Thursday, July 15, 2010

NORCAL Mutual Insurance Company: "Electronic Health Records: Recognizing and Managing the Risks"

The Insurance Industry is catching on that EMR's and other clinical IT are not exactly the cybernetic miracles they're sometimes held out to be, for example as implied in this statement from HHS in the July 13, 2010 NEJM that-

... EHRs will improve caregivers’ decisions and patients’ outcomes. Once patients experience the benefits of this technology, they will demand nothing less from their providers.

NORCAL Mutual Insurance Company, for example, produces a near-monthly publication entitled "Claims Rx." Its purpose:

The goal of Claims Rx is to help physicians better recognize their medical professional liability risks and implement strategies to minimize those risks. The topics addressed here are derived from numerous sources including closed malpractice claims analyses, emerging liability trends and current medical literature. Each issue is meticulously reviewed by an editorial board consisting of physicians, nurses, risk management specialists and attorneys with a mind toward optimal patient safety and proactive risk management.

The October 2009 issue, found at this link in PDF, is entitled "Electronic Health Records: Recognizing and Managing the Risks." It contains advice to physicians to limit their liability, reduce errors, and presents de-identified examples of HIT-related horror stories.

Some of the advice offered includes:

Electronic health records (EHRs) hold great promise of improving patient safety and decreasing medical liability exposure, but their use is creating a variety of new risk management and patient safety issues.

For some reason, the "great promises" made over the past fifty years still seem elusive...but we'll get it right some day...(perhaps when these issues among others are solved).

Some of these issues are directly associated with EHRs (e.g., providers disregard warnings generated by the EHR), but many of the risk concerns associated with EHRs are analogous to problems that currently exist in paper documentation systems.

The marketing puffery that today's EMR's are vastly superior to paper might be just that - marketing puffery.

In this month’s Claims Rx we present a number of shorter-than-usual case studies that exemplify various aspects of unsafe EHR documentation and communication practices. The scenarios are based on NORCAL closed claims, facts presented in appellate opinions, research findings and the observations of NORCAL Risk Management Specialists.

What many of the examples show is that EHRs do not eliminate many of the dangerous documentation and communication practices that have historically led to patient injury and malpractice lawsuits.

I would add "EHR's in their present form do not eliminate dangerous practices", designed as if they were clinical data inventory systems by business-IT eggheads rather than as clinical tools "of, by and for clinicians." (One might ask why we are about to spend 100+ billion dollars on them in national rollout the next few years.)

Consequently, while it is important to address new issues that arise with EHRs, many of the risk management recommendations that apply to a paper-based documentation system remain valid.

This Claims Rx will discuss the risks associated with various aspects of EHRs and will provide guidance for instituting policies and procedures designed to enhance the quality and safety of patient care, while diminishing professional liability risk.

A number of case examples then follow. I will reproduce two, but all can be read via downloading the PDF at the link above.

Saving Images in the Wrong Patient’s Chart

Just as an image can be misfiled or lost in a paper system, it can be misfiled in an electronic one. However, as the following case shows, it can be less obvious that an image has been misfiled in an electronic system.

Case Study

Patient #1 and patient #2 both presented to the Emergency Department (ED) complaining of abdominal pain. CT scans of the abdomen and pelvis were completed for both patients. A radiology tech mistakenly gave patient #2’s images the identification number assigned to patient #1 and uploaded the images into the Picture Archiving Computer System (PACS).

A short time later, the tech realized his mistake and called the on-duty teleradiologist to tell him about the mistake and request that the mislabeled images be deleted from the system. However, the on-duty teleradiologist did not have access to delete images from the PACS; this had to be done by the PACS administrator. The tech then corrected the labeling problem and sent the images out to the teleradiology service for a preliminary review and resent the correctly labeled images to the PACS.

Patient #1’s PACS file now contained both his own and patient #2’s images. A few days later the tech told his supervisor about the mislabeling, and assumed that the supervisor would remedy the problem. Pursuant to hospital policy, the tech should have immediately contacted the PACS administrator.

The teleradiology service reported that patient #1’s CT scan was normal. Patient #2’s CT scan, however, showed a large tumor (about the size of a grapefruit) on the patient’s kidney. The service faxed the reports to the radiology department at the hospital.

The next morning, the on-duty radiologist reviewed the PACS images from the night before. He disregarded the teleradiology service reports because they did not correspond to what he saw in the PACS. Because patient #2’s scan had been completed before patient #1’s, patient #2’s images were the first series in his file.

The on-duty radiologist noted the large tumor and dictated a note. Because patient #2’s images still carried patient #1’s identification number, the radiologist’s report was assigned to patient #1.

Patient #1 was subsequently seen by a number of specialists for the supposed tumor on his kidney. Seven days after the CT scan, he underwent a nephrectomy [lovely- ed.]. During the surgery, no mass could be positively identified on his kidney by his surgeons. Postoperatively, no tumor was identified in the removed kidney and pathology returned benign.

(Please note, once the filing mistake was recognized, patient #2 was notified and underwent a timely and successful nephrectomy.)

[In other words, through serendipity only one person was harmed, not two - ed.]

One wonders how many "cases" like this will arise out of the major-vendor EHR upgrade flaw that is claimed to have led patient data to go into wrong charts at not one, but several hospitals:

HIStalk Monday Morning Update 7/12/10

From Holy Smoke: “Re: Cerner. Misidentification incidents have been reported with Cerner PowerChart and Millenium in hospitals in Indiana, Michigan, and others after a Cerner upgrade. Entries are placed in the wrong electronic chart and reviewed data is for the wrong patient.” Unverified. I saw nothing in the FDA’s Maude database, so if it’s happening, customers should file an experience report.

