In Ontario's healthcare system, patients and healthcare professionals face substantial challenges that impede the efficiency and effectiveness of healthcare services. To bridge the gap between patients and healthcare professionals, we decided to create a cutting-edge, patient-oriented medical consultation application with an integrated chatbot feature.
Quick Summary
A mobile application that can allow patients and healthcare professionals to have consultations virtually, and allowing healthcare professionals to use integrated chatbot features to get important information in an effective way to speed up some process.
Sagith Kalaichelvam (Front-end) - UX Researcher & Product Manager
Zhiqi Chen (Front-end) - UX/Product Designer
Joseph Chung (Back-end) - AI/ML Developer & Product Owner
Youngho Kim (Back-end) - AI/ML Developer
Bianca DiPietro
January - April 2024
Primary & Secondary Research
Ideation
Persona Building
Content Design
Usability Testing
Market & Competitor Analysis
User Research
Presentation
Live Demo
Healthcare
AI Chatbot
Virtual Consultation
Mobile Application
Agile Methodology
Communication
Product Management
Healthcare Research
UX Writing
User Research
Presentation
App (All Screens - High-Fidelity Prototype Wireframes)
Web (AI Chatbot only - Implementation)
Problem Statement
In Ontario's healthcare system, patients and healthcare professionals are confronted with significant challenges that hinder the efficiency and effectiveness of healthcare services.
For Patients: Patients struggle with long wait times and confusing healthcare processes, despite the system's focus on well-being. Simplifying access and providing real-time support through a chatbot in a medical app can improve the patient experience and reduce delays.
For Healthcare Professionals: Ontario healthcare providers struggle with complex billing under the 'Schedule of Benefits,' needing a user-friendly tool with real-time assistance to ensure compliance and improve patient care. Our solution is a patient-focused app with an integrated chatbot to simplify these processes.
Team Front-end
Literature Review
a. Our in-depth research into Ontario's healthcare system has revealed another challenge faced by healthcare providers, particularly in managing patient consultations and understanding the reimbursement processes of the Ontario Health Insurance Plan (OHIP). According to a critical article on LinkedIn, the OHIP system faces systemic challenges that pose formidable barriers to delivering efficient and accessible healthcare services to Canadians. Moreover, data from the Ontario Health Coalition shows that clinics have wrongfully charged patients more than $42,000 for services that should have been covered by OHIP since 2018. These findings highlight the existing issues within the system, underscoring the necessity of developing an application that aids medical providers in more effectively managing patient consultations and understanding OHIP reimbursement processes.
b. In terms of app design, we referred to articles from Uptech and UXPin, which discuss key principles and challenges in designing healthcare apps for both user groups. They emphasize the importance of user-centred design, clear and intuitive interfaces, and the design of necessary and recommended features for healthcare apps. These design principles are fundamental to ensuring our application meets the needs of healthcare providers while providing a positive user experience.
Market Needs:
Patients to connect with doctors virtually
Doctors can access information easily in an accessible and user-friendly format
Book appointments virtually
Doctors to access accurate and reliable information that is up to date within a few seconds
Patients have access to services at home without paying extra
Competitor Analysis:
Dialogue
Purpose: It is a medical app, where users can connect with doctors virtually
Pros:
Order prescription drugs virtually from doctors
Doctors can easily identify health symptoms based on how the user interacts with the virtual interface of moving their face at certain angles
Book appointments on the app
Cons:
Doctors do not have a resource guide that outputs information easily
Doctors or patients do not see important news/announcements within the healthcare industry
Lack of appointment timings: Most of the time they are fully booked as this app only have doctors who work for this organization
Limited to this organization
Not centrally-design
DynaMed Plus
Purpose: Healthcare professionals uses it to inform clinical care.
Pros:
Easy to use
Thousands of photos/graphics
Micromedex drug content integration
Cons:
Less comprehensive
Not updated that much
Paid subscription for accessibility needs
Joule: CMA App
Purpose: Provides medical tips and tools to healthcare professionals
Pros:
Searchable database Infopoems and InfoPratique, where there is
Access to Clinical Key, videos and journals.
