English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The medical codes prediction problem from clinical notes has receivedsubstantial interest in the NLP community, and several recent studies haveshown the state-of-the-art (SOTA) code prediction results of full-fledged deeplearning-based methods. However, most previous SOTA works based on deeplearning are still in early stages in terms of providing textual references andexplanations of the predicted codes, despite the fact that this level ofexplainability of the prediction outcomes is critical to gaining trust fromprofessional medical coders. This raises the important question of how wellcurrent explainability methods apply to advanced neural network models such astransformers to predict correct codes and present references in clinical notesthat support code prediction. First, we present an explainable Read, Attend,and Code (xRAC) framework and assess two approaches, attention score-basedxRAC-ATTN and model-agnostic knowledge-distillation-based xRAC-KD, throughsimplified but thorough human-grounded evaluations with SOTA transformer-basedmodel, RAC. We find that the supporting evidence text highlighted by xRAC-ATTNis of higher quality than xRAC-KD whereas xRAC-KD has potential advantages inproduction deployment scenarios. More importantly, we show for the first timethat, given the current state of explainability methodologies, using the SOTAmedical codes prediction system still requires the expertise and competenciesof professional coders, even though its prediction accuracy is superior to thatof human coders. This, we believe, is a very meaningful step toward developingexplainable and accurate machine learning systems for fully autonomous medicalcode prediction from clinical notes.

Can Current Explainability Help Provide References in Clinical Notes to Support Humans Annotate Medical Codes?