Cutting medical errors means systems, money

IOM panel chair says patients should leave hospitals with fewer medical problems, not more.

By Andis Robeznieks, amednews staff. Dec. 8, 2003.

Correcting safety deficiencies in the nation's health care system will require technological solutions that will not work without universal data standards and are unlikely to be developed without some kind of federal funding.

That was the message delivered by the chair of the Institute of Medicine's Committee on Data Standards for Patient Safety, Paul C. Tang, MD, chief medical officer for the Palo Alto Medical Foundation. Dr. Tang was unveiling the IOM's latest report: "Patient Safety: Achieving a New Standard for Care."

First do no harm
Repeat offenders
Links
See related content

The release of the IOM report, in mid-November, came just two days after the U.S. Pharmacopeia released its own report analyzing the 192,477 medication error reports voluntarily submitted by 482 hospitals into the MEDMARX database.

Dr. Tang noted that the IOM's 1999 report, "To Err is Human," alerted the public to the problem of medical errors. The IOM's 2001 report, "Crossing the Quality Chasm," described how these errors were mostly the fault of weaknesses in the health care system and not necessarily the fault of individuals.

Education equals prevention

Using electronic health records can significantly shore up these weaknesses, he said, and "Achieving a New Standard of Care" offers a blueprint to step up implementation of electronic systems.

"The aviation industry takes great pains to ensure that the number of landings equals the number of takeoffs," said Dr. Tang during a press conference held in Washington, D.C. "Similarly, patients discharged from the hospital should have fewer medical problems, not more.

Improper computer entry may be responsible for more medication errors than illegible handwriting.

"The committee believes that the only way to achieve this new standard of safety is to implement patient care systems that prevent errors from occurring in the first place and enable health care providers to learn from them when they occur," he added.

Dr. Tang noted that "there are pockets of intense activity throughout the country" where systems are being developed and implemented, but the absence of both data standards and a national infrastructure prevent information from crossing regional or organizational boundaries.

As the largest payer of health care expenses and guardian of public safety, Dr. Tang said the federal government should provide the financial support needed to build a "national health information infrastructure."

In addition, the IOM report calls for individual "health care settings" to adopt a culture of safety and establish patient safety programs that include reporting and analyzing both adverse events and "near misses."

In a sense, some of this is already under way with the MEDMARX database, but even those who are directly involved with this error reporting system recognize its limitations as an anonymous, voluntary program.

"We are very cognizant of the limitations of these databases," said Diane D. Cousins, RPh, vice president of USP's Center for the Advancement of Patient Safety. "But we can learn from what we have."

For example, the statistics that were collected in 2002 include more errors reported but a lower rate of errors resulting in harm. Not only are there now more institutions participating in the error-reporting program, but also the institutions that participated before are believed to be doing a better job of reporting. So Cousins warned against reading too much into rates.

Instead, she said, the data show where to "peel back the onion" to find where errors occur to learn how to design systems to eliminate them.

Of the 192,477 errors reported, only 3,213 (1.7%) caused patient harm, and in 67,707 cases (35.2%), errors were caught before the medication reached the patient.

Human error still a factor

Other interesting findings of the USP report included evidence that technology, in some instances, may merely replace one problem with another.

The use of technology for computerized physician order entry is seen as a tool to eliminate medication errors caused by illegible handwriting but, according to the IOM report, improper computer entry may be responsible for more errors than illegible handwriting.

The study found that mistakes in computer entry accounted for 17,998 errors, or 10.3% of the total amount, and 247 errors (8.2%) that caused harm. Poor handwriting was linked to 4,924 errors (2.8% of the total amount) and 113 errors (3.7%) that caused harm.

In all, causes were identified in just over 90% of the reported errors, with "performance deficit" and "procedure/protocol not followed" topping the list and accounting for 36.7% and 16.6% of all errors respectively and 46.6% and 29.4% of errors that caused harm.

The MEDMARX medication-error reporting system has been running for four years now, and its program director, Jeffrey Silverstone, RPh, said the reporting system is expected to receive its one millionth report next year.

ADDITIONAL INFORMATION:

First do no harm

U.S. Pharmacopeia's MEDMARX database received 192,477 error reports. Statistically, the bulk resulted in no or minimal harm. Less than 1% of cases were more serious, but this included more than 600 errors that resulted in hospitalization, with about 70 causing permanent harm and 20 resulting in death.

Category	Result	Error reports
No error	Events created risk of error *	30,111	15.64%
No harm	Error didn't reach patient	67,707	35.18%
No harm	Reached patient, but no harm	79,267	41.18%
No harm	Monitoring/intervention needed	12,179	6.33%
Harm	Temporary, intervention used	2,600	1.35%
Harm	Temporary, hospitalization	514	0.27%
Harm	Permanent, life-threatening or fatal	99	0.05%
		192,477	100.00%

Note: *Reports filed because risk for error existed (such as different drugs improperly stored together), but the situation was fixed before an error occurred.

Source: U.S. Pharmacopeia

Repeat offenders

These 10 drugs, for reasons that range from giving too high of a dose to not giving the prescribed dose at all, were involved in 18.3% of the 192,477 reported errors in the MEDMARX database in 2002.

Errors
reported

Albuterol 5,862

Insulin 5,583

Morphine 3,919

Potassium chloride 3,771

Heparin 3,684

Warfarin 2,564

Cefazolin 2,538

Furosemide 2,447

Vancomycin 2,440

Levofloxacin 2,374