Oncology volume assessment program: Novel approach to determine patient volumes at Vanderbilt-Ingram Cancer Center (VICC) and Vanderbilt University Medical Center (VUMC).

Abstract

e14047 Background: There is currently a lack of easily-accessible data regarding the volume and type of cancer patients seen at VUMC, despite extensive use of an electronic medical record (EMR) with a dedicated tumor registry. This leads to errors in estimating metrics that depend on subject accrual to clinical trials. We aimed to develop a tool to determine the number of head and neck squamous cell carcinoma (HNSCC) patients, stratified by desired subsets, seen yearly at the VUMC in order to achieve a more accurate and efficacious approach to estimating future patient accrual to clinical trials. Methods: Working with Nashville Biosciences, we identified patients with HNSCC using ICD codes in combination with VUMC’s Tumor Registry data (collected for patients diagnosed and treated at VUMC). We estimated the rate of accrual of HNSCC patient blood samples (a proxy for ICD code) based on the number of specimens acquired over the past 3 years. Each patient was counted only once based on their first blood sample acquired during this period, based on a comprehensive list of laboratories and clinical procedures that require a specimen to be drawn. We then evaluated the distribution of patients by primary site, standardized to SEER registry terminology. We also investigated what fraction of patients received treatment with surgery, radiation, and/or chemotherapy (we expect that this includes treatment with biologics) using a combination of ICD9/10 and CPT coding. Results: We identified a total of 1,131 HNSCC patients. SEER summary staging information was available for 603 patients - 432 (72%) had Stage 1-3 disease and 171 (28%) had Stage 4 disease. Primary site data was available for 717 patients - 648 (90%) HNSCC of the oral cavity, 20 (3%) had HNSCC of the oropharynx, 29 (4%) HNSCC of the larynx, 20 (3%) HNSCC of the hypopharynx. Of all patients, we found that 598 had at least 1 code for chemotherapy, 540 for radiotherapy, and 825 for surgery such that the majority (96%) received cancer treatment. We have not evaluated the timing of the treatment relative to diagnosis. Conclusions: By using ICD codes and tumor registry data extracted from VUMC’s EMR, we determined the site distribution and type of treatment for HNSCC over a 3-year period. The disease site distribution did not match prior experience Refinements are ongoing. However, once improved, these methods may be used to estimate future clinical trial accrual, thus reducing start-up time, cost and poor trial site performance, while increasing accrual.