Case Study: SelectData increases production and accuracy with John Snow Labs OCR and NLP

SelectData boosts production by 10% with John Snow Labs

SelectData, a home health business, faced the challenge of managing a rapidly growing number of complex patient records stored as document images. Managers were overwhelmed by the volume of data and struggled to classify record complexity and reduce noise within clinical documentation. To address this, they partnered with John Snow Labs.

John Snow Labs implemented a solution using Spark OCR to process difficult-to-read document images and Spark NLP to extract entities and normalized values from the text. This enabled SelectData to build predictors that recommend the best next action and classify assessments by difficulty. The results included a 10% increase in overall production, a 95% accuracy rate for recommendations, and the ability to interpret millions of patient records.

SelectData