- Centers for Disease Control and Prevention
1600 Clifton Rd
Atlanta, GA 30333
TTY: (888) 232-6348
- Contact CDC–INFO
Lesson 5: Public Health Surveillance
Section 4: Identifying or Collecting Data for Surveillance
After the problem for surveillance has been identified and defined and the needs and scope determined, available reports and other relevant data should be located that can be used to conduct surveillance. These reports and data are gathered for different purposes from multiple sources by using selected methods. Data might be collected initially to serve health-related purposes, whereas data might later serve administrative, legal, political, or economic purposes. Examples of the former include collecting data from death certificates regarding the cause and circumstances of death and collecting data from national health surveys regarding health-related behaviors; examples of the latter include collecting data on cigarette and alcohol sales and administrative data generated from the reimbursement of health-care providers.
Before describing available local and national data resources for surveillance, understanding the principal sources and methods of obtaining data about health problems is helpful. As you recall from Lesson 1, the majority of diseases have a characteristic natural history. An understanding of the natural history of a disease is critical to conducting surveillance for that disease because someone — either the patient or a health-care provider — must recognize, or diagnose, the disease and create a record of its existence for it to be identified and counted for surveillance. For diseases that cause severe illness or death (e.g., lung cancer or rabies), the likelihood that the disease will be diagnosed and recorded by a health-care provider is high. For diseases that produce limited or no symptoms in the majority of those affected, the likelihood that the disease will be recognized is low. Certain diseases fall between these extremes. The characteristics and natural history of a disease determine how best to conduct surveillance for that disease.
Examples of documentation of financial, legal, and administrative activities that might be used for surveillance
- Receipts for cigarette and other tobacco product sales.
- Automated reports of pharmaceutical sales.
- Electronic records of billing and payment for health-care services.
- Laws and regulations related to drug use.
Sources and Methods for Gathering Data
Data collected for health-related purposes typically come from three sources, individual persons, the environment, and health-care providers and facilities. Moreover, data collected for nonhealth–related purposes (e.g., taxes, sales, or administrative data) might also be used for surveillance of health-related problems. Because a researcher might wish to calculate rates of disease, information about the size of the population under surveillance and its geographic distribution are also helpful. Table 5.2 summarizes health and nonhealth-related sources of data, and the box to the left provides examples of nonhealth-related data that can be used for surveillance of specific health problems.
Table 5.2 Typical Sources of Data
Health-care providers, facilities, and records
- Physician offices
- Outpatient departments
- Emergency departments
- Inpatient settings
- Animal vectors
- Sales of goods and services
Laws and regulations
A limited number of methods are used to collect the majority of health-related data, including environmental monitoring, surveys, notifications, and registries. These methods can be further characterized by the approach used to obtain information from the sources described previously. For example, the method of collecting information might be an annual population survey that uses an in-person interview and a standardized questionnaire for obtaining data from women aged 18–45 years; or the method might be a notification that requires completion and submission of a form by health-care providers about occurrences of specific diseases that they see in their practices.
Depending on the situation, these methods might be used to obtain information about a sample of a population or events or about all members of the population or all occurrences of a specific event (e.g., birth or death). Information might be collected continuously, periodically, or for a defined period, depending on the need. Careful consideration of the objectives of surveillance for a particular disease and a thorough understanding of the advantages and disadvantages of different sources and methods for gathering data are critical in deciding what data are needed for surveillance and the most appropriate sources and methods for obtaining it.(9, 14) We now discuss each of these four methods.
Examples of environmental monitoring
- Cities and states monitor air pollutants.
- Cities and towns monitor public water supplies for bacterial and chemical contaminants.
- State and local health authorities monitor beaches, lakes, and swimming pools for increased levels of harmful bacteria and other biologic and chemical hazards.
- Health agencies monitor animal and insect vectors for the presence of viruses and parasites that are harmful to humans.
- National, state, and local departments of transportation monitor roads, highways, and bridges to ensure that they are safe for traffic; they also monitor traffic to ensure that speed limits and other traffic laws are observed.
- Public safety and health departments periodically monitor compliance with laws requiring seat belt use.
- Occupational health authorities monitor noise levels in the workplace to prevent hearing loss among employees.
