HEALTH INDICATORS: Conceptual and operational considerations (Section 3)

3.3 MAIN DATA SOURCES

The main data sources include demographic censuses, vital statistics information systems, disease reporting systems, cancer registries, population-based research, and other sample-based research—which may be local (subnational) and/or periodic (regular)—and the various information systems created by the health and other sectors for administrative purposes.

Below is a description of the most frequently used data sources for the development of health indicators:

3.3.1 DEMOGRAPHIC CENSUSES

In most countries, the demographic census is the most commonly used data source on the characteristics of a population. A census is of paramount importance for preparing indicators and planning health interventions. Other demographic data sources are household censuses, civil registries, and national estimates for variables of interest. Demographic data are necessary for calculating many health-related indicators.

The data in a national census include: (a) total population, by sex, age, and ethnic origin; (b) increase in population; (c) rural and urban distribution of the population; and (d) dependency ratio. In addition, demographic censuses become secondary data sources when used to establish the denominators of many health indicators (rates, proportions, ratios): mortality; incidence and prevalence of diseases, accidents and violent acts; prevalence of risk factors for accidents and violence, as well as sequelae of these events; and ratio of hospital beds to population.

The growing need for information has made censuses an essential tool for countries' information systems. They serve a political, administrative, technical, and scientific purpose. Such data are compiled for the entire population, through personal interviews. In most countries, censuses represent periodic data that are collected every ten years, with publication of the information they yield approximately two years later.

One limitation in the use of census data is the possibility of inaccuracies in population estimates for the inter-censal years. These estimates tend to lose accuracy the further they are from the year of the census. In addition, the calculations are subject to changes as new demographic information is generated. There are various methods for making these estimates, each based on its own assumptions, but all include the basic demographic factors of fertility, mortality, and migration (1).

With major changes in the factors affecting a country's demographic structure, existing methodologies can become problematic, particularly with respect to population projections. There is the difficulty, for example, of making adequate population projections based on the 10-year demographic censuses given the steep decline in fertility rates that occurred between 1980 and 2010. Population migrations resulting from conflicts or economic factors can also affect population projections. In addition, these factors can affect estimates of health indicators since population data from demographic censuses and the corresponding projections are used as the denominator. For these reasons, if indicators are to be monitored over time, they should be recalculated retrospectively as new population projections are made.

3.3.2 HEALTH INFORMATION SYSTEMS

National health information systems provide data for health-related events; these systems also provide some census-related data. National health information systems can have sub-systems that address specific health events to include mortality, births, notifiable diseases, cancer and other disease registries.

• Mortality information systems: All of the countries in the Region of the Americas are required to report all deaths. In some countries, this requires the completion a form known as the "declaration of death," followed by registration of the event in the civil registry system where a death certificate (a legal document), is issued. WHO has established an international form of medical certificate of cause of death. It contains a minimum set of variables that should be included in a death certificate. These include the underlying, intermediate, and immediate causes of death. Most countries use the International Classification of Diseases (ICD) (2) to code the causes of death. This permits comparisons between countries and at different points in time. In some countries, especially in remote areas, the reporting of deaths is incomplete, thereby compromising the representativeness of mortality statistics. The proportion of under-reporting of deaths and the proportion of deaths due to ill-defined causes are indicators of the quality of mortality statistics (3-5).
• Birth registration systems: These are a source for census data and for the collection of vital statistics. In all of the countries in the Region of the Americas, it is compulsory for every live birth to be reported. In most of these countries, births occur in health facilities where a form known as the "declaration of live birth," is issued. Subsequently, the birth must be recorded in the civil registry where a birth certificate, a legal document, is issued. Declarations of births generate data for formulating indicators that are highly useful in monitoring maternal and child health during the prenatal, delivery, and perinatal periods, while also providing information on a population's fertility profile. Standardizing definitions, forms, and variables of interest facilitate comparisons between countries and over time. The main limitation in working with birth indicators is the fact that coverage may be incomplete, particularly in remote areas within certain countries thereby compromising the representativeness of the resulting statistics. The proportion of unregistered births and of incomplete information on important variables are indicators of the quality of birth registration (6).

• Public health surveillance systems: According to WHO, "public health surveillance is the continuous, systematic collection, analysis, and interpretation of health-related data needed for the planning, implementation, and evaluation of public health practice." (7). These systems are useful sources of data for developing morbidity indicators regarding the prevention and control of communicable diseases, noncommunicable diseases, accidents, and violence. Surveillance systems for communicable diseases play a key role in providing early warning of possible threats to public health and make it possible to monitor measures and programs for prevention and control. Effective national surveillance and response systems are therefore essential for national, regional, and global health security.

Most of the Region's countries have national surveillance systems established by their ministries of health. These differ, however, from country to country with regard to the number of diseases under surveillance, the type of information compiled, the use of electronic or hardcopy records, and the systems' coverage. Surveillance of noncommunicable diseases can be conducted by health care services (with information from health institutions) or using population-based health surveys. Additionally, surveillance can occur through disease-specific programs such as for tuberculosis and vaccine-preventable diseases. In some countries, the surveillance of certain diseases may not be integrated into the national surveillance system. Notably, surveillance case definitions can change over time in response to changes in the characteristics of an epidemic, as in the case of HIV/AIDS and H1N1. Another surveillance tool is the International Health Regulations 2005 (IHR) which calls for the monitoring of diseases that have the potential to cross borders and threaten people worldwide.

