Abstract
To assess the influence of physicians' diagnostic practices on the recognition of nosocomial infections, the authors analyzed data collected on 339,044 patients selected randomly from admissions in 1970 and 1975–1976 to 338 randomly selected hospitals representative of all acute-care US hospitals. Eight rates representing the frequency of cuftures or chest x-rays among patients with or without signs of infection were calculated. These varied widely among hospitals, were highest in teaching hospitals and in the Northeast, but increased more among small hospitals and in the South and West. Rates of performing urine cuftures and reporting colony counts were highly correlated with observed rates of nosocomial urinary tract infection. Analogous measures were moderately correlated with observed bacteremia rates and pneumonia rates but were only weakly associated with surgical wound infection rates. These data indicate that the nationwide increase in the use of these diagnostic tests increased the recognition of infectious diseases in US hospitals.
