In this paper we have presented a new method for prospective MRSA outbreak surveillance in a hospital that uses case and molecular typing data. Historically, MRSA outbreak detection in hospitals has relied on the watchful eyes of physicians and other health-care workers. However, the increasing availability of timely electronic surveillance and molecular typing data raises the possibility of earlier outbreak detection and intervention if suitable analytic methods are found.
Germany belongs to a group of Western European countries with an intermediate level of MRSA (approximately 20% of all S. aureus
diagnosed in laboratories are MRSA positive). However, the isolation rate has increased significantly in recent years [27
]. Although the MRSA laboratory isolation rate in UHM of 6.4% in 2003 is still rather low in comparison with other German hospitals, the relative risk of acquiring MRSA within this hospital facility rose significantly during the study period (). Furthermore, the absolute risk will also probably rise because of epidemiological pressure and the rising prevalence of MRSA in Germany as a whole ( and ). It is clear that control of MRSA is a pressing concern where new concepts are needed, and therefore we studied spa
typing in combination with an automatic early warning algorithm to detect MRSA clusters at UHM.
We showed that the feasibility and speed with which it was possible to carry out spa
typing was highly satisfactory. The discriminatory power, however, was lower than previously reported, probably because only a local strain collection was analyzed [12
]. Although not examined by us, the high intra- and inter-laboratory reproducibility of 100% and the robustness of the method have recently been documented (Aires-de-Sousa et al., unpublished data). Moreover, there is a high concordance of results between spa
and PFGE, microarray and MLST [11
]. The practicability of using spa
in short-term epidemiological studies has been questioned because differences in PCR amplicon sizes in related strains was thought to imply instability in the target gene [28
]. In the meantime, however, there has been a plethora of publications demonstrating the value of spa
in the investigation of MRSA outbreaks, e.g., [10
]. Moreover, it has recently been shown that spa
data not only contain information on short-term, but also long-term evolutionary events, as observed in whole repeat duplications and deletions [12
]. Because of the steady fall in the cost of DNA sequencing and an average hands-on time of only 20 min per sample (determination of both strands of DNA and processing ten samples in parallel), this technique is within the capability of even small laboratories [30
The present study has compared three early warning algorithms for the detection of nosocomial MRSA outbreaks before limited clusters of preventable MRSA transmissions develop into larger outbreaks. The evaluation of an early warning system, however, is difficult because there is no accepted “gold standard” and it is likely that no system will be completely reliable [31
]. Therefore, we chose to combine epidemiological and molecular typing data with statistical analysis to provide an objective measure of performance between the varied approaches. In this approach, by definition, the sensitivity and NPV of the frequency and clonal alert methods will always be 100%. This means that the ICP method will give results that are less sensitive, or at best, of only equal sensitivity to those of the automated methods. Infection and colonization with MRSA were given equal status since both can lead to further transmissions. The typing data accumulated since 1998 enables the significance of spa
-type clustering with respect to time to be calculated for all those occasions when there is a suspicion of an outbreak. By excluding all non-significant clusters, it was possible to reduce the likelihood that two or more MRSA with the same spa
type, coincidentally isolated on the same or related wards within the 2-wk window, would be counted as correct. A 2-wk window is approximately 1.5 times as long as the mean duration of hospitalization in the UHM hospital. A 4-wk time window yielded similar results (unpublished data).
An outbreak can be defined as (i) two or more cases of infection by a common agent that are linked epidemiologically. However, this definition has usually limited practical relevance in the identification of outbreaks, because it presupposes that detailed epidemiological and typing data are available as soon as the outbreak occurs. Thus, in more practical terms, an outbreak is often defined operationally as (ii) an increase in the number of cases above expected levels [32
]. Historical data can be used to calculate a baseline and an alert is given when the number of cases exceeds a certain threshold. Early warning systems at national levels are based on this definition of an outbreak and have already implemented this approach successfully [33
]. A similar approach has also been used in hospitals, e.g., using 2-fold standard deviation and monthly increase algorithms for detecting clusters of nosocomial infections [31
]. In using the 2-fold standard deviation algorithm, the threshold for a suspected outbreak is defined as the mean of all previous cases per unit time plus two standard deviations. The monthly increase algorithm triggers an alert if there is either a 100% increase in the number of observed cases in the current month compared to the monthly totals for the two previous months, or a 50% increase over a three-month period. In the case of MRSA, however, infections/colonizations occur infrequently and irregularly. Applying the 2-fold standard deviation and monthly increase algorithms to our data (including typing data) resulted in delayed alerts for cluster detection indicating that they were insufficiently sensitive (unpublished data). In order to improve the detection of MRSA clusters and to avoid delay, we applied the first of the two definitions of an outbreak mentioned above, i.e., two or more cases of infection/colonization that are linked epidemiologically, because discriminatory typing details of cases were rapidly available.
