We designed a low-cost system that could stop pandemics
As part of a university course, 4 people posed this problem: Aaron R., Jorge A., Rósalym A. and Pablo D. Why is this not working NOW all over the world?
1. Before we start, let’s clarify some terms and questions
Disclaimer: I am not a doctor, I am here to explain an engineering solution, the explanations are oversimplified. It contemplates a conceptual explanation without excepting the technical part. Spanish version clic here
¿What are epidemics?
Based on its multiple definitions, it basically spreads like a rapidly spreading disease in a community, and the goal of calling it an “epidemic” is to communicate risk.
¿Where epidemics start?
They can appear anywhere, for different reasons, regardless of the nation and we need to be prepared.
¿Patient zero?
In the case of covid-19, detecting the people who were in contact with the first person detected who acquired the virus is essential. However, doing so is extremely difficult, even in our time: years in which everyone carries a mobile phone in their pants pocket.
¿How to stop an epidemic?
Experts mention 7 phases in which governments must contribute to the benefit of humanity to reduce the effects of the pandemic:
(1) containment, (2) prevention, (3) diagnosis, (4) control, (5) treatment, (6) cure — vaccine — retrovirals, (7) mitigation.
The proposed system is intended for the first and second points. That is, at the beginning of a new outbreak, when the curve still appears to be a line without inclination. Currently, with the coronavirus, we are close to stage 6.
So, basically, to stop an epidemic you need to know in detail at what stage we are. That’s where data science and artificial intelligence come in, an expert branch in finding patterns, because “what do infected people share?”: Some symptoms. Of those with symptoms, some self-medicate, some seek medical help, and some just stay home.
Usually, seeking medical help in public or private hospitals is where the outbreak spreads rapidly. And what would happen if a system detected anomalies in the hospital staff? We could stop the outbreak, even with the days apart that it takes time for symptoms to appear.
¿Someone already applies it?
According to the research, it is not used in a common way, there may be particular cases as in some medical centers in China. In the case of Europe and North America, they can implement it without problems. However, the systems implemented in South America must also be low cost and high performance. In other parts of the world, it should work in a similar way.
The system has to work all over the world, it is not an option, it is a must if we want to avoid another covid-19.
2. Our problem in detail
That phrase is one of the most applied in engineering practice. This work was not the exception, we focused on a single health center, only for workers during operating hours, and so on. In the following diagram the sequence of events is conceptualized.
So we need a system that detects epidemics in medical centers automatically, quickly, reliably, and easily integrates. Based on what? in the symptoms of operating personnel; What technologies allow us such a massive tracking? Thanks to digital clouds (Clouds) and Internet of Things (IoT) technology, this is possible. And how do we make it a low cost? There are multiple free open source platforms that have the technology to track 100,000, 1,000,000, and up to 10,000,000 devices at the same time. In the case of Latin America, the operating personnel in a hospital is no more than 1,000.
3. Solution
What do we need to sense? Vital signs. Why? In most cases, biological agents cause alterations in vital signs in symptomatic patients.
3.1. Wearables (Devices that can be dressed)
There are different types of digital sensors aimed at different types of use: hospitalized patients, critical patients, among others. In the research, the devices are oriented towards patients and not towards staff, since the latter do not suffer (or should not) suffer diseases while they are part of the staff. So a device was designed that can sense 3 important vital signs:
Blood pressure, Body Temperature and Respiratory rate.
Due to the difficulty of sensing blood pressure, it is preferred to wear an additional device on the wrist. The Medikit attaches to a GoPro mount as shown in the Figure to the left. Besides being easy to acquire, it provides ergonomics certified by the said company. A little more detail of the two devices is shown in the following Figures.
The approximate cost of the components (MediKit + GoPro Harness + MediBand) is approximately $ 440. More information at the end of the post.
3.2. Web system
Regarding the web system, transparency with the staff must be paramount. That is, such personnel must be able to see the data that their associated device collects. The figure shows the page developed with source code.
The system architecture is a bit complex. In simple terms, there is a database specialized in the mass storage of this type of data, a page visible to staff, an API for third-party connections, and a control panel. Different technologies are used for each component, all oriented at low cost with enviable performance. The diagram also shows the continuous integration, that is, this type of software must be open source so that everyone knows and can contribute to its development.
The panel looks like the following Figure: an administrative staff can review the vital signs of their staff at the macro and micro level. An anomaly detection system based on artificial intelligence contributes to the solution without saturating the computing power of the servers, obtaining the data that the monitoring system receives over time.
The monitoring system also has an alert panel that can be viewed by administrative personnel and that is reported by the artificial intelligence system.
3.3. Mobile app
As for the mobile application, it is an extension of the web service, it communicates with the database through the API of the web page. We currently have the source code ready to be used.
After a simple login, you can check your vital signs, view the history of notifications issued, and change the sample time of the associated device. Synchronization is done using QR codes.
Why mobile application? For several reasons:
- Immediate notification to operational personnel.
- Immediate notification to the administrative staff of a possible outbreak.
- Easy for operating personnel to view their vital signs.
In summary, this system allows us to receive data from a large number of people. Enough for a common medical center of 100 people and distribute it on the website and mobile application. The approximate cost is $ 250 per month. Although the system can be optimized much more and reduce this amount.
3.4. Costs
Devices: $ 440. Servers: $ 250 / month. With these costs, any hospital in a country with a growing economy can implement this solution. This solution is not intended to be final, but rather the first step.
Technical part
You can review the research including the technical part here (Spanish version).
4. The Future
The dream of this work is the creation of a global epidemic detection system, in the first days of the outbreak, regardless of the nation. Of course, multiple features can be added to improve the system:
- Increase the vital signs monitored.
- Mass manufacturing of the devices.
- The decrease in device dimensions.
- Increased battery life.
- Additional support for wireless connectivity wherever they are within the health center.
- Interaction between country systems.
- Sale of data from public centers to pharmaceutical companies so that the review or prevention of drugs that the market will need.
A cool name for the system could be “Human Life Support Against Epidemics (HLSAE)” or “Global Epidemic Detector (GED)”. Is dreaming worth it? Anyway, contributing to that dream with what you can.
