< Truth / Justice / Patient Safety >

This page has moved to here

This page has moved to here

This page has moved to here

Click this link to go to the new page.

How Many Are Dying
Unnecessarily in Health Care?

IOM studied only 30,000 records
from the year 1984

The oldest study, the 1999 Institute of Medicine (IOM) report "To Err is Human" by Lucian Leape, concluded that between 48,000 and 98,000 patients die each year as a result of preventable medical errors committed in hospitals, but that was based only on data that was reported by health care professionals who don't report most adverse events and only on data from three states.*

Since that study was done, the Centers for Disease Control (CDC) has determined that 99,000 patients die each year from one single source alone - infections caught unnecessarily in hospitals. There are additional sources of unnecessary death in medicine. And hospitals are not the only place in medicine where patients die unnecessarily. For instance, most infections resulting in death are caught in healthcare settings other than hospitals. Considering just one infectious disease, C. difficile, 75% of fatalities contract the infection in nursing homes, primary care physicians' offices, and similar non-hospital settings. The 99,000 number counts only people who died as a result of only one kind of unnecessary death, infections, and only those caught in one kind of healthcare facility, hospitals.

But putting aside other facilities, considering only hospitals, there are newer studies of larger pools of data. It is time for the health care industry to stop quoting the antiquated, underestimates of the IOM study.

HealthGrades studied 37,000,000 records
from 2000 to 2002

HealthGrades, the health care quality company, looked at 37 million patient records taken from three years of Medicare data in all 50 states and D.C., approximately 45 percent of all hospital admissions (excluding obstetric patients) in the U.S. from 2000 to 2002. They found that an average of 195,000 patients in the USA died in each of those years due to potentially preventable, in-hospital medical errors. But, once again, they were working only with data that had been reported by health care professionals, and health care professionals report only 2% of adverse events accurately (see Medical.Reporting).

Annals of Internal Medicine found additional information

A study in the Annals of Internal Medicine** examined information that previous studies did not. They recognized the problem with basing studies only on the few adverse events health care professionals report and looked for where they could get unrecorded information. They conclude that 320,000 patients die unnecessarily in medicine in the USA each year. Their report, written by 10 experts with various advanced degrees (including three MDs) in the July 15, 2008 issue, sought to discover if patients' knowledge of medical errors revealed errors that the medical records did not. In the Annals study, serious preventable medical errors documented in medical records of 1,000 patients hospitalized in 2003 in Massachusetts were compared with serious preventable errors that patients themselves could recall 6-12 months after their discharge.

Only eleven serious preventable errors were documented in the medical records created by caregivers, but patients reported 21 additional ones that were confirmed (by an investigating team) that the healthcare professionals did not report. If the rate of documentation of serious preventable errors in medical records is the same as the rate of documentation of lethal medical errors in the records used by the Harvard study, a better estimate of lethal medical errors would be 110,000 x (21 + 11)/11 = 320,000 unnecessary deaths per year. That is approaching a thousand per day. And this estimate is based only on the cases that could be confirmed.

In an environment in which only 2% of adverse events get reported accurately by health care professionals, and an unknown number of events reported by patients could not be confirmed, how much larger might the fatality figure be if either health care professionals reported honestly or more of the events reported by patients could be confirmed? The routine expertness and ubiquity with which medicine erases evidence, and memory, of adverse events makes confirming anything a rarity (ask injured patients).

320,000 is more Americans than died during the entire 8 years of World War II, our deadliest war.

25,000 per month

Death By Medicine, by Gary Null, puts the figures at around 1 million patients dying unnecessarily per year. He is not the only one, and his methodology is not the only methodology arriving at such a figure, but on this site I use the figure from the Annals of Internal Medicine, sometimes rounded down to 300,000 so that it is an even 25,000 per month for the purposes of discussion.

Late in 2010, the Office of Inspector General for the U.S. Department of Health and Human Services said 180,000 Medicare recipients die each year from hospital mistakes (Link is to a site that highlights the specific lines). Medicare recipients make up only a portion of the patient population. So 300,000 is a conservative number.

Accident Pyramid

People who work in public health are familiar with the Accident Pyramid.*** It says that for every fatality there are 300 injuries (some disabling) and 300,000 unsafe acts committed to cause them. 320,000 unnecessary deaths means 96 million injuries (most innocuous) and 90 billion unsafe acts to cause them (some lethal). To try to get a perspective on that, the number of insurance claims filed in a recent year was 4 billion which means there would be about 20 to 25 unsafe acts committed for each insurance claim that was file. We could run through all of the small unsafe acts to which a patient could be exposed during a routine visit to a family doctor just to try to get a grip on how the numbers can be so large, but the point isn't the precise accuracy of the number.

The point is that the numbers are large and that lots of patients are dying because of them. They are only projections. Because of how little honest reporting there is in medicine, that's the best we can do. But what can be measured creates reason to worry that the projections might have something to valid to say. Like the fact that in a recent year the number of medication errors in hospitals was reported to be roughly equal to the number of patients admitted to hospitals. Only 6% of medication errors are reported through traditional means (according to Making Care Safer, an analysis prepared for AHRQ), which means that the measurement of how many errors were made, once again, is based on numbers that are far lower than the actual number of errors made. It creates the worry that, no matter how off the projections might be, they might be more valid than the numbers found in the record.

Many small acts by providers
with a cumulative damage to patients that is enormous

When researchers and hospitals administrators have no accurate way to learn the true numbers, what hope is there for patients to determine which operators and which facilities commit the most unsafe acts and cause the most injuries and deaths? Patients cannot get the information necessary to make choices that avoid these dangers. Fortunately, there is a way around this.

