It's the meme that health IT "certification" is a certification of safety.
I expressed concern about the term "certification" being misunderstood even before the meme formally appeared, when the term was adopted by HHS with regard to evaluation of health IT for adherence to the "meaningful use" pre-flight features checklist. See my mid-2009 post "CCHIT Has Company" where I observed:
HIT "certification." ... is a term I put in quotes since it really is "features qualification" at this point, not certification such as a physician receives after passing Specialty Boards.
The "features qualification" is an assurance that the EHR functions in way that could enable an eligible provider or eligible hospital to meet the Center for Medicare & Medicaid Services' (CMS) requirements of "Meaningful Use." No rigorous safety testing in any meaningful sense is done, and no testing under real-world conditions is done at all.
I've seen the meme in various publications and venues. I've even seen it in legal documents in medical malpractice cases where EHR's were involved, as an attempted defense.
Now the WSJ has fallen for the health IT Certification meme.
An article "There's a Medical App for That—Or Not" was published on May 29, 2012. Its theme is special regulatory accommodation for health IT in the form of opposition to FDA regulation of devices such as "portable health records and programs that let doctors and patients keep track of data on iPads."
In the article, this assertion about health IT "certification" is made:
... The FDA's approach to health-information technology risks snuffing out activity at a critical frontier of health care. Poor, slow regulation would encourage programmers to move on, leaving health care to roil away for yet another generation, fragmented, disconnected and choking on paperwork.
The process already exists for safeguarding the public for computers in health care. It's not FDA premarket review but the health information technology certification program, established under President George W. Bush and still working fine under the Obama Health and Human Services Department. The government sets the standards and an independent nonprofit [ATCB, i.e., ONC Authorized Testing and Certification Bodies - ed.] ensures that apps meet those standards. It's a regulatory process as nimble as the breakout industry it's meant to monitor. That is where and how these apps should be regulated.
It's a wonderful meme. Unfortunately, it's wrong. Dead wrong.
Certification by an ATCB does not "safeguard the public." Two ONC Authorized Testing and Certification Bodies (ATCB's) admitted this in email, as in my Feb. 2012 post "Hospitals and Doctors Use Health IT at Their Own Risk - Even if Certified". I had asked them, point-blank:
"Is EHR certification by an ATCB a certification of EHR safety, effectiveness, and a legal indemnification, i.e., certifying freedom from liability for EHR use of clinical users or organizations? Or does it signify less than that?"
I received two replies from major ONC ATCB's indicating that "certification" is merely assurance that HIT meets a minimal set of "meaningful use" guidelines, not that it's been vetted for safety. For instance:
From: Joani Hughes (Drummond Group)
Sent: Monday, March 05, 2012 1:06 PM
To: Scot Silverstein
Subject: RE: EHR certification question
Per our testing team:
It is less than that. It does not address indemnification although a certification could be used as a conditional part of some other form of indemnification function, such as a waiver or TOA, but that is ultimately out of the scope of the certification itself. Certification in this sense is an assurance that the EHR functions in way that could enable an eligible provider or eligible hospital to meet the CMS requirements of Meaningful Use Stage 1. Or to restate it more directly, CMS is expecting eligible providers or eligible hospitals to use their EHR in “meaningful way” quantified by various quantitative measure metrics and eligible providers or eligible hospitals can only be assured they can do this if they obtain a certified EHR technology.
Please let me know if you have any questions.
Client Services Coordinator
Drummond Group Inc.
The other ATCB, ICSA Labs, stated that:
... Certification by an ATCB signifies that the product or system tested has the capabilities to meet specific criteria published by NIST and approved by the Office of the National Coordinator. In this case the criteria are designed to support providers and hospitals achieve "Meaningful Use." A subset of the criteria deal with the security and patient privacy capabilities of the system.
Here is a list of the specific criteria involved in our testing:
In a nutshell, ONC-ATCB Certification deals with testing the capabilities of a system, some of them relate to patient safety, privacy and security functions (audit logging, encryption, emergency access, etc.).
What was suggested in the email below (freedom from liability for users of the system, etc.) would be out of scope for ONC-ATCB testing based on the given criteria. [I.e., certification criteria - ed.] I hope that helps to answer your question.
