Lesson 5: Public Health Surveillance

Section 3: Identifying Health Problems for Surveillance

Multiple health problems confront the populations of the world. Certain problems present an immediate threat to health, whereas others are persistent, long-term problems with relatively stable incidence and prevalence among the populations they affect. Examples of the former include influenza epidemics and hurricanes; the latter include atherosclerotic cardiovascular disease and colon cancer. Health problems also vary for different populations and settings, and an immediate threat among one population might be a chronic problem among another. For example, an outbreak of malaria in the United States in 2006 would be an immediate threat, but malaria in Africa is a chronic problem.

Selecting a Health Problem for Surveillance

Because conducting surveillance for a health problem consumes time and resources, taking care in selecting health problems for surveillance is critical. In certain countries, selection is based on criteria developed for prioritizing diseases, review of available morbidity and mortality data, knowledge of diseases and their geographic and temporal patterns, and impressions of public and political concerns, sometimes augmented with surveys of the general public or nonhealth-associated government officials. Criteria developed for selecting and prioritizing health problems for surveillance include the following: (9–12)

In the United States, the Centers for Disease Control and Prevention (CDC) and the Council of State and Territorial Epidemiologists (CSTE) periodically review communicable diseases and other health conditions to determine which ones should be reported to federal authorities by the states. Because of their greater likelihood of producing immediate, increased threats to public health, communicable diseases are the most common diseases under surveillance. Table 5.1 presents nationally notifiable infectious diseases for the United States for 2006. The Morbidity and Mortality Weekly Report (MMWR) presents a weekly and annual summary of nationally notifiable infectious diseases in the U.S. After priorities have been set, the extent to which a state or local health department can conduct surveillance for particular diseases is dependent on available resources.

Table 5.1 Nationally Notifiable Infectious Diseases — United States, 2006

Adapted from: National Notifiable Diseases Surveillance System [Internet]. Atlanta: CDC [updated 2006 Jan 13]. Nationally Notifiable Infectious Diseases United States 2006 . Available from: http://www.cdc.gov/ncphi/disss/nndss/phs/infdis2006.htm.

Defining the health problem, identifying needed information, and establishing the scope for surveillance

After a decision has been made to undertake surveillance for a particular health problem, adopting — or, if necessary, developing — an operational definition of the health problem for surveillance is necessary for the health problem to be accurately and reliably recognized and counted. The operational definition consists of one or more criteria and is known as the case definition for surveillance. The case definition criteria might differ from the clinical criteria for diagnosing the disease and from the case definition of the disease used in outbreak investigations. For example, the case definition of listeriosis for surveillance is provided in the box below. (See Lesson 1 for further discussion of case definitions and for an example of a case definition of listeriosis for outbreak investigation). CDC and CSTE have developed case definitions for common communicable diseases,(13) certain chronic diseases, and selected injuries.

Situations might exist in which the criteria for identifying and counting occurrences of a disease consist of a constellation of signs and symptoms, chief complaints or presumptive diagnoses, or other characteristics of the disease, rather than specific clinical or laboratory diagnostic criteria. Surveillance using less specific criteria is sometimes referred as syndromic surveillance.

For example, a syndromic surveillance system was put in place in New York City after the World Trade Center (WTC) attacks in 2001. Here, the objectives were to detect illness related to either a bioterrorist event or an outbreak because of concern that the WTC attack could be followed by terrorists' use of biological or chemical agents in the city. One example of non-bioterrorist syndromic surveillance is surveillance for acute flaccid paralysis (syndrome) in order to capture possible cases of poliomyelitis. This is an example where the syndrome is monitored as a proxy for the disease, and the syndrome is infrequent and severe enough to warrant investigation of each identified case.

The goal of syndromic surveillance is to provide an earlier indication of an unusual increase in illnesses than traditional surveillance might, to facilitate early intervention (e.g., vaccination or chemoprophylaxis). For syndromic surveillance, a syndrome is a constellation of signs and symptoms. Signs and symptoms are grouped into syndrome categories (e.g., the category of "respiratory" includes cough, shortness of breath, difficulty breathing, and so forth).