Easy to use
Cons:
Need to remember login (There is no option to save username and password)
Identified Gaps
Although there are several healthcare platforms and tools on the market aimed at enhancing the efficiency and accessibility of medical services, most focus primarily on patient needs rather than addressing the specific challenges faced by healthcare providers. The current system still faces the following gaps:
a. The opacity and issues with incorrect billing of medical fees lead to financial stress and decreased trust among patients. Moreover, there is a lack of a unified platform by the provincial government that could simplify the complexities of medical insurance plans (such as OHIP), allowing healthcare providers and patients to easily access and understand relevant information.
b. While virtual medical services offer convenience, there are clear gaps in guidance regarding the verification process and standardization.
c. There is a lack of research and practice in medical application design focusing on the needs of healthcare providers and accessibility. I seldom see the implementation of accessibility design in any existing tools.
Key Findings
a. The access and billing system for healthcare services requires greater transparency and simplification to reduce wrongful charges and enhance patient trust.
b. There is an opportunity to develop a medical application integrating UI/UX design principles, focusing on improving the workflow for healthcare providers and the experience for patients. Additionally, given the sensitivity and complexity of the medical field, ensuring user data security and compliance with relevant regulations is also necessary.
c. Accessible design is crucial to ensuring that all users, especially those with special needs, can effectively use the medical application.
Developments
In light of these challenges and gaps, our project aims to develop a medical consultation app for patients and another version for healthcare providers, by integrating efficient OHIP management tools and optimizing the user interface according to healthcare service app design principles. This approach is intended to improve the efficiency and quality of healthcare provision. Our application will leverage the latest technological innovations to help healthcare providers quickly answer questions about OHIP, while simplifying the patient management process through an intuitive and user-friendly interface.
Team Back-end
The back-end team embarked on a comprehensive literature review to delve into various themes pertinent to back-end development. This included exploring topics such as tokenization, word embedding, and illustrated transformers. Through academic papers, technical documentation, and tutorial videos, the team gained insights into the latest trends and technologies building up necessary technical knowledge for back-end development.
Identified Gap
Through the literature review, the back-end team identified potential challenges in existing back-end development practices. The major challenge was the integration of external API (Open AI API) with high accuracy. Due to the inherent limitations of generative AI, OpenAI has excelled in natural language generation, but it has shown weaknesses in accuracy, which are particularly important for sensitive topics such as health concerns.
Key Findings & Development
Synthesizing the research findings, the back-end team highlighted key insights and trends relevant to back-end development in the project context. Based on the findings, the backend team proposed various solutions to respond to the identified gaps and build the infrastructure for backend development.
a. Limiting Data for Answer Generation: One key finding is the importance of ensuring the reliability and accuracy of data used to train the AI model. Given the limited data available, it's crucial to focus on utilizing trustworthy sources and implementing algorithms that can effectively filter and refine the data used for answer generation.
b. Tailoring Model Parameters for Accuracy: Another key finding is the recognition that traditional Generative AI models may not be optimized for accuracy with limited data. Generally, generative AI models need to be trained on huge data and the more natural language they learn, the more accurate embedding values are derived. By customizing parameters such as token size, temperature, and sensitivity, the project aims to improve the quality of AI-generated responses tailored to the specific dataset available.
c. Construction of Vector Database for Monitoring: The last key finding involves the need for a robust monitoring system to track the fine-tuning values and intermediate processes involved in training the AI model. Building a vector database allows for efficient tracking and analysis of these parameters, facilitating optimization and refinement of the AI model over time.
Primary Stakeholders - Medical Professionals
The primary stakeholders relevant to our identified issue comprise individuals who are healthcare professionals who have multiple different responsibilities. This includes family doctors, dentists, optometrists, healthcare providers, etc. These stakeholders share a common interest in understanding and applying OHIP billing after taking care of each patient. They will be busy with different responsibilities such as scheduling, patient management, and accessing medical records. Want to understand what is covered in a patient's OHIP coverage depending on their health condition. They will very often manage their healthcare facilities.
