Improving Personal Health Quality Index

Using technology to easily track medical vitals like blood pressure, weight, body temperature, etc., and to share that information with doctors and caregivers for the elders and physically challenged.

According to WHO 2022 statistics, noncommunicable diseases (NCDs) kill 41 million people each year, equivalent to 74% of all deaths globally mostly in low to middle income nations. The same report cites that “ In terms of attributable deaths, the leading metabolic risk factor globally is elevated blood pressure (to which 19% of global deaths are attributed) (1), followed by raised blood glucose and overweight and obesity.” Detection, screening and treatment of NCDs, as well as palliative care, are key components of the response to NCDs. 

For my product, I have used technology to make it easier for users to track their medical vitals like blood pressure, weight, body temperature, etc., and be able to share that information with their doctors and caregivers. I decided to work on this project because COVID made me aware of how important it is to have automated and intelligent medical healthcare systems.

Before the stay-at-home period, many people visited the doctor’s office regularly but during the pandemic, people were scared to go to the doctor’s office, let alone walk outside their house. My grandmother is from India and during COVID she was isolated. She is not very familiar with texting and cannot read or write English. Having to keep track of her health vitals every day was a major concern. I observed that most medical device readings [Blood Pressure, Temperature, Oxygen Levels, Sugar Levels, Weight machines, etc] are in English and in the seven-segment digital display format. While at my doctor’s office the nurses always used paper and pen to note down the readings from the medical devices.

What if there was a system that could help medical experts and family members keep track of the patient’s health while the patient stays at home? Can we free up doctors’ and nurses’ time and office space so the office resources can be used more efficiently for patient care? Can the repetitive task of taking a patient's health vitals be automated? Can the daily vitals be made readily available to the immediate family or caretaker? This system would be the perfect way for medical experts to keep a close eye on their patients without having to see them in person and having the risk of getting secondary infections, especially those who are taking immunosuppressant medicines. The collected data could also provide early signs of other comorbidities. These experiences led me to believe that -

a) Everyone should get their health vitals checked periodically from home, rather than waiting for the next doctor’s appointment.

b) Keeping track of our daily health readings and providing this information to the health professional can help with early detection and cure/management of diseases.

c) The system needs to be cheap and should integrate easily with our existing devices [smart watches, phones, blood pressure meter, etc]

d) With close to 5+ billion people already using a phone with a camera, this system can easily be made available to everyone in a short period of time.

I decided to address the problem of easily tracking daily health vitals using intelligent technology.

The solution I am proposing would use an app on a phone to take pictures of the health monitoring device results or health information from smartphones or smartwatches. It would then read, interpret, and store this data. A portal can then display the information and provide the analytics on the collected data. A medical expert can check the portal or get automatic alerts and depending on the vitals the portal can automatically schedule an appointment for further examination. APIs [Application Programming Interfaces] can be written to sync the data to the doctor’s system, and could also alert urgent care in case of immediate medical attention.  This app can also help connect the patient to a specialist or a doctor remotely, this will open up medical care to remote places where there aren’t enough doctors or specialists.  The app also gives an ability to share health related data to care takers or family members so remote family members can keep an eye on the progress of their loved ones’ health.

This system enables a proactive health care service that medical professionals can use to monitor their patients remotely and use the technology to track the progress of a patient’s health. With 5+ billion people already using a phone with a camera, this system can easily be made available to everyone in a short period of time. Tracking daily vitals, trending on them, and having them accessible to medical experts can help early detection of diseases. This solution also allows people to remotely keep track of the health of their elderly family members.

The Health Quality Index will collect and trend on:

1. Blood Pressure

2. Blood Sugar reading

3. Temperature

4. Weight

5. Oximeter Reading

6. Heart Rate

7. Exercise monitor Reading, etc.

It can be further expanded to specialized tracking for Arthritis progression, skin cancer detection and much more. Being able to measure the health quality index for individuals will help governments to measure the health of the population. Currently, we have no system like this. This will help the government make the right policy decisions.

