Paul Egerman, Co-Chair, HIT Policy Committee Workgroup on Certification & Adoption
The merits of working with the FDA to help ONC craft an EMR certification program that addresses both pre- and post-market HIT patient safety issues were discussed at the HIT Policy Committee Workgroup on Certification and Adoption meeting held this week.
On one side of the debate, HIT Software Entrepreneur and Workgroup Co-Chair Paul Egerman argued that the FDA could bring some significant value to the table in crafting a program. “We could collaborate on certification criteria that improves patient safety. The ONC and FDA can work together to look at the FDA’s Quality Systems Regulation (QSR) and see what parts might work well for EHR systems.”
Egerman talked about developing a post-market surveillance process that would offer software buyers — such as hospitals, physicians and consumers —a mechanism to report vendors who were not in compliance with soon-to-be-defined certification criteria. “The certification organization can then issue a warning or revoke the vendor’s certification. That’s a powerful lever and the FDA can help us figure out the process, because the devil is in the details.”
Offering a different view, Carl Dvorak, executive vice president at Epic Systems Corporation and workgroup member, cautioned that the FDA might not be the best organization to help ONC create an environment where people felt comfortable coming forward with incident reports.
Dvorak said he also had concerns that the FDA’s QSR process “would not be as conducive as it sounds on the surface. The kinds of problems people spoke about (in recent workgroup testimony) aren’t addressed by what the FDA does. If the mission is to build a safer system, I don’t know that the FDA can contribute toward that mission.”
As to a specific instance where the FDA’s greater leverage (ONC is restricted to governing EMR certification) could be of value, Egerman gave the example of pharmacies that operate software which cannot process cancellations or give compliance data back to ordering physicians. “There are no policy levers at ONC to help us fix that.”
Dvorak asked if CMS wasn’t a better partner for ONC in its quest to address those types of issues. “Is there any more opportunity for CMS to interact directly with ONC? I don’t know the FDA has the right set of skills or processes to enforce that with the pharmacy. That might be better to give to CMS.”
Egerman responded that he still “viewed (the FDA) as potentially a valuable ally.”
Though in favor of engagement, Egerman said he was sensitive to concerns that some FDA processes might slow innovation. Going through the FDA process at one point with a company he owned, Egerman said it added “a six-figure cost” to the software which caused “very little new development” because each improvement would have triggered additional FDA approval processes.
Others thought that since the FDA usually kicked into action upon incidents of death or serious injury, ONC needed to create a system that brought, by comparison, mere unsafe conditions to light.
Workgroup Member Joan Ash, Ph.D., — Professor and Vice Chair, Department of Medical Informatics and Clinical Epidemiology, School of Medicine, Oregon Health & Science University — said any surveillance process needed to appreciate that “these devices are embedded into a larger socio-technical system.”
In an effort to move the discussion forward, Egerman added, “We have to create a certification program, and we should do our best to make it as good as it can be. This is a way we can put some patient safety concepts into it.”
When asked by Egerman, “Do we want to work with the FDA — that is the fundamental question?” HIT Policy Committee Co-Chair Paul Tang, M.D., CMIO at the Palo Alto Medical Foundation, responded, “The FDA has advice to give us, so we don’t want to not talk with them. They want to talk with us.”
Gerry Higgins says
The FDA has a poor history of dealing with patient safety issues. Their response was often punitive to vendors, yet without any recommendations for product improvement.
On the other hand, until recently, they have erred on the side of using the U.S. population as a ‘beta test’ group for new FDA-approved pharmacologics. In the drug development industry, it used to be a joke that to obtain accurate drug side-effect data, multiply the percentages in the drug label by a factor of 5.
To adopt or adapt their QSR process would be a mistake – it is designed for another purpose. However, I believe that Paul Tang is correct – communication is a good thing.
Also- what about including pharmagenomic data in the EHR? When we move to a certified that is genome-enabled, then we will have to talk with them.
DISCLOSURE: I work for an FDA subcommittee.