Monitoring the environment is critical for ensuring that it is healthy and safe (see Examples of Environmental Monitoring). Multiple qualitative and quantitative approaches are used to monitor the environment, depending on the problem, setting, and planned use of the monitoring data.
A survey is an investigation that uses a “structured and systematic gathering of information” from a sample of “a population of interest to describe the population in quantitative terms.”(15) The majority of surveys gather information from a representative sample of a population so that the results of the survey can be generalized to the entire population. Surveys are probably the most common method used for gathering information about populations. The subjects of a survey can be members of the general public, patients, health-care providers, or organizations. Although their topics might vary widely, surveys are typically designed to obtain specific information about a population and can be conducted once or on a periodic basis.
A notification is the reporting of certain diseases or other health-related conditions by a specific group, as specified by law, regulation, or agreement. Notifications are typically made to the state or local health agency. Notifications are often used for surveillance, and they aid in the timely control of specific health problems or hazardous conditions. When reporting is required by law, the diseases or conditions to be reported are known as notifiable diseases or conditions.
Individual notifiable disease case reports are considered confidential and are not available for public inspection. In most states, a case report from a physician or hospital is sent to the local health department, which has primary responsibility for taking appropriate action. The local health department then forwards a copy of the case report to the state health department. In states that have no local health departments or in which the state heath department has primary responsibility for collecting and investigating case reports, initial case reports go directly to the state health department. In some states all laboratory reports are sent to the state health department, which informs the local health department responsible for following up with the physician.
This form of data collection, in which health-care providers send reports to a health department on the basis of a known set of rules and regulations, is called passive surveillance (provider-initiated). Less commonly, health department staff may contact healthcare providers to solicit reports. This active surveillance (health department-initiated) is usually limited to specific diseases over a limited period of time, such as after a community exposure or during an outbreak.
Table 5.3 shows the types of notification and examples.
Table 5.3 Types of Notification and Examples
- Disease or hazard-specific notifications
- Communicable diseases
- World Health Organization: International health regulations require reporting of cholera, plague, and yellow fever
- National: United States and Canada specify diseases that require notification by all states and provinces, respectively
- Provincial, state, or subnational: for example, coccidioidomycosis in California
- Chemical and physical hazards in the environment
- Childhood lead poisoning
- Occupational hazards
- Firearm-related injury
- Consumer product-related injury
- Communicable diseases
- Notifications related to treatment administration
- Adverse effect of drugs or medical products
- Adverse effect from vaccines
- Notifications related to persons at risk
- Elevated blood lead among adults
- Elevated blood lead among children
Adapted from: Koo D, Wingo P, Rothwell C. Health Statistics from Notifications, Registration Systems, and Registries. In: Friedman D, Parrish RG, Hunter E (editors). Health Statistics: Shaping Policy and Practice to Improve the Population’s Health. New York: Oxford University Press; 2005, p. 82.
Use of sentinel sites has become the preferred approach for human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) surveillance for certain countries where national population-based surveillance for HIV infection is not feasible. This approach is based on periodic serologic surveys conducted at selected sites with well-defined population subgroups (e.g., prenatal clinics). Under this strategy, health officials define the population subgroups and the regions to study and then identify health-care facilities serving those populations that are capable and willing to participate. These facilities then conduct serologic surveys at least annually to provide statistically valid estimates of HIV prevalence.
Because underreporting is common for certain diseases, an alternative to traditional reporting is sentinel reporting, which relies on a prearranged sample of health-care providers who agree to report all cases of certain conditions. These sentinel providers are clinics, hospitals, or physicians who are likely to observe cases of the condition of interest. The network of physicians reporting influenza-like illness, as described in one of the examples in Appendix C, is an example of surveillance that uses sentinel providers. Although the sample used in sentinel surveillance might not be representative of the entire population, reporting is probably consistent over time because the sample is stable and the participants are committed to providing high-quality data.
