Monday Morning Update 10/19/09

“I’ve said all along that HIMSS is a highly profitable bar with a business model in which ladies (providers) drink free provided they don’t mind being groped by men (vendors) who pay dearly for the privilege.”

Mr. HIStalk, you made my Sunday morning laughter meter reach 10. This reminds me of another business model, though this business involves another “s” word, that being sick patients. This is the saddest commentary I have read on the entire HIT industry.

Some here have asked for some real examples of HIT defects. Here is such a list from the FDA MAUDE (Manufacturer and User Facility Device Experience) database:

http://tinyurl.com/hit-defects

These reports were voluntary, many apparently submitted after reports by users. How many problems are not reported via this voluntary system?

” What some have asked for is evidence that computer records result in more errors than paper records.”

Such studies have not been accomplished. “Results” arise from the hype and opinions that were purchased by the HIT trade groups. This capriciousness has deceived the Congress to persuade the purchase of unproven devices. Interested readers and users of theses devices should add to the FDA data base of adverse events and device defects, which should include user interface issues.

Re: The PCPs of 63% of respondents use an EMR, which is astoundingly high and, frankly, hard to believe as representative.

When you asked about using EMR, were you intending to ask if EMR is used at point of care? That might ellicit different responses.
Plus, I’m assuming your typical reader / survey respondent is not the average American.

Programmer moves the goalpostrs once again.

I did nothing of the kind. If you think computer records are more dangerous than paper records, it’s up to you to prove it.

RE: MEDITECH’s request for the purchase of their scanning and archiving solution; our client’s are also being told that to migrate to 6.0 the MEDITECH solution is required.

The prerequisite of Scanning & Archive (SCA) and Data Repository (DR) for Focus (C/S 6.0) is true. Since it is a rewrite and the Focus platform, they store and reference items in both, making both required for movement to Focus. In addition, there is also a hardware requirement of a SAN.

From Simon Kirke: “Re: Meditech 6.0. I hear Meditech is forcing customers to buy a scanning and archive product with Meditech 6.0, even if they have a current document imaging solution. Can you verify this?” The floor is open.

“If you think computer records are more dangerous than paper records, it’s up to you to prove it.”

Non-inferiority (to a comparable product) is the minimum proof needed to be acceptable for use in clinical medicine. It is the onus of the manufacturer of the new product to accomplish this. The HIT vendors have violated cardinal precepts of scientific study with the aid of the purchased opinion from its trade groups.

What’s more wrong, paying for an interface with a vendor or not being able to have an interface from your vendor? Obviously not all interfaces are reusable, but I agree that charging for each interface is a little much…but at least you are one of the fortunate ones that uses a vendor that interfaces or “plays well with others”. I have a hard time buying the excuses from some vendors out there that say that you have to buy alllllll of their software & tools if you want to have a properly functioning infrastructure. While interfacing is not ideal, I think it’s rather asinine that you are pigeonholed (or penalized) if you choose to get a far superior product for your hospital. This obviously won’t fly for much longer with the requirements for interoperability.

Programmer wrote:

“What some have asked for is evidence that computer records result in more errors than paper records.”

Such information would be useful as well. However, considering the lack of transparency in the health IT industry about defects and failures, close calls and patient injury, such a calculation is not currently possible. The voluntarily reported information I posted is a small fraction of some larger total.

Those who work for HIT vendors, would you care to share your company’s HIT defects and adverse outcomes/close calls events list? Or is that forbidden?

However, considering the lack of transparency in the health IT industry about defects and failures, close calls and patient injury, such a calculation is not currently possible

Why do you need health IT industry participation? That isn’t going to give you any info about the pre-IT error rates. It shouldn’t be very difficult to compare pre-IT error rates with IT error rates.

It is not possible to calculate the value of n/(unknown), where n=error rates due to paper, and (unknown)=error rates due to IT.

You don’t need to know the error rates due to IT or paper. You just need to know the overall pre-IT error rates and the post-IT error rates.

Of course, the fact that you need an IT person to tell you this is a pretty good indicator of why you need IT.

Pezman writes:

“Even a layperson knows that there were FAR more errors and points of failure in a paper-based world then there is or has been in a paperless world.”

No, those who study these issues objectively don’t know that, at all; we cannot know it since information on HIT-caused clinical mishaps is shielded by the industry.

