Kim Chulis, a clinical assistant professor at Southern Illinois University Carbondale, is developing LymeSignal.AI: Early Signal Navigator. It’s a patient-centric tool to track and document symptoms and other relevant information, aimed at helping patients and clinicians have earlier, better-informed conversations about Lyme disease. (Photo provided by Kim Chulis).
September 10, 2026
SIU professor creating AI-based portal to support earlier Lyme disease evaluation
CARBONDALE, Ill. — A visit to the doctor’s office with fever, chills, fatigue and aches could seem like the flu or another viral infection. But what happens when symptoms don’t go away? The patient is left frustrated and scared wanting to know what’s wrong, while the clinician is left puzzled, trying to pinpoint a diagnosis.
This is one challenge that patients with Lyme disease and the clinicians who treat them face. Now using artificial intelligence, Kim Chulis, a clinical assistant professor at Southern Illinois University Carbondale, is developing LymeSignal.AI: Early Signal Navigator. It’s a patient-centric tool to track and document symptoms and other relevant information, aimed at helping patients and clinicians have earlier, better-informed conversations.
“Some clinicians may live in an area where the disease is not as prevalent and do not see Lyme regularly,” she said. In addition, she said, sometimes symptoms appear well after a patient has been bitten by a tick, making it seem the symptoms have come on for no immediately apparent reason.
Lyme disease is on the rise
Every year, an estimated 31 million people in the United States are bitten by a tick, according to the Centers for Disease Control and Prevention (CDC). Lyme disease is the most common tickborne disease in the United States, with an estimated 476,000 patients treated for it each year. Other common tick-born illnesses include Rocky Mountain spotted fever and alpha-gal syndrome.
Since the mid-1990s, Lyme disease cases have grown significantly. The CDC highlights this trend in its year-by-year data. In 1996, there were 16,461 reported cases. By 2010, that number nearly doubled to 30,158 before skyrocketing to 89,470 reported cases in 2023.
Health officials note these numbers could be even higher due to unreported and undiagnosed cases.
An invisible illness
Chronic conditions with few visible symptoms are sometimes called “invisible illnesses.” Lyme disease can be difficult to recognize because it produces wide-ranging symptoms that may overlap with other infections and conditions. Some patients also experience persistent or evolving symptoms, which can contribute to delayed or fragmented care.
Developing LymeSignal AI as a tool
Chulis is creating LymeSignal AI as a tool to assist patients, caregivers and clinicians. It will likely be a mobile-based or laptop-based portal where patients can enter several different inputs, including exposure, symptom timelines, photos, travel information and more.
“The idea is to give patients a way to bring together information that might otherwise be scattered or forgotten — things like symptoms over time, possible exposures, travel or location history, and photos,” Chulis said.
With patient permission, future versions of the tool could also incorporate selected information from a phone’s health or fitness applications to help reconstruct timelines and possible exposure context.
Caregivers could also document observations, giving clinicians a more comprehensive understanding of changes the patient has experienced.
“Sometimes a caregiver or family member may notice a change that the patient doesn’t recognize or remember to mention during an appointment,” Chulis said.
LymeSignal AI will compile an evidence timeline that might be invisible or hard to put together otherwise, Chulis said. The clinician would then be able to review the information as part of evaluating the patient.
“The proposed outputs include a concise patient-visit packet, clinician-facing risk context and aggregate insights that could identify patterns in delayed or fragmented care,” Chulis said. “The goal is to organize relevant information, support earlier and better-informed conversations, and make patterns more visible to patients, clinicians and potentially public-health stakeholders.”
The tool is not intended to diagnose Lyme disease or replace clinical judgment.
Although Lyme disease is the initial use case, Chulis said the broader idea may also have applications for other complex conditions where relevant information is spread across time, providers and different sources of data.
Competition mode
This tool is being created as part of the TOPxHHS Tech Sprint for AI and Invisible Illness. The competition is a four-month sprint among industry, academia and the public, where competitors build digital tools with AI-accelerated insights for Lyme disease, invisible illness and the cost of illness.
In Phase One of the competition, prospective competitors pitched their concepts. Of more than 130 applicants, only 45 teams advanced to Phase Two — among them, Chulis and LymeSignal.AI: Early Signal Navigator.
“The sprint itself is a new experience for me, which is part of what makes it exciting,” Chulis said. “It’s an opportunity to bring together work I’ve been doing for years in AI and analytics and apply it to something that could really help people who are facing Lyme disease or other invisible illnesses, as well as their families and clinicians.”
The competition is a first for Chulis, but she has plenty of experience working with AI. She has more than 20 years of experience in analytics, data science and AI strategy, including serving as a principal at Microsoft. Through her company, Core Analytics® and Deconstructing.AI™, she’s also worked with organizations to implement AI and analytics to support better decision-making.
Phase Two of the competition is the sprint itself, where teams build, test and demonstrate a functional prototype with real users to test and evaluate the tool — A typical step before a product is refined and ready for commercial availability. The program includes up to $2 million in total prize funding. Chulis said any funding awarded to the project would support continued development and validation of the work.
For Chulis, though, the inspiration to participate goes far beyond the competition— its personal. Her mother and brother both had Lyme disease. She also carried out research about 10 years ago data mining Twitter to uncover communication trends surrounding asthma, cancer and diabetes. Chulis believes LymeSignal AI is an evolution of that past research, enhanced with today’s AI capabilities.
Take the survey
As part of the user-research phase of the competition, Chulis launched a short anonymous Lyme Patient Experience Survey to gather first-hand input from adults with experience with suspected, diagnosed or treated Lyme or another tick-borne illness, including those who have experienced persistent symptoms.
The survey takes eight to 10 minutes and does not include names or email addresses. The responses are helping to inform the design and priorities of the prototype.