I have been doing coding and robotics since 3rd grade.  I have good experience in working on hardware and coding.  Currently, I can code/use 10+ programming languages.  I am proficient in C++ and Python.  On the hardware side, I have been doing robotics for the last six years.  I have competed and won state level robotics competitions.  I am proficient in building and programming robots.  For the last two years, I have also been making robot parts in our garage and school labs.  The experience in hardware and software has helped me work on this product and I was able to build a prototype to show how it can operate.  

My inspiration to build this product came through observations of how my family and neighbors were reacting to lockdowns.  My grandmother lives with me and I help her with her day to day chores.  She has arthritis and it is a progressive disease.  We needed to do a regular checkup on her and adjust her medications accordingly.  Due to COVID, my grandmother was not able to go to the doctor's office.  She was taking immunosuppressant medicines and that made her a high risk patient for COVID.  We had to take utmost care of her during that period.  During this time, I wondered what the impact was on all individuals who were taking immunosuppressant medicines.  I wanted to build a product that could help patients avoid unnecessary visits to the doctor's office, thus reducing the chances of getting infections and also help the doctors with their schedules by reducing the number of appointments they have with their patients. This will help them to see more patients and reduce the wait time for the next appointments. 

My first step was to check and see if this was a possible solution technically, and I experimented with it with my grandmother, her friends, and my parents.

They all had feedback for me, and I have been keeping an engineering notebook to try and incorporate the feedback. This process is ongoing.

In addition, I have also sent out surveys to my close family and friends. In total, I got 30 responses as broken down by this table

Based on this response, I focused on blood pressure monitoring first as that was the highest priority. Now I am focusing on weight measurements.

I have not worked with any organizations yet, but I plan on reaching out to my local community elderly and see if this product helps them to monitor blood pressure readings at least.

Currently, there is no solution that helps patients to self-monitor their health vitals and at the same time store, trend, do predictive analysis, and share that information with doctors and loved ones.

For example, if a senior citizen or a patient with chronic illness is living on their own, it would be great to know their vitals, like blood pressure, sugar levels, body temperature, etc, regularly and not wait till the next doctor's appointment.  Currently, a patient can use medical devices to measure these vitals at home, but there is no system that collects all the information, stores it, and makes it easily available.  

In my solution, the patient's vitals are easily available to the doctors or loved ones on a periodic basis.  This will help keep everyone aware of the patient's health, impact of a change in medication, and a long term view on changes to the patient's health.  My solution can also be enhanced to use AI to predict the progressiveness of arthritis, onset of skin cancer or other diseases.  This can change the way we do health care and more importantly it gives easy access to improving health to the underprivileged and people in third world countries.

My short-term goal is to have this product available to individuals who are 65+ with certain medical conditions and are at risk of getting a secondary infection.  The impact will be to reduce the risk of these individuals getting sicker considerably, check the progression of the current disease, and keep their loved ones and doctors updated on their health.

For example, a patient with arthritis taking immunosuppressant medicines is at risk of frequent visits to the doctor's office as he/she can easily get infected with new strands of virus or bacteria.  Also, using this product, he/she can take regular pictures of the hands to capture the progression of arthritis and whether the medicines are working or not.  Since the data is owned by the patient, the patient can then easily consult with any doctors across the globe to get second opinions.  This will give access to medical care and advice to patients at remote locations where doctors and specialists are not easily available.  Additionally, patients will have the ability to share his/her health information to loved ones who are remote and those loved ones can keep an eye on the health of an aging parent or family member.

Approach to the solution:

I had multiple approaches in mind to solve this problem. I wanted it to be able to capture health readings from various devices, which are in 7-segment display format. I started with trying to get the blood pressure meter reading captured, decoded, and stored.

Potential Solution # 1 -
The first solution that I came up with was to connect a Bluetooth-based
processor [Figure 2] to the blood pressure machine and try to send the
reading to a server. This solution involved using a processor and would
cost over $200.

