PSYCHOMETRIC VALIDATION OF A PATIENT SYMPTOM ASSESSMENT-LUNG CANCER (PSALC) INSTRUMENT FOR SMALL CELL LUNG CANCER (SCLC)
Author(s)
Lei Chen, PhD, MD, Associate1, Mei Duh, MPH, ScD, Vice President1, Lucia Antras, PhD, Manager1, Mary ER. O'Brien, MD, FRCP2, Maureen P. Neary, PhD, Director, Oncology3, Joachim Von Pawel, MD, Doctor41Analysis Group, Inc, Boston, MA, USA; 2 Royal Marsden Hospital, Sutton, United Kingdom; 3 GlaxoSmithKline, Collegeville, PA, USA; 4 Asklepios-klinik, Gauting, Germany
OBJECTIVES: The PSALC is an instrument that was developed for use in patients with SCLC for assessment of nine symptoms (i.e., shortness of breath, cough, chest pain, hemoptysis, appetite loss, sleep interference, hoarseness, fatigue, interference with daily activities) scored on a scale from 1 (not at all) to 4 (very much), but it has not been formally validated. METHODS: A retrospective psychometric validation was conducted using data from a randomized multicenter trial in which 107 patients with SCLC were treated with intravenous topotecan and 104 patients with cyclophosphamide, doxorubicin, and vincristine. PSALC was administered to patients at baseline and at each of the four subsequent clinical visits at 3-week intervals. Factor analysis (FA), internal consistency, construct validity, reliability, and responsiveness were evaluated. RESULTS: Baseline FA (n=137) indicated that there was only one factor, so the PSALC total score was used for validation. Weighted Cronbach's alpha from all visits was 0.74 showing an acceptable internal consistency. Construct validity was supported by the high correlation with established measures, with lower baseline PSALC total scores associated with better Eastern Cooperative Oncology Group (ECOG) performance status (12.41, 18.24, and 20.33 for ECOG score=0, 1, and 2, respectively, p<0.0001; regression slope estimate=3.84, p<0.0001; n=135). Reliability was supported by an intraclass correlation coefficient of 0.61, calculated using PSALC total scores evaluated before any change in clinical status, and a concordance correlation coefficient of 0.72, calculated using PSALC total scores at baseline and before the first visit. The PSALC total score was responsive to tumor progression (responsiveness statistic=0.64; n=33). CONCLUSION: A retrospective analysis suggests that the PSALC is a reliable, valid, and responsive instrument for measuring SCLC symptoms. If feasible in this population, a prospective validation study could be used to further evaluate these findings.
Conference/Value in Health Info
2007-05, ISPOR 2007, Arlington, VA, USA
Value in Health, Vol. 10, No.3 (May/June 2007)
Code
CN4
Topic
Patient-Centered Research
Topic Subcategory
Patient-reported Outcomes & Quality of Life Outcomes
Disease
Oncology