- The University of Pittsburgh Medical Center (UPMC) is investing in a cloud-based healthcare operating system (hcOS) to improve healthcare interoperability and break down data silos.
The downloadable operating system is designed to make it easy for app developers to build software solutions that access clinical data and clinical systems.
“The IT problem we are trying to solve is interoperability. We are trying to access all of the healthcare data that can solve healthcare problems for healthcare participants in a way that makes it easy for them to access the data,” said Tal Heppenstall, president of UPMC Enterprises, the commercialization arm of UPMC.
“We are trying to get data out of the silos so it can be used in a variety of applications. The hcOS is a layer between existing systems and applications,” Heppenstall told HITInfrastructure.com.
“Right now, if you build an application at one of your hospitals, it only works at that hospital. The goal of hcOS is provide a set of tools that will allow people to build applications that will work across any number of hospital systems,” he added.
UPMC recently filed a trademark application for the technology with the U.S. Patent and Trademark Office. The application states that hcOS is a “downloadable cloud-based software for providing a healthcare information technology infrastructure whereby healthcare data is available to providers, consumers, and payers.”
It is a “platform as a service featuring computer software platforms that provides tools for developers to create healthcare software applications.”
Heppenstall noted that existing legacy systems, such as EHRs, were built to perform a particular function.
“We want to get the information collected by the EHR and to surface it for other purposes. There are a lot of things in healthcare that need to get done that don’t happen in the EHR. We believe that by freeing the data from its silos, we can actually use it for creative applications,” he explained.
By freeing up the data, it could be used for other purposes, such as a natural language processing tool.
“When a clinical note is in an EHR, the physician can pull it up and read it on the screen. Natural language processing tools would enable the physician to have the computer read that note and find things that are comparable in other patients. This would enable the physician to do his or her job better,” Heppenstall related.
“This is where we see hcOS being a differentiator to allow different types of data to get surfaced, whether it’s structured data, unstructured data, speech, or images. Our goal is to make sure those are available for a variety of uses,” he said.
UPMC is currently using hcOS for some solutions at its facilities and in discussions with other organizations to deploy it outside UPMC.
UPMC is also using machine learning to reduce hospital readmission rates and false positives in lung cancer screenings.
UPMC has developed an algorithm that identifies hospital patients who are at high risk of re-hospitalization within 30 days of discharge. In a pilot project using one patient unit, the algorithm was able to cut re-hospitalization in half.
In addition, UPMC has employed machine learning to significantly reduce false positive for lung cancer. UPMC said that this is the first time machine learning has been used to sorting out benign from cancerous nodules in lung cancer screening.
“We were able to rule out cancer in about a third of patients, so they wouldn’t need biopsies, they wouldn’t need PET scans or a short-interval CT scan. They just need to come back in a year,” said David Wilson, co-director of the UPMC Hillman Lung Cancer Center and an associate professor of medicine, cardiothoracic surgery, and clinical and transformational science at the University of Pittsburgh.
“The next step is to evaluate this technique in a larger population, and actually it’s started already, using about 6,000 scans from the National Lung Screening Trial,” Wilson added.