Negatives - Hardware+Software are needed. Too many active parts thatcan potentially fail. The solution is very expensive.

Positives - Less work for the patient, since the information will be automatically taken from the device and the patient doesn't have to do anything. End to End automation.

Potential Solution # 2 -
The second solution that I thought of was a software-only solution of using an app and image processing intelligent software to decode the readings. For this solution, the user would need to take the pictures using the app. The app will upload the picture to the image processing code, and the data from the photo will be reflected on the user's portal account.

Negatives - Patient needs to be able to take the picture

Positives - Cost-Effective, easy to use, fewer active gear, and hence lower chances of gear failing or needing repairs. I ultimately selected Solution #2 because it was cost-effective and could be easily deployed to
make an immediate impact.

PROJECT DESIGN
I divided the project into 3 parts -
1. Image Processing
2. Database Design
3. App Design - FrontEnd and User Interface

1. Image Processing [Rosebrock, Adrian. “Recognizing Digits with OpenCV and Python - PyImageSearch]
The first piece of software that I worked on was trying to read the image and get the readings into an integer or text. I used my machine as a python server on port 8888 to accept connections from the app on the phone.

I tried 2 software solutions -

I. Tesseract: It is an existing OCR [Optical Character Recognition] solution to
recognize text from an image. This solution did not work for me because the digital display has a unique issue. A sample digit is shown in Figure 4. When you look at an individual digit, it is not a continuous digit, there are gaps. Tesseract was not able to recognize the digits and it needed a lot of preprocessing which made it more difficult to get the right readings. So I looked at different options and decided to go with Python’s Open Computer Vision module - OpenCV2.

II. OpenCV2: I used this software ultimately to recognize the digits. The images were preprocessed because the input pictures lacked clarity. CV2 was also easier to train, compared to Tesseract. I am still having problems reading some images and numbers, but CV2 still gives me a number, unlike Tesseract which gave me symbols.

After installing the required software libraries, I coded the image processing logic in Python. The steps for the image processing code are:
1) Receive the image from the app
2) Run the image through Image Processing code [explained below]
3) Take the digits retrieved from the image and connects to the database
4) Upload the data to the database

Image Processing Logic - 

Step 1 - First, the image processing code receives an image of a device with the results of a test.
Step 2 - The image is turned black and white so that it is easy to recognize the segments, As you can see in the image below, the gaps are very prominent. To recognize number 8 these gaps need to be filled.
Step 3 - After the grayscaling, the image is then blurred a couple of times. This is to make the digitized font turn into smooth edges.
Step 4 - Once the image is blurred the code then tries to find the LCD screen of the device so it can focus on the numbers when finding the contours.
Step 5 - Next is finding the contours. These are the outlines of digits. This is why we narrowed it down to the LCD screen so when the contours are getting formed, the numbers are the only contours that meet the requirements.
Step 6 - After the contours are found, the image is sliced into 7 parts to determine which segments are on. This is done by comparing the number of pixels that are on and off within the area of the slice. The on/off segments are then compared to the reference list to determine the number.

Step 7 - The numbers are then uploaded to the Bluehost database (Hosting Server).

(1, 1, 1, 0, 1, 1, 1): 0,
(0, 0, 1, 0, 0, 1, 0): 1,
(1, 0, 1, 1, 1, 1, 0): 2,
(1, 0, 1, 1, 0, 1, 1): 3,
(0, 1, 1, 1, 0, 1, 0): 4,
(1, 1, 0, 1, 0, 1, 1): 5,
(1, 1, 0, 1, 1, 1, 1): 6,
(1, 0, 1, 0, 0, 1, 0): 7,
(1, 1, 1, 1, 1, 1, 1): 8,
(1, 1, 1, 1, 0, 1, 1): 9

2. Database Design
Data output from the image processing step needed to be captured for trending purposes that would be displayed on a portal. The minimum information I needed to capture was - User info, the medical device that is used, its readings, and the timestamp.The database needs more development work but I decided to keep it simple. I focused more on making the end-to-end prototype work.

