stepwise methods on clinical models and a symptom model; the appropriateness of the selected linear regression models was evaluated using residual diagnostics. We used one-way analysis of variance to determine possible predictors from among the categorical variables and correlation coefficients for the continuous, ordinal, and binary variables. All statistical procedures were performed using SPSS Statistical Software Program for Windows (27). All P values are two-tailed, with an a of.05. Results Patient Characteristics Table 1 presents patient demographic and disease-related characteristics by study site (n ¼ 317). Although 46% of patients had metastatic cancer, 78% had good performance status (ECOG PS ¼ 0 1) and 61% had no major comorbidities. Approximately 40% of patients were less than one year out from their cancer diagnosis, and 74% were being treated according to TCM. Patients were receiving TCM supportive care treatments for fatigue, gastrointestinal symptoms, sleep problems, and pain (46% to 18%) at time of the study. Validation of the MDASI-TCM Twelve candidate TCM-specific items (sweating, fidgeting/feeling irritable, feeling cold, increased enuresis, constipation, bitter taste, coughing, palpitations, coughing with sputum, heat in palms/soles, diarrhea, and mouth ulcers) were added to the already-validated MDASI core items to form the provisional MDASI-TCM. All 317 patients responded to the provisional MDASI-TCM, which contained all 19 MDASI core symptom and interference items, plus 12 TCM-specific candidate items). On the basis of participant responses and clinical relevance, we dropped five items for the following reasons: fidgeting/feeling irritable overlapped with the core item distress in its Chinese translation; increased enuresis was primarily relevant for elderly males; coughing with sputum was very highly correlated with coughing and therefore lacked independence; and diarrhea and mouth ulcers were rarely reported (rated 0 by > 60% of patients) (Table 2). The remaining seven TCM-specific symptom items were retained in the final MDASI-TCM (Figure 1) for psychometric analysis. Validity Convergent validity. Principal axis factoring analysis generated a two-factor structure for the MDASI core symptom severity items: gastrointestinal physical symptoms (nausea and vomiting) and general symptoms (remaining 11 core symptoms). Figure 2 presents the symptom clusters in the finalized MDASI- TCM with 20 symptom items: gastrointestinal (nausea, vomiting, poor appetite), neurological (numbness, bitter taste, drowsiness, dry mouth), pain-related (pain, constipation, heat in palms/soles, feeling cold), affective (distress, sadness), cardio/lung (shortness of breath, palpitations, coughing), and fatigue-related symptoms (fatigue, difficulty remembering, sleep disturbance, sweating). Known-group validity. The MDASI-TCM detected different severity levels based on performance status, cancer stage, and current treatment status (Table 3). Patients who had poor performance status (ECOG PS 2) reported statistically significant higher severity for all symptom and interference subscales (all P <.0001). Patients with late-stage cancer reported statistically signifcant poorer functioning (stage IV vs I III, P ¼.038). Patients undergoing cancer treatment reported more severe MDASI core symptoms (P <.01) and symptom interference (P <.05), but the differences between groups on ratings of TCMspecific items were not statistically significant. Criterion validity. The MDASI core symptoms were highly correlated with SF-36 fatigue (r ¼ 0.61, P <.0001) and the SF-36 physical component subscale (PCS; r ¼ 0.68, P <.0001) and mental component subscale (MCS; r ¼ 0.62, P <.0001). The TCM-specific symptoms were associated with SF-36 PCS ( 0.52, P <.0001) and MCS (r ¼ 0.50, P <.0001). The MDASI WAW interference subscale was highly correlated with the SF-36 physical functioning subscale (r ¼ 0.67, P <.0001) and PCS (r ¼ 0.68, P <.0001). The MDASI REM interference subscale was highly correlated with the SF-36 mental health subscale (r ¼ 0.62, P <.0001), PCS (r ¼ 0.65, P <.0001), and MCS (r ¼ 0.60, P <.0001). Total interference was associated with the single-item quality-of-life question (r ¼ 0.50, P <.0001). Reliability Internal consistency reliability. Cronbach a coefficients were.90 for the symptom severity scale (20 items: 13 core items plus seven module items) and 0.93 for the interference scale (six items), indicating a high level of internal consistency. As a further test,

