Measuring concordance with head and neck cancer survivorship algorithms.

Abstract

e21602 Background: We developed institutional algorithms for cancer survivors that are site specific. However, little data exists regarding concordance with these clinical management tools. Our aim was to evaluate concordance rates with our Head and Neck (H/N) survivorship algorithms. Methods: We reviewed electronic health records (EHR) of patients seen in the H/N Survivorship Clinic between 9/01/2011 and 8/31/2014. Eligibility included: diagnosis of primary head and neck cancer, ≥ 18 years of age, no active disease, and alive at time of abstraction. Data were collected at baseline (V1) and visits 9-15 months after prior visits (V2 and V3). Data were obtained from tumor registry, a departmental database and the EHR. Twelve data elements were grouped as: 1) lab/diagnostic tests, 2) physical and psychological late effects, and 3) history and physical exams. Concordance rate (%) for each element was defined as a clinical action or recommendation aligned to the standards outlined in the H/N algorithms. Descriptive statistics were used to summarize clinical, demographic, and concordance rates. Results: Included in this analysis were patients with primary H/N malignancies of oropharynx, oral cavity, larynx, and hypopharynx, and who had completed 3 visits (V1-3). 53 patients were evaluable for concordance analyses. Compliance ranged from 55% to 100%, with lowest rates for: CT scans (55.6%), distress screening (75.5%), T4/TSH level (78.4%) and laryngoscopy (79.2%). Highest rates were noted for full physical exam (100%), chest x-ray (97.5%), dental exam (95.9%), and dysphagia evaluation (95.1%). Assessments for lymphedema and xerostomia were similar (91.8%). Conclusions: High compliance ( > 90%) occurred when elements of the algorithm relied on the clinician’s assessment, whereas lower levels of compliance were seen with elements that required orders (CTs and labs), or reliance on others’ actions (i.e. distress screening by nursing). Concordance can be improved using best practice advisories in the EHR to remind providers of orders, and by fostering strong inter-professional communication on the elements of care.