The World Health Organization (WHO) defines surveillance as “the ongoing systematic collection, analysis, and interpretation of data in order to plan, implement and evaluate public health interventions.” There are three key elements to surveillance: (1) the continuous, cyclical collection of data in a systematic manner; (2) the timely analysis and interpretation of these data, and (3) the linkage of this information to operational activities and policies. These three elements share a symbiotic relationship, and each one is of little value without the other two. Together they provide a unique tool to track trends in overall population health and identify emerging public health concerns. Surveillance differs from other data gathering mechanisms used during disasters, such as rapid assessments and field surveys, which typically are cross sectional and provide a “snapshot” of particular health needs and outcome variables.
Disasters create circumstances in which general-population health is difficult to sustain and epidemics can easily arise. In this context, robust surveillance provides the only reliable means of monitoring public health, managing the burden of disease on health care systems, identifying emerging diseases with epidemic potential, and evaluating successes or failures in policy, programming, and disaster management.
Surveillance has been practiced in some form since the 1348 bubonic plague epidemic. However, in the seventeenth century, Leiniz in Germany and Graunt in England advanced the notion of applying numerical value to disease and tracking death counts. At the same time, Sydenham developed a classification for diseases, providing a uniform recognition of disease definitions and the basis upon which statistical analyses could be understood. By the nineteenth century, efforts were being made to nationalize health information, collect vital statistics, perform analyses, and initiate a reporting mechanism. Chadwick in England and Shattuck in Massachusetts noted the links between poverty and disease and promoted the idea of adding socioeconomic information, as well as geographic and occupational information, to death notices. By the turn of the century, all U.S. states and most European countries mandated the reporting of specific infectious diseases. In the United States the Great 1918 Influenza Pandemic prompted a national mortality reporting requirement.
Events occurring during the 1950s and 1960s helped to hone the definition of the term surveillance and to elucidate its role in public health practice. Langmuir at the U.S. Centers for Disease Control and Prevention (CDC) emphasized the need for the systematic collection of pertinent data and its timely analysis and dissemination to policy makers. Meanwhile, Raska at the WHO stressed that surveillance should apply not only to the control of communicable diseases but also to their prevention. In 1968 the World Health Assembly broadened the role of surveillance beyond the realm of communicable diseases. Since then, surveillance has been applied to lead poisoning, injury, substance misuse, congenital malformations, behavioral risk factors, and disasters, along with a comprehensive list of other public health issues.
The routine use of surveillance in humanitarian relief response followed the massive population displacement into Goma, Zaire, in 1994. When over one million people rapidly moved onto uninhabitable land with inadequate public health infrastructure, an epidemic of cholera, followed soon after by dysentery, swept through their camps, killing 50,000. The crude mortality rates recorded by the humanitarian community during the first month of the epidemic were among some of the highest in recent history: between 19.5 and 31.2 deaths per 10,000 population per day, far above the emergency threshold of 1 death per 10,000 per day. The crisis led to an initiative to develop standards of practice in field operations: in 1996 a group of experts from 228 humanitarian organizations inaugurated the Sphere Project, a consensus-driven document elaborating minimum technical and ethical standards in all sectors of humanitarian aid work. The resulting Sphere Handbook emphasizes that “the design and development of health services should be guided by the collection, analysis, interpretation, and use of relevant public health data.”
Likewise, the contemporaneous Médicins sans Frontières refugee health handbook identified surveillance as one of the top-ten priorities in an emergency and “an integral part of all relief activities.”
The cholera epidemic that followed the 2010 Haiti Earthquake focused attention on postdisaster surveillance. Prior to the earthquake, the Haitian Ministry of Health tracked only six indicators, all communicable diseases (acute hemorrhagic fever, acute flaccid paralysis, suspected bacterial meningitis, suspected diphtheria, suspected measles, and bite by animal suspected of having rabies). Within weeks following the event, a national sentinel surveillance system had expanded to 25 indicators, including watery diarrhea, bloody diarrhea, tuberculosis, and tetanus, as well as earthquake-related noncommunicable diseases, such as renal failure (from crush injuries), trauma and fractures, mental health, infected wounds, and concussion, among others. When cholera not endemic to the country emerged 9 months later, a rapidly scalable cholera-specific surveillance system was needed to capture daily cases in all areas of the country. The implementation of Haiti’s National Cholera Surveillance System was built on the rigorous postdisaster national system that by then also critically included the displaced populations.
The role of ongoing, pertinent health information that is iterative and available to those who must intervene in the context of humanitarian crises is now a widely accepted and essential component of disaster response and rehabilitation.
During and immediately after a disaster, health information systems and national surveillance systems may be underperforming, being either overwhelmed quickly and disabled or destroyed. To provide timely data on the primary needs of a disaster-affected population in the early days of the emergency, the WHO, in 2012, published guidance on the establishment of an Early Warning Alert and Response Network (EWARN). The first steps are to identify the EWARN network (responsible for collecting and analyzing data and disseminating information to relevant organizations) and to establish objectives for the surveillance system. The WHO outlines six objectives of surveillance applicable in emergencies: (1) identify public health priorities, (2) monitor the severity of the emergency by collecting and analyzing mortality and morbidity data, (3) detect outbreaks and monitor the response to these, (4) track trends in incidence and case fatality from major disease, (5) monitor the effects of specific health interventions, and (6) provide information to the Ministry of Health and other implicated organizations in the affected area to assist in planning and implementing health programs and mobilizing resources. To achieve these objectives, the surveillance system should be simple and clear in its focus, easily understood by all who receive its directions, and flexible enough to respond to new health problems and program activities.
