Use of strain typing data to estimate bacterial transmission rates in healthcare settings

OBJECTIVE: To create an affordable and accurate method for continuously monitoring bacterial transmission rates in healthcare settings. DESIGN: We present a discrete simulation model that relies on the relationship between in-hospital transmission rates and strain diversity. We also present a proof of concept application of this model to a prospective molecular epidemiology data set to estimate transmission rates for Pseudomonas aeruginosa and Staphylococcus aureus. SETTING: Inpatient units of an academic referral center. PATIENTS: All inpatients with nosocomial infections. INTERVENTION: Mathematical model to estimate transmission rates. RESULTS: Maximum likelihood estimates for transmission rates of these two species on different hospital units ranged from 0 to 0.36 transmission event per colonized patient per day. CONCLUSIONS: This approach is feasible, although estimates of transmission rates based solely on strain typed clinical cultures may be too imprecise for routine use in infection control. A modest level of surveillance sampling substantially improves the estimation accuracy (Infect Control Hosp Epidemiol 2005;26:638-645).