Overview

Overview

We developed an AI-based medical care solution for a well-known healthcare organization. Installed in their hospitals and used by doctors, this solution helps the medical practitioners extract major drug-related data from patients’ profile and predict the side effects of new or current medication, if any. With its help, doctors can prevent probable drug allergies or consequences for their patients.

Domain
Medical

Challenges

Challenges
  • According to WAO (World Allergy Organization), 10-15% of hospitalized patients while 3-6% of all patients suffer from drug allergy.
  • Drug allergies cause higher medical cost and risk for the patients.
  • It may increase the time taken in healing for a patient.
  • Previously consumed drugs and all the drug allergies cannot be recalled by the patient.
  • May result in legal actions against the practitioner if a patient files a lawsuit and collects enough pieces of evidence that it was the doctor’s fault.
  • The manual process of drug allergy prediction is very time-consuming and not really precise. It needs doctors and nurses to go through previous drug history and the medical history of the patient. Even then, the prediction need not be correct.

Our Solution and its Capabilities

Our Solution and its Capabilities
  • A data collection module, which records the medical history of the patient on every visit. Alongside, it keeps a track of problems, diseases, allergies, symptoms, etc.
  • An arrangement for saving medical prescriptions and drug details.
  • A drug discovery module to read and compare the properties of drugs given to the patient previously and prescribed ones.
  • A setup to compare the medical properties of drugs and confirm they won’t react adversely. This module will also check if any prescribed drug isn’t suitable for the patient as per his or her medical history.
  • An OCR (Optical Character Recognition) tool for converting the text from uploaded medical documents into digital data.
  • A Natural Language Processing module for text analytics and extracting the important information from the medical documents.
  • An aid for converting final reports into PDF.

Outcomes

The AI-based Medical care solution, we’d crafted, is in use currently. This medical care solution alerts the doctors when their newly-prescribed drug shows the probability of allergy in the patient as per his/her history. So with its help, doctors using our solutions can prevent probable drug allergies or consequences for their patients. It operates dynamically and is extensible enough in case the client wants to extend its functionalities in the future. The client is extremely satisfied with its performance as the medical staff has reported a 60% decline in drug allergies among patients from the time it has been deployed. Its OCR system has saved hospital staff a significant amount of time for other operations, causing a benefit of $1000 PER DAY (which means $1.08 Million is 3 years) for this healthcare group.

Technical Stack

Front end

  • React js
  • Javascript/CSS
  • HTML

Backend

  • Node
  • Express js

AWS Service Stack

  • AWS Lamda
  • API Gateway
  • DynamoDB
  • Appsync

PMS and Communication Tool

  • Slack
  • Jira

AWS ML Stack

  • OCR
  • Tensorflow API
  • Python
  • Amazon Comprehend
  • Amazon Extract
  • Advanced ML algorithms for data mining
  • AI-based text analytics
Team
acr

1

React Developer

acr

1

NodeJS Developer

acr

1

DevOPS Developer

acr

1

Python Developer

acr

1

AWS ML Specialist