AHN Becomes First Health Network in Western Pennsylvania to Utilize New Data Tool to Identify Rare Symptoms and Diseases

Wednesday, August 14, 2024

Press Releases Autoimmunity Institute

AHN Becomes First Health Network in Western Pennsylvania to Utilize New Data Tool to Identify Rare Symptoms and Diseases

Epic's 'Look-Alikes' technology matches patients with hard-to-diagnose conditions to similar cases from around the world, using records from its massive Cosmos dataset

PITTSBURGH (August 15, 2024) - When patients have unusual health symptoms or a rare medical condition, their doctors are often left searching for similar cases by checking medical journals or by chasing down peer specialists across the country. So-called "rare disease odysseys" can be frustrating for caregivers and patients alike - the process of narrowing the clinical possibilities, visiting different specialists, obtaining a diagnosis, and settling on a treatment plan can drag on for years.

But now, Allegheny Health Network (AHN) has a new technology at its disposal for tackling these rare diseases: Epic's "Look-Alikes" tool, which is embedded into Epic's electronic health record (EHR) system. Look-Alikes, in turn, scans Epic's "Cosmos" research dataset, which houses billions of pieces of de-identified patient data, drawing on 12.2 billion patient encounters gleaned from more than 257 million patients, 1,500 Epic-affiliated hospitals and 352,000 physicians worldwide.

Using the Look-Alikes tool, AHN doctors can now query the Cosmos dataset to find doctors who have treated patients whose symptoms mirror the unusual cases they are seeing locally. Whether across the state or across the country, if there's a strong match between patients and symptoms, AHN's physician can connect with the doctors who managed the similar cases, compare notes, and potentially arrive at a diagnosis and treatment plan much more quickly.

The tool is bi-directional, meaning AHN doctors who have already diagnosed a patient with one of the dataset's rare diseases can also be contacted by other physicians who have patients with similar symptoms or the same diagnosis. For now, the Look-Alikes system flags symptoms related to 37 rare diseases - conditions such as erythropoietic protoporphyria (a metabolic disorder that causes severe pain when a patient is exposed to sunlight), Beckwith-Wiedemann syndrome (which causes overgrowth of limbs and organs), and Friedreich ataxia (which causes neurological degeneration).

In June, AHN became the first health system in the region, and one of the first in the U.S., to start using Epic's Look-Alikes technology. Epic is one of the world's largest providers of electronic health records technologies, and its EHR products are broadly considered the industry's gold standard. AHN has been utilizing Epic's EHR and patient-charting technologies since 2013, and also utilizes Epic's patient-facing records portal, known as MyChart.

"When a patient has a rare or unusual condition, finding the correct diagnosis and developing an appropriate treatment plan is a bit like searching for a needle in a haystack," said Susan Manzi, MD, MPH, chair of the AHN Medicine Institute, and director of the AHN Autoimmunity Institute's Lupus Center of Excellence. "With the Look-Alikes tool, we have the ability to search for that needle far more efficiently - or better yet, we can consult with someone who has already found it."

The Look-Alikes tool will prove especially useful for the Autoimmunity Institute, and for patients with autoimmune diseases. Autoimmune conditions, some of which are rare to begin with, can be notoriously difficult to diagnose - symptoms can come and go, mimic those caused by other diseases, vary from person to person, and affect a variety of bodily systems simultaneously.

"AHN is a national referral center for many difficult-to-diagnosis autoimmune conditions," Dr. Manzi said. "This tool allows us to connect with other physicians in the country that have similar patients but, more importantly, it provides a tool for those physicians to find us."

According to Global Genes, an advocacy organization advocacy for people fighting rare and genetic diseases, the average time from disease onset to accurate diagnosis for a rare disease is 4.8 years. And even when a rare disease is quickly diagnosed, the scarcity of similar patients often means there is little clinical evidence to draw from when determining the best course of treatment for a patient.

All AHN physicians - including primary care and specialty providers - are able to utilize the tool. Many of the diseases currently being tracked by Look-Alikes exhibit a combination of neurological, musculoskeletal, hematologic and autoimmune symptoms.

An AHN doctor utilizing the tool enters the patient's unusual symptoms and other health indicators, known allergies, and demographic information. Look-Alikes then scans the Cosmos dataset for doctors treating similar patients and allows the physicians to connect with each other. The system is intelligent: If no matches are found, the system suggests similar searches that might find matches.

And when an AHN doctor diagnoses the condition, the patient's de-identified record can be "tagged" with the disease in question, making it easier for other physicians to seek AHN's expertise.

"Through this tool from Epic, AHN physicians will be able to augment rare disease research, physician collaboration, and clinical decision support capabilities while still protecting patient privacy," said John J. Gabrick, AHN's vice president, Clinical Information Systems.