The term, as used in the UnitedStates, often refers to observing emergency department visits for multiple syndromes (e.g., "respiratory disease with fever") as an early detection system for a biologic or chemical terrorism event. The advantage of syndromic surveillance is that persons can be identified when they seek medical attention, which is often 1–2 days before a diagnosis is made. In addition, syndromic surveillance does not rely on a clinician's ability to think of and test for a specific disease or on the availability of local laboratory or other diagnostic resources. Because syndromic surveillance focuses on syndromes instead of diagnoses and suspect diagnoses, it is less specific and more likely to identify multiple persons without the disease of interest. As a result, more data have to be handled, and the analyses tend to be more complex. Syndromic surveillance relies on computer methods to look for deviations above baseline (certain methods look for space-time clusters). Emergency department data are the most common data source for syndromic surveillance systems.

You might use syndromic surveillance when:

• Timeliness is key either for naturally occurring infectious diseases (e.g., severe acute respiratory syndrome [SARS]), or a terrorism event;
• Making a diagnosis is difficult or time-consuming (e.g., a new, emerging, or rare pathogen);
• Trying to detect outbreaks (e.g., when syndromic surveillance identified an increase in gastroenteritis after a widespread electrical blackout, probably from consuming spoiled food); or
• Defining the scope of an outbreak (e.g., investigators quickly having information on the age breakdown of patients or being able to determine geographic clustering).

Syndromic surveillance is a key adjunct reporting system that can detect terrorism events early. Syndromic surveillance is not intended to replace traditional surveillance, but rather to supplement it. However, evaluation of these approaches is needed because syndromic surveillance is largely untested (fortunately, no terrorism events have occurred that test the available models); its usefulness has not been proven, given the early stage of the science and the relative lack of specificity of the systems. Criticism and concern have arisen regarding the associated costs and the number of false alarms that will be fruitlessly pursued and whether syndromic surveillance will work to detect outbreaks (See below for a possible scenario).

After a case definition has been developed, the persons conducting surveillance should determine the specific information needed from surveillance to implement control measures. For example, the geographic distribution of a health problem at the county level might be sufficient to identify counties to be targeted for control measures, whereas the names and addresses of persons affected with sexually transmitted diseases are needed to identify contacts for follow-up investigation and treatment. How quickly this information must be available for effective control is also critical in planning surveillance. For example, knowing of new cases of hepatitis A within a week of diagnosis is helpful in preventing further spread, but knowing of new cases of colon cancer within a year might be sufficient for tracking its long-term trend and the effectiveness of prevention strategies and treatment regimens.

Another key component of establishing surveillance for a health problem is defining the scope of surveillance, including the geographic area and population to be covered by surveillance. Establishing a period during which surveillance initially will be conducted is also useful. At the end of this period, the results of surveillance can be reviewed to determine whether surveillance should be continued. This approach might prevent the continuation of surveillance when it is no longer needed.

References (This Section)

9. Protocol for the evaluation of epidemiological surveillance systems [monograph on the Internet]. Geneva: World Health Organization [updated 1997; cited 2006 Jan 20]. Available from: http://whqlibdoc.who.int/hq/1997/WHO_EMC_DIS_97.2.pdf.
10. Hopkins RS. Design and operation of state and local infectious disease surveillance systems. J Public Health Management Practice 2005;11(3):184–90.
11. Doherty JA. Establishing priorities for national communicable disease surveillance. Can J Infect Dis 2000;11(1):21–4.
12. Rushdy A, O'Mahony M. PHLS overview of communicable diseases 1997: results of a priority setting exercise. Commun Dis Rep CDR Suppl 1998;8 (suppl 5):S1–12.
13. Centers for Disease Control and Prevention. Case Definitions for Infectious Conditions Under Public Health Surveillance. MMWR 1997;46(No. RR-10):1–55.