Differential impact of cognitive computing augmented by real world evidence on novice and expert oncologists.

INTRODUCTION: Cognitive computing point-of-care decision support tools which ingest patient attributes from electronic health records and display treatment options based on expert training and medical literature, supplemented by real world evidence (RWE), might prove useful to expert and novice oncologists. The concordance of augmented intelligence systems with best medical practices and potential influences on physician behavior remain unknown. METHODS: Electronic health records from 88 breast cancer patients evaluated at a USA tertiary care center were presented to subspecialist experts and oncologists focusing on other disease states with and without reviewing the IBM Watson for Oncology with Cota RWE platform. RESULTS: The cognitive computing "recommended" option was concordant with selection by breast cancer experts in 78.5% and "for consideration" option was selected in 9.4%, yielding agreements in 87.9%. Fifty-nine percent of non-concordant responses were generated from 8% of cases. In the Cota observational database 69.3% of matched controls were treated with "recommended," 11.4% "for consideration", and 19.3% "not recommended." Without guidance from Watson for Oncology (WfO)/Cota RWE, novice oncologists chose 75.5% recommended/for consideration treatments which improved to 95.3% with WfO/Cota RWE. The novices were more likely than experts to choose a non-recommended option (P < .01) without WfO/Cota RWE and changed decisions in 39% cases. CONCLUSIONS: Watson for Oncology with Cota RWE options were largely concordant with disease expert judged best oncology practices, and was able to improve treatment decisions among breast cancer novices. The observation that nearly a fifth of patients with similar disease characteristics received non-recommended options in a real world database highlights a need for decision support.

A Patient-Reported Outcome Instrument to Assess Symptom Burden and Predict Survival in Patients with Advanced Cancer: Flipping the Paradigm to Improve Timing of Palliative and End-of-Life Discussions and Reduce Unwanted Health Care Costs.

BACKGROUND: Discussions regarding palliative care and end-of-life care issues are frequently delayed past the time of usefulness, resulting in unwanted medical care. We sought to develop a patient-reported outcome (PRO) instrument that allows patients to voice their symptom burdens and facilitate timing of discussions. SUBJECTS, MATERIALS, AND METHODS: A seven-item PRO instrument (Cota Patient Assessed Symptom Score-7 item [CPASS-7]) covering physical performance status, pain, burden, and depression was administered (September 2015 through October 2016) with correlation to overall survival, correcting for time to complete survey since diagnosis. RESULTS: A total of 1,191 patients completed CPASS-7 at a median of 560 days following the diagnosis of advanced cancer. Of these patients, 49% were concerned that they could not do the things they wanted; 35% reported decreased performance status. Financial toxicity was reported by 39% of patients, with family burdens noted in 25%. Although depression was reported by 15%, 43% reported lack of pleasure. Pain was reported by 33%. The median CPASS-7 total symptom burden score was 16 (possible 0-112). With a median follow-up of 15 months from initial survey, 46% had died. Patients with symptom burden scores <29 and ≥29 had a 6-month overall survival rate of 87% and 67%, respectively, and 12-month survival rates of 72% and 50%. A one-point score increase resulted in a 1.8% increase in expected hazard. CONCLUSION: Patients with advanced cancer with higher levels of symptom burden, as self-reported on the CPASS-7, had inferior survival. The PRO facilitates identification of patients appropriate for reassessment of treatment goals and potentially palliative and end-of-life care in response to symptom burden concerns. IMPLICATIONS FOR PRACTICE: A seven-item patient-reported outcome (PRO) instrument was administered to 1,191 patients with advanced cancers. Patients self-reporting higher levels of physical and psychological symptom burden had inferior overall survival rates. High individual item symptom PRO responses should serve as a useful trigger to initiate supportive interventions, but when scores indicate global problems, discussions regarding end-of-life care might be appropriate.

Development of a Precise, Clinically Relevant, Digital Classification Schema for Cancer.

PURPOSE: Health care transactions depend on the efficiency of digital codes. The International Classification of Diseases and Related Health Problems (ICD) coding system, which is the most commonly used digital system, fails to capture the complexity of oncologic diseases. Because important prognostic information such as cancer stage and genomic markers are missing, the potential for ICD codes to define and compare patient cohorts is severely limited. A more precise, clinically relevant, digital classification schema that incorporates prognostic elements would address these needs. METHODS: Working with cancer disease-specific experts, a new digital classification scheme, known as the Cota Nodal Address (CNA) system, was developed. The CNA has six components that define the disease of interest and incorporate all standard-of-care prognostic and predictive markers related to the particular cancer, including patient features. RESULTS: Properly sorted into homogeneous groupings of patients with similar prognostic characteristics, the CNA system facilitated big data analytic approaches, such as evaluations of population health, identification of variation in treatment decisions, and the enablement of value-based payment models. The schema has been applied to patients with breast cancer at a large tertiary cancer care hospital and a regional community cancer care network and has facilitated the creation and application of value-based payment models. CONCLUSION: The development and potential uses of a prognosis-based classification system are reviewed herein. Compared with ICD coding, the greater precision of the schema permits improved analyses of variance in treatment, outcomes, and costs in cancer care management.

Genomic Profiling of Advanced Non-Small Cell Lung Cancer in Community Settings: Gaps and Opportunities.

BACKGROUND: National guidelines have advocated broad molecular profiling as a part of the standard diagnostic evaluation for advanced non-small cell lung cancer (NSCLC), with the goal of identifying driver mutations for which effective therapies or clinical trials are available. However, adherence to genomic testing guidelines could present challenges to community oncologists. PATIENTS AND METHODS: We performed a retrospective review of genomic testing patterns in patients with nonsquamous NSCLC treated by 89 oncologists at 15 sites throughout New Jersey and Maryland from January 2013 to December 2015. RESULTS: A total of 814 patients (89% with stage IV; 11% with stage IIIB) were identified in the COTA Inc database. Of the 814 patients, 479 (59%) met the guideline recommendations for EGFR (epidermal growth factor receptor) and ALK (anaplastic lymphoma kinase) biomarker testing; 63 (8%) underwent comprehensive genomic profiling for all 4 major types of alterations (point mutations, indels, fusions, and copy number amplifications). Gender, age, race, site of care (referral vs. community center), and practice size did not influence comprehensive genomic profiling frequency. Active smokers and patients who died within 30 days were tested less frequently (P < .05). Among those not tested for EGFR and ALK, 52% received chemotherapy without documented reasons for no testing, 32% did not receive antineoplastic therapy, and 13% had insufficient tissue for genotyping. CONCLUSION: Genomic testing presents multiple logistical challenges for the community-based oncologist, including coordination of sample handling, long turnaround times, test reimbursement, access to targeted therapies, insufficient tissue, and patient harm from the repeat biopsies necessary if the tissue sample is insufficient. Opportunities exist for improvement in guideline adherence, possibly through new technologies such as "liquid biopsies," which obviates the need tissue biopsy samples in select settings.