Picture This: A Tech Tool That Helps Physicians Make More Accurate Diagnoses Through Imagery

HealthLeaders takes a closer look at VisualDx, an innovation designed to reduce diagnostic errors.

Imagine an app that uses pictures and graphics to help pinpoint one of thousands of diagnoses. VisualDx, a clinical decision support tool, does just that, and much more to help physicians make more accurate diagnoses, save time, and improve the patient experience.

Originally developed as a resource to help primary care, emergency department, and urgent care physicians diagnose dermatological conditions, VisualDx has now expanded its focus to general medicine and includes opportunities to explore drug reactions, infectious diseases, diagnosis of ophthalmologic issues, and assistance interpreting information in x-ray images, among other uses.

This article takes a close look at the innovation developed by the Rochester, New York–based company. In separate article, 5 Ways Health Systems Can Use a Clinical Decision Support Tool to Provide Better Care, HealthLeaders examines the way physicians at CoxHealth in Springfield, Missouri, and Anne Arundel Medical Center in Annapolis, Maryland, use the tool. Among the highlights:

1. Keeps Physicians on Top of Current Protocols
    
2. Helps PCPs and ED Docs Diagnose Dermatological Conditions and Reduce Referrals
    
3. Improves the Patient Experience
    
4. Offers a Rapid Way to Diagnose Unusual Conditions and Diseases
    
5. Guides Diagnostic Processes in a Way That EMRs Cannot
    

VisualDx reflects a new way of practicing medicine, says Art Papier, MD, co-founder and CEO of the technology company. "We're moving from a generation that was trained in medical school to read books and … hold [all knowledge] all in your brain … without referencing anything, to a future state where [physicians] grab information at the point of care to answer patients' questions." Technology is helping the profession make that transition, he says.

How It Works
 

The innovation combines clinical search with a medical image library, plus knowledge from medical experts to help with diagnosis, treatment, self-education, and patient communication. The tool can be accessed via an online portal or smartphone.

Physicians enter key symptoms, such lower quadrant pain, fever, an enlarged liver, or joint pain. The system guides the user to through the differential diagnosis process, narrowing the options from about 3,000 possibilities.

The program relies heavily on photographic images depicting various patient conditions, as well as building graphics, known as "Sympticons" that enable users to refer at a glance to visual references as they work their way through a diagnosis.

The tool now uses SMART on FHIR technology (Substitutable Medical Applications, Reusable Technologies on Fast Healthcare Interoperability Resources) which directly integrates into the EHR. By pulling relevant information from the EHR, practitioners don't have to input information such as age, sex, medications, allergies, and more, speeding the diagnostic process.

Reduces Diagnostic Errors
 

One of the primary advantages of VisualDx is its ability to reduce diagnostic errors.

According to a study published September 23, 2014, by BMJ Quality & Safety, approximately 12 million U.S. adults are improperly diagnosed in outpatient settings each year, an error rate of 5.08%, or about one in 20 adults. "Based upon previous work," the authors say, "we estimate that about half of these errors could potentially be harmful."

According to VisualDx, in a randomized, blinded study conducted by the company at the University of Rochester, the tool improved diagnostic accuracy more than 120% for complex medical cases compared with standard textbooks and atlases. This improved outcome, involving of emergency physicians, internists, family physicians, and dermatologists was achieved after only five minutes of training.

In a study published March 15, 2011, in Dermatology Online Journal conducted at Harbor-UCLA Medical Center and University of Rochester Strong Memorial Hospital, physicians using VisualDx were over four times more likely to suggest the correct diagnosis for patients admitted to the hospital for serious infections. Without VisualDx, admitting physicians made diagnostic errors 28% of the time.

The Latest Advances
 

Since VisualDx was founded in 1999, the product has gone through numerous evolutions. A new app, VisualDx Plus DermExpert, employs artificial intelligence (AI) through Apple's new machine learning technology, Apple Core ML2. It features an advancement that allows physicians to take pictures of a patient's rash or lesion, and the image accelerates the ability to produce a differential diagnosis directly on the provider's smartphone. To ensure patient privacy and HIPAA compliance, the image is not saved.

"Our definition of AI is not artificial intelligence, it's augmented intelligence," says Papier. "The idea is that the computer-based information plus your brain is going to make a better decision than your brain alone. We're arguing that we can use this machine learning to augment that doctor, but he's still in control and still has to look at these possibilities and confirm or deny because the artificial intelligence is never 100 percent; it's just a very sophisticated tool."

In addition, the company has also released a consumer iPhone app, Aysa, providing diagnostic guidance to help individuals determine whether or not they need to see a physician for certain dermatological conditions.

Also of interest: VisualDx has partnered with an insurance company which is offering discount on professional liability insurance for physicians who use the system. In a separate initiative, continuing medical education credits are offered to physicians using VisualDx.