Early detection of nosocomial outbreaks caused by rare pathogens: a case study employing score prediction interval.

OBJECTIVES

Nosocomial outbreaks involve only a small number of cases and limited baseline data. The present study proposes a method to detect the nosocomial outbreaks caused by rare pathogens, exploiting score prediction interval of a Poisson distribution.

METHODS

THE PROPOSED METHOD WAS APPLIED TO THREE EMPIRICAL DATASETS OF NOSOCOMIAL OUTBREAKS IN JAPAN: outbreaks of (1) multidrug-resistant Acinetobacter baumannii (n = 46) from 2009 to 2010, (2) multidrug-resistant Pseudomonas aerginosa (n = 18) from 2009 to 2010, and (3) Serratia marcescens (n = 226) from 1999 to 2000.

RESULTS

The proposed method successfully detected all three outbreaks during the first 2 months. Both the model-based and empirically derived threshold values indicated that the nosocomial outbreak of rare infectious disease may be declared upon diagnosis of index case(s), although the sensitivity and specificity were highly variable.

CONCLUSION

The findings support the practical notion that, upon diagnosis of index patient(s), one should immediately start the outbreak investigation of nosocomial outbreak caused by a rare pathogen. The proposed score prediction interval can permit easy computation of outbreak threshold in hospital settings among healthcare experts.