I had noted that:
... My question was certainly answered [by the ATCB responses]. ONC certification is not a safety validation, such as in a document from NASA on aerospace software safety certification, "Certification Processes for Safety-Critical and Mission-Critical Aerospace Software" (PDF) which specifies at pg. 6-7:
In order to meet most regulatory guidelines, developers must build a safety case as a means of documenting the safety justification of a system. The safety case is a record of all safety activities associated with a system throughout its life. Items contained in a safety case include the following:
• Description of the system/software
• Evidence of competence of personnel involved in development of safety-critical software and any
• Specification of safety requirements
• Results of hazard and risk analysis
• Details of risk reduction techniques employed
• Results of design analysis showing that the system design meets all required safety targets
• Verification and validation strategy
• Results of all verification and validation activities
• Records of safety reviews
• Records of any incidents which occur throughout the life of the system
• Records of all changes to the system and justification of its continued safety
A CCHIT ATCB juror, a physician informatics specialist, has also done a guest post in Jan. 2012 on HC Renewal about the certification process, reproducing his testimony to HHS on the issue. That post is "Interesting HIT Testimony to HHS Standards Committee, Jan. 11, 2011, by Dr. Monteith." Dr. Monteith testified (emphases mine):
... I’m “pro-HIT.” For all intents and purposes, I haven’t handwritten a prescription since 1999.That said and with all due respect to the capable people who have worked hard to try to improve health care through HIT, here’s my frank message:ONC’s strategy has put the cart before the horse. HIT is not ready for widespread implementation.... ONC has promoted HIT as if there are clear evidence-based products and processes supporting widespread HIT implementation.But what’s clear is that we are experimenting…with lives, privacy and careers.
... I have documented scores of error types with our certified EHR, and literally hundreds of EHR-generated errors, including consistently incorrect diagnoses, ambiguous eRxs, etc.
As a CCHIT Juror, I’ve seen an inadequate process. Don’t get me wrong, the problem is not CCHIT. The problem stems from MU.EHRs are being certified even though they take 20 minutes to do a simple task that should take about 20 seconds to do in the field. [Which can contribute to mistakes and "use error" - ed.] Certification is an “open book” test. How can so many do so poorly?For example, our EHR is certified, even though it cannot generate eRxs from within the EHR, as required by MU.To CCHIT’s credit, our EHR vendor did not pass certification. Sadly, our vendor went to another certification body, and now they’re certified.MU does not address many important issues. Usability has received little more than lip-service. What about safety problems and reporting safety problems? What about computer generated alerts, almost all of which are known to be ignored or overridden (usually for good reason)?The concept of “unintended consequences” comes to mind.All that said, the problem really isn’t MU and its gross shortcomings, it is ONC trying to do the impossible:ONC is trying to artificially force a cure for cancer, basically trying to promote one into being, when in fact we need to let one evolve through an evidence-based, disciplined process of scientific discovery and the marketplace.
In that situation, the producer had been simply advising users (in critical care environments, no less) to "work around the defects" that could indicate incorrect recommended dosage values of critical meds, including a drug dosage up to ten times the indicated dosage, as well as corrupt critical cardiovascular monitoring data. As I observed:
... I find a software company advising clinicians to make sure to "work around" blatant IT defects in "acute care environments" the height of arrogance and contempt for patient safety.
Without formal regulatory authority to take actions such as this FDA recall, "safeguarding the public" is a meaningless platitude.
It's also likely the ATCB's, which are private businesses, would not want the responsibility of "safeguarding the public." That responsibility would open them up to litigation when patient injuries or death were caused, or were contributed to, by "certified" health IT.
I have in the past also noted that the use of the term "certification" might have been deliberate, to mislead potential buyers exactly into thinking that "certification" is akin to a UL certification of an electrical appliance for safety, or an FAA approval of a new aircraft's flight-worthiness.
The WSJ needs to clarify and/or retract its statement, as the statement is misinformation.
At my Feb. 2012 post "Health IT Ddulites and Disregard for the Rights of Others" I observed:
Ddulites [HIT hyper-enthusiasts - ed.] ... ignore the downsides (patient harms) of health IT.
This is despite being already aware of, or informed of patient harms, even by reputable sources such as FDA (Internal FDA memo on H-IT risks), The Joint Commission (Sentinel Events Alert on health IT), the NHS (Examples of potential harm presented by health software - Annex A starting at p. 38), and the ECRI Institute (Top ten healthcare technology risks), to name just a few.
In fact, the hyper-enthusiastic health IT technophiles will go out of their way to incorrectly dismiss risk management-valuable case reports as "anecdotes" not worthy of consideration (see "Anecdotes and medicine" essay at this link).
They will also make unsubstantiated, often hysterical-sounding claims that health IT systems are necessary to, or simply will "transform" (into what, exactly, is usually left a mystery) or even "revolutionize" medicine (whatever that means).
Health IT is a potentially dangerous technology. It requires meaningful regulation to "safeguard the public." How many incidents like this and this will it take before that is understood by the hyper-enthusiasts?
I've emailed the ATCB's that had responded to my aforementioned query for clarification on the WSJ assertion about their role, being that the statement is in contradiction to their earlier replies to me. I also advised them of the potential liability issues.
However, if it turns out to be true that the ONC-ATCB's do intend themselves as the ultimate watchdog and assurer of public safety related to EHR's, that needs to be known by the public and their representatives.