(POSTER PRESENTATION) THE ASSOCIATION BETWEEN AGE AND GENDER AND THE FINANCIAL BURDEN IN CANCER PATIENTS: A POOLED ANALYSIS OF CLOSED EORTC CANCER TRIALS. 
Chantal Quinten, John Maringwa, Francesca Martinelli, Corneel Coens, Quality of Life, EORTC, Brussels, Belgium, Charles S. Cleeland, Symptom Research, University of Texas, Houston, Texas, Henning Flechtner, Child and Adolescent Psychiatry and Psychotherapy, University Magdeburg, Magdeburg, Germany, Carolyn Gotay, School of Population and Public Health, University of British Columbia, Vancouver, British Columbia, Canada, Eva Greimel, Obstetrics and Gynecology, Medical University Graz, Graz, Austria, Madeleine King, Psycho-oncology Cooperative Research Group, University of Sydney, Sydney, Australia, David Osoba, Quality of Life, Quality of Life Consulting, West Vancouver, British Columbia, Canada, Martin J. Taphoorn, Neurology, VU Medical Center/Medical Center Haaglanden, Amsterdam/The Hague, Netherlands, Bryce B. Reeve, Cancer Control and Population Science, National Cancer Institute, Bethesda, Maryland, Jolie Ringash, The Princess Margaret Hospital, University of Toronto, Toronto, Canada, Joseph Schmucker-Von Koch, Medical Ethics, University of Regensburg, Regensburg, Germany, Joachim Weis, Psychooncology, University of Freiburg, Freiburg, Germany, Andrew Bottomley, Quality of Life, EORTC, Brussels, Belgium 
 
 AIMS: The objective of this analysis was to investigate the influence of age and gender on the financial burden of cancer patients, as evaluated by the European Organisation for Research and Treatment (EORTC) QLQ-C30. METHODS:  In 30 EORTC Randomized Controlled Trials patients completed the EORTC QLQ-C30 at baseline. This includes an item about financial difficulties caused by physical condition or medical treatment, scored from 1 (not at all) to 4 (very much). Clinical data including age (<=60 vs. >60), gender (men vs. women), distant metastasis (no vs. yes), World Health Organization (WHO) performance status (PS) (WHO 0-1 vs. WHO 2-3) and cancer site were incorporated into a model to examine the association of age and gender with financial burden, adjusted for WHO PS, distant metastasis and cancer site. The magnitude of financial burden was calculated by dividing the actual scoring difference by its scale and reported in percentage. RESULTS: Data were available for 4,666 male and 2,749 female cancer patients, with 65% older than 60 years. Financial impact of the disease was highly associated with age, with older patients reporting a lower financial burden than younger cancer patients (1.51 vs. 1.28, 7.6%, p=<.0001). Men reported a slightly higher financial burden than females (1.45 vs. 1.42, 0.6%, p=0.006). These gender findings were consistent for both age groups as interaction between age and gender was found not significant (p=0.44). Additionally, low PS was associated (1.42 vs 1.64, 7.3%, p=<.0001) with greater financial burden, but this differed with age (interaction with PS, p=0.005) and gender (interaction p=0.021). CONCLUSIONS: Age, gender and performance status are determinants of cancer patients' financial burden. Less financial burden was experienced by older people and women with poor performance status. These findings could be important for supportive care when treating cancer patients, although more work is needed to interpret the size and importance of these differentials.  
 
 
 
 
 
 

(POSTER PRESENTATION) MINIMAL IMPORTANT DIFFERENCES FOR HEALTH RELATED QUALITY OF LIFE SCORES FROM THE EORTC QLQ-C30 IN LUNG CANCER PATIENTS: ANALYSIS OF POOLED DATA 
John Maringwa, Chantal Quinten, Francesca Martinelli, Corneel Coens, Quality of Life, EORTC, Brussels, Belgium, Charles S. Cleeland, Symptom Research, University of Texas, Houston, Henning Flechtner, Child and Adolescent Psychiatry and Psychotherapy, University Magdeburg, Magdeburg, Germany, Carolyn Gotay, School of Population and Public Health, University of British Columbia, Vancouver, Canada, Eva Greimel, Obstetrics and Gynecology, Medical University Graz, Graz, Austria, Madeleine King, Psycho-oncology Co-operative Research Group, University of Sydney, Sydney, Australia, David Osoba, Quality of Life, Quality of Life Consulting, West Vancouver, Canada, Martin J. Taphoorn, Neurology, VU Medical Center/Medical Center Haaglanden, Amsterdam/The Hague, Netherlands, Bryce B. Reeve, Cancer Control and Population Science, National Cancer Institute, Bethesda, Jolie Ringash, The Princess Margaret Hospital, University of Toronto, Toronto, Canada, Joseph Schmucker-Von Koch, Medical Ethics, University of Regensburg, Regensburg, Germany, Joachim Weis, Psycooncology, University of Freiburg, Freiburg, Germany, Giuseppe Giaccone, Medical Oncology Branch, National Cancer Institute, Bethesda, Pieter Postmus, Pneumology, Vrije Universiteit Medisch Centrum, Amsterdam, Netherlands, Egbert F. Smit, Pulmonary Diseases, Vrije Universiteit VUMC, Amsterdam, Netherlands, Jan Van Meerbeeck, Thoracic Oncology, Universiteit Gent, Gent, Belgium, Andrew Bottomley, Quality of Life, EORTC, Brussels, Belgium 
 