4 Z. Li et al. 51 Table 1. Patient demographic and clinical characteristics by study site* Beijing Cancer Hospital (n ¼ 107) Xiyuan Hospital (n ¼ 73) China Japan Hospital (n ¼ 81) Guang anmen Hospital (n ¼ 57) Overall (n ¼ 317) Patient characteristic, mean (SD) Age, y (10.02) (11.12) (11.17) (16.04) (11.82) Education level, y (3.21) (3.61) (3.23) (3.72) (3.39) Patient characteristics, No. (%) Elderly (65 y) 21 (19.63) 14 (19.18) 10 (12.35) 20 (35.71) 65 (20.50) Female 64 (59.81) 38 (52.05) 50 (61.73) 18 (32.14) 170 (53.63) >1 y from diagnosis 76 (71.03) 35 (47.95) 44 (54.32) 35 (62.50) 190 (59.94) ECOG PS, No. (%) 0 13 (12.15) 10 (14.08) 34 (42.50) 35 (62.50) 92 (29.30) 1 74 (69.16) 34 (47.89) 36 (45.00) 11 (19.64) 155 (49.36) 2 16 (14.95) 24 (33.80) 7 (8.75) 5 (8.93) 52 (16.56) 3 2 (1.87) 3 (4.23) 2 (2.50) 5 (8.93) 12 (3.82) 4 2 (1.87) 0 1 (1.25) 0 3 (0.96) Cancer diagnosis, No. (%) Lung cancer 32 (29.91) 8 (10.96) 18 (22.22) 17 (30.36) 75 (23.66) Breast cancer 32 (29.91) 7 (9.59) 26 (32.10) 1 (1.79) 66 (20.82) Colorectal cancer 17 (15.89) 20 (27.40) 11 (13.58) 0 (0.00) 48 (15.14) Lymphoma 3 (2.80) 2 (2.74) 0 (0.00) 36 (64.29) 41 (12.93) Other gastrointestinal 12 (11.21) 7 (9.59) 9 (11.11) 0 (0.00) 28 (8.83) Others 11 (10.28) 29 (39.73) 17 (20.99) 2 (3.57) 59 (18.61) Cancer stage, No. (%) I 16 (15.09) 9 (14.29) 10 (13.33) 6 (13.95) 41 (14.29) II 16 (15.09) 4 (6.35) 18 (24.00) 8 (18.60) 46 (16.03) III 21 (19.81) 21 (33.33) 9 (12.00) 17 (39.53) 68 (23.69) IV 53 (50.00) 29 (46.03) 38 (50.67) 12 (27.91) 132 (45.99) Receiving TCM-based cancer care, No. (%) Yes 76 (71.03) 62 (84.93) 60 (74.07) 36 (64.29) 234 (73.82) Cancer treatment (chemotherapy, radiotherapy, surgery), No. (%) No 70 (65.42) 38 (52.05) 50 (61.73) 42 (75.00) 200 (63.09) Yes 37 (34.58) 35 (47.95) 31 (38.27) 14 (25.00) 117 (36.91) Receiving TCM-based symptom control, No. (%) Pain 11 (10.28) 23 (31.51) 13 (16.05) 11 (19.64) 58 (18.30) Fatigue 21 (19.63) 52 (71.23) 44 (55.70) 28 (50.00) 145 (46.03) Disturbed sleep 6 (5.61) 24 (34.29) 18 (23.08) 16 (28.57) 64 (20.58) Gastrointestinal symptoms 5 (4.67) 31 (44.29) 25 (30.86) 13 (23.21) 74 (23.57) Neurological symptoms 0 (0.00) 4 (5.88) 3 (3.95) 2 (3.57) 9 (2.93) Comorbidities, No. (%) No 72 (67.29) 51 (69.86) 57 (70.37) 14 (24.56) 194 (61.01) Yes 35 (32.71) 22 (30.14) 24 (29.63) 43 (75.44) 124 (38.99) Anemic (hemoglobin 12) 15 (14.02) 24 (32.88) 10 (12.35) 30 (53.57) 79 (24.92) Albumin (15.89) 26 (35.62) 17 (20.99) 34 (60.71) 94 (29.65) *ECOG PS ¼ Eastern Cooperative Oncology Group performance status; TCM ¼ Traditional Chinese Medicine. we deleted each single symptom item in turn and recalculated the Cronbach a coefficient. We found that the coefficients were consistently similar to the overall coefficient for that factor, which verified that each symptom contributed to the factor and should remain in its group. Test-retest reliability. The MDASI-TCM showed excellent testretest reliability, with intraclass correlation coefficients of.93 for core symptoms,.91 for TCM-specific symptoms, and.91 for interference. Cognitive Debriefing All 317 patients participated in cognitive debriefing of the MDASI-TCM. Most (87%) reported that the MDASI-TCM was understandable and easy to complete; 23% found the instrument to be burdensome or to have too many questions. Almost all (92%) were comfortable completing the MDASI-TCM and reported that the 0 10 rating scale was understandable and easy to use. Most patients, including the elderly, completed the MDASI-TCM in three to five minutes. Symptom Severity and Prevalence Table 2 also presents the comparison of means and standard deviations of MDASI-TCM items for the entire sample and provides an indication of floor and ceiling effects and the prevalence of moderate to severe symptoms (rated 5). Approximately 61% of patients rated at least one symptom as moderate to severe. Fatigue was the most severe symptom

