MOSA

MOSA (Medicine Outcome Simulation Algorithm) aims to provide a way of predicting all possible outcomes and side effects, at any given time, after applying medication to a human body. 

MOSA´s purpose is to predict and simulate possible outcomes, such as side-effects and level of protection, of a vaccine (and hopefully any medicine, in the future) in the human body.

To accomplish this task, MOSA will be built as a machine learning algorithm. By using datasets of outcomes of vaccines that have been already produced, we are seeking to train the algorithm to predict results by taking information from the vaccine and predicting all the possible ways it could interact with a human body.

Because of the nature of the interaction, we consider MOSA should be applied in an individual way because not everyone would have the same immune response. However, in the initial stages of MOSA, we will be looking into only applying to healthy adults and later on to make it more individualized.

The world has seen a difficult challenge since the start of the pandemic caused by the Covid-19 virus. Multiple strategies and restrictions have been put in place to avoid the excessive contagion of the population to not overwhelm hospital capacities, and multiple medical institutions started the development of vaccines to protect individuals from falling severely ill due to the virus, dramatically reducing the possibilities of dying.

When the vaccines were available to the majority of civilians in early to mid-2021, we saw a lot of uncertainty from some people debating whether they wanted to get vaccinated or not. In some cases, some individuals were primarily worried about the long-term effects that the vaccine would have on their bodies; doubting specifically the fast approval and production of these vaccines. Evidently, fast production was necessary to provide a swift solution to the crisis the world is facing. 

However, for some people, this doesn't entirely remove the concern, and fear of the long-term effects and it could be one of their main reasons as to still not getting vaccinated. MOSA could provide a way to address this concern, and encourage vaccination.

While that was the initial motivator of this solution. We are hoping to develop MOSA beyond solving this specific problem. We believe that it could address the concern for parents that don´t want to get their children vaccinated; It could help for faster production and approval of vaccines if another pandemic happens in the future; amongst many other benefits.

The primary population that we would like to serve, in these initial stages of MOSA, are healthy individuals that haven´t yet been vaccinated against Covid-19 because they are concerned about the efficiency and safety of the vaccine. MOSA would work case-by-case, and it could show these individuals what they can expect (in terms of side-effects after getting vaccinated, long-term effects of the vaccine, and level of protection that their body would have) after getting vaccinated.

This solution comes to mind because we are aware that while there are a lot of data and information that ensures the safety of these vaccines from the FDA, the institutions developing the vaccines, WHO, amongst others. Misinformation and mistrust heavily impact the decisions of some people; providing a new source of information that is personalized to their body´s reaction could potentially help and reassure the safety of the vaccines.

We have talked amongst the people we know about their concerns about getting vaccinated. The reasons why they would be fearful of the vaccines and what things would encourage them to get vaccinated.

A good majority of the response was that they were concerned about the fast development of the vaccine. A good majority of the people that had this concern, also said that they would wait some months to get vaccinated to see if anything concerning about the vaccines would come up.

Most of the people that gave us a response are currently fully vaccinated. However, for some of them, the reason to get vaccinated came with an external incentive (like government policies that restrict unvaccinated from traveling, or the companies they work at gave them an incentive such as a fully paid day off to get vaccinated). Those that have not yet gotten vaccinated, told us that they do not want to get vaccinated because they do not trust the vaccines, they are afraid of their own bodies' reaction to it, or they simply do not trust western medicine.

There has not been further research on the population we would like to serve. We are also considering what we would like to see and how this product would serve us if we were thinking of ourselves as patients that have doubts about medications and vaccines.

So far we have been doing research about what information we need in the two areas that MOSA has. First the technological part of it, we have been looking at machine learning models. We have taken supervised, semi-supervised and unsupervised models, and conducted a much smaller task to see how efficient and appropriate the model is for the purpose of our solution.  

As the team members are mainly experienced in the technological side of the solution, we have spent the majority of the time undertaking research on the medical side of MOSA. This has helped us further understand what machine learning model we can use, as well as, the type of results we are expecting.

We are following the paper "Methods for predicting vaccine immunogenicity and reactogenicity" by Patrícia Gonzales-Díaz, et al. (can be accessed here), as the initial guide to developing our machine learning algorithm, as it outlines how to accomplish this task. We are following other similar papers to ensure a deep understanding of combining technology and medicine.

