The study concludes that the rate is higher because hospitals tend to be short-staffed on weekends, with fewer supervisors and other experienced personnel on the duty roster. A report on the findings appears in the Aug. 30 issue of the New England Journal of Medicine.
Drs. Chaim Bell and Donald Redelmeier of Sunnybrook and Women's College in Toronto reviewed the medical records of almost 3.8 million emergency room admissions in Ontario between 1988 and 1997, analyzing those for mortality rates for the 100 most-common fatal conditions.
They also compared death rates for three predetermined conditions -- a ruptured aortic aneurysm in the abdomen, a lung clot and acute epiglottitis (a severe throat infection) -- and three control conditions -- heart attack, brain bleeding and hip fracture.
The researchers chose the predetermined conditions because they are fatal but treatable, and patients are cared for outside the emergency room and intensive care units. The control conditions are all treated in the ER and ICU, which are generally immune to variability in staff size, they say.
Indeed, the day of admission didn't seem to affect mortality rates for the control conditions, the researchers say. But for the three predetermined conditions, arrival at the emergency room between midnight Friday and midnight Sunday significantly raised the odds of death.
The in-hospital death rate for weekends compared with weekdays was 42 percent vs. 36 percent for the ruptured artery; 13 percent vs. 11 percent for the lung clot; and 1.7 percent vs. 0.3 percent (a more than fivefold increase) for the throat infection.
For 23 of the 100 most commonly fatal conditions, including breast cancer, leukemia and kidney failure, patients admitted over the weekend were significantly more likely to die in the hospital than those admitted on a weekday, the researchers found.
Since hospital infrastructure doesn't change from day to day, Bell says the mortality difference must lie with the workforce. It doesn't help that the weekend shift is considered the province of rookies, he says.
"Our message to the people in charge of hospitals is they should be more creative in creating an environment where the weekends are as attractive as the weekdays" to work, says Bell.
Joanne Turnbull, executive director of the National Patient Safety Foundation, says: "Research on the frequency and type of error, that is the epidemiology of health care error, is clearly important. However, addressing the numbers in one study is not an answer to the patient safety problems in today's health care culture."
Alicia Mitchell, a spokeswoman for the American Hospital Association, calls the results "inconclusive" and emphasizes that they don't translate to the health care system the United States.
Bell and Redelmeier say their study might contain a few important flaws. They're not sure, for instance, if people who come to the hospital on Saturday and Sunday aren't sicker than those admitted during the week.
The study also didn't account for national holidays, which might have affected the death rates. Nor does it cite any data showing that medical errors are more common on the weekends, something that might reflect staff inexperience.
On the other hand, they say, they might have underestimated the true difference by not including deaths recorded by emergency crews outside the hospital, which are known to be more common on the weekends.
Overall, 26.5 percent of patients were admitted on the weekends. Patients were roughly half male and had an average age of 51. Ten percent were children, and one in five was treated at a teaching hospital.