Life Saving Errors


On March 28, 1979 the Three-Mile Island Unit-2 nuclear power plant experienced a feed system failure which prevented the steam generators from removing heat from the plant. The reactor automatically shutdown but, without the feed system to cool the primary, the pressure in the primary system (the nuclear portion of the plant) began to increase. In order to prevent that pressure from becoming excessive, a relief valve opened. The valve should have re-closed once the pressure dropped by a small amount, but it didn’t. The only indication available in the control room showed the valve in the closed position, but that indication was erroneous, representing only that the signal to close the valve (pressure below a set value) had been sent to the valve. Nothing in the system verified the actual valve position. This stuck-open valve caused the pressure to continue to decrease in the system (and ultimately provided a path for spewing thousands of curies of radioactive material into the atmosphere), but the false shut indication prevented the operators from taking actions to mitigate their severe loss of coolant accident.

The primary relief valve design had a history of sticking. That same valve had been involved in at least nine other minor incidents prior to the TMI incident. Most notably, eighteen months before TMI, a similar incident had occurred in another nuclear plant involving a loss of feed and rising temperatures shutting down the plant. In that incident, the plant was just starting up after a maintenance shutdown, so the power level and temperature of the system were not as dangerously high as at Three-Mile Island.

The plant supervisors reported this incident to the NRC–even going so far as to note the disastrous consequences that might result if such an incident occurred during operation at power–but the TMI-2 operators were not aware of the earlier incident.

Several factors contributed to the severity of the TMI disaster, but the false indication of that pressure relief valve was arguably the worst contributor.

I became intimately familiar with the TMI incident because I was in training as a US Naval reactor operator at the time of the event. The Naval Reactors Office (a safety oversight office reporting to the Department of the Navy and the DOE) drilled us on every aspect of the TMI design and operations that led to or aggravated the TMI incident. One aspect that was clearly never going to happen in a Naval reactor, however, was a failure to learn from the earlier stuck-valve incidents.

The Naval Nuclear Power program had long had a model incident reporting system in operation. Any incident that occurred in a Naval nuclear plant, whether it had dire consequences or just resulted in a close call, went out to every plant operator in the Navy. I was a submariner, and it often seemed that every incoming communication included another set of incident reports. We were required to read and sign every new incident report–no matter how minor–that had any chance of impinging on our watch stations or maintenance responsibilities. Naval Reactors was quick to respond to incidents that demonstrated systemic and procedural flaws by sending out action change notices that went into effect until formal revisions to the operating manuals could be published.

After TMI, the NRC took a cue from Naval Reactors and initiated a simple incident reporting system.

Today, errors that result in plant damage and near-misses are collected and published to operators across the country.

Recently the Institute for Healthcare Improvement (IHI) Open School has been encouraging medical personnel to share their mistakes with their peers in the hopes that their errors, too, might prevent future disasters. One of the IHI videos I saw, IHI: Perspectives: The Mistake (part 1) includes a talk by Michael Leonard, the Physician Leader for Patient Safety at Kaiser Permanente, about an incident of his
own in which he accidentally grabbed the wrong vial and re-injected a patient with a paralyzing agent when he had intended to bring the man out of anesthesia. Dr. Leonard says that, as head anesthesiologist of his hospital, he was in a position to enact a change in procedure to prevent anyone else making the same mistake ever again. When I saw that video, my first thought was, “What about anesthesiologists in the other thousands of hospitals and surgical centers in the US? What’s to prevent their anesthesiologists from making the same mistakes?” The IHI Open School program–which addresses healthcare students–is a good one, but I think it could be much more far-reaching and effective. My proposal is twofold.

First, let’s establish a national medical incident reporting system. People like Don Berwick, Atul Gawande, and Kent Bottles have taken steps to bring medical error out of the closet and into the mainstream where they can have positive effects. A comprehensive system for collecting and deliveringthose errors could have a vaccinating effect across the country. Every iatrogenic infection, injury, or death reported could result in untold infections and injuries prevented and lives saved.