The ability of the current procedures in hospitals to prevent nosocomial infections and to recognize nosocomial outbreaks often depends on the manual review of laboratory results and surveillance by the ICP. However, this review process is resource-intensive because duplicate isolates must be eliminated, results must be correlated with patients' charts, patient locations within the hospital must be tracked, and related events must be correlated and monitored [6
]. Not surprising, many minor transmissions of MRSA infections amenable to intervention go undiscovered. In this study, only five of the 13 “true” clusters were detected as clusters by visual screening of laboratory reports by the ICP ( and ). However, the ICP alerts had the highest PPV (81.8%) because of their high specificity. On the other hand, frequency alerts with a sensitivity of 100% detected every cluster. However, the high number of false-positive alerts giving the lowest PPV (27.4%) clearly demonstrates that this method is unsuitable (also shown by the significant differences in specificity and PPV compared with clonal and ICP alerts). Clonal alerts combine the best of both methods, i.e., high specificity and high sensitivity with a PPV comparable to that of ICP (no significant difference). In comparison to ICP, only a few more false-positive alerts were triggered, and more clusters, especially the smaller ones, were detected (). The data also indicated that surveillance conducted in the laboratory has the advantage in that clusters occurring throughout the hospital can be identified at a single, central data collection point. Further advantages are its speed (<3 d after detection of MRSA) and the portable nature of data generated by spa
typing permitting the differentiation between outbreaks and pseudo-outbreaks and the central coordination of a suitable response in real-time [14
]. Cost-benefit analyses have demonstrated that the cost of MRSA infections far exceeds those costs involved in active surveillance and isolation procedures in a hospital [35
]. Whether the expenditure required for spa
typing is less than that for the labor-intensive manual review of patients' charts and laboratory results needs to be determined.
Since the second half of 2004, when the study was finished and data analyzed, the Ridom StaphType software v. 1.5 beta that features an automated early warning based on clonal alerts came into routine use in our laboratory. A data-driven “re-admission alert” triggered by a hospital information system, which identifies the re-admission of any patient previously colonized or infected with MRSA, could enhance the accuracy of such a system [37
All methods described above are based on underlying rules or seek predefined patterns. The advantages of these hypothesis-based methods are the high sensitivity and specificity achievable. However, a rapid method with a high discrimination involving gene typing is necessary to attain such a high specificity. Due to the predefined rules, unusual patterns of outbreaks might go undetected (e.g., retarded epidemics). A different approach is employed by data mining, i.e., knowledge discovery in databases [6
]. Data-mining uses techniques based on computer science and statistics to search large event spaces (data warehouses) for interesting patterns that would otherwise have gone undetected by traditional analysis. These “discovery” models are independent of an underlying hypothesis, but are usually less sensitive and specific [38
There are a number of limitations within the proposed method. There is no definitive proof available, with any method, to authenticate a MRSA transmission event. Furthermore, the “gold standard” used incorporates elements of the diagnostic test under study. Finally, if the epidemiological pressure of a certain clone changes rapidly, temporal clustering could fail and false-positive (e.g., as possible in the case of the “Rhine Hesse” MRSA clone) or false-negative clusters could be recorded.
Examining the occurrence of MRSA clusters in the way described here not only provides a useful early warning system, but can also be used to model infection dynamics and estimate important epidemiological parameters, e.g., cross-transmission rates [39
]. Time series and typing data was used by Grundmann et al. with scan test statistics and risk factor analysis to show that the incidence of infection can be related to staffing levels [41
]. In addition, molecular typing data can also be incorporated into geographical information systems combined with space-time scan statistics analysis on a regional and national level [42
In conclusion, a surveillance method based on spa typing and automated alerts is useful as an early warning system in a hospital and is at least comparable to classical epidemiological approaches. We have shown that the combined use of medical informatics and molecular laboratory techniques makes intervention possible before limited clusters of preventable MRSA transmissions expand into outbreaks.
Everyone carries many types of bacteria on or in their bodies; Staphylococcus aureus is a normal bacteria for people to carry. About 25% to 30% of people have it, usually in the nose. It is usually harmless; however, this bacterium can also cause infections—especially in people who are otherwise unwell, or who have surgery. These infections need to be treated with antibiotics. Methicillin-resistant S. aureus (MRSA) is an increasing problem in much of the developed world because, unlike other types of these bacteria, MRSA cannot be killed by most of the usual antibiotics that are used, such as methicillin. Without treatment, staphylococcal infection can become very severe.
Why Was This Study Done?
MRSA is a particular problem in hospitals, where there is a need to be able to identify infected and colonized people quickly and isolate and treat them. These researchers wanted to test for the best way of identifying early clusters of MRSA outbreaks, which are more serious than just single cases and are an indication of hygiene deficiencies.
What Did the Researchers Do and Find?
Between 1998 and 2003 the researchers analysed 557 MRSA strains from staff and patients admitted to one German university hospital. They collected information about the characteristics (in space and time) of these people, and genetically identified each of the strains. They then looked for the most efficient way to identify an outbreak, including assessment of the risk by specially trained hospital staff, with and without genetic analysis. They also assessed a specially designed computer programme (developed by some of the authors), which combined the genetic type of the MRSA as well as details about the outbreak, such as the characteristics of the patients infected. They found that the most efficient and reliable method to identify outbreaks was to combine the genetic type of the MRSA with details about the outbreak, using the computer programme tested.
What Do These Findings Mean?
The computer programme seems to be more efficient than other methods tested here in identifying when an outbreak is likely to occur. However, this is the first test of this method, and before being adopted more widely, further testing is needed in different settings and by other researchers.
Where Can I Get More Information Online?
Medline Plus has many links to pages of information on different staphylococcal infections:
The Centers for Disease Control in the United States has a patient information sheet on MRSA:
The Health Protection Agency in the United Kingdom has a leaflet on MRSA aimed at patients: