Surveillance

HAIs occur in all veterinary facilities; if HAIs are never recognized, it is because surveillance has been ineffective at detecting them. All comprehensive infection control programs should have a surveillance component because the information collected serves as sensory input to guide ongoing efforts so that they are focused and efficient.

Surveillance Options and Strategies

Theoretically, an optimal surveillance system would allow real-time detection of all HAIs. However, pursuing this goal would be unrealistic for reasons of practicality and would be an inefficient use of resources. Although all HAIs are important relative to patient well-being, some are more common and more important than others. This is because the impact on patient health is much more significant for some diseases, as well as because highly contagious diseases are more likely to affect a larger number of patients. Further, regardless of whether there is a contagious etiology, HAIs that affect a large number of patients can have a tremendous impact on the cost of care and the overall burden of morbidity for patients. Focusing surveillance on these high-impact HAIs is logical, as a greater proportion of these occurrences are also likely to be preventable in comparison to more sporadic problems.

Determining which specific HAIs were likely preventable is a difficult and potentially contentious task. Rather than focusing on individual infections, infection control efforts in human hospitals have benefited by focusing on comparing rates of infection and estimating the proportion that might be preventable by referencing some accepted standard. During the past 30 years, human hospitals have made great efforts to characterize rates of HAIs that can be expected even under the best of circumstances. By estimating rates for these “nonpreventable” infections, it has then been possible to identify hospitals with higher-than-average rates for HAIs. Because of the tremendous differences among hospitals and patient populations, it has become standard to focus on more restricted, high-risk patient groups for which there is better comparability (e.g., neonatal or cardiac intensive care patients). It is also far more feasible to perform surveillance for HAIs in these high-risk patient subgroups than it is in the general population, which has much lower risks of HAIs. Performing surveillance with standardized case definitions²⁰ over time allows identification of specific risk factors, which could then be used to prospectively refine management of patients (i.e., a high-risk patient could be identified and subjected to more aggressive control measures as a preventive strategy). Unfortunately, there is very limited information available on rates of disease related to HAIs at individual hospitals let alone estimation of rates in multiple hospitals; this information would be required in order to obtain interhospital comparisons that would be useful in identifying latent problems at specific facilities. Although making standardized comparisons between hospitals would require tremendous cooperative effort, individual facilities could make significant progress in identifying important changes in rates for HAIs if efforts were made to benchmark their infection rates over time.

When designing a formal surveillance effort to aid infection control programs, it is important to tailor efforts to a specific facility or practice. A wide variety of design possibilities are available for hospital surveillance systems, and the specific focus and methods should be carefully matched to the needs and resources of each establishment. Diseases that are the highest priority for surveillance must be specified along with standard­ized case definitions that will be used to identify these cases. Efforts may target disease caused by HAIs that are related to specific procedures or organ systems, or just as commonly they may target specific infectious agents of interest. For example, systems might be established to look for surgical or intravenous catheter site infections, methods may target surveillance for respiratory tract infections, or efforts may specifically target detection of infections with specific etiologic agents such as Salmonella or MRSA. It must also be determined whether clinical disease will be the outcome of interest or whether it is important to look for animals with subclinical infections. This will generally vary by disease and likely will be determined by the natural history or pathophysiology of the disease (i.e., Does shedding occur in the absence of clinical signs? Is there a chronic carrier state? What risk factors are related to the likelihood of shedding?). Further, if clinical disease is the outcome of interest, it is important to consider whether the case definition will include some type of etiologic confirmation, or whether diagnoses will be syndromic in nature and based on clinical signs (i.e., surgical site infections could be defined based on recovery of a bacterial agent in the presence of clinical disease, or it could be defined solely on the presence of a predetermined combination of clinical signs such as erythema and drainage). As discussed, another consideration is whether surveillance will target specific subgroups of the population or include all patients.

Another important consideration is how data will be gathered for this effort. Significant personnel compliance is needed for active culture-based surveillance of large numbers of patients. Similarly, any type of chart review system for benchmarking expected rates of HAIs requires a significant time commitment. Surveillance efforts are greatly aided by computer-based search mechanisms, and it is strongly recommended that practices consider using electronic medical record systems to maximize the ability to perform surveillance and that medical record systems be designed to facilitate collection and recovery of infection control-related data. For example, something as simple as being able to monitor the number of febrile or leukopenic patients daily could be extremely powerful as an aid to infection control efforts. Even financial databases can be useful in surveil­lance efforts. For example, they may be very useful for benchmarking antimicrobial prescriptions or specific procedures, such as intravenous catheterization, surgeries, or submission of samples for diagnostic tests.

