New research reveals unreported hospital errors

Paul Sharek

Adverse events in neonatal intensive care units across North America occur more frequently and are more severe than previously reported, according to a study led by researchers at the School of Medicine and Lucile Packard Children's Hospital.

In the largest study of its kind, the authors reviewed records from 15 NICUs from throughout North America. Their intent was not to cast blame, but rather to identify ways in which the naturally high-risk NICU environment can be made safer for patients through research and changes in clinical practice.

"Many of the harmful events we identified are not currently preventable," said Paul Sharek, MD, chief clinical patient safety officer at Packard Children's Hospital and assistant professor of pediatrics. As an example, Sharek cited a life-saving medicine with unavoidable adverse side effects. "But it's still vitally important to document these events so we can establish future priorities." In Sharek's example, a lengthy rap sheet for a particular treatment or drug may spark physicians and drug developers to look for new alternatives.

Sharek and his collaborators were able to identify previously unreported adverse events—defined as harmful results of hospital care—by using a recently developed method of chart review known as the trigger tool method. In this method, trained medical staff members scan patients' charts for key events, medications or labs—called triggers—that often occur in conjunction with an adverse event.

Despite their name, triggers aren't necessarily the cause of the event; sometimes they are the result. For instance, the drug naloxone is used to rapidly reverse the respiratory depression that can occur as a result of the use of narcotics, such as morphine. Its use, documented in an infant's chart, triggered a more intensive review of that child's NICU treatment that sometimes revealed an adverse event related to narcotic administration.

Participation in the study by the 15 NICUs was voluntary. The researchers limited their search to infants who spent at least two days in a NICU between November 2004 and January 2005. They found that about half of all the adverse events were preventable, and about 40 percent resulted in substantial and sometimes permanent harm. Only 8 percent of the adverse events had been previously identified using traditional reporting methods.

The frequency of adverse events in the NICU was similar to those found in recent studies of adult intensive care units, but the repercussions of the events were more severe.

"The NICU is a very high-risk environment," explained Sharek. "These infants are exceptionally medically fragile." The authors corrected the results for the infants' birth weights and gestational age in an effort to take into account the fact that some institutions, such as Packard Children's, treat higher acuity cases that may require more sophisticated interventions. The research is published in the October issue of the journal Pediatrics.

Each institution received a report of its results. While nearly all struggled with similar issues, some inter-site variations were enlightening. For instance, although many physicians and NICU staff believe that hospital-borne infections are inevitable in their highly fragile population, a few institutions had no such infections during the study.

"This doesn't mean that some hospitals are inherently safer than others," said Sharek, who cautions against any direct comparisons. "These data provide a starting point for local and national efforts to decrease those adverse events with the highest frequency and greatest risk through patient safety interventions in the NICU." A similar study among 21 pediatric intensive care units, spearheaded by Sharek and pediatric fellow Swati Agarwal, MD, is under way.

Sharek's collaborators on the NICU study include investigators from Harvard, Dartmouth, the University of Vermont, Children's Hospital Los Angeles, the Medical University of South Carolina and the University of Utah.