 AIMS: The objective is to determine changes in Health Related Quality of Life (HRQOL) scores on the EORTC QLQ-C30 scales (with range 0-100) which correspond to the minimal important difference (MID). METHODS:  Two closed European Organization for Research and Treatment of Cancer (EORTC) randomized controlled trials enrolling in total 812 advanced NSCLC cancer patients were jointly analyzed. WHO performance status (PS) and weight loss (WL) were chosen as clinical anchors for 6 HRQOL scales of the QLQ-C30 questionnaire; physical (PF), social (SF) and role (RF) functioning, global health status (GH), fatigue (FA) and pain (PA). Changes in PS formed 3 groups; improvement, no change, and deterioration, while changes in weight were grouped as weight loss and no change. Patients gaining 5% or more weight were excluded. Analysis of variance was used to compare HRQOL scores between groups. The differences in the mean of HRQOL score changes between adjacent groups were used to estimate the MID. RESULTS: 485 (for PS) and 311 (for WL) patients who had both HRQOL and anchor values on at least 2 maximally separated time points were included for the analysis. For PS, statistically significant differences (p<0.05) in HRQOL across groups were noted for all 6 scales, while for WL significant differences appeared for SF, FA and PA only. Based on the 2 anchors, the ranges of adjacent group differences were; PF (4.6-9.9), SF (4.9-10.1), RF (3.7-12.3), GH (3.0-10.2), FA (5.9-15.7) and PA (2.2-14.8). CONCLUSIONS: Our results suggest that in patients with advanced NSCLC undergoing treatment, the functioning scales can be interpreted using a change in score of 4 to 12 as the MID. For fatigue a change of 6-16 and for pain 2-15 can be used. These MID estimates are similar to previous reports. They provide guidelines for classification of patients by changes in HRQOL and symptoms over time, and may aid sample size determination for future studies. Further validation in cancer patients with other diagnoses and other anchors is planned. 
 
 
 
 
 

(ORAL PRESENTATION) A SYMPTOM INDEX FOR CANCER PATIENTS BASED ON THE QLQ-C30 
Francesca Martinelli, Chantal Quinten, John Maringwa, Corneel Coens, Quality of Life, EORTC, Brussels, Belgium, Charles Cleeland, Symptom Research, MD Anderson Cancer Center, Houston, Texas, Henning Flechtner, Child and Adolescent Psychiatry and Psychotherapy, University of Magdeburg, Magdeburg, Germany, Carolyn Gotay, School of Population and Public Health, University of British Columbia, Vancouver, Canada, Eva Greimel, Obstetrics and Gynecology, Medical University Graz, Graz, Austria, Madeleine King, Psycho-oncology Co-operative Research Group, University of Sydney, Sydney, Australia, Tito Mendoza, Toru Okuyama, Symptom Research, MD Anderson Cancer Center, Houston, Texas, David Osoba, Quality of Life, Quality of Life Consulting, West Vancouver, Canada, Bryce Reeve, Cancer Control and Population Science, National Cancer Institute, Bethesda, Maryland, Jolie Ringash, The Princess Margaret Hospital, University of Toronto, Toronto, Canada, Joseph Schmucker-Von Koch, Medical Ethics, University of Regensburg, Regensburg, Germany, Qiuling Shi, Symptom Research, MD Anderson Cancer Center, Houston, Texas, Martin Taphoorn, Neurology, Medical Centre Haaglanden, The Hague, The Netherlands, Xin Wang, Symptom Research, MD Anderson Cancer Center, Houston, Texas, Joachim Weis, Psychooncology, University of Freiburg, Freiburg, Germany, Young Yun, Quality of Cancer Care Branch, National Cancer Center, Goyang, Korea, Andrew Bottomley, Quality of Life, EORTC, Brussels, Belgium  
 
 AIMS: The aim of this study was to construct a clinically-relevant symptom index for cancer patients from the EORTC QLQ-C30. METHODS:  The factor structure of 17 symptom-related items from the QLQ-C30 was retrospectively explored in 2 MD Anderson datasets: N=252 Japanese cancer patients (J) and N=177 Korean cancer patients (K). QLQ-C30 and MD Anderson Symptom Inventory (MDASI) data were available for all patients. The "square root of the off-diagonal elements of the residual matrix" was used to select factors, and items with loadings above 0.5 on selected factors were used to build the index. The reliability of the index was measured with Cronbach's alpha (a). Student's t-test was used to compare index scores between subgroups of patients by metastatic status (no vs yes) and ECOG performance status (PS, 0-1 vs 2-3-4). Concurrent validity was assessed by computing Pearson's correlation coefficient (r) between the index and MDASI items. An external pooled dataset including N=5302 cancer patients from 32 countries who partecipated in EORTC clinical trials was used to assess reproducibility and generalizability of the findings. RESULTS: 11 items met selection criteria and were summed to construct the index: pain, need to rest, weakness, lack of appetite, nausea, vomiting, tiredness, difficulty concentrating, worry, irritability and depression. High reliability (a=0.86 for the J dataset, a=0.91 for the K dataset) and ability to discriminate by PS (J, p<0.0001; K, p<0.0001) were seen. The index could discriminate by metastatic status in the J patients only (p<0.0005). Correlations with most MDASI items were high: r=0.55-0.75 for the J dataset and 0.54-0.80 for the K dataset. In the external EORTC dataset, the index showed high reliability (a=0.86) and was able to discriminate patients by metastatic status (p<0.0001) and WHO PS (p<0.0001). CONCLUSIONS: A brief, reliable and valid symptom index can be derived from the QLQ-C30 and used in analysis of clinical trials. Additional research is planned to understand why some symptoms did not appear in the index and how those symptoms should be assessed.