The main path to achieve the purpose of MOSA is to use Machine Learning Algorithms and Artificial intelligence to create software that is able to simulate possible reactions of a human body when its given medications. As the solution is aimed to provide a very specific outcome to each user, it is expected to be able to help everyone around the world, with any medication that will interact with their body. However, it would be ideal to test the initial states of MOSA with healthy adults, who are yet to receive any dose of the Covid-19 vaccine.

At the moment MOSA does not serve any person as it has not been launched yet. However, the initial stages of MOSA will serve people who have not yet been vaccinated against Covid-19. Later stages of the solution aim to serve more people to predict what outcome would a medication have in their bodies.

This year we are expecting to launch MOSA for its initial purpose of predicting and simulating the side-effects and protection levels of the Covid-19, to healthy adults that have not yet been vaccinated. 

The initial purpose and motivation for MOSA is to help those who are uncertain to get vaccinated, to give them assurance. The first impact therefore will be to help our initial target population to get vaccinated. This will hopefully help to increase the vaccination rate and help with the current stage of the pandemic.

A similar goal for MOSA, after accomplishing that initial goal is to proceed with predicting outcomes of vaccines for babies, this will help give reassurance to parents about the reactions that their baby could have and the safety the vaccines provide to their baby.

 Later stages of MOSA will seek to expand the functionality of the simulation from only vaccines to medications in general. One of the first medications we would like MOSA to create a prediction to is birth control pills, this could help girls and women to understand the effects of the pill, and if needed, recommend alternative birth control methods. 

At these later stages, we would like to also provide a way to test and simulate efficiency for medications that are being developed or are in medical trials. Hopefully, MOSA would serve as a way to reduce the time for the approval of new medications, by making testing a lot more efficient. MOSA could only achieve this stage when it has successfully achieved the previous goals, as we would like to ensure that results are 100% accurate and reliable. 

For the technological side, we can easily measure the accuracy of MOSA by using cross-validation and producing a confusion matrix if we are using a supervised or semi-supervised machine learning algorithm. This will help us see how far we are from getting an accurate and reliable model. 

We would then need to compare the results that are given by MOSA to real-life examples to make sure the simulation and prediction are what we are expecting. For this we would need to find volunteers, in this case, we might use a different vaccine, or look for an already existing dataset related to Covid-19 vaccines to achieve this goal (we could even use this same dataset to train the model).

We would then look at the next goal for MOSA and follow a similar process.

The main barrier that we worry about is that we don't count the sufficient technological devices to successfully train the model that we want, on the scale that we want. If we were to train our model, with the current technology we own, it would take a lot of time to successfully do it. 

Likewise, time is a big constraint for the team members, as we are university students, that are also working. Therefore, we are very limited to working on this project during our own free time. 

Another barrier we have is that we lack access to a dataset or information or a general lack of knowledge of the medical side of MOSA; which means, as previously stated, we are focusing our time on research at the moment. 

Tatiana is a Software Engineer student, specializing in Robotics and Artificial Intelligence at the University of Canberra. And worked at the ACT Health Department (in Canberra), as part of the Covid-19 response team from May 2020 to August of 2021, where she had an administrative role at Covid-19 testing centers and vaccination centers.

Juan Diego is a Mechatronics Engineer, currently finishing his second degree in Computer Science at the National University of Colombia. As of now, he is working at a shipping company called Mercado Libre as a Front-End Engineer.

Both team members are passionate about technology, software and have a passion for topics such as AI, AR, Blockchain, and Robotics. We both like to further develop our skills in the technology industry and make a positive impact on our community and the world.

As Tatiana worked at Covid-19 Vaccination centers for about 6 months in 2021, she was really interested in the input of both patients and nurses that were administrating the vaccines. This experience is one of the main reasons why this team would like to develop MOSA. Additionally, as a woman that is taking birth control pills, Tatiana is also looking forward to making MOSA useful for that type of medication and helping other women in their research and selection of birth control methods.

We are not partnered with any organization at the moment.

We do not qualify for The HP Girls Save the World Prize

Although the initial goal of MOSA is to address health concerns about Covid-19 Vaccines. One of our main goals after launching MOSA is to adapt the algorithm to predict and simulate reactions of the human body when is taking medication. 

The biggest goal in this area would be to provide girls and women an idea of how their bodies could react to birth control medication, so they can make an informed decision, and look for better alternatives if necessary. This could be extended to any medication that girls and women take throughout their life.