National Institutes of Health (NIH)

NIH

Chief Information Officer: Andrea Norris

AI Automates and Accelerates Grant Processing

The National Institute of General Medical Sciences (NIGMS), one of the 27 institutes and centers at NIH, uses Natural Language Processing (NLP) and Machine Learning (ML) to automate the process by which it receives and internally refers grant applications.
The approach ensures efficient and consistent grant application referral, and liberates Program Managers from the labor-intensive and monotonous referral process. It allows them to focus more on high-value work.

Prior to the introduction of Machine Learning (ML) and Natural Language Processing (NLP), each of the NIGMS funding opportunity announcements (FOA) had various scientific points of contact (POCs), who were responsible for receiving and referring (i.e., assigning) the applications made to that FOA to the relevant organizational Program Managers (PM).

This referral of research grant applications is a laborious manual process that can take an average of two to three weeks to complete. To address some of the process’s pain points, the NIGMS first piloted a machine learning approach to the receipt and referral process.

An NLP ML algorithm was created capable of receiving and assigning an incoming grant. Using thousands of historical records of grant applications, the NIGMS AI team trained the NLP ML algorithm.

The NIGMS AI team then identified an accuracy measure to allow them to compare the manual or “human-executed” referral process to the accuracy of the algorithm’s output.

The NLP ML algorithm is now functionally integrated with existing NIGMS grant application web-based systems and is run nightly to process new applications. Thanks to this integration, the average referral time for an application has been cut from 2-3 weeks to less than one day.

Recognizing the administrative gains and high level of accuracy achieved from NLP ML algorithms, the NIGMS’ senior management team expanded usage of the novel system to five major FOAs that cover about 80 percent of all NIGMS research applications requiring referral.