Preventing Patient Harm: The Real Story

The date is July 17th, 2014. It is 10am in the Dirksen Senate building, and the congressional subcommittee on health and aging is about to focus on patient harm.  The educating will be done by some of the leaders in the medical field,  Ashish Jha and Tejal Gandhi from Harvard, Peter Pronovost from Johns Hopkins. The star of the proceedings is John James, a toxicologist, a PhD from Texas, and the founder of Patient Safety America.

The tone is set from the beginning by none other than Bernie Sanders.  In somber tones, he relays that hospitals can make patients worse, and that a recent study suggests medical errors is America's third leading cause of death behind only heart disease and cancer. Hospitals are killing patients, and something needs to be done about it.  The panelists then go on to speak strongly about the ongoing epidemic of patients dying in hospitals, and re-enforce the staggering numbers introduced by Bernie Sanders.

Headlining the proceedings is an unassuming gentleman named John James.  He has a Ph. D in pathology, and he worked as a Chief Toxicologist at NASA.  He is at the congressional proceedings, and is one of the lead activists in patient safety because of personal tragedy.  His 19 year old son died in the summer of 2002 due to "uninformed, careless, and unethical" care by cardiologists.   He proceeded to write a book, "A Sea of Broken Hearts" that details the errors he believes cardiologists made in his son's care that lead to his death.  Of note 2 cardiologists that were sought by Dr. James' lawyers believe the care his son got did not violate the standards of care.  A further 2 appeals to the Texas Medical Board also rendered two opinions from two other separate cardiologists that the standards of care in this case were not only met, but exceeded.  Dr. James, armed with information he has carefully selected from a number of different sources, strongly disagrees.

Dr. James is now a crusader for patient rights.  He writes of a broken health care system on his website, and more importantly wrote a paper in 2013 in the Journal of Patient Safety that estimated 400,000 patient deaths per year that were due to medical error.  No physicians on the panel or elsewhere seem to have any issue with this number, and this has become fairly widely accepted.  Even Captain 'Sully' Sullenberger, the hero pilot who landed a plane in the Hudson, noted that this was the "equivalent of three jumbo jets going down every day with no survivors."

As a busy clinician who spends much of his time in the hospital, it doesn't feel like patients are dying daily because of medical errors.  But of course, data necessarily must trump feelings.  So, I decided to read John James's landmark paper.

The paper reviews four original articles that reported on medical error causing patient harm.  The first study was a pilot trial of 278 patients that examines one week in August of 2008.  The second trial examined 838 patients in October of 2010. The third trial was a very similar study that looked at 795 patients in October of 2004.  The most robust (because it was the largest over the longest duration of time) examined 100 hospital admission per quarter in North Carolina.  Dr. James's paper combines all four trials, but weighs the North Carolina trial the heaviest.  He simply divides the total lethal adverse events found in all four trials (38) and divides by the total number of records reviewed (4252) to come up with a lethal event rate of 0.89%.  He estimates that, of the harms found in these trials, 69% were noted to be preventable.  Since there were 34.4 million hospital admissions in 2010, simple multiplication (34,400,000 x 0.69 x .0089) reveals a number of 210,000 preventable harms that resulted in the death of a patient.  Dr. James isn't done here, however.  He notes that the tools used to find patient errors are imperfect.  He notes that failing to follow guidelines, errors not documented in the medical record, and a failure to make life saving diagnoses would necessarily add to these numbers.  He believes that at a minimum, this should increase the actual number of deaths related to medical error by a factor of 2.  That's it.  No statistical modeling for how many patients a year present with heart failure and don't leave on a beta-blocker, no examination of the number of young patients dying due to a missed diagnosis of long QT.  He just comes up with a factor of 2 because that feels about right.  And there we have it, 400,000 patients, 3 jumbo jets a day crashing, the third leading cause of death in the United States.  Dr. James notes this is likely an underestimate.  Good thing, otherwise the health care system would be the biggest killer of patients.

