National surveillance for infectious diseases began in 1878 when the United States Congress passed the Port Quarantine Act authorizing the U.S. Marine Hospital Service to collect cases of cholera, smallpox, plague, and yellow fever from U.S. Consulates overseas for quarantine purposes to prevent introduction and spread in the United States. In 1893, the Interstate Quarantine Report expanded to weekly publication to include cases from states and municipalities. In 1902, the U.S. Congress reorganized the U.S. Marine Hospital Service to the U.S. Public Health Service (USPHS) and authorized the U.S. Surgeon General to provide standard forms for collection, compilation, and reporting at the national level. In 1912, State and Territorial Health Authorities and the USPHS agreed to formalize state to federal reporting of immediate telegraphic reporting of five infectious diseases and monthly reporting by letter of an additional 10 diseases. Subsequently, at the 1950 annual meeting of U.S. State and Territorial Health Officials, it was agreed that state and territorial epidemiologists would determine with the USPHS the list of national notifiable diseases at an annual meeting. This was the initiation of the U.S. Council of State and Territorial Epidemiologists. In 1961, the Centers for Disease Control and Prevention (CDC) assumed responsibility for the collection of these nationally notifiable disease data from physicians and laboratories and the publication in the CDC’s Morbidity and Mortality Weekly Report. These data are currently used for measurement of incident endemic disease, outbreak recognition, assessment of prevention/control measures, allocation of public health resources, and contribution to epidemiological understanding of new and emerging pathogens.¹

Surveillance of healthcare-associated infections (HAIs) has been identified as a key strategy in preventing infections in healthcare settings. In the 1970s, the Study on the Efficacy of Nosocomial Infection Control (SENIC) identified that the impact of hospitals developing intensive surveillance and prevention programs could prevent up to onethird of HAIs.² Following this report’s findings, the CDC also recommended surveillance as a key strategy for prevention of HAIs. The CDC’s National Healthcare Safety Network (NHSN) is the current national surveillance system for HAIs. It was formed in 2005 as a combination of the National Nosocomial Infections Surveillance System (NNIS), the National Surveillance System for Healthcare Workers, and the Dialysis Surveillance Network. The NNIS was initiated in 1970 in order to describe the epidemiology of HAIs and provide national level comparative rates for these outcomes. Submission of HAI surveillance data into the NNIS and eventually NHSN was voluntary until 2006-2007, when many states began requiring reporting of HAIs,³ and later became publicly reportable as part of the U.S. Centers for Medicare and Medicaid Services (CMS) quality assurance programs. As part of this requirement, healthcare facility-specific surveillance data for certain HAIs must be reported to CMS; these include the following: central line-associated bloodstream infections (CLABSI, as of 2011 for intensive care unit [ICU] events, 2015 for events in medical and surgical wards), catheter-associated urinary tract infections (UTI) (CAUTI, 2012 for ICU events, 2015 for events in medical and surgical wards), surgical site infections (SSIs) after colon surgery and abdominal hysterectomy (2012), Clostridioides difficile infections (CDI, 2013), and methicillin-resistant Staphylococcus aureus (MRSA) bacteremia (2013).⁴ This performance is now available on the Hospital Compare Web site.⁵ The full cohort of surveillance components that are currently available within NHSN include biovigilance component (eg, transfusion-associated events), dialysis component, healthcare personnel (HCP) safety component (eg, HCP exposures and immunizations), long-term care facility component, outpatient procedure component, as well as the patient safety component, which has modules for device-associated infections (eg, CLABSI, CAUTI, ventilator-associated events [VAE]), procedure-associated module (eg, SSI), antimicrobial use and resistance module, and the multidrug-resistant organisms and C difficile infection module.⁶

NHSN is considered the premier surveillance system in the United States for HAIs, defined broadly as a localized or systemic condition that results from adverse reaction to the presence of an infectious agent(s) or its toxin(s) and that was not present or incubating at the time of admission to the healthcare system.⁷ The system offers a secure, Internet-based surveillance system for submission of both
patient and HCP safety surveillance systems. Standardized definitions are provided by the CDC, and data are collected and reported actively by infection preventionists. The CDC publishes the data as standardized infection ratios (SIRs) annually. NHSN provides infection rates or ratios with some risk adjustment primarily for device-associated infections and SSIs. Reporting into a national database allows acute care hospitals to benchmark their data to other hospitals, provides information on the scope of healthcareassociated pathogens, and provides data on the frequency and antimicrobial susceptibility of pathogens.

Surveillance of infectious diseases is used for public health/population health purposes and also for clinical/individual patient uses. Analysis of surveillance data across person, place, and time can provide critical information for planning, implementing, and evaluating public health practice (Fig. 2-2). Disease clusters in time can be reviewed to monitor temporal trends of diseases, identify unexpected increases in cases, and assess effectiveness of prevention and control measures. Epidemic curve graphs and histograms of case frequency by time can be useful
to visualize time trends. By assessing the geographic locations where diseases occur, epidemiologists can identify clusters of cases by location; identify risk factors, vehicles, or vectors for diseases; and target control measures to specific locations. Geographic information system mapping can aid in analysis and interpretation by place. Collecting surveillance data that allows for assessment of the types of persons affected by the disease can be important for identifying risk factors for disease and targeting prevention messages.