Generalized, nontargeted surveillance to look for potentially pathogenic bacteria that can be found in health care environ­ments is widely considered of no real benefit to infection control efforts. This includes sampling surfaces in surgical theatres to catalog the different bacteria that can be found. However, there is utility in targeted environmental cultures where there is biological plausibility as to their relevance and a plan to act on the results (and where the action differs if results are positive or negative). For example, monitoring for contamination of veterinary hospitals with Salmonella has been a useful adjunct to patient surveillance in comprehensive infection control programs at veterinary hospitals.^36,64,76,77 In addition, in some situations, such as when attempting to control ongoing out­breaks, there is no substitute for culturing the hospital environ­ment to detect important environmental reservoirs.^1,3,4 Use of electrostatic household wipes has been extremely useful for sampling to detect Salmonella in routine surveillance as well as in the face of outbreaks.

The use of rapid and sensitive diagnostic tests is particularly important for the management of highly contagious diseases. As such, polymerase chain reaction (PCR) and antigen detection tests, which can be much more rapid than traditional culture systems for bacteria and viruses, have become more widely used for both patient and environmental surveillance efforts. However, these tests cannot fully replace culture-based assays because analysis of microbes is essential to fully understand the epidemiologic implications for each recovery. For example, if on a single day three patients were found to be shedding Salmonella after use of PCR testing, it would not be possible to know whether these were unrelated events. In contrast, if shedding were detected using culture and the Salmonella isolates were then shown to be from different serogroups, this would make it unlikely that shedding was related to exposures from a single source.

Monitoring Personnel Compliance

Success of an infection control program depends on universal participation within a hospital or other animal care facility that is born from a well-developed sense of responsibility among individuals. Actively engaging all personnel in infection control programs becomes increasingly difficult as organizations become larger and more complex. Unfortunately, the most cautious actions of a majority of personnel can be for naught if even one individual neglects to take appropriate precautions under just the right (wrong) circumstances. Although it is very important for administrators to engage in surveillance to detect systematic noncompliance with policies, it is just as important to empower and expect all personnel to monitor for individual acts of noncompliance. Remembering that most noncompliance is born from a natural tendency for people to revert to the most convenient practices, which are not necessarily the most “safe” practices from the infection control perspective, it is important to couple surveillance efforts with education so that all personnel fully understand why they are being inconvenienced. In addition, providing useful feedback and communication about issues that arise in the hospital will help to reinforce the need for compliance. For example, monitoring and reporting bacterial shedding or environmental contamination detected as a part of surveillance programs can help to keep hospital personnel aware of the potential hazards of reduced infection control efforts.

Measurements Related to Health and Disease

Although outcomes for individual animals may be the most important bottom line for many clinicians, it is important to remember that infection control by its very nature often has a larger population-based perspective. Thus although diagnostic test results for an individual animal obviously have great relevance for that animal, for the population a positive test result may not have great meaning unless it is interpreted contextually. It is therefore critical to remember that in addition to tracking numerators (e.g., numbers of infections or animals with disease caused by HAIs), some type of denominator information is needed to provide relevance over time or among different subpopulations. Examples of relevant denominators are patient admission totals for a given period, patient-days of hospitalization, numbers of procedure events such as surgeries or catheter placements, and numbers of samples submitted for culture.

Developing a Comprehensive

Surveillance Strategy

As mentioned, it is unrealistic to expect to detect all occurrences of HAIs or all animals shedding agents of interest through any hospital surveillance system. Rather, veterinarians should work toward developing strategic surveillance programs that will allow reliable estimation of rates for important events if they occur with any significant frequency, as well as rapid detection of unexpected important events. It is likely that a mix of strategies described previously will be used for the various agents and disease problems. For example, if S. enterica shedding occurs at some low-level yet regular frequency in a hospital (e.g., 3% of all hospitalized equine patients) and is considered an important potential hazard for other patients, it is reasonable to develop an active surveillance program that allows detection of clinical and subclinical shedding of Salmonella as a patient management tool and detection of important trends over time so that outbreaks of HAIs can be rapidly identified and stopped. In contrast, Clostridium difficile or EHV-1 may not have been detected with any regularity in patients at a particular hospital or may be more difficult to actively monitor because of available testing methodologies, yet they may still be considered important because of their significance as pathogens and their potential for contagious spread. For these pathogens, it may be more reasonable to use a reliable passive surveillance strategy to call attention to patients when clinical signs or diagnostic testing indicates that these agents may be present. In addition, maintaining some level of surveillance for compliance with infection control procedures and policies is necessary to ensure that an infection control program is functioning properly.