Maintaining registries is a method for documenting or tracking events or persons over time (Table 5.4). Certain registries are required by law (e.g., registries of vital events). Although similar to notifications, registries are more specific because they are intended to be a permanent record of persons or events. For example, birth and death certificates are permanent legal records that also contain important health-related information. A disease registry (e.g., a cancer registry) tracks a person with disease over time and usually includes diagnostic, treatment, and outcome information. Although the majority of disease registries require health facilities to report information on patients with disease, an active component might exist in which the registry periodically updates patient information through review of health, vital, or other records.
Reanalysis or Secondary Use of Data
Surveillance for a health problem can use data originally collected for other purposes — a practice known as the reanalysis or secondary use of data. This approach is efficient but can suffer from a lack of timeliness, or it can lack sufficient detail to address the problem under surveillance. Because the primary collection of data for surveillance is time-consuming and resource-intensive if done well, it should be undertaken only if the health problem is of high priority and no other adequate source of data exists.
Table 5.4 Types of Registries and Examples of Selected Types
- Vital event registration
- Birth registration
- Marriage and divorce registration
- Death registration
- Registries used in preventive medicine
- Immunization registries
- Registries of persons at risk for selected conditions
- Registries of persons positive for genetic conditions
- Disease-specific registries
- Blind registries
- Birth defects registries
- Cancer registries
- Psychiatric case registries
- Ischemic heart disease registries
- Treatment registries
- Radiotherapy registries
- Follow-up registries for detection of iatrogenic thyroid disease
- After-treatment registries
- Handicapped children
- Disabled persons
- Registries of persons at risk or exposed
- Children at high risk for developing a health problem
- Occupational hazards registries
- Medical hazards registries
- Older persons or chronically ill registries
- Atomic bomb survivors (Japan)
- World Trade Center survivors (New York City)
- Skills and resources registries
- Prospective research studies
- Specific information registries
Adapted from: Koo D, Wingo P, Rothwell C. Health Statistics from Notifications, Registration Systems, and Registries. In: Friedman D, Parrish RG, Hunter E (editors). Health Statistics: Shaping Policy and Practice to Improve the Population’s Health. New York: Oxford University Press; 2005, p. 91.
Weddell JM. Registers and registries: a review. Int J Epid 1973;2:221–8.
State funding for a childhood asthma program has just become available. To initiate surveillance for childhood asthma, the staff is reviewing different sources of data on asthma. Discuss the advantages and disadvantages of the following sources of data and methods for conducting surveillance for asthma. (Figure 5.12 in Appendix C indicates national data for these different sources.)
- Self-reported asthma prevalence and asthmatic attacks obtained by a telephone survey of the general population.
- Asthma-associated outpatient visits obtained from periodic surveys of local health-care providers, including emergency departments and hospital outpatient clinics.
Major health data systems
Data regarding the characteristics of diseases and injuries are critical for guiding efforts for preventing and controlling those diseases. Multiple systems exist in the United States to gather such data, as well as other health-related data, at national, state, and local levels. These systems provide the “morbidity and mortality reports and other relevant data” for surveillance, as described by Langmuir, and examples of such systems are listed in Appendix E. Remember, however, that surveillance is an activity — the continued watchfulness over a disease by using data collected about it — and not the data about a disease or the different data systems used to collect or manage such data.
Surveillance for communicable diseases principally relies upon reports of notifiable diseases from health-care providers and laboratories and the registration of deaths. Because the most common use of surveillance for communicable diseases at the local level is to prevent or control cases of disease, local surveillance relies on finding individual cases of disease through notifications or, where more complete reporting is required, actively contacting health-care facilities or providers on a regular basis.(10) At the state and national level, the principal notification system in the United States is the National Notifiable Disease Surveillance System (NNDSS). State and local vital registration provides data for monitoring deaths from certain infectious diseases (e.g., influenza and AIDS).
More About the National Notifiable DiseaseSurveillance System
A notifiable disease is one for which regular, frequent, and timely information regarding individual cases is considered necessary for preventing and controlling the disease.
The list of nationally notifiable diseases is revised periodically. For example, a disease might be added to the list as a new pathogen emerges, and diseases are deleted as incidence declines. Public health officials at state health departments and CDC collaborate in determining which diseases should be nationally notifiable. The Council of State and Territorial Epidemiologists, with input from CDC, makes recommendations annually for additions and deletions. However, reporting of nationally notifiable diseases to CDC by the states is voluntary. Reporting is mandated (i.e., by legislation or regulation) only at the state and local levels. Thus, the list of diseases considered notifiable varies slightly by state. All states typically report diseases for which the patients must be quarantined (i.e., cholera, plague, and yellow fever) in compliance with the World Health Organization's International Health Regulations.
