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

Objectives: To describe the development and validation of a zero-shotlearning natural language processing (NLP) tool for abstracting data fromunstructured text contained within PDF documents, such as those found withinelectronic health records. Materials and Methods: A data abstraction tool basedon the GPT-3.5 model from OpenAI was developed and compared to three physicianhuman abstractors in terms of time to task completion and accuracy forabstracting data on 14 unique variables from a set of 199 de-identified radicalprostatectomy pathology reports. The reports were processed by the softwaretool in vectorized and scanned formats to establish the impact of opticalcharacter recognition on data abstraction. The tool was assessed forsuperiority for data abstraction speed and non-inferiority for accuracy.Results: The human abstractors required a mean of 101s per report for dataabstraction, with times varying from 15 to 284 s. In comparison, the softwaretool required a mean of 12.8 s to process the vectorized reports and a mean of15.8 to process the scanned reports (P < 0.001). The overall accuracies of thethree human abstractors were 94.7%, 97.8%, and 96.4% for the combined set of2786 datapoints. The software tool had an overall accuracy of 94.2% for thevectorized reports, proving to be non-inferior to the human abstractors at amargin of -10% ($\alpha$=0.025). The tool had a slightly lower accuracy of88.7% using the scanned reports, proving to be non-inferiority to 2 out of 3human abstractors. Conclusion: The developed zero-shot learning NLP toolaffords researchers comparable levels of accuracy to that of human abstractors,with significant time savings benefits. Because of the lack of need fortask-specific model training, the developed tool is highly generalizable andcan be used for a wide variety of data abstraction tasks, even outside thefield of medicine.

Validation of a Zero-Shot Learning Natural Language Processing Tool for Data Abstraction from Unstructured Healthcare Data