Even the Joint Commission in their Sentinel Events Alert on HIT stated “There is a dearth of data on the incidence of adverse events directly caused by HIT overall” ( http://www.jointcommission.org/SentinelEvents/SentinelEventAlert/sea_42.htm ) They are not saying that idly, or because such events don’t happen.

This type of statement about layperson wisdom reflects a lack of scientific understanding and/or of logical reasoning. The fallacy is called Argumentem Ad Populum (http://en.wikipedia.org/wiki/Argumentum_ad_populum). An argumentum ad populum (Latin: “appeal to the people”), in logic, is a fallacious argument that concludes a proposition to be true because many or all people believe it; it alleges, “If many believe so, it is so.”

This lack of scientific understanding and logical reasoning, alien to medical science and culture, is a form of professional dyscompetence. It is why many IT personnel are dangerous in clinical environments unless overseen by domain experts.

By way of anecdotal experience, (with the caveat that anecdotal experience does not make data), in my own clinical experience and that of many of my colleagues, we in fact cannot recall a medical error being committed due to problems in the paper record, including unavailability or illegibility. I, in fact, did my pulmonary and neurology fellowships at a hospital that in that time period did not staff the records room and kept it locked on weekend evenings, Hahnemann in Philadelphia. No medical errors occurred as a result; we took out own histories and physicals and that was quite adequate to assure decent care. I don’t recommend that system, but it reflects a very possible irrational underconfidence in paper and equally irrational overconfidence in computing as HIT exists in 2009.

On the other hand, many of the events described in the MAUDE reports are already at the level of affecting patient care adversely. MAUDE represents the one-gram tip of the iceberg.

we cannot know it since information on HIT-caused clinical mishaps is shielded by the industry.

I already pointed out how this can be determined without HIT industry help. Do you really need the Stats for Dummies explanation?

Should have added this to the above comment as an example of the troubling populist belief that “doctor + computer = Marcus Welby.”

“High Rates of Adverse Drug Events in a Highly Computerized Hospital” ( http://archinte.ama-assn.org/cgi/reprint/165/10/1111.pdf ) that show that irrational exuberance and sticking one’s head in the sand is inappropriate for medicine. Evaluation studies in HIT are anything but simple A/B comparisons.

The best data in any study, of course, is primary data, not surrogate. The primary data on HIT-caused clinical close calls and mishaps is sorely lacking. This is why I have written that HIT needs regulation akin to that of pharma IT.

From the report posted above.

“It appears from
the shift in distributions of error-associated ADEs that
the VA computerized interventions worked to reduce error-
related ADEs almost exactly as designed.”

“Because the VA’s computerized patient
record system likely makes ADEs more visible than
would a paper-based system, this study does not support
the interpretation that the VA computerized patient
record system induces ADEs.”

They go on to recommend decision support, which really is the point. It’s possible with a well implemented HIT system, no so much on paper. Its also worth noting some of the reasons VA hospitals don’t have decision support, its mainly political.

It is a complex issue, it depends on the individual implementation, but I have still to see a valid study that shows HIT increases error rates.

By the way, the individual implementation issue is also a problem for the vendors. A patient safety issue in one hospital with a certain vendor, may not be in another hospital with the same vendor.

I have a hard time believing that if in truth that ADEs are increased because of implementing an HIT given the “facts” you presented. I presented an equally compelling counterargument before you even responded within my original post. If you look at some of the most automated healthcare establishments like the Cleveland Clinic or Mayo Clinic, you can clearly see that their workflow processes have greatly changed since they went from paper to paperless. Of course when initially starting out they worked around their paper workflows that they were used to and were only met with strife and benefit realization. Only through decision support, efficiency studies, and using the system they purchased were they able to realize what it is like to truly be in a paperless environment. Typically what you and most of the clinical community sees is a national healthcare landscape that is no more than 8% fully automated. There are tons of paper workflow processes still in effect mixed in with paperless “bolt on” applications.

I also like your definition of argumentum ad populum. I know and understand this…but do you not find that statement entirely ironic since it was posted on a blog for HIT.

Also, there is transparency in the US HIT vendors out there…just as much as there is across the pond. So, I fail to see your point that the grass is greener “internationally.”

“I have a hard time believing that if in truth that ADEs are increased because of implementing an HIT given the “facts” you presented.”

In fact it does not matter what you or I or others “believe” about HIT and adverse events; more data in the form of studies are needed considering the risks evidenced, for example, in MAUDE.

