MediDate

Notifications Page
File Upload Page
Prescription Description Page
Final
Before

Inspiration

One of the 6 most common medication problems in an ageing population comes from the scheduling and burden of taking several medications several times a day. At best, this can be a hassle and an annoying process. However, it is often more likely than not that many may simply forget to take certain medication without supervision and reminders which may result in further deterioration of their health. In order to address this issue to make living healthy a smoother process for the ageing population as well as provide better support for their healthcare providers, MediDate was born. Designed with the user in mind, the UI is simple and intuitive while the hardware is also clean and dependable. The diversity of features ensures that all aspects of the medication and caretaking process can be managed effectively through one comprehensive platform. A senior citizen now has a technological solution to one of their daily problems, which can all be managed easily by a caretaker or nurse.

What it does

MediDate is a combination of hardware components and a web application. The hardware aspect is responsible for tracking dosage & supply for the patient as well as communicate issues (such as low supply of medication) to the web application and the caretaker. The web application is made up of several different features to best serve both patient and caretaker. Users are first brought to a welcome page with a daily schedule of their medications as well as a simulation of the pillbox below to keep track of total medication supply. When the web app detects that supply is below a certain threshold, it will make calls to local pharmacies to reorder the medication.

Along the side navigation bar, there are several features that the users can take advantage of including a notifications page, monitoring pharmacy orders, uploading new subscriptions, and descriptions of their current medication. The notifications page is pretty self-explanatory, it keeps track of any notifications patients and/or caretakers should be aware of, such as successful prescription uploads, low medication supply, and errors in uploads. The upload page allows users to take photos of new prescriptions to upload to the web app which will then make the appropriate processes in order to add it to both the schedule and the explanation bar through RX numbers, dates, etc... Finally, the prescription pages offer quick shortcuts for descriptions of the medication to make understanding meds easier for users.

In order to be as accessible as possible, an Alexa skill has also been created to support functionality from the web application for users to interact more directly with the caretaking solution. It currently supports limited functionality including querying for today's prescription, descriptions of different medication on the patients' schedules, as well as a call for help function should the need arise. This aspect of MediDate will allow more efficient service for a larger population, directly targeting those with vision impairment.

Another feature was integrated using Twilio's SMS API. For the convenience of the user, a notification text would be sent to a registered Pharmacy phone number with details of prescription requirements when the current pill inventory fell below an adjustable threshold. Pharmacies could then respond to the text to notify the user when their prescription was ready for pick-up. This enables seamless prescription refills and reduces the time spent in the process.

How I built it

Hardware Powered by an Arduino UNO, buttons were attached to the bottom of the pillbox to act as weight sensors for pills. When pills are removed, the button would click "off", sending data to the web application for processing. We used CoolTerm and a Python script to store Arduino inputs before passing it off to the web app. This aspect allows for physical interaction with the user and helps to directly manage medication schedules.

Google Cloud Vision In order to turn images of prescriptions into text files that could be processed by our web app, we used Google Cloud Vision to parse the image and scan for relevant text. Instead of running a virtual machine, we made API calls through our web app to take advantage of the free Cloud Credits.

Backend Scripting was done using JavaScript and Python/Flask, processing information from Cloud Vision, the Arduino, and user inputs. The goal here was to send consistent, clear outputs to the user at all times.

Frontend Built with HTML, CSS, bootstrap, and javascript, the design is meant to be clean and simple for the user. We chose a friendly UI/UX design, bright colours, and great interface flow.

Alexa Skill Built with Voiceflow, the intents are simple and the skill does a good job of walking the user through each option carefully with many checks along the way to ensure the user is following. Created with those who may not be as familiar communicating with technology verbally, MediDate is an excellent way to integrate future support technologies seamlessly into users' lives.

Twilio SMS The Twilio SMS API was integrated using Python/Flask. Once the pill inventory fell below an adjustable pill quantity, the Twilio outbound notification text workflow is triggered. Following receipt of the text by pharmacies and the preparation of prescriptions, a return text triggers a notification status on the user's home page.

Challenges I ran into

Flask proved to be a difficult tool to work with, causing us many issues with static and application file paths. Dhruv and Allen spent a long time working on this problem. We were also a bit rusty with hardware and didn't realize how important resistors were. Because of that, we ran into some issues getting a basic model set up, but it was all smooth sailing from there. The reactive calendar with the time blocks also turned out to be a very complex problem. There were many different ways to take on the difference arrays, which was the big hurdle to solving the problem. Finding an efficient solution was definitely a big challenge.

Accomplishments that I'm proud of

Ultimately, getting the full model off the ground is certainly something to be proud of. We followed Agile methodology and tried (albeit unsuccessfully at times) to get a minimum viable product with each app functionality we took on. This was a fun and challenging project, and we're all glad to have learned so much in the process.

What's next for MediDate

The future of MediDate is bright! With a variety of areas to spread into in order to support accessible treatment for ALL users, MediDate is hoping to improve the hardware. Many elderly also suffer from tremors and other physical ailments that may make taking pills a more difficult process. As a result, implementing a better switch system to open the pillbox is an area the product could expand towards.

Built With

Submitted to

McHacks 7
- Winner Novartis

Created by

I worked on the front-end and the design of our web-app, using HTML, CSS, and Bootstrap. I used JavaScript to update the DOM using the information provided by the Python/Flask back-end. I also integrated the use of Twilio's SMS API using Python to automatically send notification texts when pill inventory fell below a minimum threshold.

Allen Lu
Backend Team Lead @ Hack the North
I worked on the Arduino build and programming for this project, constructing our cardboard monstrosity and it's C++ coding on the Arduino IDE. I also developed the reactive calendar for the Home page, which was built through HTML/CSS and JavaScript.

Tailai Wang
I worked on the back end configuring Google Cloud Vision to pick out key words from images using Python. Integrated multi-threading to read real-time input from Arduino Uno while continuously relaying that information to the client using FLASK and AJAX. Used JavaScript to format and update information on the client side.

Dhruv Shah
Ada H