3. AppDesign - FrontEnd, and User Interface 

A. FrontEnd - App on the Phone 
The app is created using MIT app inventor 2. It allows the user to take a photo of a device with the results after testing. Here are
the quick steps -
1. Take the blood pressure reading
2. User logs on the app
3. The user takes a picture of the reading on a device
4. The picture is then transferred to a server that does Image Processing

B. User Interface -
These are the consideration for the Portal Design:
1. Find a hosting service to host the web server
2. Determine an easy way to build a portal
3. The portal should have the ability to maintain user privacy [unique access to each user]
4. Build a database to store user information
5. Ability to display analytics of user data when a user logs on

The web server is hosted on Bluehost hosting services. The portal is built using WordPress. The portal allows a user to create a userid/password and displays user-related information.

With an app that can take a photo of a user’s device results, an image processing code that can read the medical device results, and a portal that can store user data securely and display the analytics, this solution can easily and economically help everyone to proactively take care of their health without leaving the comfort of their own home. Doctors and nurses will also have an efficient way of tracking their patients' health and record the data for their check ups, thus freezing up time and space during the in-person visit.

My solution is still in the early stages.  I have experimented with this with few family and friends.  Also, I have used MIT App Inventor to create the app.  

Within a year, I would like to - 

1. Complete building the app for Apple IOS and Android OS.

2. Build a backend engine and database to store the user data.

3. Understand and build user privacy features in my app.

4. Make the app available to a limited user base as a pilot, up to 10,000 users, to test its features and resolve any bugs.  Depending on the success of the app, I can expand it to a larger user base.

5. Someone to help with finances - getting adequate funds and managing a sustainable business model.

Below are few of the barriers that I am currently anticipating - 

Financial - I will need funds to host this app and the data on some hosting server.  Additionally, I will need funds to buy a lot of medical devices: blood pressure meters, glucose measuring meters, etc. Additionally, I will also need to hire a few resources to help me build this product.

Technical - I will need to hire resources who have expertise in User Experience, AI/Machine Learning, managing big data, image processing, XCode, and Android programming.

Legal - I will help in getting the patent/copyrights for this app.  Also, I will need to understand the HIPAA laws and local country privacy laws to ensure my app is compliant.

Cultural - I don't anticipate a big issue here but I do understand that many places don't have the medical devices to measure vitals at home.  But, the app could be used to take pictures and can be used for predictive health features like how progressive arthritis is or whether a particular growth on the skin is skin cancer or not.  So depending on the local economic conditions, the app can be used in different ways to improve an individual's Health Quality Index.

Market Barriers - In my research I have not seen any product like mine in the market.  We have health measuring devices like watches etc but there is no integration of that data with the system that the doctors use or that information is not available to the loved ones.

This is a self starter product.  I do not have any partner that I work with.

Health is very important but it gets neglected due to everyone's busy life.  My goal is to make it easy for individuals to maintain good health by keeping an eye on their Health Quality Index.  Few simple steps one can take to ensure health is measured on a periodic basis and we are able to do predictive analysis on the collected data related to the health of the individual.

My business model is to provide this app for free through smartphone apps.  Measuring and tracking one's health should not be a financial burden on anyone.  

The immediate beneficiary of this product is the end user, but it will also help the doctor's office and the staff as it will take some work out of them since the user will be keeping a constant check on their health.  This will also help fewer visits to doctor's offices and fewer cases of secondary infections in the patients who are taking immunosuppressant medicines.  Another help will be to the local community and the government as this app will help find health issues at their onset and will avoid/reduce huge future medical bills as they manage or treat diseases in their early stages.

Since the product is going to help reduce costs to doctors and the government, I am thinking of getting grants from governments as the product matures and shows impact.

To jump-start this product, I will be seeking funding from investment capital, individual donations, etc.