Merck’s leaders, for example, believed VIOXX quite safe until more postmarketing surveillance data came in.

My approach is one of medical caution: it can be summarized by this: “I see many red flags, and conflicting evidence as well on the benefits of HIT and these need to be investigated rigorously before we spend $20 – $50 billion on it nationally.”

This is in line with my medical training, where we were taught a simple rule that has served me well: “critical thinking always, or your patient’s dead.” (Victor P. Satinsky, MD, Hahnemann Medical College, 1972).

If people take issue with that, that’s their right, but it seems some who take issue with it have inherent conflicts of interest. I do not work for or have any investments in any HIT, IT, or technology company or affiliate.

My fear is the cyber equivalent of the “Libby Zion” incident (google it) that will result in forceful governmental intervention. I’d rather see regulation of HIT occur (which it most certainly will, as is occurring in the EU) under less catastrophic circumstances. Also, if litigation occurs over a known HIT defect that causes injury, the plaintiff lawyers will be asking the vendors why the defect was not publicly revealed. It is in the industry’s own best interests to be forthcoming. Pharma learned that to the tune of tens of billions of dollars.

“I also like your definition of argumentum ad populum. I know and understand this…but do you not find that statement entirely ironic since it was posted on a blog for HIT.”

Not sure I understand what you mean, but it’s not my definition. A whole page of logical fallacies is at this link: http://en.wikipedia.org/wiki/Category:Logical_fallacies . Learned most of them in high school debating. Required reading for my students.

“Also, there is transparency in the US HIT vendors out there…just as much as there is across the pond. So, I fail to see your point that the grass is greener “internationally.”

Not sure what you’re referring to here, either. However, if there were transparency, we’d see far more entries in MAUDE – as a simple example of what transparency means.

With that, I end my participation in this thread.

Mr HISTalk, can you maybe start a PaperRecordsTalk site where some of these folks can spend their time so we can focus here on the main issues for the future?

That’s a good point Mr HISTalk. I was just reporting back on the finding of a report posted to show CPOE is a bad thing, and to make sure there was no confusion over its conclusions.

However there are still 1.5 Million injuries caused by prescribing errors, 7000 deaths, every year that are not caught down the line. Resulting in 3.5 Billion in costs per year, and $887 million to Medicare alone.

Many are caused by “sounds like” errors, many by interaction problems, many by allergic reactions. These are human errors, that can be easily reduced using IT. Can anybody come up with a paper based system to reduce these errors, or should we just live with them?

Of course a Physician who is sloppy in the paper world is likely to be sloppy in an electronic world as well. So HIT is no magic bullet. But when we have accurate data, we can protect the Physician and the patient from mistakes. There are no valid and credible studies that show otherwise.

http://docs.google.com/gview?a=v&q=cache:d1RqItJzBuAJ:www.iom.edu/File.aspx%3FID%3D35943+National+Academies+of+Science's+Institute+of+Medicine+(IOM),+preventable+medication+mistakes+also+injure+more+than+1.5+million+Americans+annually&hl=en&gl=us&pid=bl&srcid=ADGEESigSrFxzh2QkKmcdNvVbVMVBENnxmFaCPA6f5htqpKZ1JMCSeEVZoCOI_4kra5_xgbs8xJYUUtGoOfBMY96-MU3yIdlRHo9dP-J-kAMFymcVNe3BXVjJxsf1ui7l6ePKmfvfEMa&sig=AFQjCNHC7wPOpHeTNsNEtppbjQ4lmuytPw

Mr Histalk and Inga,

Thank you for this resource.

I write to state that the data is not there to support using current HIT iterations on patients with complex illness and whose lives depend on pinpoint accuracy. MAUDE data cause distrust for these systems. You never know when the unexpected will rear its ugly head. Once you know where to look for patient care errors in this technology, they are abundant and do reach the patient.

Certainly, there will not be billions in savings as concocted by the US Government who based opinion on HIT industry sponsored “study”.

If the vendors wish to experiment to improve their devices to make them user friendly, safe and effective, then declare all current deployments and “go lives” (or deads) as experiments. Then get IRB approval and determine and record the dangers and injuries to patients, explore each unexpected death; and report the data to the FDA.

It has been suggested by other commenters on this dynamic blog for users to report problems of these computerized ordering machines to the FDA.

While you are at it, write to your Congressmen (and women) to educate them that CCHIT Certification is not a stamp of safety and efficacy.