5 52 J Natl Cancer Inst Monogr, 2017, Vol. 2017, No. 52 Table 2. Descriptive statistics for the MDASI-TCM, 0 10 scale, in rank order (n ¼ 317) % of patients rating item as: % missing Mean SD LCL UCL 0* 1 4 (mild) 5 6 (moderate) 7 10 (severe) Baseline Test-retest Core symptoms Fatigue Difficulty remembering Sleep Dry mouth Poor appetite Shortness of breath Distress Sadness Numbness Drowsiness Pain Nausea Vomiting Module items Sweating* Fidgeting (in Chinese, irritated ) Feeing cold* Increased enuresis Constipation* Bitter taste* Coughing* Palpitations* Coughing with sputum Heat in palms/soles* Diarrhea Mouth ulcers Interference items Work Enjoyment of life Walking Mood Activity Relations with others *Retained in the final MDASI-TCM. LCL ¼ lower 95% confidence limit; MDASI-TCM ¼ MD Anderson Symptom Inventory Traditional Chinese Medicine module; UCL ¼ upper 95% confidence limit. for patients undergoing active treatment and for cancer survivors, followed by difficulty remembering, disturbed sleep, dry mouth, and poor appetite. Sweating and feeling cold were the most severe TCM-specific symptoms. Coughing with sputum, heat in palms/soles, diarrhea, and mouth ulcers were the least severe of the TCM-specific symptoms and were rated 0 by more than 60% of patients. Increased enuresis, although rated as the fourth most severe TCM symptom in our study, is typically more prevalent in elderly men (age > 65 years). Missing data rates were low for the final 26-item MDASI- TCM: 2.8% (nine of 317) for the core symptoms; 0.3% (one of 317) for the seven TCM-specific items (a single patient missed the item sweating ); and 0.6% (two of 317) for the interference items. Symptom Impact on Daily Functioning In the multivariable linear regression model examining the impact of clinical variables and individual symptom items on the mean of the six interference items, poor ECOG PS was related to higher mean interference scores (P <.0001). Among the eight most severe symptoms, having moderate to severe fatigue (P ¼.002), sweating (P ¼.03), distress (P <.0001), and poor appetite (P ¼.001) was predictive of more severe interference. Discussion The MDASI-TCM was developed to measure the severity of multiple symptoms and their interference with functioning in cancer patients receiving TCM-based care in mainland China. Our results show that the MDASI-TCM has satisfactory validity and reliability, indicating that it is internally stable and sufficiently sensitive. That the MDASI-TCM was minimally burdensome for patients in the target population was indicated by the short administration time, negligible amount of missing data, and patient-reported ease of use. The MDASI-TCM module reflects the most critical TCMassociated symptom burden, either from disease or treatment experience, and its impact on daily functioning in cancer patients or survivors. Additional characteristics make the MDASI-TCM useful for evaluating treatment effectiveness in clinical practice and clinical trials and for estimating symptom

6 Z. Li et al. 53 Date: Participant Initials: Participant Number: Institution: Hospital Chart #: MD Anderson Symptom Inventory - Traditional Chinese Medicine (MDASI-TCM) Part I. How severe are your symptoms? Patients frequently have symptoms that are caused by their disease or by their treatment. We ask you to rate how severe the following symptoms have been in the last 24 hours. Please select a number from 0 (symptom has not been present) to 10 (the symptom was as bad as you can imagine it could be) for each item. 1. Your pain at its Not Present As Bad As You Can Imagine Your fatigue (tiredness) at its 3. Your nausea at its 4. Your disturbed sleep at its 5. Your feeling of being distressed (upset) at its 6. Your shortness of breath at its 7. Your problem with remembering things at its 8. Your problem with lack of appetite at its 9. Your feeling drowsy (sleepy) at its 10. Your having a dry mouth at its 11. Your feeling sad at its 12. Your vomiting at its 13. Your numbness or tingling at its Page 1 of 2 Copyright 2000 The University of Texas M. D. Anderson Cancer Center All rights reserved. MDASI-TCM - English 2016 Figure 1. The Traditional Chinese Medicine module of the MD Anderson Symptom Inventory (MDASI-TCM).

