Concept of human voice and AI based medical prescription system

Author: Iqbal Hossain
Date: 2017-06-01

Abstract:

The concept of this paper aims to increase efficiency of electronic prescription by introducing artificial intelligence to analyze audio conversation between doctor and patient. conventional paper based medical prescription system which is written or printed by physicians using pen or printer might have several challenges including time consuming to create correct prescription and may include human error in medical trial procedure mentioned in prescription. In this work I aim to make an efficient system thus it can automatically detect doctor's and patient's voice, analyze the voice and give better suggestive treatments as well as analyze patient's past medical record. This concept is under development and focusing to integrate as a tool into electronic medical record management system.

Key Words: doctor and patient conversation, voice analysis, artificial intelligence, electronic prescription

Introduction:

A usual hand written prescription was convenient communication media between doctors and patients in last decades. Now a day technology made it more convenient converting it to digital form. Digital form refers to electronic prescription which has a broad term of meaning. Electronic prescription includes different stakeholders i.e. doctor, patient, pharmacist and health plan. [1] Every stakeholder has their specific different roles. In this article only doctor and patient role were discussed.

Usually doctor start diagnosis through basic conversation with patient and listen carefully patient’s problem. The doctor reviews the past history and determine what kind of pathological test and medicine would be preferable for patient. After considering many medical issues doctor writes prescribed information to paper or any software pads or in computer oriented application. It is usual to write prescription in paper by hand in developing countries. If there is a computer system, doctor may print the prescription and give it to patient.

Figure-1: Steps of Doctor-Patient conversation and prescription generation by doctor

The steps mentioned in Figure-1 is very common pattern in any healthcare service. In several review and survey we found that in 90% of cases patient could not understand hand written. Even there are few cases that pharmacy sales person mistook to read medicine name and patient took wrong medicine with wrong doses! Although this kind problem is now almost eliminated by using electronic prescription but can we make it more efficient? Because doctor has to think and consider many issues. And after a suitable decision write the prescription. In fact, doctor should make it understand to patient which is time. As well as doctor may miss many important things during conversation with patient. Even the patient may mistakenly provide wrong information in conversation intentionally or non-intentionally and usually conversation is not kept anywhere. Thus there are chances to provide wrongful prescription.

An artificial intelligence (AI) system which use machine learning (ML) to understand the voice of doctor and patient and automatically extract important keywords for analyzing them to predict best medical procedure for patient. This assistance by a device can greatly help to decide what kind of best clinical trial should provide to patient after analyzing patient's history and other medical trial cases. Primarily it is assumed that mentioned benefits could avail from this system. 1) Automatic detection of patient's medical history, 2) Better analysis of medical record and suggestive medical trial, 3) Time savings for both doctor and patient, 4) Efficiency of prescription information archiving and analyzing.


Figure-2: A concept flow diagram of Voice synthesis based medical prescription generation system



Figure-3: Steps of voice detection and generating efficient prescription