(See my July 11, 2010 post '
Health IT and 'High Regulatory Standards': Criminal Negligence for Implementing Defective Systems That Put Data in the Wrong Charts?').

Next from NORCAL:

Checking the Wrong Box

In the following case, the appearance of the computer screen probably played a role in the medication error.

Case Study

A patient presented to his primary care physician (PCP) for the treatment of headaches and episodes of altered consciousness. The PCP prescribed amitripyline at 10 mg nightly. The PCP told the patient to escalate the dosage by 10 mg every three to four days until the pain was relieved, but not to exceed 50 mgs without consulting him.

When creating the prescription, the PCP intended to check off the 10-mg box in the computerized physician order entry (CPOE), but inadvertently checked the 100-mg box, which was right above it. In the medication instructions section, he indicated that five pills could be taken per night, so the patient would not have to return to the pharmacy and pay an additional co-pay if he ultimately needed the larger dose.

The pharmacist had noticed that the dose seemed high and requested that a call be made to the PCP prior to it being dispensed. A nurse at the PCP’s office picked up the call, and because she was very busy that day, told the pharmacy to dispense the medication as it had been ordered — she did not check the dose. Three days later, the patient took five of the 100-mg pills together. Early the next morning, the PCP was contacted by an emergency department (ED) physician who reported that the patient was in the ED reporting dizziness, an altered state of consciousness, an inability to coordinate his movements and a rapid heartbeat.

He was further informed by the ED physician that the patient had taken five 100-mg amitripyline tablets. The PCP then checked the patient’s record and realized his mistake.

[It's serendipitous that the patient was not in the morgue when the mistake was realized - ed.]

Read the rest of the cases and the explanations at the Electronic Health Records: Recognizing and Managing the Risks "Claims Rx" document from NORCAL here (PDF).

In these cases, both technological and "people" issues were responsible for the malpractice events.

However, healthcare is an extremely complex endeavor requiring exquisite attention to detail (or your patient's dead).

These IT systems could have been designed to provide cognitive, ease-of-use, and known-error revision (or at least known-error flagging) support to clinicians that could have helped prevent these errors.

HIT designers need to do their part if they want to be considered part of the clinical team.

If they don't do their part voluntarily and stop with the mission-hostile,1970's-paradigm "inventory system of widgets" health IT, as I've written before, they will increasingly find themselves part of that team in an involuntary manner - as defendants in litigation.

Finally, due to issues such as: an utter lack of governmental regulation of the HIT industry and a lack of defect and error reporting requirements; contractual gag and hold-harmless clauses (as raised by Penn researchers Koppel and Kreda in JAMA here); physician fear of hospital retaliation such as sham peer review for HIT whistleblowing especially now that more physicians are becoming hospital employees; and other causes, there is no reliable data on the incidence of EHR-related medical errors and malpractice.

As in my paper "Remediating an Unintended Consequence of Healthcare IT: A Dearth of Data on Unintended Consequences of Healthcare IT", there is also no reliable data on "near misses" or potential IT related mishaps that were averted by serendipity. These near misses expose patients to risk. With no reporting, there is no systematic data that can be used for remediation and prevention.

I thus present once again the following thoughts from my post
Science or Politics? The New England Journal and "The 'Meaningful Use' Regulation for Electronic Health Records":

I believe we should hold off national Health IT roll outs until we:

  • learn sufficiently from failures such as the UK CfH and our own military's AHLTA debacle on how to avoid same, which can injure and kill patients and wastes massive money and resources healthcare can ill afford, and more importantly that can be better used elsewhere - such as care of the poor;
  • improve the technology's usability, safety and efficacy through the years of Medical Informatics and other disciplinary research needed, that was short circuited through the invention of the ONC office by Bush (although national HIT then remained a goal, not a mandate), and the 'militarization' of ONC under Obama whereby HIT was unilaterally declared a proven technology and mandated for national rollout;
  • end the contractual and fear-based censorship of information on health IT problems, and patient injuries and deaths related to the devices; and
  • meaningfully regulate these devices that have increasingly become governors of care delivery.
I have written extensively on these topics at this blog, at my academic website on health IT failure, and other sources (see list at end of my bio).

When there are significant doubts about a medication or medical device, we ought not push for national rollout.

Health IT devices have gotten special accommodation, and it's not on the basis of any rigorous science I am familiar with.

-- SS


Anonymous said...

These devices remain experimental. They are devoid of meaningful usability and have not any proven safety and efficacy. They should be recalled, but who is to do that when the government is creating financial incentives for them to be out there?

Anonymous said...

Gee, these devices are not only meaningfully unusable and meaningfully risky, but meaningfully unmanageable when something goes wrong. Just a few minor human factors.

Anonymous said...

Insidious risks to the patients are promoted by CPOEs. Their meaningful unusability equates to errors, even by the most expert clinicians. You have described the primary etiology of their errorgenicity, ie error promoting features. The companies know but do nothing. Trial lawyers take heed.

Australian Health said...

When deciding to use different medical technologies, sometimes the decision becomes the lesser of two evils.

InformaticsMD said...

Australian Health said...

When deciding to use different medical technologies, sometimes the decision becomes the lesser of two evils.

As I point out in this post and at a many others, the incidence of HIT-related morbidity and mortality, as well as HIT-related "near misses" averted by serendipity, is unknown.

Therefore it cannot be determined which is the lesser of two evils - HIT vs. paper.

To make nation-level social engineering decisions based on assumptions and guesswork about health IT is inconsistent with both science and ethics.

-- SS