Second, for greatest efficiency, immediacy, and impact, this incident reporting system should take the form of a medical error Wiki. Every medical professional should have the right to log on and share her or his errors. That Wiki should be supported by alerts to email but just for the same of rapidity of promulgation, through social media such as LinkedIn, Facebook, and Twitter.

Life Saving Errors is an official panel proposal for SXSW Interactive 2011. If you like his ideas, why not take a moment and vote for his proposal at the SXSW site. Dennis, a negotiator and investigator for MedicalBillDog.com (a medical billing advocacy service), spent a decade improving documents for IBM. As liaison to other computer companies, Dennis’s win-win negotiating style won him the nickname “Switzerland” from his work team. He notes “with a little luck, my panel proposal will be approved for SXSW. If not, I’ll still be looking to start a wiki for the purposes described in my proposal.”

An avid scuba diver and martial artist, Dennis enjoyed an active, healthy lifestyle until a martial-arts accident jarred loose a damaged disc in his lower spine. Chronic pain from degenerating discs in his lower back led to one medical procedure after another-X-rays, MRIs, epidural injections, spinal fusion surgery, physical therapy, and lots of medication. Hoping to find an explanation for the unbelievable size of his medical bills, Dennis studied to pierce the fog of codes and abbreviations. What he found astonished him: over $5,000 in errors in the first two pages alone.

15 replies »

  1. That seems like a great collection of insights, although would this lead to repercussions for their errors in a way? On a different light, if a victim/patient got hold of this, it might add a valid weight to a possible litigation that might occur.

  2. Our organization, The Empowered Patient Coalition has an established adverse event reporting survey to capture medical events from the patient’s point of view. While it would be a great asset to have a reporting system that would be used robustly by providers we have worked to be sure that the patient voice is heard as well. AHRQ is working on establishing a national patient reporting system and until such a system is fully implemented we will continue to collect patient data. Our survey is open to all respondents, including providers and can be taken anonymously. We do not collect IP addresses. A majority of people taking the survey have shared their contact information.
    Current data is available on our website at http://www.EmpoweredPatientCoalition.org under the “report an event” section. We have 585 responses to date that chronicle over 1400 events.
    Julia Hallisy
    The Empowered Patient Coalition

  3. The short answer for everyone has two parts:
    1) Go to http://panelpicker.sxsw.com/ideas/view/6873 and vote yes.
    2) Come to Austin in March and enjoy the SXSW Interactive conference.
    The only-slightly-longer answers for some:
    Michelle W: I never foresaw this as a government-run organization. Like the AHRQ WebM&M, this would have to be an all volunteer operation (I’m your first)–at least at the outset.
    Rick: I swear, anonymity was one of my first concerns. This system would have to be done via secure log-in.
    David: I’m not sure I want a Hyman G. Rickover running the hospitals. He was efficient, but he was a little bit nuts. Among other problems, he didn’t trust electronics. I’m pretty sure computers and the Internet would have scared Hell out of him.
    Bobby G: I love the idea of transparency, and I would love to see medical personnel learn to admit and learn from mistakes. Still, if half of my salary went to malpractice insurance, I’d want anonymity, too.
    All answers are presumptuous, though, until after SXSW. Thanks for responding, folks.

  4. jd: So, I can assume you support and encourage name calling as a part of healthy debate and discussion?

  5. “Second, for greatest efficiency, immediacy, and impact, this incident reporting system should take the form of a medical error Wiki.”
    That is exactly what many have been saying. Why has it not happened with HIT?
    The FDA already does this at MAUDE and MedWatch. Great idea to report. The HIT vendors will prevent.
    Send your adverse event reports to the FDA until Wiki gets it going!

  6. Excellent suggestion to share medical errors on wiki. As rightly pointed out, sharing error related info across hospitals and even national borders can immensely reduce fatalities and saves considerable amount of time.

