![[ Visit the Fifth Decennial online ]](images/banner.gif)
Friday, March 19, 2010
Grand Hall (Hyatt Regency Atlanta)
Background: Published data on the prevalence of healthcare-associated infection (HAI) in patients who die in the hospital are sparse and generally not well adjusted for severity of illness on admission. Thus, the extent to which HAI is simply a marker of severe underlying illness or a significant contributor to hospital mortality is unclear.
Objective: To determine the prevalence and attributable mortality of HAI in patients who die in the hospital.
Methods: We selected a random sample of 150 of the 496 patients who died in a large public teaching hospital during calendar year 2008. We measured severity of illness on admission using the Mortality Probability Model II at zero hours (MPM II0), which has been validated prospectively in critically ill patients. We identified HAI using the most recently published Centers for Disease Control and Prevention (CDC)/National Healthcare Safety Network (NHSN) surveillance definitions.
Results: Sixty-seven of the 150 patients survived more than 48 hours in the hospital and were thus at risk for development of HAI. Twenty-two (33%) of these 67 patients actually acquired one or more HAI prior to death. Twelve (18%) had more than one HAI. There were 18 lower respiratory tract infections, 11 urinary tract infections, 5 bloodstream infections, 3 Clostridium difficile infections, and 2 other infections. The severity of illness on admission was unexpectedly lower in the patients who developed HAI compared to patients who did not (mean MPM II0 score 38% versus 50% respectively, p = 0.11). Length of stay (LOS) in the hospital was longer in patients who developed HAI compared to those who did not (mean 20 days versus 8 days respectively, p < 0.0001). Interestingly, however, HAI tended to occur earlier rather than later in the hospital stay with a mean time to first HAI of only 10 days compared to the mean LOS of 20 days.
Conclusions: HAI is frequent in patients who die in the hospital, occurring in one-third of the patients in our study. Patients with HAI actually had lower predicted mortality scores compared to those who did not develop HAI. This suggests that HAI is not simply a marker of severe underlying disease. Additionally, patients who developed HAI had significantly longer LOS; however, HAI occurred earlier rather than later in their hospital stay. This suggests that the HAI added to LOS rather than being caused by prolonged LOS. Together these findings suggest that HAI is a significant contributor to hospital mortality. We conclude that both the prevalence and attributable mortality of HAI are high in patients who die in the hospital, perhaps much higher than previously thought.
Objective: To determine the prevalence and attributable mortality of HAI in patients who die in the hospital.
Methods: We selected a random sample of 150 of the 496 patients who died in a large public teaching hospital during calendar year 2008. We measured severity of illness on admission using the Mortality Probability Model II at zero hours (MPM II0), which has been validated prospectively in critically ill patients. We identified HAI using the most recently published Centers for Disease Control and Prevention (CDC)/National Healthcare Safety Network (NHSN) surveillance definitions.
Results: Sixty-seven of the 150 patients survived more than 48 hours in the hospital and were thus at risk for development of HAI. Twenty-two (33%) of these 67 patients actually acquired one or more HAI prior to death. Twelve (18%) had more than one HAI. There were 18 lower respiratory tract infections, 11 urinary tract infections, 5 bloodstream infections, 3 Clostridium difficile infections, and 2 other infections. The severity of illness on admission was unexpectedly lower in the patients who developed HAI compared to patients who did not (mean MPM II0 score 38% versus 50% respectively, p = 0.11). Length of stay (LOS) in the hospital was longer in patients who developed HAI compared to those who did not (mean 20 days versus 8 days respectively, p < 0.0001). Interestingly, however, HAI tended to occur earlier rather than later in the hospital stay with a mean time to first HAI of only 10 days compared to the mean LOS of 20 days.
Conclusions: HAI is frequent in patients who die in the hospital, occurring in one-third of the patients in our study. Patients with HAI actually had lower predicted mortality scores compared to those who did not develop HAI. This suggests that HAI is not simply a marker of severe underlying disease. Additionally, patients who developed HAI had significantly longer LOS; however, HAI occurred earlier rather than later in their hospital stay. This suggests that the HAI added to LOS rather than being caused by prolonged LOS. Together these findings suggest that HAI is a significant contributor to hospital mortality. We conclude that both the prevalence and attributable mortality of HAI are high in patients who die in the hospital, perhaps much higher than previously thought.