Data in the National Notifiable Disease Surveillance System(NNDSS) are derived primarily from reports transmitted to CDC by the 50 states,two cities, and five territorial health departments.
National Notifiable Diseases Surveillance System[Internet]. Atlanta:CDC [updated 2006 Jan 13]. Available from: http://www.cdc.gov/ncphi/disss/nndss/nndsshis.htm.
Surveillance for chronic diseases usually relies upon health-care–related data (e.g., hospital discharges, surveys of the public, and mortality data from the vital statistics system). Given the slow rate of change in the incidence and prevalence of these diseases, data for surveillance of chronic conditions need not be as timely as those for acute infectious diseases.
Surveillance for behaviors that influence health and for other markers for health (e.g., smoking, blood pressure, and serum cholesterol) is accomplished by population surveys, which might be supplemented with health-care related data. The Behavioral Risk Factor Surveillance System (BRFSS), the Youth Risk Behavior Surveillance System (YRBSS), the National Health Interview Survey (NHIS), and the National Household Survey on Drug Abuse are all surveys that gather data regarding behaviors that influence health. The National Health and Nutrition Examination Survey (NHANES), probably the most comprehensive survey in the United States of health and the factors that influence it, gathers extensive data on physiologic and biochemical measures of the population and on the presence of chemicals among the population resulting from environmental exposures (e.g., lead, pesticides, and cotinine from secondhand smoke). Data from NHANES have been used for approximately 40 years to monitor the lead burden among the general public, demonstrating its marked elevation and then substantial decline after the mandated removal of lead from gasoline and paint.
Assume you work in a state in which none of the following conditions is on the state list of notifiable diseases. For each condition, list at least one existing source of data that you need for conducting surveillance on the condition. What factors make the selected source or data system more appropriate than another?
Listeriosis: A serious infection can result from eating food contaminated with the bacterium Listeria monocytogenes. The disease affects primarily pregnant women, newborns, and adults with weakened immune systems. A person with listeriosis has fever, muscle aches, and sometimes gastrointestinal symptoms (e.g., nausea or diarrhea). If infection spreads to the nervous system, such symptoms as headache, stiff neck, confusion, loss of balance, or convulsions can occur. Infected pregnant women might experience only a mild influenza-like illness; however, infections during pregnancy can lead to miscarriage or stillbirth, premature delivery, or infection of the newborn. In the United States, approximately 800 cases of listeriosis are reported each year. Of those with serious illness, 15% die; newborns and immunocompromised persons are at greatest risk for serious illness and death.
Spinal cord injury: Approximately 11,000 persons sustain a spinal cord injury (SCI) each year in the United States, and 200,000 persons in the United States live with a disability related to an SCI. More than half of the persons who sustain SCIs are aged 15–29 years. The leading cause of SCI varies by age. Motor vehicle crashes are the leading cause of SCIs among persons aged <65 years. Among persons aged ≥65 years, falls cause the majority of spinal cord injuries. Sports and recreation activities cause an estimated 18% of spinal cord injuries.
Lung cancer among nonsmokers: A usually fatal cancer of the lung can occur in a person who has never smoked. An estimated 10%–15% of lung cancer cases occur among nonsmokers, and this type of cancer appears to be more common among women and persons of East Asian ancestry.
References (This Section)
- 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.
- Hopkins RS. Design and operation of state and local infectious disease surveillance systems. J Public Health Management Practice 2005;11(3):184–90.
- Parrish RG, McDonnell SM. Sources of health-related information. In: Teutsch SM, Churchill RE, editors. Principles and practice of Public Health Surveillance, 2nd ed. New York: Oxford University Press; 2000, pp. 30–75.
- Groves RM, Fowler FJ, Couper MP, Lepkowski J, Singer E, Tourangeau R. Survey methodology. New York: John Wiley; 2004.