Pre-Consultation Assistant

Filled Input - Section 1
At-home care Recommendations - Section 2
Appointment Preparedness - Section 3
Positive energy through a Haiku - Section 4

Inspiration

Patient Activation and Interaction: Many patients find themselves at a loss when it comes to effectively interacting with healthcare providers. The intricacies of articulating their medical history, symptoms, and concerns often pose significant barriers to productive communication. This lack of patient activation can hinder the delivery of optimal care and impede the establishment of a collaborative patient-provider relationship.

Inadequate At-home Care Guidance between the start of symptoms and appointment: In the intervals between medical appointments, patients are often left without clear guidance on how to manage their health and well-being. The absence of structured support and resources for at-home care can leave individuals feeling overwhelmed and uncertain about the steps they should take to address their health concerns.

Our app empowers patients by bridging the gap between them and healthcare providers. We offer intuitive tools for better patient activation and informed decision-making. With comprehensive at-home care guidance, we aim to enhance patient confidence in managing their health between appointments, fostering improved outcomes and stronger relationships.

What it does

The app is divided into 3 sections:

Patient History and follow-up Chat: Patient History and Follow-up Chat lets users record their medical history and ongoing health issues. The chatbot ask questions as per user symptoms and history. They can update information in symptoms and history widgets, and receive follow-up questions accordingly, simplifying communication with healthcare providers.
At-home Care Recommendations & Consultation Preparation: Offers personalized at-home self-care advice and prepares users for medical appointments. It provides guidance and checklists to empower users to manage their health effectively. Based on the user's symptoms, medical history, and basic information provided following recommendations are created:
- At-home self-care recommendations.
- At-home basic tests, if needed.
- Dietary recommendations.
- Top 5 questions to ask the doctor during the appointment.
Positivity Dose: Delivers uplifting Haiku to users, ensuring it aligns with their emotional state and doesn't conflict with their well-being. It offers tailored doses of positivity to promote mental wellness and resilience.

How we built it

The app is built with AWS Partyrock utilizing all three capabilities of widgets:

LLM-based text Generation.
Intuitive chatbot widget to ask follow-up questions.
Image generation from Haiku to engage patients more with a dose of positive energy.

Challenges we ran into

The obstacles we faced revolved around handling free-form text input and leveraging Language Models (LLMs), making imposing constraints on input and output a formidable task. For this we added conditions in our prompts i.e:

In at-home recommendations, we added to our prompt that it should not recommend any medicine to the patient as during interaction with LLM, we aren't sure if the patient is hallucinating or have any other underlying conditions that are still not presented to LLM.

Accomplishments that we're proud of :

We take pride in several accomplishments, including the successful implementation of innovative features, our app's positive impact on users' lives, and the recognition and appreciation received by getting a user experience survey done by colleagues.

The app is aimed toward pre-consultation assistance but could be used for everyday home care recommendations and the haiku section also brings engagement.

What we learned

Throughout the process of making the partyrock app, we've learned invaluable lessons about user engagement, the importance of adaptability in technology, the significance of user feedback, and the intricacies of healthcare applications. We learned that LLMs fail in most of the cases where the patient is hallucinating by getting a review on our app from a Doctor. These lessons have helped us refine our approach and prompts, enhanced our understanding of user needs, and continually improved our app to better serve the user.

What's next for Pre-Consultation Assistant

Next for Pre-Consultation Assistant involves following:

Integrating advanced AI for further personalized recommendations.
A feature to trigger some other APIs or call an ambulance in case of emergency
Generating a report from the collected history, through patients' input widgets and chatbot answers, which they can bring at the time of appointment with them.