A new report that doubles the previous high estimate of patient deaths from medical errors-to 200,000 annually-relies on a new and imprecise measure of hospitals' inability to catch preventable complications that accounts for most of the difference in the number of deaths.
The study by Health Grades, a Lakewood, Colo.-based healthcare-quality rating and advisory services company, used a patient-safety indicator first put forward by the Agency for Healthcare Research and Quality to isolate 187,000 incidences of a hospital's "failure to rescue" a patient from certain life-threatening complications acquired during an inpatient stay.
In all, the study tallied 264,000 deaths among Medicare patients from 2000 to 2002 that could be attributed to 16 types of patient-safety lapses identified by the AHRQ as measurable with enough validity using available hospital administrative data. The failure-to-rescue measure accounted for 70% of the deaths.
Extrapolating to the entire U.S. population, an average of 195,000 people per year-or 575,000 over the three-year period-died from "potentially preventable" lapses in safety of care with an associated cost of $6 billion per year, the Health Grades study concluded.
Critics of the report and its reliance on the failure-to-rescue indicator said the death count did not adjust for cases in which hospitals recognized and aggressively treated the targeted complications to no avail. The report used data from Medicare to assume a cause-and-effect relationship between lethal complications and a failure of adequate vigilance, whether or not more detailed data in a patient's chart would have made that tie less clear-cut.
The report highlighted a new concept of patient-safety responsibility that isn't broadly used in the healthcare industry and "would not be familiar to most hospitals," said Nancy Foster, senior associate director of policy for the American Hospital Association.
Medical errors usually are viewed as "blatant, obvious things" such as giving the wrong dose of medication or operating on the wrong part of the body, said Samantha Collier, Health Grades' vice president of medical affairs. But new measures are focusing on mortality from situations that hospital vigilance should prevent or minimize. Besides the failure-to-rescue indicator, the study used an AHRQ indicator charting unexpected deaths of patients admitted for low-risk conditions.
"It's a soft, at times hard-to-understand concept," said Kenneth Kizer, president and chief executive officer of the National Quality Forum, or NQF, which brokers agreement in the healthcare industry on quality measures.
But hospitals may well have to deal with the new patient-safety measure soon. A variant of the failure-to-rescue indicator used by Health Grades-limited to deaths from post-surgery complications-last year became part of a consensus set of standards on nursing-care performance endorsed by the NQF. Other NQF measure sets have found their way into performance measurement initiatives, such as that of the Leapfrog Group (Oct. 20, 2003, p. 8).
Already the NQF's nursing-care standards are being embraced by the Joint Commission on Accreditation of Healthcare Organizations. The JCAHO plans to incorporate the 15 nursing performance measures, including the failure-to-rescue indicator, among its revised accreditation standards for staffing effectiveness that go into effect in mid-2005.
The assumption in the failure-to-rescue concept is that complications such as blood-borne infections or post-surgery pneumonia should not result in death for patients coming in for elective surgery, heart failure or other unrelated problems, Collier said. Death in these circumstances is likely the result of missing important indications of a turn for the worse and "failure to diagnose and treat in time," she said.
"It's fair to expect hospitals to track patients for the signs and signals of adverse consequences," Foster said. But classifying that a patient "went downhill" because of tracking failures, sometimes regardless of what caregivers did, is likely to meet resistance from clinicians, she said.