The next step is to define the population of interest and its relationship to the health sector in the disaster area. What is the demographic makeup of the population (age, sex, ethnicity, etc.)? Where do people live before and after the disaster? Where do people access health services, if at all? Where do health events occur? What are the barriers to universal health coverage? What is the structure and capability of the national health information system and which governmental, nongovernmental (NGOs), and international agencies are working in the health sector?
To answer these questions, surveillance systems track indicators , primarily quantitative measures that describe the overall health of the population and the process and outcome of health services. These give an indication of the efficacy and efficiency of a program, system, or organization. Their inherent value in a surveillance system is their ability to be tracked over time and to be compared with a baseline. Mortality, morbidity, and nutrition rates are the most common quantitative indicators followed during the emergency phase.
A surveillance methodology is built based on the initial rapid assessments made during the first phase of the emergency response. The outputs of those assessments should focus on the health problems that produce the highest morbidity and mortality rates, especially if the population affected is displaced. Mortality rates are the most important indicator for identifying a population under stress. Most commonly, the crude mortality rate (CMR) is used in emergencies to track the effects of disaster-generated communicable diseases or injuries. The practical initial quantitative baseline data captured at the onset of the disaster are the following:
Population size: specifically the total population affected, disaggregated by age and gender (common at-risk groups include women and children less than 5 years old), in-migration and out-migration as a result of the disaster or conflict
CMR, in units of deaths per 10,000 persons per day
Mortality rate for children under 5 years old (U5MR), in units of deaths of children under 5 years per 10,000 population of children under 5 years per day
Case fatality rates (CFRs), the percentage of persons diagnosed as having a specified disease who die because of that disease within a given period
Nutritional status, specifically weight-for-height ratios of children aged 6 to 59 months, expressed as a percentage of population with a Z-score less than − 2 (moderate global acute malnutrition) and less than − 3 (severe global acute malnutrition)
In the emergency phase, surveillance activities should specifically focus on those diseases most likely to cause death or significant morbidity, namely diarrheal illnesses (especially in a malnourished, displaced population), acute respiratory infections, measles, and meningitis. As the emergency evolves into the postdisaster phase, other communicable and noncommunicable diseases are added to surveillance monitoring, and, depending on context, may include HIV/AIDS, tuberculosis, sexually transmitted diseases, chronic diseases, injury, immunization coverage (e.g., the Expanded Program on Immunization), and health care access. Surveillance for the displaced, disaster-affected population should eventually be integrated into the national health information system of the country, which is discussed later in this chapter.
Initial sector-specific indicators will identify population needs in relation to critical services, specifically water, sanitation, and hygiene (WASH), food, and health services. Early critical sector indicators, expressed as ratios of populations, include the following:
Access to sanitation services and adequate access to potable water
Access to health care services
Capacity or level of function of health care services (numbers of providers for the population)
Health care infrastructure (functioning hospitals and clinics)
Food security (physical and economic access to food)
As the disaster response becomes established and transitions into the postdisaster phase, qualitative indicators that examine the relief effort will guide resources and programming. These may include the following:
Factors that affect the ability of the population to access health care services
Quality of services (degree to which services received are adequate for population needs)
Equity in distribution of health resources (may explain why one subgroup of the population fares worse than another)
Main program activities (for individuals or groups of relief agencies involved in the delivery of care)
In terms of program performance, field epidemiologists recommend that indicators have SMART attributes: in other words, they should be “ s pecific, m easurable, a ccurate, r ealistic and t ime-bound.” Indicators must also be sensitive enough to assess the effect of interventions on health problems and determine whether the effort is having a tangible effect on the population or whether new strategies are needed. Once the emergency phase is over, the WHO emphasizes the importance of “re-integrating EWARN back into the national surveillance system.”
Methods for data collection depend on the type of surveillance used and the preexisting status of national health information gathering systems. Active surveillance involves a proactive seeking out of cases through some type of sampling mechanism in defined regions; this type of surveillance is used more frequently during the emergency phase and usually relies on direct reporting from households or, more commonly, health facilities. Passive surveillance systems rely on reports from the data sources, typically, as cases present themselves to health facilities. Passive systems are more suited to long-term, ongoing surveillance during the postemergency phase, once health systems and national surveillance programs have been restored. Both have practical pros and cons. Although active surveillance produces data that are more accurate, it is far more labor intensive and thus more costly to implement. However, the need for accurate mortality rates and incidence of critical diseases justifies the cost and effort during the first phase of the emergency. Passive systems by definition represent a self-selected sample and may not reflect accurately a health issue within the greater population. Because passive systems are less costly and require less training, they are more practical for long-term surveillance. In either case, the surveillance system must have high sensitivity in detecting disease (i.e., the ability to detect “true” cases of illness or injury). Another frequently used method of surveillance is sentinel surveillance , in which selected data collection points (e.g., designated health facilities within the system and in strategic geographic locations) are tasked with identifying and reporting specific diseases or health events. Designated segments of the population or particular locales in the disaster-affected population may be specially chosen for monitoring certain indicators (e.g., the parameter weight-for-height in children aged 6 to 59 months). Sentinel health events refer to conditions with the potential to affect the health stability of the population, often a warning signal that the current level of preventive and curative care needs attention. The occurrence of one of these events should prompt immediate action. One case of meningitis in an overcrowded camp for displaced persons demands an instantaneous response from health agencies to avert an epidemic catastrophe.