7 54 J Natl Cancer Inst Monogr, 2017, Vol. 2017, No. 52 Date: Participant Initials: Participant Number: Institution: Hospital Chart #: 14. Your problem with sweating at its Not Present As Bad As You Can Imagine Your problem with feeling cold at its 16. Your constipation at its 17. Your problem with bitter taste at its 18. Your coughing at its 19. Your problem with palpitation (racing heartbeat) at its 20. Your problem with heat palms at its Part II. How have your symptoms interfered with your life? Symptoms frequently interfere with how we feel and function. How much have your symptoms interfered with the following items in the last 24 hours? Please select a number from 0 (symptoms have not interfered) to 10 (symptoms interfered completely) for each item. Did Not Interfere Interfered Completely General activity? 22. Mood? 23. Work (including work around the house)? 24. Relations with other people? 25. Walking? 26. Enjoyment of life? Page 2 of 2 Copyright 2000 The University of Texas M. D. Anderson Cancer Center All rights reserved. MDASI-TCM - English 2016 Figure 1. Continued.

8 Z. Li et al. 55 Nausea Vomiting Lack of appetite Numbness Bitter taste Drowsiness Dry mouth Pain Heat in palms/soles Constipation Feeling cold Distress Sadness Shortness of breath Palpitations Coughing with sputum Disturbed sleep Difficulty remembering Fatigue Sweating Figure 2. Cluster analysis of core and Traditional Chinese Medicine module items. prevalence in epidemiological studies. The MDASI-TCM is a brief instrument, which renders it especially suited for use in longitudinal studies that require frequent symptom assessment (28,29). Despite its brevity, as a multisymptom assessment tool, the MDASI-TCM is also comprehensive: its 20 symptom items encompass those commonly experienced by patients with cancer, regardless of cancer site or stage. Because the MDASI-TCM concisely captures the symptom burden that is most important to patients and most relevant to disease and treatment (a feature that is often lacking with broader health-related quality-of-life assessment tools), it is optimized for use in integrative oncology clinical trials that measure multiple symptoms simultaneously. Its simple 0 10 numeric scale is familiar to respondents and is readily integrated into modern presentation modes (eg, the Internet, smartphones, web-based clinic portals). This flexibility allows patients to report symptom severity and interference in real time, whether they are in the clinic or at home. We also characterized the most severe symptoms reported by patients and survivors and examined symptom clustering to establish symptom profiles that could be useful in designing future clinical trials. Approximately 61% of patients reported one or more symptoms as being moderate to severe (rated 5 on the MDASI-TCM s 0 10 scale). Moderate to severe fatigue, sweating, poor appetite, and distress were predictive of a considerable burden that could severely affect daily functioning and limit treatment tolerability. As expected (1,28,30), fatigue was the most severe symptom for patients and survivors. Sweating was the second most severe symptom, suggesting that the TCM-specific module captures TCM theory driven symptoms that are meaningful to patients and TCM practitioners. Our study had limitations. First, we did not conduct qualitative patient interviews to identify candidate TCM-specific symptom items; we instead used a previous study that explored candidate symptom items in cancer patients under TCM care. Notwithstanding, we did solicit patient input on the comprehensiveness of the MDASI-TCM during cognitive debriefing interviews. Second, although defining moderate to severe as a rating of 5 or higher on a 0 10 scale has proven useful for indicating clinically significant pain and fatigue in patients with cancer, additional studies should be conducted to define cutpoints for moderate and severe levels of individual symptoms, including pain and fatigue, experienced by patients receiving TCM-based cancer care. Third, the MDASI-TCM was not evaluated and validated in Chinese patients without cancer or in other countries. Our results demonstrate that the MDASI-TCM is a useful instrument for measuring common symptoms in Chinesespeaking cancer patients and survivors. The TCM-specific items

