Latest News

AI-Driven Diagnostic Hub Reimagines Primary Care

By Deborah Borfitz

October 8, 2019 | Ativa Medical created buzz at last month's Next Generation Dx Summit in Washington, D.C. with a presentation about its workhorse point-of-care (POC) device being developed expressly for primary care. Founder and CEO David Deetz says the Ativa Diagnostic Workstation has been a 10-year labor of love to get it properly designed for the "coming revolution in blood diagnostics."

Deetz says his moonshot project will create a fluidic operating system that will serve as the "unifying intermediary" for many different types of diagnostic tests. Out of the gate, the POC workstation is designed to do all the fundamental tests—complete blood count (CBC), urinalysis, and lipid and chemistry panels—needed by about 70% of all primary care patients. Doctors would get results in about five minutes and in the same way they would from any other POC device, says Deetz.

During routine testing, the Ativa Diagnostic Workstation will automatically screen for anemia and sepsis, he notes. Unbeknownst to users, the device will line up almost a million blood cells, using lasers to study the properties and nuances of each one.

At launch or shortly thereafter, the POC device will also use its unique data generation capability in conjunction with machine learning algorithms to do what Deetz calls a three-dimensional blood smear. When anemia is detected by the CBC panel, a flag will be automatically generated indicating the type of anemia. The results, he says, will be of high enough sensitivity and specificity to make the flags actionable. Ativa Medical is collaborating with a physician at Johns Hopkins and they're closing in on a rapid screen for sepsis using biosensors to recognize its telltale cellular and protein patterns.

Eventually, the device will flag many different diseases after mining two million unstructured data points in the bloodstream and urine for signs of those conditions. Even the most intelligent people can only deal with about five variables at once, Deetz says, while a single process in the human body generally has dozens of biomarkers.

Once the device is being used to flag multiple diseases, Ativa will add traditional artificial intelligence (AI) capabilities, Deetz says. This would include a "test selector" application to help primary care physicians order the correct blood test, which currently happens only about 35% of the time.

As envisioned, a nurse would enter patient information into the device to learn which test to run and create the standing orders. The doctor would receive a full report inclusive of the differential diagnosis and logic for it, based on best practices, before seeing the patient, says Deetz.

The ability of the POC device to comprehensively screen for various types of serious diseases—aka "digital triage"—will be of great value in primary care, Deetz says, given that current cancer screening tests each run between $500 and $1,000. "Ativa can be the missing link… connecting crucial tests to the people who need it."

The Ativa Diagnostic Workstation will likewise improve the efficiency of retail clinics and allow them to diagnose over 50% of all medical conditions, up from 5% currently, Deetz says. It would also open a worldwide market to pharmacies with telemedicine capabilities.

Providers will need to decide if they want more information than they requested, since positive results would demand a clinical response. "It's not going to cost us more to do a test or not do a test," says Deetz. "We're going to have the data; we just have to apply our algorithms." With the touch of a screen button, clinicians or patients will be able to opt out of any tests they don't want.

Longer term, the device may be programmed to add patient demographics and symptomatology to the consolidated dataset, he adds. The same information would automatically populate the electronic health record, since the systems will be integrated.

Back to the Future

By late next summer, Ativa Medical should be ready to submit the device for approval to the U.S. Food and Drug Administration, Deetz says. It represents more than 200 patents, the "vast majority" licensed from Honeywell to bring big-system functionalities to the micro scale.

The workstation falls in a new category of products that centralize POC testing, says Deetz. A couple other companies, including Beckman Coulter, do CBC testing and use machine learning to help them do some additional tests. But none have the "data power or reach" of the Ativa system, Deetz insists, which "pulls out much more information from the bloodstream than has ever been able to be detected before… and deals with all that complexity."

The patterns derived by machine learning will effectively become the new biomarkers, since they can detect diseases before people have symptoms—and turns diagnostics into a "software subscription business," Deetz says. One day, providers will routinely be using POC devices to sign up for the disease flags of their choosing much as they do fee-based apps on their smartphone, he predicts.

"This is going to be a major shift and people don't realize it yet." Ativa Medical forecasts that after several years the revenue it generates selling disease-detection subscriptions will surpass that of its test cartridge business.

"We've done this before and we know it's hard," says Deetz, noting his background includes both AI and medical devices. He helped develop JPMorgan's AI-powered wealth engine. He also founded two medical device companies, one of which went public and the other that is now part of Danaher's Radiometer POC product line.

"Our approach is very different than that of [blood-testing startup] Theranos," which spent almost a billion dollars going it alone before being shuttered last year under a cloud of scandal, says Deetz. "We have spent $50 million [on the workstation] and are three for three when it comes to bringing difficult technologies to market."

Corporate Partners

Microfluidic processing is a foundational element of the Ativa Diagnostic Workstation, which made Honeywell a key partner in the early R&D days, says Deetz. Many of Ativa's early patents were also transferred to Honeywell to give the intellectual property the protective clout of a big company. "We know this market is huge, $140 billion a year, and we will have competitors once people know what we're doing is possible."

Ativa Medical is working with Bosch, a major infrastructure provider in the AI and digital heath arenas, and is considering setting up labs in conjunction with Bosch's R&D division in Sunnyvale, California, he says.

The partnership with Bosch is what will turn the Ativa Diagnostic Workstation into a "smart diagnostic hub" that other POC devices can plug into, Deetz says, thereby streamlining connectivity to the electronic health record—or the information system used by pharmacies and retail clinics. The larger pool of data will in turn enhance machine-based disease detection. Ativa may eventually work with several top-tier diagnostic companies to improve the system's digital triage competences.

LabCorp, meanwhile, is offering consultation and guidance on the needs of clinical customers, Deetz adds.

3M Company is the critical resource when it comes to materials and processing to keep the single-use test cartridges affordable, continues Deetz. Precision injection molding would have cost upwards of $2 per cartridge, so Ativa opted to instead use low-cost, three-dimensional structured tapes. "One reason we're in Minnesota is because 3M is here and there are also about 30 to 40 tape processing companies."

At full production, cartridges for the Ativa Diagnostic Workstation will run about one-eighth the cost of the Abbott i-STAT cartridge, says Deetz. The price of Abbott’s silicon-based cartridges hasn't come down much after 20 years on the market because raw materials make them expensive to manufacture.

The workstation itself will run between $18,000 and $20,000, Deetz says, but "in many cases we will give the instrument away for people who are committed to buying a certain number of cartridges."

No "Black Magic"

All of this sounds a "bit like black magic," Deetz concedes, which is why Massachusetts General Hospital recently gave the Ativa Diagnostic Workstation a test run for more than 30 days on some of its most difficult samples—including those with many blast cells indicative of leukemia. The device's blast capability performed better than most big laboratory machines, he says.

About 20 new workstations will be built under good manufacturing practices document control, which will be used in more formal clinical studies, Deetz says.

The Ativa Diagnostic Workstation has greater computing horsepower than a typical POC device to give it the structure needed for the future of digital health, although the company knows the path to success requires a "steppingstone approach," he says.

Ativa Medical has pursued a testing capability that's "inexpensive, easy to use and broadly available to patients throughout the world" because it's a mission-driven company, Deetz notes. The end goal is earlier detection and treatment of serious diseases "resulting in improved well-being for all people."