Published: Aug. 3, 2006
Updated: Sept. 28, 2006
Reporters & producers can visit Duke Medicine News and Communications for contact information.
By Duke Medicine News and Communications
DURHAM, N.C. – Patients who undergo surgery late in the afternoon are more likely to experience unexpected adverse events related to their anesthesia than are patients whose operations begin in the morning, a new analysis by Duke University Medical Center researchers suggests.
In the more than 90,000 surgeries analyzed, only a small percentage of the adverse events reported actually caused harm to the patients, the researchers said. The vast majority of events involved such serious though lesser problems as those related to pain management requiring additional attention to patients' pain and postoperative nausea and vomiting.
"This is one of the first studies to show that there is a difference in patient outcomes depending on the start time of surgery," said Melanie Wright, Ph.D., a human factors specialist in the Duke University Human Simulation and Patient Safety Center. Human factors specialists study how people behave physically and psychologically in different environments. Previous studies, she said, have examined the effects on patient outcomes of such factors as fatigue, sleep deprivation and circadian rhythms among health care workers.
In addition to spotting problems related to anesthesia, Wright and her colleagues also found that surgery patients experienced a significant increase in "administrative delays" during late afternoon, which might contribute to the increase in adverse events that occur during this time. The delays included waiting for laboratory test results, doctors running late, transporters not being available to move patients and rooms not being ready on time.
The team published its findings in the August 2006 issue of the journal Quality & Safety in Health Care. The research was supported by the Anesthesia Patient Safety Foundation.
Based on their findings, Wright and her colleagues suggest a number of factors that might contribute to variations in health outcomes. These factors include fatigue among health care providers, swings in the circadian rhythms that influence a person's natural ups and downs over the course of a day, and institutional work schedules.
"Health care is a 24-hour-a-day business, and it is not unexpected that factors such as fatigue, circadian rhythms, personnel shift changes and scheduling may affect patient care over the course of a day," Wright said. "We believe that identifying the specific periods when problems are most likely to occur is an important step in the overall process of making surgery safer and ensuring that patients have a good experience."
For their analysis, the researchers drew on a database of all of the 90,159 surgeries performed at Duke Hospital over a four-year period beginning in 2000. Maintained by the Department of Anesthesiology, the database contains a record of each surgical patient's course of treatment, including any adverse events experienced, from hospital admission to discharge.
Wright's team divided all reported problems into one of three categories: "error," "harm" and "other adverse events."
The researchers identified 31 instances of error. These involved problems related to inserting tubes into patients' throats to maintain respiration and by improper dosing of patients with anesthetic agents.
They found 667 instances of harm, which included such events as prolonged sedation, wound infection and postoperative nausea and vomiting. Postoperative nausea and vomiting accounted for 35 percent of the harm events.
They assigned 1,995 events to the "other" category. These events included potentially dangerous changes in blood pressure and operating room equipment problems. About half these events were problems related to adequate management of patients' pain through anesthetic techniques and pain medication during surgery and immediately afterward.
The team then matched each adverse event with the time the patient's surgery began and conducted statistical analyses to identify differences in the rates of events over various times of day.
"We found that adverse events were most common for operations starting between 3 p.m. and 4 p.m.," Wright said. "Furthermore, the predicted probability of an adverse event in the "other" category increased from a low of 1 percent at 9 a.m. to a high of 4.2 percent at 4 p.m."
Wright said that many factors, involving both patients and hospitals, may contribute to increased rates of adverse events late in the afternoon. For example, patients may be more susceptible to either pain or post operative nausea and vomiting in the late afternoon. We don't know if issues such as not having eaten all day or spending a stressful day waiting in the hospital may have an influence on this, Wright said.
Late afternoon also is a time when changes in the teams that administer anesthesia during surgery coincide with natural circadian rhythm lows, Wright said. The circadian rhythm serves as the body's internal clock that regulates sleep, brain wave activity and other bodily functions. Circadian lows occurring around 3 p.m. to 5 p.m. and again at 3 a.m. to 5 a.m may affect human performance of complex tasks such as those required in anesthesia care. Changes in anesthesia care teams usually occur around 7 am and again between 4 pm and 6 pm. End of day fatigue, a circadian low point, and changes in care team are all occurring around 3 pm to 6 pm and may be interacting in a way that affects patient care, she said.
The team's analysis also found 9,497 administrative delays that were not categorized as adverse events but may have an influence on them. "We found a significant increase in administrative delays in the late afternoon. It is possible that there is a relationship between these delays and the increase in adverse events," Wright said.
Wright cautioned that the study was a retrospective analysis of past operative cases based on self-reports by doctors and nurses of events that occur. A prospective study with unbiased observer documentation of events as they occur is needed to determine exactly what steps in the delivery of health care are responsible for the adverse events and how such events can best be avoided.
Wright and her team are now planning such a study to compare each step in the delivery of care for patients enrolled for surgery during two time periods: 9 a.m. to noon and 3 p.m. to 6 p.m. These times were identified in the current analysis as when the incidence of adverse events is lowest (first thing in the morning) and highest (late afternoon).
Also, Wright recently has received a grant from the National Institutes of Health to develop new processes of collecting and displaying information on patients undergoing surgery in order to improve patient safety.
Other Duke researchers involved in the study were Barbara Phillips-Bute, J. B. Mark, Mark Stafford-Smith, Katherine Grichnik, B. C. Andregg and Jeffrey Taekman.