Main menu


Ai medical technology in United States hospitals

Ai medical technology in United States hospitals

How US hospitals apply AI medical technology?

The Artificial intelligence  AI ​​medical industry has always been the focus of capital attention. 

Relevant industry analysts believe that  by 2026 the US medical system can save 150 billion US dollars per year.

But from the perspective of real users, there are almost no cases of hospitals applying machine learning. 

As a policy-intensive industry, any changes in the health care system, including employee insurance coverage, hospital management policies, etc., will directly affect the development of the industry. 

Therefore, it is particularly important to discuss the actual application of AI medical /ML (Machine Learning) in hospitals.

This article will focus on the analysis of the use of AI medical and other technologies in the top five hospitals in the United States (Massachusetts General Hospital, Mayo Clinic, Cleveland Clinic, Massachusetts General Hospital, Johns Hopkins Hospital). 

Answer three questions:

1. What types of machine learning and applications are currently being used in hospitals?

2. Can these innovative applications become the future trend of the healthcare industry?

3. How well are hospitals investing in machine learning and emerging technology innovation?

Application commonality

At present, the top five hospital AI medical  applications mainly revolve around three aspects:

1. Predictive analytics Monitor patients and handle patient emergencies by analyzing key indicator data.

2. Chatbots automate physician consultations, arranging appropriate specialists for referrals.

3. Predictive Health Tracking Monitor patient health by collecting real-time data.

AI medical: Mayo Clinic

In January 2017, the Center for Personalized Medicine at Mayo Clinic partnered with Ai medical  startup Tempus to explore personalized cancer care through machine learning.

The collaboration is based on "molecular sequencing and analysis of 1,000 Mayo Clinic patients enrolled in immunotherapy studies" conducted by Tempus on multiple cancer types including "lung cancer, melanoma, bladder cancer, breast cancer and lymphoma." analyze". 

The Ai medical collaboration is currently in the research and development stage. 

The Mayo Clinic hopes to use the results of the analysis to provide more customized treatment options for Mayo cancer patients.

"Tempus is trying to build a database large enough to make recommendations for cancer patients," said Tempus co-founder and CEO Eric Lefkofsky.

In addition to the partnership with Tempus, heart disease has become the leading cause of death in the United States, according to the Centers for Disease Control and Prevention (CDC). 

March 2017, Mayo Clinic with medical device manufacturers

Omron Healthcare co-funds $30M D for heart health care startup AliveCor round of investment. 

Kardio Pro by AliveCor is an AI medical -driven platform designed for clinicians with early detection capabilities for monitoring atrial fibrillation in patients.

 AI medical:Cleveland Clinic

an other exemple for application of Ai medical, in September 2016, Microsoft partnered with the Cleveland Clinic, and researchers used Microsoft's AI digital assistant, Cortana, to perform predictions and analysis to help medical centers "identify potential high-risk patients in the ICU."

In 2014, Cortana was integrated into the Cleveland Clinic's Hospital system. 

Currently, Cortana monitors "100 beds in six ICUs" from 7pm to 7am. 

Monitor patients at high risk for cardiac arrest. 

If a patient has sudden cardiac arrest, they will need to be given vasopressors, which are also part of the Pulseless Sudden Cardiac Arrest Management Protocol, but vasopressors can also raise blood pressure. 

The researchers aimed to determine whether a patient needed a blood pressure booster.

The data store collected by the ICU, such as patient life cycle and laboratory data, will be housed in Microsoft's Azure SQL database to build predictive analytics data models.

AI medical: Massachusetts General Hospital

The Massachusetts General Hospital Clinical Data Science Center announced in April 2016 that it had become a "Foundation Technology Partner" with NVIDIA. 

The center is designed to act as a hub for AI medical for "the detection, diagnosis, treatment and management of disease".

At the 2016 GPU Conference, NVIDIA officially proposed the NVIDIA DGX-1, known as the "deep learning supercomputer". 

The machine will learn from Massachusetts General Hospital's database of "10 billion medical images" for use in radiology and pathology. 

The center aims to later expand into electronic health records (EHR) and genomics.

NVIDIA DGX-1 will alleviate some of the challenges facing the field:

"If we can seamlessly capture relevant data in a highly structured, thorough, granular approach, it can ease the burden on physicians.,and get the data needed for AI medical to work in new ways." Co-founder and RemedyHealth CEO Will Jack said.

AI medical: Johns Hopkins Hospital

An other application of AI medical , in March 2016, Johns Hopkins Hospital announced the launch of the Hospital Command Center, which, in partnership with GE, designed the Judy Reitz Capacity Command Center, which can receive 500 messages per minute and will connect data from "14 different hospital ITs" System” data is integrated into 22 high-resolution, touch-screen computer monitors.

As a result, command teams are better able to identify and mitigate patient risk and make appropriate scheduling. "Prioritize patient situations and trigger interventions to expedite patient flow."

Since the launch of the command center, Johns Hopkins Hospital said it has seen a 60% increase in its capacity to admit patients with "complex conditions" from the surrounding region and the entire country, a 30% increase in the rate of bed allocation in the emergency department, and a 30% increase in the number of patients discharged by noon. an increase of 21%.

AI medical:UCLA Medical Center

In March 2017, UCLA researcher Dr. Edward Lee presented a virtual interventional radiologist (VIR) design at a conference. 

Essentially, VIR is a chatbot that "automatically communicates with referring clinicians and quickly provides evidence-based answers to frequently asked questions."

Currently, VIR is in beta mode, and the first VIR prototypes are already being used by a UCLA panel of health professionals, including "hospital physicians, radiation oncologists, and interventional radiologists."

VIR is based on more than 2,000 example data points that reflect questions that often arise during radiologist consultations. 

Responses are not limited to text formats, but can also include "websites, infographics, and custom programs".

The research team used the IBM Watson AI system to integrate VIR with natural language processing capabilities. 

In the tradition of service chatbots, if VIR is unable to provide a response to a particular query, the chatbot provides the corresponding radiologist contact information. 

Now, the researchers intend to expand the app's capabilities to serve general practitioners who interact with other specialists, such as cardiologists and neurosurgeons.

summative thinking

Compared to e-commerce, the application of machine learning in healthcare faces strong constraints and regulations. 

At present, the application of AI medical technology in these top hospitals has not produced obvious results. Health technology investor Dr. 

Steve Gullans analyzed the implementation of AI medical  technology in an interview. 

He believes that many expert physicians' fear of AI medical tools is one of the factors that has led hospitals to adopt them too slowly. 

When asked how hospitals and healthcare organizations can get around these barriers (assuming the technology will actually improve patients' lives), he argues that we should remain skeptical of the technology's application until quantifiable results are validated.

Like almost every other industry infused with AI, machine learning in healthcare is generating a flood of "technical signals" (hype about "AI medical" for attention and pressure, when not actually improving organizational outcomes). 

It is mutually beneficial for an AI vendor like NVIDIA or Microsoft to award a top hospital with "new, super-best-in-class" AI medical  technology and the title of "Founding Technology Partner" to that hospital. 

These kinds of events will almost certainly attract media attention and (potentially) benefit both parties—whether AI medical applications deliver results for hospitals (efficiency) or patients (better health outcomes).