  7. Practice admin, have you read MD as HELL, ExhaustedMD and others on this blog? They are not only perpetually aggrieved, but frequently insulting and wildly scathing in their accusations. In addition, how does language like this even remotely prevent “honest and open dialogue” whether from the aggrieved docs or from those of us who sometimes vehemently disagree with them?
    Rick owes no apology. Welcome to free speech on the internet.

  8. Thank you Dennis for a very interesting story. Before med school I earned a BS EE/CS (Elect Engr/Comp Sci) and worked in industry. The real issue is poor plant design: a valve position indicator should show the true position of the valve and not the signal sent to the valve. The idea that the plant designers did not have this philosophy in mind when designing the plant is troubling. The change in procedures in this case is simply making up for poor plant design.
    Still the procedures put in place by the Navy are always absolutely neccessary and we have the very famous father of the nuclear Navy, Admirmal Hyman Rickover to thank. He is by training an engineer (MS Elect Engineering).
    I can assure you that Admiral Rickover would never have approved of the design of that valve indicator in the TMI plant.
    If only we could have someone like him running our hospitals and outpatient clinics.
    We need someone like him who would approve of the designs of hospitals and clinics.

  9. I would commend to everyone Dr. John Toussaint’s et al’s book “On The Mend” (which I reviewed on my blog). He’s a firm advocate of transparency across the board, and decries the still-predominant “shame and blame” culture in health care, which causes people to hide errors.

  10. The nucleus of the idea is great, the execution may need to be given some thought.
    Ideally, a reporting site like this should be run by an independent Federal agency, and it should protect the anonymity of those submitting reports, and all entries should not be admissible in a court of law, should not be given to insurers, boards, health grades, administrators or anybody else, and reporting should not be optional (or voluntary), and action should be taken to learn and disseminate the lessons learned to all physicians. There should be no penalty associated with such reporting. None. If anything, reporting an error to such database should provide protection to the erring clinician.
    Anything short of all of the above will fail.
    There is a HIT incident database maintained by the FDA (MAUDE), which is voluntary and not quite anonymous and I am really not sure what action, if any, the FDA takes following those reports. So we do have a model of how not to do it.

  11. “I might as well point it out before one of the perpetual-victim MD’s does”
    Sad to see comments such as this on this blog. Honest and open debate cannot exist with these types of posts. I would think an apology should be forthcoming.

  12. I might as well point it out before one of the perpetual-victim MD’s does: an important adjunct to such a reporting system would be deciding how it interacts with the torts system and malpractice insurance. Does reporting an incident raise the provider’s malpractice insurance rate? Does a reported incident lead to a lawsuit, or does good-faith reporting inoculate the provider from a suit? It probably should be the latter, but by the same token, we don’t want to give a get-out-of-jail-free card to providers who are chronic producers of error reports.
    The good news, though, is that if this system works as advertised, these questions should become more and more moot over time.

  13. I like the idea of a wiki and reporting system as well, though I hesitantly point out a few problems. For one, who will run this system? The government is the first answer, and then you have to establish what agency will do it (or create one for the job). Wiki users have the ability to delete information off a page: should every doctor have that permission? Would some users have higher level permissions than others, and who would judge who had what permission? Would this information be open to the public? Would anonymous submissions be allowed, or would doctors have to use their real name and location (which gets back to that question about the site being open to the public)?
    Obviously this whole thing is theoretical at the moment, and these are concerns for the panelists and/or gov to decide. They’re questions to consider, though, especially if you want to get positive support and involvement amongst the medical community.

  14. Excellent, excellent idea, Dennis. There are many existing organizations (JC comes to mind, if it would be effective) through which this idea could be implemented, and the establishment of a methodology should be cause for a Wiki in itself – but the important thing is, just do it. This is a great post. Let the arguments begin. (:
    I also enjoyed the explanation of 3MI, which I had never understood to this level of detail.