Surveillance data can be useful on a population level in order to plan, implement, and evaluate public health practices, but also because it provides informative data to clinicians as they assess the likelihood or pretest probability that their patient has a disease or condition based on the epidemiology of diseases (ie, person, place, and time). Based on the epidemiology of influenza season, they may choose not to test a patient with cough and fever who presents to their clinic in June for influenza since the influenza season historically ends in March. They may choose to not test a 43-year-old non-Hispanic patient with fever and rash for rubella when they have data demonstrating that the only recent outbreaks of rubella in their state were among nonimmunized Hispanics. Similarly, a clinician who examines a patient with encephalitis who lives in the piedmont/coastal region of a state may be unlikely to suspect La Crosse virus, a mosquito-borne virus that appears mostly in the mountain region of the state.

Surveillance begins with the collection of data that can happen in one of two ways. Passive reporting occurs when case reports are initiated by the reporter usually on preprinted forms. In this case, the epidemiologist seeking the surveillance data is waiting to receive information from sources who become aware of cases through their contacts in healthcare settings or laboratories. In active reporting, the reporter is contacted at regular intervals and specifically asked about the occurrence of the disease under surveillance. This method increases accuracy and sensitivity of the case reports. Active reporting can also be used to describe infection preventionists actively reviewing microbiology results daily to identify communicable disease exposures or HAIs within their facility.

The types of data that can be requested or reviewed for surveillance can be described as diagnostic or syndromic. Diagnostic data rely on patients to be assessed and diagnosed by clinicians before reporting cases. A case of meningococcal meningitis would be defined diagnostically as Gram-negative diplococci in the blood or cerebrospinal fluid (CSF).⁹ Syndromic data are available earlier in the patients’ continuum of care and would be based on a constellation of signs and symptoms that are suspicious for the disease of interest. Syndromic surveillance for meningococcal meningitis would query for purpura fulminans, a rash characteristic of meningococcal meningitis or bacteremia. A presentation of this type of rash in a chief complaint at an emergency department could trigger a report of this potential case of meningococcal meningitis or bacteremia to the surveillance system. While syndromic surveillance allows for earlier identification of cases, it is nonspecific and can lead to “alert fatigue” and to inefficiencies in following up on false-positive cases.¹⁰ Other syndromic alternatives, prediagnostic, have been proposed and studied, including examining data on prescription drug use, absenteeism data, and even tweets and Internet searches for symptoms. When designing a new surveillance system or examining an existing system, it is necessary to carefully consider the balance between active and passive reporting, which may also incorporate the consideration for whether the surveillance system is designed to detect specific disease cases or monitor patterns of a disease.

Another surveillance type is sentinel surveillance. Sentinel surveillance refers to the selection of a targeted group of locations where surveillance is performed. Sentinel surveillance can be a reasonable compromise when the data requested are new; reporting burden is unknown
or suspected to be high (eg, influenza); or additional epidemiologic, clinical, or laboratory data are desired (eg, CDC’s Active Bacterial Core surveillance system). For example, before carbapenem-resistant Enterobacteriaceae was added by the U.S. Council of State and Territorial Epidemiologists to the list of reportable public health diseases, several large hospitals were selected to perform laboratory-based surveillance.¹¹ For different types of HAI surveillance, facilities should consider the scope or breadth of their surveillance activities (ie, will surveillance occur equally across the entire facility or be focused in targeted locations). In addition, facilities should consider which approach (ie, patient-based, laboratory-based) will be used for detection of HAI cases. While some smaller facilities can examine individual patient’s medical records to assess for signs/symptoms of infections, other facilities may use reports based on positive microbiological findings to investigate patients’ records for identification of HAIs.

Although surveillance systems can be optimized or developed based on a disease under surveillance, most often a surveillance system is implemented that covers a range of diseases that are targeted for various reasons (eg, NHSN or NNDSS). An ideal surveillance system would have all of the following characteristics: simple, sensitive, specific, flexible, acceptable, representative, and cost-effective (Table 2-2).

A simple surveillance system would be both simple for reporters to submit cases to and simple for the epidemiologists to collect, analyze, interpret, and disseminate information. Features of simple surveillance systems would include simple case definitions that are straightforward to apply to readily available data and have a user interface that allows for electronic transfer of key data on the case.

Sensitive surveillance systems can readily and correctly identify patients with the condition. Sensitivity is the proportion of people who are detected by the surveillance system among all those who have the disease. Sensitivity is especially important with diseases with a high communicability (eg, measles) so that few cases are missed if complete follow-up of exposed individuals is necessary to prevent an outbreak. A specific surveillance system is a surveillance system that correctly does not include patient without the condition. Specificity is the proportion of healthy patients without the disease who test negative for it. Specificity may be more important for diseases that require direct contact tracing or have stigma associated with the disease. A surveillance system for partner notification after sexually transmitted disease exposure should be optimized for specificity to avoid a patient who is improperly identified with a sexually transmitted disease and then has partners inappropriately notified of potential exposure.