9 56 J Natl Cancer Inst Monogr, 2017, Vol. 2017, No. 52 Table 3. Known-group validity: comparison of MDASI-TCM symptom and interference subscale scores and known factors* By performance status MDASI-TCM subscale ECOG PS No. of patients Mean SD Diff LCL UCL Effect size P Mean core þ TCM-specific symptoms Good <.0001 Poor Mean core symptoms Good <.0001 Poor Mean TCM-specific symptoms Good <.0001 Poor Mean interference Good <.0001 Poor WAW Good <.0001 Poor REM Good <.0001 Poor By cancer stage MDASI-TCM subscale Stage No. of patients Mean SD Diff LCL UCL Effect size P Mean core þ TCM-specific symptoms I III IV Mean core symptoms I III IV Mean TCM-specific symptoms I III IV Mean interference I III IV WAW I III IV REM I III IV By treatment status MDASI-TCM subscale Currently being treated No. of patients Mean SD Diff LCL UCL Effect size P Mean core þ TCM-specific symptoms Yes No Mean core symptoms Yes No Mean TCM-specific symptoms Yes No Mean interference Yes No WAW Yes No REM Yes No *ECOG PS ¼ Eastern Cooperative Oncology Group performance status; LCL, lower 95% confidence limit; MDASI-TCM ¼ MD Anderson Symptom Inventory Traditional Chinese Medicine module; REM ¼ composite of the MDASI-TCM items for interference with relations with other people, enjoyment of life, and mood (mental health/ social functioning); TCM ¼ Traditional Chinese Medicine; UCL ¼ upper 95% confidence limit; WAW ¼ composite of the MDASI-TCM items for interference with work, activity, and walking (physical functioning). are a quantitative measure of TCM-relevant symptom burden, but they do not stand alone in this; in contrast, they combine integrally with the 13 MDASI core items to form a complete, validated tool for assessing symptoms in TCM-based integrative care. The quantitative psychometric validation study and the qualitative cognitive debriefing of patients who had used the MDASI-TCM verify that the questionnaire represents the symptomatic experience of patients with cancer receiving TCMbased care (14,16). Using the MDASI-TCM, whether in practice or as an outcome measure in a clinical trial, could help TCM practitioners maximize treatment effectiveness and lessen patient suffering. Finally, the ability to derive quantitative data on the symptomatic benefit of TCM-based therapy, as provided by the MDASI-TCM, should greatly promote the introduction and communication of effective symptom control methods and goals between East and West. Funding This work was supported by an award to Pingping Li from Integrative Medicine Research, Beijing Chinese Medicine Bureau. Notes We gratefully acknowledge the editorial assistance of Jeanie F. Woodruff, BS, ELS. No competing financial interests exist.

10 Z. Li et al. 57 Author contributions Conceptualization: HKL, PL, ZL, QS, XSW; methodology: HKL, PL, ZL, QS, XSW; software: HKL, QS; validation: PL, ZL, QS, XSW; formal analysis: HKL, QS; investigation: BH, LJ, HL, JL, ML, PL, ZL, YY; resources: HKL, PL, ZL, XSW; data curation: BH, LJ, HKL, HL, JL, ML, PL, ZL, YY; writing original draft: BH, LJ, HKL, HL, JL, ML, PL, ZL, QS, XSW, YY; review/editing: PL, ZL, QS, XSW; visualization: PL, ZL, XSW; supervision: PL, ZL, XSW; project administration: PL, ZL, XSW; funding acquisition: PL, ZL, XSW. XSW and PL are responsible for the overall content as guarantors. References 1. Cleeland CS, Mendoza TR, Wang XS, et al. Assessing symptom distress in cancer patients: The M.D. Anderson Symptom Inventory. Cancer. 2000;89(7): Borden EC, Parkinson D. A perspective on the clinical effectiveness and tolerance of interferon-alpha. Semin Oncol. 1998;25(1 suppl 1): Cleeland CS. Cancer-related symptoms. Semin Radiat Oncol. 2000;10(3): Fink P, Toft T, Hansen MS, et al. Symptoms and syndromes of bodily distress: An exploratory study of 978 internal medical, neurological, and primary care patients. Psychosom Med. 2007;69(1): Kirkova J, Davis MP, Walsh D, et al. Cancer symptom assessment instruments: A systematic review. J Clin Oncol. 