Symptom tracking and mapping in low resource areas of Boston
In low-resource areas of Boston, vulnerable populations such as undocumented immigrants and those experiencing homelessness may not have the luxury of visiting hospitals or clinics for baseline COVID-19 screenings. Areas such as Chelsea and Everett have a much higher rate of positive COVID-19 tests than the Boston and Cambridge area which is extremely disconcerting. These numbers may not even be fully representative of the number or severity of cases. Furthermore, visiting the hospital can put these populations even more at risk, which may lead to fear of getting tested. Due to this fact, the severity of all symptoms and cases may go unmonitored. Access to information about COVID-19 and how to address symptoms may be very limited in these areas and false information could quickly spread.It is extremely important to monitor low-resource areas for symptoms and cases in a way that is accessible, such as a texting service that can be used on basic and prepaid phones, as well as smartphones. Although many mobile sites and web applications have recently appeared to track symptoms, many people do not download them. Smartphone applications and websites also exclude people that have basic or prepaid phones or limited internet access. Those with an income less than $30,000, little or no higher education, and minorities are less likely to own a smartphone as well. App-based interventions are not an equitable option when gathering information and an effective alternative must be found.
How will SMS systems increase symptom tracking of COVID-19 in low-income and displaced populations in Boston? How would it improve communication of information between displaced people and Boston-based governmental agencies and NGOs?
What it Does
This SMS system collects and monitors symptoms in low income and at risk populations in Boston. It can be used on basic and prepaid phones. We are examining whether usage of non-app-based phone functions can improve COVID-19 symptom tracking and information flow within these populations for priority testing and preparedness. When people report through this service, symptomatic people will also be connected to testing centers and further resources, eventually in their native language, that can supply advice and information regarding the pandemic. This could be especially beneficial for those who are not proficient in English.
This data project will identify at-risk areas for priority testing and preparedness, hopefully preventing a spike in cases in those areas. The mapping will be accessible to relevant governmental agencies and NGOs. No identifiable information will be collected by the app. We also wish to track how many individuals reuse the service, stay in connection with the system, and whether it is used in neighboring households.
Our main goal is to focus on the following benefits:
- Allows for efficient healthcare delivery and preparedness in areas that exhibit spikes in symptoms
- Equitability in technology-based symptom tracking
- Prevents spread of miseducation through health recommendations regarding specific symptoms
- Reduce time and money spent looking for information on mobile data
- Gather information about populations that popular testing methods may not reach
- Support equity in symptom tracking by having questions in accessible languages and local vernacular (we have not yet accomplished this)
How we built it
- a simplified version of CDC guidelines
- Vue.js framework
- Google Map API
- Heat Layer (to demonstrate intensity and location of data)
- GeoJSON layer (to show zip code divisions in Boston)
- Firebase Firestore: storage of user responses
- Google Cloud functions: receiving, sending, and manipulation of sensitive user data
- Written in Python
- Uses Flask, Firestore REST API’s, and many more
- Twilio SMS: sends and receives user responses and information, communicates via webhooks with our cloud functions
Challenges we ran into
Most of the challenges were in the SMS symptom tracker portion. The Twilio documentation proved to be difficult to navigate due to our inexperience with Flask and non-Web based applications. Creating the proper flow of information with data structures was challenging when creating the symptom tracker as well. The initial steps of implementing the map in the data visualization was also difficult since we needed a high level of customization that wasn’t available with current plugins, meaning we had to figure out how the Google Maps API interacted with Vue.js and create our custom maps from there.
Accomplishments were proud of
The map on the data visualization portion was challenging, but incredibly rewarding because of its ease of use and utility for policy decisions. Our team also did not have much experience with cloud functions, particularly those native to Google Cloud rather than firebase, so being able to create several functions to handle user data is definitely one of our best accomplishments.
In terms of project improvements, we want to streamline and finalize the symptom tracker. Adding more features to the visualization would also provide even more information to Boston. Finally, implementing internationalization by implementing the symptom tracker in Spanish and Mandarin would increase the range of the project, providing resources to those most vulnerable and reporting this vital information to policy makers.
We plan to partner with organizations and institutions working with people living in low-resource areas of Boston, such as Boston Healthcare for the Homeless, MIRA, and the Boston Center for Refugee Health and Human Rights. We also hope to translate this system into other common languages spoken in Boston. We are also hoping that, through the Initiative on Cities, this service could be utilized by the City of Boston to cover the costs and further reach the largest number of people. In the future, utilizing this logic on messaging apps frequently utilized by minority and immigrant communities such as WhatsApp, Facebook Messenger, and WeChat could be beneficial in reaching an even larger population. We would analyze all data collected in order to determine the areas most in need. Governments in the greater Boston area could then use this information to allocate resources to these areas.