- CAMBRIDGE—Healthcare blockchain currently has three good use cases — physician credentialing, data integrity, and data sharing consent — observed Beth Israel Deaconess Medical Center (BIDMC) CIO John Halamka in a keynote address at the MIT Sloan Healthcare and Bioinnovations Conference held here Feb. 22.
“Every year we have to send to insurance companies documents proving that we went to medical school, did a residency, haven’t committed a felony, and all the rest of that stuff,” said Halamka, who is also a professor of medicine at Harvard Medical School.
Blockchain could be used as the “source of truth” where insurance companies and others can go to verify provider information, such as demographic data and physician credentials.
The Synaptic Alliance is as example of this blockchain use case. The alliance includes Aetna, Ascension, Humana, MultiPlan, Optum, Quest Diagnostics, and UnitedHealthcare. The first project of the Synaptic Health Alliance aims to put provider demographic information on a permissioned blockchain for members to access and maintain.*
Blockchain could also be used to ensure data integrity for clinical trials and medical records. Halamka noted that sometimes there are accusations that data was falsified after a clinical trial was completed. Blockchain could be used to record all the observations in the notebook in real-time, then the blockchain could prove that the notebook wasn’t altered.
Another blockchain use case is data sharing consent. “If I had data sharing consent preferences, I could use blockchain to enumerate my consent perferences. Then those who wanted to exchange my data would look at the blockchain to find out my consent preferences,” he related.
“There are lots of start-ups and entrepreneurs working through those kinds of use cases,” he observed. “Be careful about using blockchain. It’s not a database, it’s not an analytic tool, it’s not going to solve all the world’s problems. The use case needs to be very specific. The first question you should ask is, do you really need blockchain? For most use cases, you don’t,” he added.
Halamka also identified some new digital tools that are having an impact on healthcare this year: Internet of Things (IoT), artificial intelligence (AI) and machine learning, big data analytics, and EHR APIs and services.
“We are starting to see the emergence of enough good healthcare Internet of Things devices … but there is a question of what you’re going to do with all of the data,” Halamka observed.
Physicians will be overwhelmed with all the data coming from healthcare wearables and other IoT devices, he said. They also can’t be sure of the quality of data coming from those devices.
“Internet of Things devices are exploding in the marketplace. They could be useful for wellness care, but figuring what to do with that data and medical liability around the data are still unknown,” he said.
AI and machine learning could be used to allocate operating room (OR) times based on patient characteristics. Halamka said that his facilities employed machine learning for this use case and was able to free up 30 percent of the OR schedule. “The OR is the most expensive place in the hospital. If you could free up 30 percent of your capacity, instead of building more ORs, you could see more patients and deliver higher quality at lower cost,” he said.
“We have partnered with Amazon and Google to build algorithms around simple business process flow questions, looking at the patients of the past to inform the care of the future,” he said.
Halamka cautioned that healthcare data quality is “terrible,” so there could be errors, biases, and incompleteness in the data. “So be careful. Do not use raw data from a healthcare institution in a machine learning algorithm. We send the data from all of our sites of care to a central data science repository where 85 data scientists remove the erroneous data; only then do we use it for a narrow purpose,” he said.
There are many companies developing APIs and third-party services to work with EHRs made by Epic, Cerner, and others. The large EHR providers “will always be there as the core transactional engine, but increasingly they will be surrounded by apps and services that enhance EHR usability,” he said.
On the policy side, “if I have the Office of Civil Rights fining me $1 million for every privacy breach, I’m going to have a disincentive to share data. What I need are rules of the road to tell me when and how to share data,” Halamka said.
“If I can convince my board that providing more data to more people for more purposes is what we need for value-based purchasing, then they will say ‘yes, go ahead and do that.’ Obviously, hospital administrations are risk adverse,” he said.
“It’s not often a technology issue, it’s a psychiatry issue. I don’t want to share my data. There’ll be a penalty, or I’ll violate somebody’s privacy and I’ll get in trouble. But value-based purchasing requires us to share, if we are going to stay in business because we are paid to keep people healthy,” he concluded.
*This article has been updated to make clear that the Synaptic Alliance pilot project is focused on using blockchain to maintain provider demographic data.