OpenAlex · Aktualisierung stündlich · Letzte Aktualisierung: 13.03.2026, 12:52

Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.

Concordance assessment of a clinical decision support software in patients with solid tumors.

2018·4 Zitationen·Journal of Clinical Oncology
Volltext beim Verlag öffnen

4

Zitationen

9

Autoren

2018

Jahr

Abstract

e18584 Background: IBM Watson for Oncology (WFO) is a clinical decision support software that uses natural language processing (NLP) to assist providers to make expeditious treatment decisions. This study expands on a pilot assessing the degree of concordance between treatment recommendations provided by oncologists at Bumrungrad International Hospital (BIH) and WFO at the point of care (Concordance assessment of a cognitive computing system in Thailand 2016, Suwanvecho et al). This study incorporates additional cancer types and new cancer cases from patients seen and treated in 2017. Methods: Clinical data from patients treated at BIH from January to December 2017 were entered into WFO. 264 retrospective cases were entered after a treatment plan had been determined and 326 prospective cases were entered during patients’ treatment planning sessions at the point of care. WFO treatment recommendations were provided in 3 categories: “Recommended”, “For Consideration” and “Not Recommended.” Concordance was analyzed by comparing oncologists’ decisions to WFO recommendations. Concordance was met when the oncologist’s treatment decision was in the WFO “Recommended” or “For Consideration” categories. Results: A total of 590 cases were assessed for concordance. The overall concordance rate for the retrospective cases was 81%; 78% for bladder, 70% for breast, 56% for cervical, 90% for colon, 82% for gastric, 93% for lung, 79% for ovarian, 82% for prostate and 94% for rectal. Similarly, the overall concordance rate for the prospective cases was 80%; 75% for bladder, 72% for breast, 78% for cervical, 95% for colon, 71% for gastric, 87% for lung, 63% for ovarian, 86% for prostate and 76% for rectal. Conclusions: The results are similar to those observed in the 2015-2016 pilot. Both showed a high degree of concordance between WFO treatment options and the decisions made by BIH oncologists. Discordance, in order of frequency, from highest to lowest, was attributable to BIH oncologists' preferences, including favoring non-U.S. guidelines for certain cancers, patient preferences, including toxicity, cost and logistic concerns. Other reasons for non-concordance included recommended U.S. treatment options not available in Thailand.

Ähnliche Arbeiten