Healthcare outsider developing software to improve diagnosis

Worried about a coming shortage of physicians?  Don’t be, says Cameron Powell, CEO of Physician Cognition.  Low-cost providers, such as nurse practitioners, and consumers themselves, he says, will fill in nicely.

Empowering them will be what Powell calls “MD-mimicking technology.” Powell’s company is building just such a technology.  The program, called Xebra Pro, is being groomed to advise healthcare providers and patients about the signs, symptoms, diagnoses and treatment of diseases.

Xebra is being groomed initially for use by healthcare providers.  A Web-based “clinician app” has been in testing since April. It can still be accessed online.

Lately the development team has turned its attention to a mobile app, which they hope to make faster and more user-friendly than the current Web-based one.  This app is scheduled for release in a few weeks on iOS and Android, according to Powell.  Apps for the consumer will come out in a few months.

The company behind Xebra, Physician Cognition, is Powell’s fifth startup since graduating from Harvard Law School 23 years ago.  It’s his first in healthcare, a company he co-founded with Dr. Vipindas Chengat, a specialist of internal medicine. The others addressed law or business.

“My outsider perspective allows me to see clearly what’s wrong with the processes and technology in healthcare,” he says.  “And I’m not afraid to speak plainly about it.”

The medical community is already paying attention to technologies that might empower patients.  A peer reviewed paper published in BMJ (formerly called the British Medical Journal) looked into 23 “symptom checkers.” The research team was less than enthusiastic.

“Symptom checkers had deficits in both triage and diagnosis,” the authors wrote.  “Triage advice from symptom checkers is generally risk averse, encouraging users to seek care for conditions where self care is reasonable.”

But Physician Cognition’s software is not a symptom checker, which Powell asserts is an obsolete technology.  Rather, Xebra Pro employs a cognitive process that Powell describes as being similar to that of a human clinician. Rather than allowing input of one or just a few symptoms, Xebra can handle “an unlimited number of symptoms, and all their qualifiers,” he says, “as well as an unlimited number of physical signs, lab results, medications, and patient history inputs.”

Xebra Pro performed well on numerous tests, Powell says, including ones involving the same clinical cases featured in the Harvard study of 23 online symptom checkers in the BMJ.

The current Web and the soon-to-be-released mobile apps are primarily intended for providers.  These would include nurse practitioners, physician assistants and triage nurses. The cost of these versions would be about a third of the cost that the company plans to charge MDs.  

Powell has not yet sought clearance for Xebra Pro from the FDA.  A 510(k) clearance “is the most we think we will need even for the lay, or consumer, version coming out later,” he says.  “But the professional version requires that professionals use their own considerable judgment, and it doesn’t provide any final diagnoses or treatment recommendations.  Also, our user agreement includes a pretty thorough set of disclaimers, in addition to those we pepper around the site and mobile (version).”

The patient-oriented app, which is further from market and has yet to be formally named, will empower patients by documenting symptoms and signs to discuss with their physicians and suggesting “the most appropriate tests and labs to narrow down the diagnosis quickly and cost-effectively,” Powell says.  “If a lab (or imaging exam) suggested by the physician is not on that list, the patient can push back.”

The program might even suggest where in the medical hierarchy the patient should go first. “Is this a nurse practitioner issue; is this for a GP [general practitioner]; is this for a certain specialist?” Powell says. 

This is what constitutes real patient engagement – not just access to records or alerts, he asserts.  Xebra, he says, will allow patients to “meaningfully challenge their clinicians on what symptoms and conditions to consider and what workup investigation to do.”

Powell describes Xebra as a disruptive technology:  “Imagine medicine without cognitive errors, without implicit bias, without financial conflicts of interest – and yet with a powerful advocate in every would-be patient’s corner.”

The clinician apps, which Powell expects will enter the marketplace first, will generate revenue through subscriptions and upgrades, paid by clinicians and other healthcare providers.  Apps designed for the general public may contribute to the company’s top line by any of several means, including advertising and treatment referrals.

In the meantime, the software will go through substantial evolution, culminating in the integration of artificial intelligence.  Currently, its logic is deterministic, Powell says.  Algorithms dictate suggestions based on collected medical knowledge (e.g., medical texts and publications) that has been and continues to be adjusted by beta users.

“We have had almost 1000 users who have been getting on our Web-based app since we opened it up in April in over 50 countries,” he says.  “They have been giving us feedback and improvements in the algorithms since then.”

What drive Powell, he says, are memories of his late mother’s suffering, physically and mentally, following numerous misdiagnoses and late diagnoses.  Xebra is intended to reduce those kinds of mistakes. It is an ambitious goal.

 “We will do for clinical intelligence what Google did for information generally – making it freely available to everyone in the world, regardless of race or religion, income or gender,” he says.