Shoo Lee and colleagues used anonymously linked data to determine the rates and causes of death among all 19 265 infants admitted to 17 Canadian neonatal intensive care units (NICUs) between January 1996 and October 1997.
Overall, there were 795 deaths (4%); 40% of which occurred within 2 days after admission to the NICU. Common conditions associated with death were birth at a different hospital than the NICU (340 [42%]), congenital anomalies (270 [34%]), infection (108 [14%]), and hypoxic-ischemic encephalopathy (128 [16%]). The risk-adjusted mortality rates differed significantly among the 17 NICUs (range 1.6% to 5.5%).
In a related commentary Jon Tyson and Kathleen Kennedy praise the study as being of “exemplary quality” but point to the need for careful risk adjustment in studies on the differences in patient outcomes among hospitals.
“. . . risk adjustment remains an imperfect process for neonates as for older patients,” warn the authors. “These realities need to be well understood by journalists, consumer advocates, government officials and other groups that publish hospital ‘report cards’ or ratings of patient outcome as measures of quality of care.” In another related commentary, John Hoey and colleagues briefly review the history of quality improvement studies in facilities such as health care institutions. They challenge the idea of maintaining institutional anonymity when publishing studies to important issues such as variations in mortality occurring at different public facilities.
p. 173 Variations in mortality rates among Canadian neonatal intensive care units — S. Lee, et al
p. 191 Variations in mortality rates among Canadian neonatal intensive care units: interpretation and implications — J. Tyson, K. Kennedy p. 193 What’s in a name? Reporting data from public institutions — J. Hoey, A.M. Todkill, K. Flegel
Journal
Canadian Medical Association Journal