Abstract: |
<p>Background: Geriatric assessment (GA) can be used to identify patient-specific deficits and facilitate interventions to improve patient outcomes. In this randomized pilot clinical trial, we hypothesized that geriatric assessment with or without geriatrician comanagement can reduce the risk of hospitalization for older adults with head and neck cancer receiving concurrent chemoradiotherapy (CRT). Patients and Methods: Patients aged >= 65 years with newly diagnosed or postoperative mucosal head and neck squamous cell carcinomas were eligible for enrollment and received 6 to 7 weeks of CRT. All patients completed an electronic GA and were randomized to either a geriatric consultation for comanagement of geriatric-specific deficits or management of geriatric deficits by the radiation oncology team without geriatric consultation. All patients were seen twice weekly in the radiation oncology clinic to enhance supportive care. The primary endpoint was the rate of hospitalization during CRT. Results: A total of 30 patients were randomized, with 14 ultimately eligible to proceed in each study arm. Median age was 75 years (range, 65-81) and the median number of geriatric deficits was 4 (range, 1-10). The most common geriatric deficits were symptoms of depression, social activity limitation, and lack of social support. The overall hospitalization rate was 42.9% (12/28; 95% CI, 24.5%-62.8%), with 64.3% (9/14; 95% CI, 35.1%-87.2%) and 21.4% (3/14; 95% CI, 4.7%-50.8%) of patients hospitalized in the geriatric comanagement and GA-guided supportive care arms, respectively. Accumulated geriatric deficits (AGDs) and history of falls were both strongly predictive of the risk of hospitalization (P=.0061 and P=.044, respectively). No deaths occurred during treatment or within 30 days of completing CRT. Conclusions: AGDs and history of falls were strong predictors of hospitalization risk. Both geriatric comanagement and GA-guided supportive care were implemented with high compliance, but GA-guided supportive care involves a lower resource burden for widespread implementation.</p> |