From UCLA, Department of Physical Sciences
Despite its presence in more than 80 countries, Lyme disease remains one of the least understood diseases. An estimated 476,000 people are diagnosed annually in the US alone.
This bacterial infection spreads through the mites bite poses a unique challenge due to the wide range of symptoms it produces. Lyme disease, known as the “great mimicr,” often mimics other conditions, complicating diagnosis and treatment.
Because misdiagnosis is common, the actual number of cases can be far greater than reported numbers. In addition to these challenges, climate change has exacerbated the problem even further by expanding mites' habitat.
New research funded by the Department of Defense (DOD) aims to tackle these challenges in a very efficient way. It's only $750,000 less than many multi-million-dollar research grants. The project leverages existing patient data, advanced machine learning, and a team of experts to develop better diagnostic tools and treatment strategies.
Small grants, big impact
The three-year initiative led by UCLA Mathematics Professor Deanna Needell is a perfect example of how less investment can have a big impact when combined with an innovative approach.
Professor Niedel collaborated with longtime collaborators from Professor Jana Geweltz of the University of New Jersey and Lorraine Johnson, CEO of Lymedisease.org. Using their combined expertise in mathematics, data-driven research and patient advocacy, the team is confident in their ability to shed new light on this bewildering illness.
Professor Nieder has worked with lymediase.org for over a decade. When Johnson launched Mylymedata to enroll patients eight years ago, Niedel saw an opportunity to use the data in a new type of study. Today, the registry includes more than 19,000 patients reporting diagnosis, symptom severity and treatment outcomes.
Over the years, Needell and Johnson have co-authored five papers and analyzed this data by blending statistical and machine learning approaches. Their previous studies have investigated key factors that influence the antibiotic response and treatment success in chronic Lyme disease.
Persistent neurological symptoms of Lyme disease
Based on previous success, the current project focuses on persistent neurological symptoms of Lyme disease, which affect the nervous system. Patients with these symptoms may experience problems such as neuropathy, trace amounts, memory loss, cognitive impairment, sleep disorders, and psychiatric symptoms that are associated with functional and structural changes in the brain. These symptoms can disrupt daily life.
By using topic modeling, machine learning techniques Niedel, Johnson, and Gewertz aim to uncover trends and patterns of patients' responses to antibiotics. These insights can directly inform treatment strategies. Recent advances in AI and new machine learning methodologies allow teams to design better algorithms than ever before, defining and understanding neurologically-based Lyme disease.
“Simply put, government grants are essential to solving healthcare challenges,” Niedel said.
Programs such as the Congressional Command and Medical Research Program (CDMRP) allow universities to partner with experts, patient advocates and organizations to turn research into meaningful changes.
Better care for Lyme patients
Their hope is that by understanding the nature of the persistent neurological symptoms of the disease, clinicians can study ways to better diagnose and treat diseases. In doing so, it highlights the potential for precision medicine for Lyme disease and shows that data-driven research is essential for this type of patient care.
“We launched MylyMedata for rapid research into persistent Lyme disease by empowering thousands of patients to provide their own health data,” Johnson said. “This collaboration shows how patient-equipped research can help promote personalized care for the patients that need it most.”
These results from this study will also be notified of future projects focusing on other symptom clusters, such as musculoskeletal symptoms, to help develop more personalized treatments by understanding why certain symptoms prevail in some patients.
This article was first published at UCLA Physics Science Division Website.