2006;24(9): de Haes JC, van Knippenberg FC, Neijt JP. Measuring psychological and physical distress in cancer patients: Structure and application of the Rotterdam Symptom Checklist. Br J Cancer. 1990;62(6): Portenoy RK, Thaler HT, Kornblith AB, et al. The Memorial Symptom Assessment Scale: An instrument for the evaluation of symptom prevalence, characteristics and distress. Eur J Cancer. 1994;30A(9): Cleeland CS. Symptom burden: Multiple symptoms and their impact as patient-reported outcomes. J Natl Cancer Inst Monogr. 2007;37: Dueck AC, Mendoza TR, Mitchell SA, et al. Validity and reliability of the US National Cancer Institute s Patient-Reported Outcomes Version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE). JAMA Oncol. 2015;1(8): Wang XS, Wang Y, Guo H, et al. Chinese version of the M. D. Anderson Symptom Inventory: Validation and application of symptom measurement in cancer patients. Cancer. 2004;101(8): Rosenthal DI, Mendoza TR, Chambers MS, et al. Measuring head and neck cancer symptom burden: The development and validation of the M. D. Anderson Symptom Inventory, head and neck module. Head Neck. 2007; 29(10): Fadol A, Mendoza T, Gning I, et al. Psychometric testing of the MDASI-HF: A symptom assessment instrument for patients with cancer and concurrent heart failure. J Card Fail. 2008;14(6): US Food and Drug Administration. Guidance for industry. Patient-reported outcome measures: Use in medical product development to support labeling claims. GuidanceComplianceRegulatoryInformation/Guidances/UCM pdf. Accessed December 5, Patrick DL, Burke LB, Powers JH, et al. Patient-reported outcomes to support medical product labeling claims: FDA perspective. Value Health. 2007;10(suppl 2):S125 S Zhao Y, Wang XS, Li PP. Investigation of common symptoms of cancer and reliability analysis. Chin J Integr Med. 2007;13(3): Turner RR, Quittner AL, Parasuraman BM, et al. Patient-reported outcomes: Instrument development and selection issues. Value Health. 2007;10(suppl 2): S86 S Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5(6): Sloan J, Symonds T, Vargas-Chanes D, et al. Practical guidelines for assessing the clinical significance of health-related quality of life changes within clinical trials. Drug Inf J. 2003;37(1): Sloan JA, Vargas-Chanes D, Kamath CC, et al. Detecting worms, ducks and elephants: A simple approach for defining clinically relevant effects in quality-of-life measures. J Cancer Integr Med. 2003;1: Cleeland CS, Ryan KM. Pain assessment: Global use of the Brief Pain Inventory. Ann Acad Med Singapore. 1994;23(2): Gorsuch RL. Factor Analysis. 2nd ed. Hillsdale, NJ: L. Erlbaum Associates; Mulaik SA. The Foundations of Factor Analysis. New York: McGraw-Hill; Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF- 36). I. Conceptual framework and item selection. Med Care. 1992;30(6): Serlin RC, Mendoza TR, Nakamura Y, et al. When is cancer pain mild, moderate or severe? Grading pain severity by its interference with function. Pain. 1995;61(2): Mendoza TR, Anderson KO, Wang XS, et al. Phone it in or fill it in? Comparing symptom assessment using the interactive voice response system with the traditional paper and pencil method. Poster presented at: International Society for Quality of Life Research 8th Annual Conference; November 2001; Amsterdam, the Netherlands. 26. Wang XS, Zhao F, Fisch MJ, et al. Prevalence and characteristics of moderate to severe fatigue: A multicenter study in cancer patients and survivors. Cancer. 2014;120(3): Statistical Package for the Social Sciences (computer program). Version Chicago: SPSS, Inc.; Wang XS, Fairclough DL, Liao Z, et al. Longitudinal study of the relationship between chemoradiation therapy for non-small-cell lung cancer and patient symptoms. J Clin Oncol. 2006;24(27): Gabrilove JL, Perez EA, Tomita DK, et al. Assessing symptom burden using the M. D. Anderson symptom inventory in patients with chemotherapyinduced anemia: Results of a multicenter, open-label study (SURPASS) of patients treated with darbepoetin-alpha at a dose of 200 mg every 2weeks. Cancer. 2007;110(7): Wang XS, Cleeland CS, Mendoza TR, et al. Impact of cultural and linguistic factors on symptom reporting by patients with cancer. J Natl Cancer Inst. 2010; 102(10):