Health IT Roundup—Google, IBM team with industry groups on AI standards

Google, IBM, consumer tech groups developing best practices for AI in healthcare

Big names in technology, including Google, IBM, AT&T, and Fitbit, are teaming up with the Consumer Technology Association (CTA) to develop standards and recommend best practices around the use of artificial intelligence in healthcare.

More than 30 organizations, including vendors and industry groups like AdvaMed and the American Telemedicine Association, have joined the effort to drive industry consensus and standardization on definitions and characteristics of healthcare AI.

“CTA recognizes the need to address complex issues associated with the use and application of AI solutions in health care. The effort will serve as a platform for stakeholders across the tech and health care industry to create common terminology and best practices for management and oversight of data,” the organization said in a statement.

Through its work in providing best practices, the group aims to ultimately enhance health outcomes, improve efficiencies and reduce health care costs. As part of the initiative, a working group will focus on such topics such as trustworthiness, ethics and bias.

“This unique working group represents a diverse set of stakeholders across the ecosystem, including clinicians, manufacturers, regulators, public policy and civil rights organizations. The work produced will provide an informed framework for the use of AI in the context of health care,” said Rene Quashie, vice president, policy and regulatory affairs, digital health, CTA. (Press release)

UPMC researchers use machine learning to reduce lung cancer false positives

Researchers at the University of Pittsburgh and UPMC Hillman Cancer Center have found a way to sort out benign from cancerous nodules in lung cancer screenings using machine learning without missing a single case of cancer, which significantly cuts down false positive rates.

Lung cancer is the leading cause of cancer deaths worldwide. Screening is key for early detection and increased survival, but the current method has a 96 percent false positive rate. In a study published in the journal Thorax, UPMC researchers said they applied AI to lung cancer screenings from 218 patients and were able to rule out cancer in about one-third of patients, saving those patients from unnecessary biopsies, PET scans or short-interval CT scans.

A low-dose CT scan is the standard diagnostic test for lung cancer for those at high risk. Nationwide, about a quarter of these scans turn up a positive result, but fewer than 4 percent of those patients actually have cancer. 

“For the 96 percent of people who have benign nodules, these procedures are unnecessary. So, we try to mine the data to tell which are benign and which are malignant,” study coauthor Panayiotis Benos, Ph.D., professor and vice chair of computational and systems biology and associate director of the Integrative Systems Biology Program at Pitt. (Study)

Consumers concerned about healthcare organizations’ cyber defenses

Almost half of consumers (45%) say they believe that their health information is more secure on their personal electronic devices, such as their iPhone or laptop, than on the electronic devices within their healthcare provider, according to a recent survey from Morphisec.

Half of consumers also say they don’t know if a cyber attack has hit their healthcare providers, despite HIPAA laws requiring providers to notify patients when their information has been compromised. Consumers are more fearful of a health data breach (59%) than hackers gaining access to an internet-connected medical device (41%).

Consumers believe web and endpoint defenses are providers’ weakest links in protecting their data, according to the survey.

Healthcare portal use by consumers has increased to 42% this year from 28% in 2018, increasing the need for providers to protect Internet-facing websites, which hackers are actively targeting, according to the report. (Report)