Real Time Medical, of Mississauga, Ont., has announced a partnership with Google that not only positions the Real Time Medical platform on the Google Cloud, but also creates a working partnership to create new, AI-driven solutions for clinicians in Canada and worldwide.
Real Time Medical is the creator of a peer-review, quality and education system for radiologists and other clinicians. The solution checks the quality of the readings done by radiologists and other specialists by sharing a portion of their work among a group of peers on the Real Time Medical network.
Their identities, however, are anonymized, so there is no fear of embarrassment or repercussion. If mistakes or “discrepancies” are found, the clinician is alerted so he or she can modify reports and brush up on skills for the future.
“We’ve conducted seven radiology quality audits across four provinces, for ministries of health, hospital networks and physician insurance organizations,” said Dr. Nadine Koff, a radiologist and President of Real Time Medical. “We know that performance varies not only by individual, but also by exam type and sub-speciality. It’s for this reason that our system, unlike so many others, provides for per user, per subspecialty, anonymized and individualized peer learning and review.”
Indeed, while a discrepancy rate of one to two percent is considered normal, independent reviews have found the rate of discrepancies for some radiologists run as high as 20 percent. Nevertheless, many hospitals and health regions still haven’t implemented quality assurance systems for their clinicians.
Those that have, say executives at Real Time Medical, haven’t always installed effective systems. Ian Maynard, CEO of the company, says that some hospitals are using peer review systems that may check only one exam reading per day for a radiologist.
By contrast, Real Time Medical’s system can be set to check any number – even 100% of a particular exam type, if so desired.
“Many systems amount to nothing more than an arbitrary one case per user, per day, regardless of the number of diagnoses generated,” said Maynard. “That’s really nothing more than a “check the box”, we do peer review approach. Other systems have users comment on a historical exam while interpreting the current. Not only is that approach objectively invalid because the historical report is not anonymized, but more importantly, historical exams can be 6 or 12 months old. The patient is being rescanned for a reason, likely a worsening of symptoms. These solutions are behind the times, and they fall far short of what we’re capable of as a country.”
Maynard noted that Real Time Medical’s technology, which is used in hospitals and regions across Canada, offers a more thorough analysis of readings.
Moreover, the intelligence component of its system can spot where clinicians may need extra help with readings – for example, discrepancies may be found more often when they’re reading head and neck CTs, or certain types of ultrasound scans. The system can alert them to this.
Real Time Medical’s alliance with Google is one of the first instances of a Canadian health technology company putting its solution on the Google Cloud. By doing so, it lowers the cost of using the system for hospitals and clinics, as they don’t have to acquire their own infrastructure – they simply make use of the computing power of the cloud.
Dr. Koff observed there are no issues of sovereignty or privacy involved, as Google has a high-security data centre in Montreal.
The platform now available offers shared peer review and quality assurance for radiologists and pathologists. Other specialities will be added in the future.
So, right now, hospitals and clinics that join the system benefit from the expertise of radiologists and pathologists outside their own facilities. It’s a system of shared expertise that spots discrepancies and raises quality, no matter the size of the organization that uses it.
“It doesn’t matter if you’re in a clinic with five radiologists,” said Enzo Costanza, Vice President of Product Development and Implementation at Real Time Medical. “Our Canada-specific deployment allows for radiologists and pathologists in hospitals and independent healthcare facilities throughout this country to connect and gain the benefits of large-scale, AI enhanced peer review, regardless of their size, or number of practitioners.”
He added that, “Other systems are site-limited, or tied to site specific, EHR, RIS, PACS or voice reporting solutions and offer very little in terms of a critical mass of peers with the same subspecialties.”
On a cloud-based system, by contrast, users are connected to a far bigger community of peers. And they’re not constrained by the type of PACS or EHR they are using.
The partnership with Google will allow Real Time Medical to draw on the expertise that Google has in artificial intelligence – Google is a leader in AI and Deep Learning, and the company has established AI centres of excellence in Toronto and Montreal. It is also a partner in the Vector Institute, an AI think-tank and accelerator in Toronto.
At the same time, Real Time Medical will be helping Google develop its own healthcare applications. “They approached us and since then things have progressed quickly from a partnership perspective,” commented Dr. Koff.
Indeed, in conjunction with Google, Real Time Medical plans to develop four different AI-driven applications. The first of these was announced by the company at the recent Healthcare Information Management and Systems Society (HIMSS) conference, held in Las Vegas in March – it consists of the ‘intelligent sampling’ app, which spots areas in which a particular radiologist may need to get up to speed.
As well, at HIMSS, the company announced a powerful alliance with Client Outlook, the maker of the eUnity viewer. The company’s next AI-powered solution will likely be announced at the Radiological Society of North America (RSNA) conference in November.