Study finds much higher rate of annual errors

Research examines three years of Medicare records

According to a new study of 37 million patient records by HealthGrades, a Lakewood, CO-based health care quality company, an average of 195,000 people in the United States died due to potentially preventable, in-hospital medical errors in each of the years 2000, 2001, and 2002.

The HealthGrades Patient Safety in American Hospitals study finds nearly double the number of deaths from medical errors found by the 1999 Institute Of Medicine (IOM) report, To Err is Human, with an associated cost of more than $6 billion per year.

The discrepancy in numbers should not be the focus of attention, however, insists Samantha Collier, MD, HealthGrades’ vice president of medical affairs.

"A lot of what IOM did was chart review and input from clinicians and physicians, asking them if there was a mistake or not," she explains.

"Any time you have health care workers pore over charts, there is some bias. That’s one possible explanation for the lower numbers in IOM. It’s hard to make an apples to apples comparison, but the new thinking is it’s 98,000 up to 195,000," Collier adds.

No matter which number you accept, "The magnitude is horrendous. We need to focus on ways to improve our delivery system to make sure patients have safe outcomes when they come to our hospitals," she emphasizes.

Other findings of the report include:

  • About 1.14 million patient-safety incidents occurred among the 37 million hospitalizations in the Medicare population during 2000 to 2002.
  • Of the total 323,993 deaths among Medicare patients in those years who developed one or more patient-safety incidents, 263,864 or 81% of these deaths were directly attributable to the incidents.
  • One in every four Medicare patients who were hospitalized from 2000 to 2002 and experienced a patient-safety incident died.
  • The 16 patient-safety incidents accounted for $8.54 billion in excess inpatient costs to the Medicare system over the three years studied. Extrapolated to the entire country, an extra $19 billion was spent and more than 575,000 preventable deaths occurred from 2000 to 2002.
  • Patient-safety incidents with the highest rates per 1,000 hospitalizations were failure to rescue, decubitus ulcer, and postoperative sepsis, which accounted for almost 60% of all patient-safety incidents that occurred.
  • Overall, the best performing hospitals (hospitals that had the lowest overall patient safety incident rates of all hospitals studied, defined as the top 7.5 % of all hospitals studied) had five fewer deaths per 1,000 hospitalizations compared to the bottom 10th percentile of hospitals. This significant mortality difference is attributable to fewer patient-safety incidents at the best performing hospitals.
  • Fewer patient safety incidents in the best performing hospitals resulted in a lower cost of $740,337 per 1,000 hospitalizations as compared to the bottom 10th percentile of hospitals.

The first statistic is one of the key findings of the study, Collier says. "The 1.14 million events just in the Medicare population alone in the three-year study average out to be just over 300,000 a year, which equates to $2.85 billion a year in excess Medicare costs.

"Those are incidents; they did not all turn out to cause death. The 195,000 is how many deaths were potentially preventable, which corresponds to IOM’s 45,000 to 98,000," she notes.

The second key finding, she says, is that the better hospitals had five deaths per 1,000 less than the hospitals identified in the bottom 10th percentile.

"This also equated to significant cost savings, because [the better hospitals] had fewer incidents," Collier explains.

"There’s a quality economic impact, and most people believe these statistics, at minimum, help make the business case for improving," she says.

By extrapolation, if all the Medicare patients who were admitted to the bottom 10th percentile of hospitals from 2000 to 2002 instead were admitted to the best hospitals, approximately 4,000 lives and $580 million would have been saved, Collier notes.

Building on past research

The HealthGrades study applied the mortality and economic impact models that were developed by Chunliu Zhan and Marlene R. Miller in a research study published in the Journal of the American Medical Association in October 2003.1

That study supported the IOM 1999 report conclusion, which found that medical errors caused up to 98,000 deaths annually and should be considered a national epidemic.

Whereas the IOM study extrapolated national findings based on data from three states, and the Zhan/Miller study looked at 7.5 million patient records from 28 states over one year, HealthGrades looked at three years of Medicare data in all 50 states and Washington, DC.

This Medicare population represented approximately 45% of all hospital admissions (excluding obstetric patients) in the United States from 2000 to 2002.

HealthGrades examined 16 of the 20 patient-safety indicators defined by the Agency for Healthcare Research and Quality (AHRQ), from bedsores to postoperative sepsis, omitting four obstetrics-related incidents not represented in the Medicare data used in the study.

Of these 16, the mortality associated with two, failure to rescue and death in low-risk hospital admissions, accounted for the majority of deaths that were associated with these patient safety incidents.

"If we could focus our efforts on just four key areas — failure to rescue, bed sores, postoperative sepsis, and postoperative pulmonary embolism — and reduce these incidents by just 20%, we could save 39,000 people from dying every year," Collier points out.

Laying the groundwork

Although it would appear at first glance that not much has changed since the IOM report, "the good groundwork of change has begun," she adds.

"We have gotten people over the denial stage and into the acceptance stage, which allows us to advance to the change stage." It takes time to change culture, Collier concedes.

"[However], unlike other health care issues, this is one in which every stakeholder has a significant stake. For the first time, we all have a similar agenda, a shared vision."

That shared vision involves running a health care organization like a public company, she concludes. "In other words, you meet certain metrics or you shut down. We have to get out of this box that health care is different.’"

Quality managers "must know what their adverse event rates are, whether they use the indicators in our study or others," and get beyond looking at medical errors as only commission and omission, Collier explains.

"You can’t change what you don’t know," she adds. Once you have your information, "You need to share it with your board and employees. When we start sharing information, that’s when we make changes."


1. Zhan C, Miller MR. Excess length of stay, charges, and mortality attributable to medical injuries during hospitalization. JAMA 2003; 290:1,868-1,874.

Need More Information?

For more information, contact:

• Samantha Collier, MD, Vice President, Medical Affairs, HealthGrades, Lakewood, Co. Phone: (303) 716-6548. E-mail: Web site: