Improving Quality and Safety In Your Endoscopy Unit: Live and Virtual

04_Keswani_Data Automation

Pdf Summary

The document discusses the current limitations of data collection in the field of endoscopy and the available options for collecting data. The limitations of quality metrics are highlighted, including the need for high volume to calculate certain metrics and the cumbersome nature of calculation. Lower volumes of therapeutic endoscopy procedures also limit the metrics that can be measured. Measuring outcomes in endoscopy is controversial and there is a lack of confidence in evaluating low volume providers.<br /><br />Various methods for collecting data are discussed, including manual chart review, data registries, data warehouses, and natural language processing. Manual chart review is time-consuming and requires a large number of calculations. Data registries, such as GIQuIC, provide immediate feedback and benchmarking. Data warehouses integrate data from multiple resources into one location. Natural language processing allows computers to "read" and analyze large amounts of data, such as colonoscopy reports, to determine quality metrics.<br /><br />The document also mentions the use of video-taped assessments of procedure skill and the correlation with outcomes. Machine learning and artificial intelligence are suggested as potential solutions for improving endoscopy quality. Advances in deep learning have led to computer-aided polyp detection and classification systems, which can improve detection and classification during colonoscopy.<br /><br />The document concludes by discussing the need for better registries and novel measures for collecting endoscopy quality metrics. It emphasizes the importance of commitment to an outcomes registry and the potential for automated or semi-automated data transfer to reduce the barriers to data collection. Overall, the document highlights the challenges and potential solutions for automating data collection and analysis in the field of endoscopy.

Keywords

limitations

data collection

endoscopy

quality metrics

volume

data registries

natural language processing

machine learning

artificial intelligence

automated data transfer