Enhancing Laboratory Efficiency: Implementation of Revvity Transcribe AI for Automated Data Entry in Newborn Screening
Laboratory Genetics and Genomics
-
Primary Categories:
- Public Health Genetics
-
Secondary Categories:
- Public Health Genetics
Introduction:
Manual data entry from handwritten documents into informatics systems is a time-consuming process in many laboratories, particularly in Newborn Screening facilities utilizing Dried Blood Spot (DBS) cards. To address this challenge, we developed "Revvity Transcribe AI," an innovative solution leveraging Optical Character Recognition (OCR) and Machine Learning (ML) to convert handwritten text into a digital format. The primary objective of this product was to increase the processing speed of DBS cards and optimize laboratory staff time allocation.
Methods:
Revvity Transcribe AI was implemented at the Pittsburgh Revvity Omics laboratory in conjunction with Revvity Instrument Hub to process images of DBS cards to reduce manual input work. Two key metrics were evaluated: accuracy and processing speed improvement. Accuracy was determined by comparing Transcribe AI output with human-validated data. For each card, a total of ten fields containing demographic information were processed. The fields included input types such as handwritten names, birthdates, and checkboxes. Processing speed improvement was assessed by measuring the time taken to pre-populate data entry fields using Transcribe AI and subsequent human validation of the text. First measurement was done right after implementation of the solution, second one three weeks after daily use.
Results:
Over a three-week period, Transcribe AI demonstrated a total field accuracy score of 73.3% over 1,913 cards processed. In terms of processing speed, the system showed significant improvements. During the first week of implementation, processing speed increased by 36%, from 36 cards per hour per person to 50 cards per hour. After two weeks of use, the processing speed further improved to 57 cards per hour per person, representing a 58% overall improvement in processing speed.
Conclusion:
Revvity Transcribe AI has proven to be a valuable tool in significantly enhancing the efficiency and accuracy of laboratory test request form processing. By automating data entry, the system allows for more effective utilization of human resources, enabling staff to focus on other critical tasks. The implementation of Transcribe AI presents opportunities for laboratories to substantially increase their productivity. Future developments and optimizations of the system hold promise for even greater benefits to laboratory operations, potentially revolutionizing workflow efficiency in newborn screening and other laboratory settings.
Manual data entry from handwritten documents into informatics systems is a time-consuming process in many laboratories, particularly in Newborn Screening facilities utilizing Dried Blood Spot (DBS) cards. To address this challenge, we developed "Revvity Transcribe AI," an innovative solution leveraging Optical Character Recognition (OCR) and Machine Learning (ML) to convert handwritten text into a digital format. The primary objective of this product was to increase the processing speed of DBS cards and optimize laboratory staff time allocation.
Methods:
Revvity Transcribe AI was implemented at the Pittsburgh Revvity Omics laboratory in conjunction with Revvity Instrument Hub to process images of DBS cards to reduce manual input work. Two key metrics were evaluated: accuracy and processing speed improvement. Accuracy was determined by comparing Transcribe AI output with human-validated data. For each card, a total of ten fields containing demographic information were processed. The fields included input types such as handwritten names, birthdates, and checkboxes. Processing speed improvement was assessed by measuring the time taken to pre-populate data entry fields using Transcribe AI and subsequent human validation of the text. First measurement was done right after implementation of the solution, second one three weeks after daily use.
Results:
Over a three-week period, Transcribe AI demonstrated a total field accuracy score of 73.3% over 1,913 cards processed. In terms of processing speed, the system showed significant improvements. During the first week of implementation, processing speed increased by 36%, from 36 cards per hour per person to 50 cards per hour. After two weeks of use, the processing speed further improved to 57 cards per hour per person, representing a 58% overall improvement in processing speed.
Conclusion:
Revvity Transcribe AI has proven to be a valuable tool in significantly enhancing the efficiency and accuracy of laboratory test request form processing. By automating data entry, the system allows for more effective utilization of human resources, enabling staff to focus on other critical tasks. The implementation of Transcribe AI presents opportunities for laboratories to substantially increase their productivity. Future developments and optimizations of the system hold promise for even greater benefits to laboratory operations, potentially revolutionizing workflow efficiency in newborn screening and other laboratory settings.