gms | German Medical Science

50. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds)
12. Jahrestagung der Deutschen Arbeitsgemeinschaft für Epidemiologie (dae)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie
Deutsche Arbeitsgemeinschaft für Epidemiologie

12. bis 15.09.2005, Freiburg im Breisgau

Estimation of the within Ward Transmission of Resistant Bacteria in an Intensive Care Unit

Meeting Abstract

  • Ulrich Sagel - Ev. Krankenhaus Bielefeld (EvKB), Bielefeld
  • Rafael T. Mikolajczyk - Universität Bielefeld, Bielefeld
  • Reiner Bornemann - Universität Bielefeld, Bielefeld
  • Alexander Krämer - Universität Bielefeld, Bielefeld

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. Deutsche Arbeitsgemeinschaft für Epidemiologie. 50. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds), 12. Jahrestagung der Deutschen Arbeitsgemeinschaft für Epidemiologie. Freiburg im Breisgau, 12.-15.09.2005. Düsseldorf, Köln: German Medical Science; 2005. Doc05gmds525

Die elektronische Version dieses Artikels ist vollständig und ist verfügbar unter:

Veröffentlicht: 8. September 2005

© 2005 Sagel et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen ( Er darf vervielfältigt, verbreitet und öffentlich zugänglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.




Nosocomial infections are an increasing challenge for infectious disease epidemiology. Bacteria with resistance or multiresistance to antibiotics are of special concern [1]. One practical issue is the early detection of clusters and outbreaks [2], another crucial question for the infection control is the distinction between within hospital transmission of the pathogens and new introduction by admission of already colonized patients or mutation of formerly sensible bacteria. The answer allows the choice of an appropriate intervention, since the within hospital transmission can be influenced directly by improved hygienic measures. The other causes cannot be avoided by hygiene and require either emphasis on screening at admission or early detection of new cases to contain secondary spread. We propose an approach to estimate the within hospital transmission in a low endemicity setting.

Material and Methods

We used data collected routinely according to PIA (German Protection against Infection Act [3]) in a 14-bed surgical intensive care unit (ICU) of a 893-bed tertiary level hospital in the state North Rhine-Westphalia, Germany. We analysed three single pathogens with highest incidence: methicillin-resistant Staphylococcus aureus (MRSA), multiresistant Acinetobacter baumannii (complete resistance except for carbapenems; MAB) and Imipenem-resistant Pseudomonas aerugionosa of serotype O:11 (or serotype indeterminate; IRPA) during a study period between 01.01.2002 and 31.12.2004. Intermediate results in the resistograms were considered as resistant.

The microbiological monitoring of tracheal aspirate and urine was routinely performed twice a week in all patients considered to be at a higher risk of infection. All three pathogens appeared in one or more clusters on the epidemic curve during the study period suggesting that within hospital transmission occurred to some degree. 75 (6 %) of all patients with S. aureus had MRSA, 22 (32%) with Acinetobacter baumannii had MAB and 36 (15%) with Pseudomonas aeruginosa had IRPA.

We analysed the distribution of intervals between first bacteriological cultures containing the given pathogen in consecutive patients. In the absence of transmission the occurrence of a pathogen should be governed by a homogeneous Poisson process resulting in exponentially distributed intervals between the events. On the opposite, the transmission events would be expected to cumulate in shorter time intervals. We fitted an exponential distribution to the observed intervals between the subsequent events, removing iteratively the shortest intervals. The time periods before the first case and after the last case till the end of the observation period were treated as additional units of observation. Finally, we calculated the expected number of events for the homogenous Poisson process and compared it to the overall number of observed cases during the study period.


We estimated the proportion of MRSA cases due to within hospital transmission at 32 % including all observed cases which increased to 80 % after intervals below 20 days between diagnosed persons were removed. MAB occurred only in one outbreak with 21 of 22 cases with less then two weeks between consecutive diagnoses. We were not able to fit a distribution reflecting the occurrence of MAB in the absence of an outbreak, but because of the qualitative information we assume the within hospital transmission to be above 90 %. For IRPA we estimated 26 % of cases being transmitted within hospital based on all cases and this number increased to 60 % after intervals below 20 days were excluded.


We found an increased proportion of subsequent cases in short periods after an index case was diagnosed with a given pathogen. This is inconsistent with the absence of transmissions and a solely homogenous random process of either admission of colonized cases or development of newly resistant cases. The estimate increases considerably after short intervals between subsequent cases were removed. The exclusion of short intervals is motivated by contamination of the information through transmissions, but further research is needed for the evaluation of the cut-off point. At present, the cut-off of 20 days is based on clinical plausibility.

The proportions of within hospital transmissions differed between the three investigated pathogens. This is not surprising since different biological properties are known to lead to some distinct epidemiological behaviour of bacteria.

Apart from statistical aspects of estimation, the actual number of diagnoses is probably to low due to incomplete screening and especially sporadically cases of admission are likely not to be diagnosed. This diagnostic bias may lead to an overestimation of the within hospital data in our analysis.

Estimation of within ward transmission of resistant bacteria from routine data appears to be feasible and may contribute to our understanding of the epidemiology of nosocomial infection.


Vincent JL. Nosocomial infections in adult intensive-care units. Lancet 2003; 361: 2068-77
Sagel U, Mikolajczyk RT, Krämer A: Using mandatory data collection on multiresistant bacteria for internal surveillance in a hospital. Methods Inf Med 2004; 43: 483-5
Anonymous. Surveillance of nosocomial infections and recording of pathogens with specific resistances and multiresistances [German]. Bundesgesundheitsbl-Gesundheitsforsch-Gesundheitsschutz 2000; 43: 887-90