For every hour that physicians spend on direct patient care, they spend 2 hours on administrative tasks and EHR documentations. This causes growing dissatisfaction and burnout amount physicians, decreasing the quality of care that they are able to provide to patients.
The goal of the PhenoPad project is to provide convenient and efficient clinical note taking during physician-patient interactions through the use of AI.
PhenoPad is an application that is easy to use and allows clinicians to capture patient notes in realtime. It provides AI-based note taking services which improve the quality of patient-physician interaction through:
When a doctor writes down a line of script, Phenopad immediately captures the line and recognizes it using medical language model. Once done analyzing, it displays a text bar above the script with the recognized text. Alternative candidates for each recognized word are also provided for the doctor to correct any spelling or semantic mistakes.
At the same time, the NLP detects any potential abbreviations and returns the extended forms based on local context. NLP will also search for key medical terms and displays all candidate phenotypes found in the text as togglable tags. The doctor can choose to add or delete phenotypes or switching their states as Yes/No, which indicates whether the patient presents the existence of the specific phenotypes.
Instead of jotting down the sentences, the doctor can also choose to record the session. Phenopad provides convenient speech audio service which processes and identifies different speakers using deep learning approach, then displays the transcript on side for quick reference. The NLP will also search through the transcript to detect any key medical terms and suggests all candidate phenotypes. All transcripts of past conversations are also saved locally, which allows the doctor to recapture any missed patient information through a single click.
When the doctor changes the state of a candidate phenotype to "Y", Phenopad automatically performs analysis and provides a list of related phenotypes as further investigate suggestion. Meanwhile, it also determines the diseases that are most commonly found based on the collected phenotypes, and list them out as a clinical diagnosis reference to the doctor. The doctor has the option to view detailed information of each phenotype and disease through an external extension where all information is retrieved from authorized database APIs.
Phenopad also provides the add-in feature which allows the doctor to add a widget by drawing a square or a rectangle on the note page. The add-in widget gives three different options that support drawing pedigree diagrams, inserting local images and taking photos for annotations. All add-in widgets also support resizing and repositioning, and can be minimized into side bar if the note page has run out of writing space.
jixuan.cs@gmail.com
Send Email D.L. Pratt Building #286
University of Toronto - St. George
6 King's College Road, Toronto
ON, M5S 3H5