Data collection tools, especially in the emergency phase, should be short, easy to use, and readily understandable. Moreover, they should remain consistent over time. These tools are designed to collect only the minimal, most-essential information in a clear and unambiguous fashion. The use of simple, uniform case definitions for communicable diseases is critical. Such definitions can be found on the websites of both the Centers for Disease Control and Prevention (CDC) and the WHO. All health providers involved in surveillance should know, for instance, that a generalized rash for more than 3-days duration and a temperature greater than 38 °C and one or more symptoms of cough, rhinorrhea, or red eyes define a case of measles. Common diarrhea is classified as three or more liquid stools per day. In countries where malaria is endemic, a temperature greater than 38.5 °C in the absence of other infection is indeed malaria for data purposes. From these simple yet sensitive definitions, trends can be established and monitored. Epi Info software, available free of charge from the CDC website, assists in both generating data forms and analyzing data.
Data sources are most often health facilities, specifically the health care providers who daily record hospital emergency departments and clinic visits. In a displaced population living in temporary camps, registration systems provide demographic data. Household surveillance may be necessary to identify basic needs. Other sources of demographic and health data include vaccination cards (often carried by mothers), burial records, networks of community health workers, and distribution lists of nongovernmental agencies responsible for other sectors of the disaster response. Nontraditional sources of information should also be remembered, including police, fire, aid agencies, pharmacies, gravediggers, and others. With the global spread of Internet access, mobile communications, and social media, digital communications are proving valuable assets to public health informatics supporting surveillance data collection and analysis.
Surveillance systems should be judged by their simplicity. Data collection should be an easy process with a logical format. Indicators should represent communicable and noncommunicable diseases of the population of concern that are relevant to both the disaster and the phase of the emergency (e.g., diarrhea is more likely to be a problem than hookworm infestation in an emergency). Data outputs should be produced in a timely fashion to identify outbreaks and be reliable (using standardized case definitions, for example). Data collection must occur in an ongoing way at regular intervals but be flexible enough to adapt to new health problems or sudden program changes.
Analysis, Interpretation, and Dissemination
During the emergency phase, data should be collected daily and reported weekly to all government and nongovernment relief agencies working in the health sector. Such reports are often collated by the United Nations Office for the Coordination of Humanitarian Affairs (OCHA) and published on their website ( www.unocha.org ), as well as ReliefWeb ( www.reliefweb.int ). In the postemergency phase, most likely, data will be collected daily but analyzed monthly. During all phases, surveillance data should be sent in either paper or digital form to a central location at the field level for analysis. Most countries have some type of ongoing surveillance system, with personnel tasked for inputting new data, rapidly analyzing it, and tracking disease burden. Usually this is a component of a broader health information system housed within a Ministry of Health with an established means for data analysis (in the United States, state and local public health offices and the CDC). A country’s health information system, assuming it remains intact after the disaster, should provide baseline predisaster health information on communicable and noncommunicable diseases. This information provides important contextual information, such as seasonal variability of endemic diseases and the burden of chronic diseases and periodic health trends, information that is useful for program planning and conducting comparisons with newly acquired postdisaster data. In the event where a national surveillance and health information system is rendered disabled from the disaster, the WHO’s EWARN and the CDC provide technical analytical assistance working with the affected country’s health ministry and participating health NGOs to establish emergency-phase surveillance and to support ongoing postemergency-phase surveillance and its transition to a formalized health information system.
The analysis of surveillance data should focus on the clustering of events over time, within specific subgroups of the population, emphasizing trends in time, place, and person. The interpretation should highlight the needs and priorities and, where possible, be referenced against “normal” values or indicator thresholds. What matters most is how indicators change over time and how these changes may be linked to programming. Finally, the flow of surveillance data must ensure that those persons responsible for program planning and implementation receive the information in an actionable format. The use of geographic information system (GIS) software for mapping and geopatial graphic analysis makes interpretation more meaningful to stakeholders.
If an alert threshold is crossed during data collection or analysis or if the analysis demonstrates an increasing trend, an alert verification goes out to point persons within the health information system at the regional and national levels (designated by the EWARN framework), to initiate an outbreak investigation and control measures. The 2005 International Health Regulations (IHR) provide a framework for WHO epidemic alert and rapid-response activities in collaboration with national EWARN systems.