Setting/Problem

Lyme disease cases have been steadily increasing in the U.S. since the early 1990s, but surveillance and prevention efforts have lagged behind. Lyme disease is the most common tick-borne illness in the U.S. and a growing concern in Canada as well. Lyme disease can lead to serious health problems especially if it is not treated.

The incidence of Lyme disease in the United States has nearly doubled since 1991, from 3.74 reported cases per 100,000 people to 7.21 reported cases per 100,000 people in 2018.
(https://www.epa.gov/climate-indicators/climate-change-indicators-lyme-disease)

Furthermore, the lack of timely and accurate public health surveillance tools makes it difficult for officials and the public to prepare for outbreaks. The COVID-19 pandemic further strained resources, causing reported data for Lyme disease cases to become outdated, exacerbating the need for an up-to-date dashboard of disease outbreaks and areas of concern.

This project aims to develop an early warning Lyme dashboard that utilizes a proactive surveillance algorithm based on social media data (Twitter, Facebook, Google Trends, etc.). This dashboard will provide near real-time information on potential cases and clusters for public health officials to investigate and respond to effectively. Ultimately, this project aims to provide timely alerts to public health authorities, healthcare professionals, and the general public to implement preventive measures, promote awareness, and facilitate the early detection/treatment of Lyme disease.

Importance of the study

Lyme disease causes serious health and economic problems, not only in the United States and Canada but also worldwide. Given that the incidence of Lyme disease has been increasing in recent years, it is even more imperative that steps are taken to reduce Lyme disease cases.

According to the John Hopkins Lyme Disease Dashboard, only 10% of 476,000 Lyme disease cases in the United States get officially reported. Increasing the percentage of cases that are reported would be extremely helpful as it will provide a more accurate picture of the true impact of Lyme disease. Just like COVID-19 pandemic, knowing the true number of cases geographically is a significant first step to controlling an outbreak.

Lyme disease results in an increase of $3000 per patient per year in healthcare spending. Therefore, Lyme disease increases healthcare costs by around $1.3 Billion every year. The health impacts of Lyme are more important but the financial costs should also not be ignored.

Early diagnosis and proper antibiotic treatment of Lyme disease can help to prevent late Lyme disease. Although Lyme disease is rarely life-threatening, delayed treatment can result in more severe disease. 

The goal of this project is to create a Lyme disease surveillance prediction algorithm using AI. The results of this predictive algorithm will be used to create a live-streaming dashboard to help spread awareness of and offer information on Lyme disease.

We would also like to perform some correlation between the number of predicted cases from the algorithm and the number of actual reported cases to cross-validate our findings.  Testing the accuracy of our model in real-world scenarios is important to build confidence in the model's predictions.