The experience of patients can help to identify the problems. Unfortunately, not only is there no place for the experiences of injured patients to be collected, injured patients discover that even when they have evidence and witnesses and determination they cannot get their injuries noted in the record. No matter how much they want to, and no matter how hard they try, they cannot even become statistics. If you doubt that, see Janice M. Scully, MD. Even when the victim of an adverse event is a physician she cannot get it put in the record. This is the rule, not the exception. It is shameful that anyone in medicine does not know that. And I've never met anyone in medicine who does.


The CDC reports that 99,000 patients die each year from infections contracted during the course of treatment. That means that the number who are infected but do not die is millions, possibly as high as 29,700,000 according to what the Accident Pyramid would project. I, myself, have a hard time understanding how the number could be that large. Most of the infections either are innocuous or are easily cured, but some multiple of the deaths are disabled and some number larger than that are not disabled, but are never the same, and that number is in the millions. This can be, and elsewhere has been, brought down to zero. Elsewhere on this site is discussion about why our health care professionals do not want to do that (at least not enough to do it) and what could be done about that (for instance, see Semmelweis).

For a more narrow focus to help understand how the numbers can be so large, the CDC estimates that for one specific kind of procedure, central-line catheters, infections are caused in 250,000 patients annually, costing $25,000 each and claiming the lives of one in four of those infected patients. Focusing on this one single procedure could bring that number to zero.

Only Errors

All of these studies examined only errors. None included unfriendly practices. No one is addressing the problems of exploitation, abuse and murder in medicine. Nowhere in health care is there an adequate mechanism for addressing such problems. Not only is nothing less likely to be recorded than unfriendly practices, but nothing is more likely to cause the medical community to crush the victim in every way possible to cover it up. Janice M. Scully, MD was not injured on purpose and so had it easy compared to the victims of unfriendly practices.

Any patient safety initiative that does not include provisions for dealing with unfriendly practices should be shredded and the people who wrote it should be spanked. To imagine that any facility does not need to protect patients from unfriendly practices is to believe in comfortable absurdities.

"If we believe absurdities, we shall commit atrocities."
- Voltaire

In our experience, the faith of health care professionals in their own objectivity is so strong that when statistics like these do not agree with their perceptions, they disbelieve the statistics. They do not consider the possibility of having a self-interested bias that causes them to interpret their experience in self-interested ways. We do not hear even lip service given to the idea that a self-interested group-think and herd-mentality have more to do with what they believe than the evidence of their senses and the evidence in the studies.

This suggests, and our experience suggests, that they live in a fog of self-serving delusion oblivious to these numbers and to the outcomes they produce for their own patients. But suggesting that ends discussion. So, decade after decade, the unacceptable rate of unnecessary death does not improve and everyone in medicine believes him or herself to be "one of the good ones" doing the best that can be expected in spite of maintaining the routines and habits that make medicine the most dangerous and financially ruinous place Americans go.

The "disinclination to believe the monstrous is constantly strengthened by [he/she] . . who makes sure that no reliable statistics, no controllable facts and figures are ever published, so that there are only subjective, uncontrollable, and unreliable reports . . . "
- Hannah Arendt



*The original Institute of Medicine study was based on physician examination of the medical records of 30,000 patients receiving care in New York hospitals in 1984. Of the 30,000 patients studied by the Harvard group, 87 died as a result of preventable medical errors committed while they were hospitalized. These data, when extrapolated to all admissions in U.S. hospitals in 1997, yield 98,000 preventable deaths nationwide in 1997 when there were 34.6 million admissions. But each year the number of admissions increases. In 2002 there were 37.8 million hospital admissions. For that year the estimated total is 110,000 preventable deaths per year based on evidence in medical records alone.
But, as this site continually points out, people in medicine do not record most of what should be put in the record. Where else to get the information? From patients.

**From an article called "Comparing Patient-Reported Hospital Adverse Events with Medical Record Review: Do Patients Know Something That Hospitals Do Not?"
by Joel S. Weissman, PhD; Eric C. Schneider, MD, MSc; Saul N. Weingart, MD, PhD; Arnold M. Epstein, MD, MA; JoAnn David-Kasdan, RN, MS; Sandra Feibelmann, MPH; Catherine L. Annas, JD; Nancy Ridley, MS; Leslie Kirle, MPH; and Constantine Gatsonis, PhD
in The Annals of Internal Medicine, 15 July 2008 | Volume 149 Issue 2 | Pages 100-108
The abstract is viewable at: http://www.annals.org/cgi/content/abstract/149/2/100

*** Accident pyramid or Safety pyramid
In 1931, H.W. Heinrich theorized that for every major accident there are 29 minor accidents and 300 near misses. This theory has been reevaluated several times, including by Conoco Phillips in 2003, where it was determined that for every fatality there are 30 lost day injuries, 300 recordable injuries, 3,000 near misses, and 300,000 unsafe acts.

Fatalities often are considered freak rarities and as such are not considered events to be recorded and learned from. But the causes of fatalities are different from the causes of injuries. Learning how to prevent one does not lead to preventing the other. Both must be studied. [Dan Petersen, 2nd edition, Safety Management] Unfortunately, in medicine, neither are accurately recorded more than 2% the time. And it is difficult to find anyone in medicine who even is aware of that.


In many cases when someone dies as the result of an infection acquired in a hospital, the infection is not listed on the death certificate as the cause of death.