Job Stories:
1. As a healthcare provider in Ontario, I need a streamlined way to manage consultations and patient interactions. I want an online platform that enables me to offer timely medical advice and consultations efficiently, reducing the need for in-person visits and minimizing administrative overhead.
2. As a family doctor, I require an efficient system for understanding and applying OHIP billing codes for various services. I need a tool that simplifies this process, offering quick access to billing information to ensure accurate and timely compensation for my services.
3. As a medical professional unfamiliar with the latest changes in OHIP coverage, I need a straightforward, interactive tool to help me stay updated on coverage details. I want a chatbot that can instantly provide me with accurate information on OHIP coverage, enabling me to make informed decisions about patient care and billing.
4. As a healthcare provider juggling multiple responsibilities, I need a comprehensive solution for scheduling, patient management, and accessing medical records. I want a unified app that integrates these functionalities, making it easier for me to manage my workload and enhance patient care.
5. As a healthcare provider seeking to improve patient education on health and wellness, I need a platform that allows me to share tailored educational content. I want an application that enables me to provide my patients with resources that are relevant to their health needs and lifestyle, fostering better health outcomes.
Design Requirements:
1. Convenience: The solution must prioritize convenience by being usable and easily accessible to users without imposing a greater obstacle to their busy lifestyle.
2. Wait Time Reduction & Online Virtual Appointments: The solution must minimize wait times for in person visits by enabling individuals to schedule and meet with a doctor virtually for general and other questions that can easily be solved at home without the need to go in person.
3. Accuracy and Medical History/Reports: The solution must include guidance on how to improve their health condition based on their recent medical report. The solution must allow users to track and visualize their health progress to manage it.
4. Easier to apply for Direct Billing to OHIP covered Patients: The solution must be easy to navigate that can allow healthcare professionals to easily charge OHIP covered patients to the government.
Deployment
Key milestones
a. Development Environment Setup: Setting up the development environment involves configuring essential tools and frameworks for back-end development. This includes selecting programming languages, frameworks, and database systems.
b. Deployment of Monitoring System: Implementing a monitoring system is vital for tracking application performance, especially during training in AI projects on given documents. A comprehensive monitoring tool was necessary for efficient tracking system performance facilitated proactive management and ensured smooth operation.
c. Integration with OpenAI API: Integrating the OpenAI API into the backend system involves seamlessly incorporating its functionality. The system had to be compatible and reliable in both local and network environments, which required proper authentication, error-handling mechanisms, and data transmission protocol.
d. Prototype Development: Prototyping entails creating initial versions of the application to test functionalities and gather feedback for iterative refinement. To maximize resource utilization, focus areas were carefully selected to meet the minimum viability for demonstrating project foresight.
e. Algorithm Refinement: Algorithm refinement is an iterative process aimed at optimizing AI response accuracy and reliability. This involves continuous evaluation and adjustment of model parameters based on performance feedback and experimentation. A comprehensive monitorization system and development environment set in previous steps were utilized to leverage performance
f. Integrate Test Feedback: Updates in UI integration based on user testing feedback. This involves analyzing user input, understanding their needs, planning and implementing changes, and testing the updates. This iterative process aims to enhance the user experience and ensure the interface meets evolving user expectations.
Technologies employed
a. Development Environment: Utilizing Visual Studio as the integrated development environment (IDE) for back-end programming tasks. JavaScript will serve as the primary programming language for implementing backend logic and functionality. Github was utilized for version control and progress synchronization.
b. Monitoring System: Employing Pinecone Vector Database for monitoring system implementation. Pinecone's vector database technology offers efficient tracking and analysis capabilities for fine-tuning values and intermediate processes during AI model training.
c. Model Integration: Integrating the OpenAI API into the back-end system to facilitate natural language generation tasks. Leveraging the OpenAI API as the core model for generating AI-driven responses based on inputs from the OHIP dataset.
d. Data Source: Incorporating OHIP's "Schedule of Benefits Physician Services Under the Health Insurance Act" as the primary data source for training and refining the AI model. This dataset will provide the foundational information necessary for building the prototype of the project.
