Abstract
This chapter presents a voice-enabled mobile application that provides preliminary medical diagnosis of non-fatal diseases. It operates in Urdu language, asking questions from the user and responding in the form of a voice-based dialogue. It addresses the problem of lack of hospitals and medical facilities, especially in rural areas. Furthermore, due to the ongoing pandemic of COVID-19, people are wary of visiting the over-crowded government hospitals. The mobile application developed by the authors provides a quick and inexpensive preliminary diagnosis and medical advice. In case of emergency, it recommends a nearby hospital/doctor. As the system is in Urdu language and voice based, it is well-suited to the Pakistani population's low literacy rate. The application user interface is engineered to be intuitive and simple to use.
TopIntroduction
The purpose of the project described in this chapter is to create a virtual healthcare assistant mobile application for an average Pakistani (literate/illiterate) who converses in Urdu. The healthcare assistant mimics a doctor and talks to its user just like a doctor. It asks certain questions from the user regarding his health and records the user’s responses. Based on these responses, the assistant predicts the disease the user might be suffering from. The aim is to create a user-friendly application which will use an audio-visual interface for easy communication with the user. Urdu voice-based conversation and natural language understanding technologies are used in its implementation.
There are inadequate healthcare facilities in the rural areas of Pakistan. So, in case of a medical emergency, the patients often have to travel very long distances to seek medical attention, even for minor diseases. Early diagnosis of a disease can greatly increase the chances of survival of a critically ill person. Also, lack of medical awareness further worsens the patient’s condition while travelling. Furthermore, hospitals are already very crowded, given the ongoing COVID-19 pandemic. People have to wait in long queues to see a doctor. This also poses the risk of spreading the pandemic and increases the workload on the doctors. The purpose of creating this application is provide basic healthcare to the average Pakistani using the internet and mobile phone, otherwise free of cost. There is no such system currently available in the market that operates in Urdu language, and that considers the conditions of rural areas of Pakistan.
The mobile application has a chat type interface with access to the microphone and speaker. The user is queried by the chatbot in Urdu about the symptoms which he/she might be facing. The user may either reply verbally through speech using the microphone, or interact with the application through on-screen buttons with possible answers being provided on the user screen. Every next question that the assistant would ask from user depends on the user’s answer to the previous question. As a result, the system provides the predicted disease of the user with the corresponding prescription and treatment for that disease in Urdu. By using this application, a person can avoid the following difficulties:
- •
See a doctor for common, non-fatal diseases.
- •
Pay medical care fees for mild diseases.
- •
Travel long distances to a hospital.
- •
Wait for one’s turn to see a doctor.
The system offers the following advantages to its users:
- •
Receive free healthcare attention.
- •
Save a lot of time.
- •
Save money.
- •
Get instant and preliminary diagnosis.
- •
Verbal conversation in Urdu for natural interaction
This chapter provides a detailed description of the development of an Urdu voice-based disease diagnosis mobile app. The background section provides a review of the previous work in this domain. Details of technology and algorithms used in the proposed system are discussed in the methodology section. The system design and architecture section describes the overall system and modules developed by the authors. Evaluation of this system is presented in the experiments and results section. Finally, a conclusion is provided at the end of this chapter.
Key Terms in this Chapter
Rural Healthcare: The study of providing healthcare systems and services in a rural environment. There are aspects of policy making as well as technology.
Automatic Speech Recognition: A technology that allows a human being to converse with a computer in order to give some commands or generate a written transcript.
Natural Language Processing: A field of computer science related to automatic processing of human generated language data. It can be used for information retrieval as well as command and control.
COVID-19: A contagious disease which is caused by severe acute respiratory syndrome coronavirus 2. Currently as of 2021, it is a worldwide pandemic.
Machine Learning: It is a branch of artificial intelligence that uses large amounts of data to learn how to classify it or gain some insights from it. In some aspects, it strives to mimic a human brain’s learning from data.
Urdu: It is the national language of Pakistan. It is primarily an Indo-Aryan language, with influence of middle eastern languages.
Medical Diagnosis: It is the practice of determining which disease or ailment a particular person is affected with.