Medical Brain

Machine learning reads and combines the patient's data, including free text in EHR notes, patient conversations and monitoring feeds, to produce the Intelligent Medical Summary, detect risks and drive the conversational assistant, and the company says the system learns from its users' daily clinical experience. Those models apply clinical modules built from guidelines by more than 50 full time physicians.

The guideline modules and the physician team carry value of their own, so the models power the core assistant rather than everything the company sells.

The assistant talks to patients directly, around the clock, and communicates with providers only when it judges that necessary, alerting doctors, pharmacy and approved caregivers when something changes. The company says it combines its AI with advice from its dedicated physicians, and clinicians receive summaries and risk alerts to act on.

What triggers an escalation to a clinician, and how a physician reviews the assistant's patient guidance, is not described.

The system is described as AI and machine learning clinical decision support that aggregates and understands structured and free text data from EHRs, patient conversations and remote monitoring, continuously updated with evidence based guidelines, and that learns from its users' clinical experience. Its chief technology officer and a director of medical engineering lead the build.

No model, training data or accuracy figure is named.

healthPrecision builds the Medical Brain with its own engineering, DevOps and medical engineering teams, and the physician team writes the clinical modules. The website privacy policy says third party service providers may help run the service.

Those providers, and any outside model, are not named.

A research report by Maimonides Medical Center's maternal fetal medicine department, published in the American Journal of Obstetrics and Gynecology, reports a 90.7 percent reduction in adverse obstetrical events with the Medical Brain, which the company also describes as a 91 percent reduction in red never events. Named partnerships include Atlantic Health System, OrthoCarolina, One Health and New York City's Administration for Children's Services, and Gartner cited the product in its Market Guide for Clinical Decision Support Solutions.

The obstetrics study was run at one center by its own clinicians, and no results are published for the other modules.

The assistant handles symptoms, medications, EHR records and monitoring data for patients through a mobile app and for providers on desktop. The public privacy policy covers the website, asks visitors not to send protected health information through it, and says personal information is not sold or shared beyond service providers.

How patient data in the product is retained, protected or used for the learning the company describes is not stated.

Medical Brain processes EHR data and patient conversations for provider organizations such as Maimonides, Atlantic Health System and OrthoCarolina. The published privacy policy and terms cover the website and are governed by New York law, and the policy refers visitors to HIPAA's definition of protected health information only to ask them not to send it.

No business associate statement or HIPAA position for the product is published.

The company runs back end, DevOps, mobile and QA functions under named directors, and the public site describes the product's clinical design rather than its security. The website privacy policy is the only security related document published.

No certification, audit report, security page or trust center is published.

Medical Brain gives patients guidance on symptoms, medications and chronic disease control through a conversational assistant, and gives clinicians decision support and risk alerts across obstetrics, cardiology, diabetes, sepsis and other conditions. Those are clinical functions aimed at both clinicians and patients.

No intended use statement or regulatory pathway is published for any module.

Clinical content comes from guidelines and more than 50 full time physicians, a director of medical QA oversees quality assurance, and the modules are updated as guidelines change. The system also learns from the daily clinical experience of its users.

No governance process for the learning loop, monitoring practice or performance across patient groups is published.

Clinical content rests on guidelines and a physician team, and the assistant routes changes in a patient's condition to doctors, pharmacy and caregivers. The website terms disclaim responsibility for reliance on third party content and are governed by New York law.

No accuracy figure, warranty or remediation commitment covers the assistant's guidance to patients or its alerts to clinicians.

Medical Brain aggregates data from EHRs and other sources, including free text notes, and delivers an Intelligent Medical Summary and risk alerts back to clinicians, with information kept continuously updated for the care team. It runs alongside the record systems of partners such as Atlantic Health System and OrthoCarolina.

Which EHRs it connects to, and by what interface, is not stated.

Patients use a mobile app, and clinicians get summaries and alerts on the office desktop and on mobile, with data drawn from EHRs, patient conversations and remote monitoring. Deployments with health systems and practices are set up as partnerships, and the modules can be customized to a practice's workflows.

Hosting, regions and tenancy are not stated.

Providers, health systems and employers request a demo, and the company frames the value around value based care contracts, earlier action before costly exacerbations and revenue from virtual care activities. Partnerships such as One Health's plan to reach 100,000 patients show the population scale it is sold at.

The company lists no price or contract structure.

Modules span obstetrics, orthopedics, cardiology (heart failure, arrhythmias, valvular disease), hypertension, diabetes, hyperlipidemia, COPD, renal failure, infectious diseases including sepsis, radiology appropriateness and lifestyle change, across ambulatory and acute care. Named settings include maternal fetal medicine at Maimonides Medical Center, visiting nurse and transition of care patients at Atlantic Health System, orthopedic practice at OrthoCarolina and independent primary care at One Health, and New York City's Administration for Children's Services used a COVID version for employees.

Results beyond obstetrics are not reported.