Integration
During the UX Research step, we have all worked together on ideating ideas that will help us to deliver a successful product based on the needs and wants of selected stakeholders that our MVP primarily focuses on. Then, the UX Researcher of this project created a variety of different personas, user journey mapping, design requirements, job stories, competitor analysis, market needs, etc.
The UX Designer of the project followed all the criteria that were done through the UX Research steps to ensure the product can be delivered in a user-friendly and accessible environment that meets the needs of the stakeholders. The UX Research steps help the UX Designer to follow and organize their steps to ensure the product aligns with the MVP main idea, and scope
UI integration is a pivotal aspect of software development that bridges the gap between the frontend user interface and the backend system. The seamless connection and synchronization of frontend design elements and functionalities with the backend system and data ensures a cohesive and intuitive user experience. Smooth communication and close collaboration between the backend and frontend teams was essential in this process
The primary objective of the integration process was to enable users to interact with the application without being aware of the underlying complexities of the system. The system and design were both analyzed through rigorous testing to achieve optimal resource utilization while still offering essential features and functions. Ultimately, the integration efforts were geared towards delivering a user-friendly interface that prioritized simplicity for demonstrating the future of the project.
Figma Prototype
Quick Snapshot of AI Chatbot Functionality
This is a project that I was very interested to work with as its related to healthcare, which is something that is important right now in the province of Ontario. After researching about patient wait times at healthcare professional offices, I realized that many patients like me have a feeling of waiting too long for a health concern or other medical related actions.
I have gained a more experience in User Research by ensuring research is aligned with the current market and ensuring the team's progress on the technical side is aligned with the research that I have been working on.
Overall, this project have made me to learn so many greatful things while researching and collaborating with my team members. I had the opportunity work on a 4 months project that felt similar to a real project that you will do in the professional work industry. I had a fun time attending lots of weekly stand-up weekly meetings on MS Teams with my team members, and supervisor that is involved in this short-term project. This experience have made me to expand my interest of pursuing a career within UX Research and Product Manager.
Reference
Chandrasena, C. (2018). Medical Apps for Physicians and Patients. Presented at the OntarioMD EMR Every Step Conference, Toronto. Retrieved from https://www.ontariomd.ca/documents/events%20info/esc%20-%20toronto%20-%202018%20-%20presentations/medical%20apps%20for%20physicians%20and%20patients%20cchandrasena%20ontariomd%20emr%20every%20step%20toronto%202018.pdf
Government of Ontario. (n.d.). Ontario Launches New Tool to Connect People to Nurses and Other Health Services from Anywhere, at Any Time. Retrieved from https://news.ontario.ca/en/release/1002095/ontario-launches-new-tool-to-connect-people-to-nurses-and-other-health-services-from-anywhere-at-any-time
Najafi Mehri, S. (n.d.). A Critique of OHIP Canada: Systematic Challenges. Retrieved from https://www.linkedin.com/pulse/critique-ohip-canada-systematic-challenges-soheil-najafi-mehri/
Ontario Health. (n.d.). Digital and Virtual Services. Retrieved from https://www.ontariohealth.ca/getting-health-care/digital-virtual-services
Ontario Health. (n.d.). Virtual Visits Verification. Retrieved from https://www.ontariohealth.ca/system-planning/digital-standards/virtual-visits-verification
Ontario Health Coalition. (n.d.). Over $42K in Wrongful Fees Charged to Ontario Patients by Clinics Since 2018. Retrieved from https://www.ontariohealthcoalition.ca/index.php/over-42k-in-wrongful-fees-charged-to-ontario-patients-by-clinics-since-2018/
UXPin. (n.d.). 7 Best Practices for Designing a Healthcare App. Retrieved fromhttps://www.uxpin.com/studio/blog/healthcare-app-design/
Uptech. (n.d.). UX Design in Healthcare: A Comprehensive Guide. Retrieved from https://www.uptech.team/blog/ux-healthcare
Virtual Health Clinic. (n.d.). Common Visits in Ontario. Retrieved from https://www.virtualhealthclinic.com/common-visits-ontario