Modeling intrafacility spread of KPC-producing bacteria and impact of cohort strategies in long-term acute care hospitals in the Chicago region, USA
Methods: Data on room occupancy, admission cultures, and every-other-week point prevalence cultures were available from four LTACHs in the Chicago region from June 2012 until June 2013. Three different cohort strategies were adopted at the LTACHs: a pure cohort (all KPC-positive patients on one floor), single rooms for KPC-positive patients, and a mixed cohort (all KPC-positive patients on one floor, supplemented with KPC-negative patients). A data-augmented Markov chain Monte Carlo method with Metropolis-Hastings algorithm was developed to model the transmission process in the LTACHs and to study the effect of different cohort strategies. The transmission process was described by the background transmission rate α (including transmissions independent of the colonization pressure, such as endogenous selection) and the patient-dependent transmission rate β (including transmissions dependent on the colonization pressure: β * ward prevalence).
Results: The average point prevalence of KPC among patients as calculated by the model was 35%. The overall estimates were 0.0022 for α and 0.011 for β. 18% of patients were colonized on admission to the LTACHs and sensitivity of the screening process to detect KPC was 81%. The number of acquisitions per 1000 patient days was lowest in the LTACHs with a pure cohort ward or private rooms for colonized patients compared to mixed cohort wards.
Conclusion: The prevalence of KPC-producing bacteria in LTACHs is high, primarily due to a high admission prevalence and the resultant impact of high colonization pressure on risk of cross-transmission. Use of a pure cohort or single rooms for KPC-positive patients in LTACHs seemed to limit transmission compared to use of a mixed cohort.
M. Y. Lin, None
S. Weiner, None
D. Blom, None
K. Lolans, None
N. Moore, None
R. D. Lyles, None
R. A. Weinstein, None
M. Bonten, None
M. K. Hayden, None
See more of: Poster Abstract Session