Mental health is increasingly in evidence and, due to the pandemic in which we find ourselves, rates of mental disorder are increasing every day. According to the World Health Organization (WHO), approximately 10% of the world population has mental disorders [1]. With the arrival of the novel coronavirus, these numbers are expected to rise. it is also known that several social problems have been intensified in this period, in addition to those already mentioned, there is also domestic violence and femicide [2], suicide [3], violation of the rights of homeless people [4].
According to a global company dedicated to maintaining the psychological well-being of the population, in 2020 there was an increase of 7% of north-americans who reported a persistent feeling of sadness, that is the same as three out of five people [5].
The social isolation imposed by the pandemic is a big factor for the feeling of loneliness and, according to the Commonwealth Fund, they are linked to many problems, not only mental but also physical, among them: anxiety, depression, diabetes, heart problems, etc. Biologically speaking, social isolation and loneliness act as potentiating agents for these diseases, leaving the immune system weaker. For example, loneliness can alter the tendency of cell inflammation, which could help cancer cells to spread faster or promote inflammation of the brain, which can even lead to Alzheimer's [6]. These facts increase the risk of early death by 26%, which would be equivalent to smoking 15 cigarettes a day [7].
Another important factor to be considered is how the brain reacts to isolation. The region responsible for learning and memory (hippocampus) decreases in size when exposed to long periods of isolation and loses its ability to grow. This reduces our ability to interact in social situations. At the same time as this occurs, the amygdala regulates our feeling of fear and anxiety and these sensations can be intensified at times like a pandemic [8].
So, do people get depressed because they feel alone, or do they isolate themselves socially because they are depressed? In this sense, our team chose the challenge that seeks to combat social isolation in some way and help people not to reach such emotional states. We developed a feature that will be present in the main social media, so that people can have one more option to distract from bad times and to connect with other people who have the same tastes and, hopefully, do not reach very high levels of stress. After all, the coronavirus already causes several deaths around the world and we want to try to reduce the factors that cause so many other indirect deaths during this period of crisis.
The team analyzed all of the proposed challenges and we chose three final themes. Everyone discussed possible executions for each topic and we did some voting. In the end, we came to the conclusion that we would like to develop something to impact people directly and, although the solution was initially not the most obvious, we wanted to address a socially relevant topic, not only in pandemic times, but also in our daily lives, in addition to explore the potential of all team members. So, we chose “The Isolation Solution”.
Thus, for the development of the project, we divided the group into tasks: development of the motivation and scope of the project, mock-up of the application, chatbot and emotion classifier.
Project Development
For the development of the project, we sought several sources on the topic of social isolation and loneliness and its complications, statistical data and articles related to mental health and the problems that can be caused by its poor care. We seek to develop a solid basis for the construction of the problem that we would like to solve and make it clear. We discussed presentation forms, as we would not have a pitch to talk about our idea and decided that it would be important for the text to reflect all of our work. In addition, we made a demonstrative video a little more commercial, with a striking tone and not only explanatory.
Application Mock-up
We used Figma for rapid prototyping of mobile applications and we made a mock-up, since we didn't have any mobile developers on the team. The mock-up facilitated the visualization of how the feature would work within the social network. It consists of a small button of a book with a heart in the middle, right by the side of the messaging button.
Chatbot
For the chatbot, the Telegram API was used, which provided interaction with future users of the feature. First we used the DialogFlow interface to test some interactions with users and see if the form was appropriate and not too much time consuming. The next step was to use Telegram as the platform to ask the questions and get the inputs from the users. We integrated it with the Heroku host, so we could save the answers, analyze them and then return with a suggestion of chat group.
Sentiment Analysis Classifier
In the development of the emotion classifier, we used data from a NASA astronaut report [9]. This report provided possible solutions for greater acceptance of isolation, since astronauts are trained to stay long periods isolated from other people. The classifier analyzes the open question that is made by the chatbot, where users should answer “what’s happening?”. We used a neutral question. This analysis is the one that redirects users to the a specific chat group. It also classifies what people are saying in them, so eventually it can provide people with insight of what NASA astronauts would do in a similar situation.
Problems and Achievements
Due to the great problem that the chosen subject portrays, some obstacles were found during the development of the application. 2020 data related to the mental health of the world population associated with the pandemic was not easily found, due to the fact that we couldn’t spend so much time on a topic. Instead, we changed the focus of the research by reading articles on general topics and not only specific data to specific problems (such as the consequences of the pandemic in mental and physical health) and compared them with some data provided by WHO.
With the ground resources collected, we had some problems trying to form a robust idea and ended up adding many features that would not be possible to develop within the hackathon period. Thus, we chose to keep it simple and proceeded with the development of a chatbot, previously created affinity groups and only one group with the bot that indicates activities based on data from NASA astronauts.
We also had problems with the integration of the chatbot and the classifier. We needed to store the data in some kind of host and we used Heroku, but we could save all the data. We tried different tutorial, but it didn’t work. So we opted to save just the three more relevant data for our analysis (country, state and the open question answer), and instead of saving the category of the group chat though the API, we used DialogFlow to get this information.
Tools and Coding Languages
Future Work and Improvements
We researched on how to produce a more elaborated questionnaire that could be applied for a deeper analysis of users' emotional states. This would be an improvement for the project, but it requires further study in relation to the interaction of social network users with this type of research.
We would like to have a supervisory bot in all chat groups regarding the emotional state of users. This would be relevant to refer them to professionals trained to deal with issues related to mental health.
The mock-up of the application already provides for a game within the platform, for people to play with each other. We want to develop a simple game, like the google dinosaur game.
We also formatted and analyzed demographic data from different regions of Brazil (age, gender, demographic density). With this data, we would be able to cross-check with the saved data from the chatbot form. Thus it will be possible to know who would access the feature, which age group and, in the future, invest in new features developed for specific groups of users. The data can be found here.
You can see a video of our demo here.
How it works
It consists of a small button of a book with a heart in the middle, right by the side of the messaging button.
There is a button on the top right corner of the screen, where it is possible to simulate it. To restart the simulation just click out of the phone and in the bottom right corner of the screen, there is a restart button. You can see the mock-up here.
[1] UNASUS. WHO warns that 10% of the global population has mental health disorder. Available only in Portuguese.
[2] ONUBR. ONU chief warns of increased domestic violence amid coronavirus pandemic. Available only in Portuguese.
[3] ROLLINGSTONE. Suicide fires in quarantine against coronavirus - luckily, the search for help too! Available only in Portuguese.
[4] UOL. Homeless people: they live so many risks, the coronavirus is one more. Available only in Portuguese.
[5] [8] TECMUNDO. How social isolation affects health. Available only in Portuguese.
[6] NIH. COVID-19 is an emerging, rapidly evolving situation.
EURO DATA CUBE SERVICES FOR COVID-19 SPACE APPS CHALLENGE
Cyber-Abuse Research Initiative (CARI)
Psychological Stress Associated with the COVID-19 Crisis
The Coronavirus Anxiety Project
Coronavirus Anxiety Scale: A brief mental health screener for COVID-19 related anxiety