I was stunned.  This was the evidence?  4 trials.  One of the trials took place over one week.  All four trials did use the same error reporting tool, but were simply added together, with no regard to the varying settings the different trials took place in.  The smallest trial did not even report what percentage of cases were preventable.  One of the trials (Classen 2011) was a trial designed to test the efficacy of a patient harm reporting tool, and did not report preventable harm events.  Dr. James, inexplicably in his review, notes that 100% of the harms found in this trial were preventable.  Another trial, The Office of the Inspector General (OIG) analysis, notes a 44% preventable harm rate, but does not note which deaths were clearly preventable.  Far and away, the best quality trial is the North Carolina study from the New England Journal of Medicine.  Of the 2341 cases reviewed, there were 588 total harms identified.  364 of these harms (63%) were deemed preventable and 9 of these resulted in patient deaths (0.4%).

Dr. James' problem is in how he aggregates the data.  He includes trials which did not state preventable deaths, and in one case assumes that all the deaths were preventable.  If you only used the two trials that published data on preventable harms, the preventable harm rate is 58%, not the 69% listed by Dr. James.  Of course, I don't think one should use this number to estimate the preventable lethal death rate, because this assumes that the total preventable harm rate is equivalent to the lethal preventable harm rate.  Why is that a safe assumption?  Luckily, since Dr. James's statistics to arrive at his estimate uses multiplication, and I just happen to be reviewing this subject with my 6 year old, I can generate my own number. The lethal preventable death rate is .384% (9 preventable deaths/2381 total cases reviewed).  34,400,000 x  .004 = 130,000 patients. Using the sounds about right Dr. James factor of 2, that brings us to 260,000 patients.  That is still a lot of patients, but a lot less than 400,000 patients.

Medical errors are a serious problem, that is no doubt deadly, and needs attention.  We in the healthcare community need to work hard locally and nationally to combat this issue.  I applaud Dr. James and the other physicians that have shed light on this important issue.  Perhaps, the actual numbers don't matter, perhaps it's missing the point to focus on the actual number, perhaps it doesn't matter that Bernie Sanders thinks medical errors are the third most common cause of death.  Except, it does.

Bernie Sanders frequently talks about the broken health care system, and in support uses this to buttress his claim.  Others (including Dr. James) go further.  They specifically point to physicians as the problem.  We are the captains of this ship, and we are steering this ship into an iceberg.  It generates distrust among the public and foments anger against physicians when patients do not do well.  Michael Davidson, a cardiothoracic surgeon at Brigham and Women's, and more importantly a husband to a pregnant wife and three children, was shot to death by his patient's son.  The assailant's mother, Marguerite Pasceri, 78 years old, had recently died while she was in Dr. Davidson's care.  She had multiple medical comorbidities and her death was ruled as being related to these severe comorbidities.  Unfortunately,  fueled by the internet, Steven Pasceri became obsessed with the idea that use of the drug amiodarone had caused her death.  He confronted the doctor, the scene is described in chilling fashion by the Boston Globe:

{Right away, Pasceri told Davidson to open the Internet, go to Drugs.com, and look up amiodarone.

“Are you aware that this drug is extremely toxic?” Pasceri asked, St. Jean said, pointing to the website. “Do you see all of the warnings on Drugs.com?”

Davidson explained he was aware of all the side effects but said Marguerite Pasceri did not react badly and was being monitored. Any drug, he explained, even an antibiotic, has potentially dangerous side effects.

“Well, my mother died because of this,” Pasceri said, his face twisting into a snarl.}
Minutes later, Dr. Davidson was shot three times.

You would be a fool not to connect the relentless drum beat of the media, congress, and the public about the horrid broken down medical system and the even more horrid, incompetent doctors that are killing patients in hospitals to an event like this.

The facts are that 34 million patients are arriving at hospitals with an illness.  They are presenting in distress, in need of help.  Dr. James and many leading members of our profession have unfortunately whipped our representatives and the general public into a frenzy.  Three jumbo jets are going down every day!  The facts are that 0.4%, 4 out of a thousand patients, die from a medical error. Work needs to be done to reduce this rate, but it is unlikely ever to be 0.

The idea that every error in the hospital is a preventable one, the impression that physicians are by and large an incompetent group that is killing patients needs to strongly be repudiated.  The reason that it doesn't 'feel' like patients are dying on a daily basis in large numbers due to medical errors is because they are not.  Even Dr. James notes that there is no statistically rigorous way to arrive at a number.  The number we use today, 400,000, is a made up number. It is based on a feeling.  Using this to falsely indict, demoralize, and create a toxic environment for millions of medical providers is in no one's best interest.