Choosing geriatric assessment tools for elderly in clinical trials

Abstract

4728 Background: Over 65% of cancer patients are > 65, yet elderly are underrepresented in clinical trials. To evaluate representative elderly in clinical trials, a few brief, reliable and sensitive measures of functional status will be critical. We piloted measures of physical performance used in geriatric assessment on a sample of breast and prostate cancer patients. We present here results regarding the potential value of selected measures. Methods: An on-going study of elderly cancer patients enrolled any breast (BrCa) or prostate (CaP) patient scheduled for adjuvant cytotoxic or hormonal therapy. Assessments included self-report and directly observed measures (DOM) of performance. Patients with age > 50, MMSE ≥ 23 and no previous cancer diagnosis or chemotherapy were consented. Assessments were performed at baseline, but before therapy, at 6 months and are planned at 2 years. Results: Complete baseline data are available on 47 BrCa patients, 27 with 6-month followup; 21 CaP baseline and 11 with 6-month followup. The 2 groups did not differ by age (mean 69.7; 69.8), MMSE (mean 27.9; 28.7), self-reported IADL dependency (mean 13; 14) or number of co-morbidities (mean 2.6; 2.7). DOM included the Berg Balance Scale BBS (56 points, < 40 = fall risk) and Timed Up and Go TUG (nl = 15 sec. or less) either raw time or scored (1–4). At baseline 9/47 BrCa and 11/21 CaP scored high risk for falls on the BBS; at 6-mo, 4/27 BrCA AND 5/11 CaP were fall risks. TUG scores in sec 5/46 BrCa at baseline and 3/27 at 6-months had slowed gait speed. Among CaP 2/21 at base and 1/11 at 6-months had slowed gait. A paired analysis revealed no change in individuals’ classification over the 6 months of observation. The TUG takes about 1 minute, the BBS about 15 min to complete, requiring a stop watch, tape measure, chair and little other equipment. Published norms for age and sex are available. Conclusions: DOM identified 10–48% of patients with functional impairments, whereas self-report measures identified none in this small sample. Classification remained stable over time indicating either good treatment tolerance or test insensitivity to small changes. Larger samples are needed to assess the sensitivity and predictive value of these measures for elderly cancer patients in clinical trials. No significant financial relationships to disclose.