Garcia CA, et al. - Authors gauged frailty assessment via the Risk Assessment Index (RAI) in cancer patients and its ability to predict hospital utilization and 90-day mortality. Findings suggested the probability of RAI to be utilized for predicting 90-day mortality in cancer patients. Unplanned hospitalizations were more common in the patients at greatest risk of 90-day mortality. The difference in RAI scores was statistically significant between patients with and without unplanned ED visits/hospitalizations, but this difference was too small to be clinically meaningful.

Methods

• RAI was completed by new patients with solid tumor malignancies at the UPMC Hillman Cancer Center from September 15, 2017 to January 31, 2018.
• They abstracted the age, cancer type and stage, and treatments.
• The Wolff Center at UPMC provided ED visits, hospitalizations and mortality data.
• They performed Mann-Whitney U and likelihood-ratio tests comparing RAI, ED usage/hospitalizations, and mortality.
• They excluded the elective hospital admissions.

Results

• As per data, 273 patients completed the RAI.
• Findings suggested 58.6% of pts to be male and average age to be 62.7 (21-88), and the average RAI score to be 39.9 (IQR 35-44).
• Melanoma (22.7%), prostate (14.3%), and lung (11.4%) were the most common cancer types.
• Stage IV cancer was seen in 43.6% of patients.
• Within 90 days after RAI, at least one unplanned ED visit or hospitalization were experienced by 51 patients (18.7%), and 10 patients (3.7%) were deceased.
• Compared to an average RAI of 39.4 (IQR 35-43) for patients without an ED visit/hospitalization, patients with an ED visit/hospitalization had an average RAI of 42 (IQR 37-46): a 2.6-point difference (p=0.008); 90-day mortalities had an average RAI score of 47.5 (IQR 42-53) compared to 39.6 (IQR 35-44) for non-mortalities: a 7.9-point difference (p=0.003).
• Mortality was experienced by 6 of 242 (2.5%) patients without a hospitalization whereas 4 of 31 (12.9%) patients with a hospitalization experienced mortality (p=0.017).