In recent years, surveillance systems have been used to monitor a broad range of health conditions, risk factors, and other public health issues. Some countries have implemented surveillance registries for noncommunicable diseases (cervical cancer, acute myocardial infarction, violence, diabetes, etc.) or for health problems caused by toxic substances in the environment. These registries compile information on all cases of these diseases through the health care network or through sentinel surveillance units. Such systems are a source of data on morbidity indicators that are useful in disease prevention and control efforts (8).

The collection, analysis and interpretation of surveillance data and indicators on immunization were essential in formulating strategies to control and eliminate vaccine-preventable diseases in the Region. Notably, PAHO is the first WHO region certified as polio-free and the first to have interrupted the endemic transmission of measles and rubella.

• Population-based cancer registries: These are the gold standard for providing information on this disease. Cancer registries compile, classify, analyze and report information on all cancers occurring in a geographically defined population. The information is collected from a multiplicity of sources, including hospitals, clinical analysis laboratories, and departments of vital statistics. Routine calculation of rates (per 100,000 inhabitants) produced by population-based cancer registries provides information that can assist public health officials to better understand the disease, its treatment, and to evaluate cancer prevention and control programs. This information provides a solid basis for planning and implementation of programs to reduce the cancer burden as well as provide input for research.

3.3.3 ROUTINE DATA FROM HEALTH FACILITIES

As part of their normal activities, health facilities generate highly useful data on the delivery of services (hospitalizations, consultations, surgeries, insertion of prostheses, etc.), coverage of health interventions, resources (number of physicians, nurses, and beds, number of vaccines given, health sector expenditures, etc.), and disease patterns. For a number of indicators-such as access to health services and the cesareansection rate-data from health facilities are the sole source. Moreover, health facilities are important sources of subnational data, by province, region, or district that have a direct bearing on management decisions.

Within the Region, particularly in the Caribbean, the countries have shown an increasing interest in implementing national information systems on hospitalizations, ambulatory and emergency services, and in the use of electronic medical records. Some countries, such as Belize and Dominica, have made major investments in these areas.

Since these data sources tend be created for administrative purposes rather than for monitoring of health events per se, the epidemiological data they generate may have biases that affect their coverage, quality, and usefulness. In some of the countries, these systems are fragmented, and information exchange between the public and private sectors is difficult, leading to late, incomplete, and/or incorrect reports. Despite these challenges, there is reason for optimism, given the growing introduction of new information technologies (electronic medical records, etc.) with the potential to substantially improve the quality and use of data for decision-making (9).

Data on human resources in health can be obtained through professional associations. However, the data can be distorted if the information on deaths, migration, type of training, and current occupation, etc., are not periodically updated.

3.3.4 POPULATION SURVEYS

Population surveys have become an important source of information on health status and the social determinants of health. They can provide data for several purposes such as risk factors; access to (and use of) services; medication availability and use; morbidity, mental health; violence and injuries; disability; drug abuse; reproductive health; work conditions; lifestyles, etc.

These surveys can estimate prevalence rates for diseases, risk factors, behavior, prior conditions, situations of vulnerability, knowledge, attitudes and habits regarding various health-related practices, use of health services, etc. The target population can be a general population defined by its geographic area, or a specific population defined by some attribute of interest such as adolescents, schoolchildren, women of childbearing age, males in specific age groups, persons deprived of liberty, etc.

An example of a population survey is the Global Tobacco Surveillance System (GTSS). This is an exhaustive survey that monitors tobacco use and the policies enunciated in the WHO Framework Convention on Tobacco Control (WHO/FCTC), though that system is not designed to facilitate comparisons with other factors. An example of a multi-thematic survey is the WHO STEPwise approach to surveillance, which has the characteristics of a general survey. This survey encompasses a multiplicity of issues for monitoring within-country trends. It permits the compilation of biological and physical measures, though it provides only limited possibilities for close monitoring of the entire range of issues, a constraint that particularly affects the ability to monitor policy implementation (10).

Survey data can be gathered through various types of communication, including personal interviews (household surveys), telephone interviews, and questionnaires completed by the subjects themselves. Population surveys make it possible to monitor the population over time using probabilistic samples—a method that involves lower cost and simpler fieldwork—or using a more complex and costly structure. Surveys conducted in schools have been used globally and are simple and relatively inexpensive way of surveying a population. This is in contrast to household surveys, which involve a more complex structure and process.

Household surveys can be used as data sources to estimate indicators regarding determinants of health, health status, and trends in health expenditures at the national and individual levels.

Constraints in using surveys as data sources include the following:
a) people provide information based on their own evaluation of their health and illnesses; b) in terms of response rates, the success of a survey depends on its planning and on interviewees' willingness to provide information; and c) representativeness and disaggregation of data can considerably increase the cost of a survey (10-11). Additionally, sustainability due to a scarcity of resources or political will can be an obstacle.

Many countries in the Region have conducted household surveys on different areas of interest. Some countries establish national health surveys as a key source of information on various health-related issues. Nevertheless, it is important that countries continue their efforts to strengthen national capacities, create sustainable surveillance systems, and provide adequate budgetary allocations to ensure the production of reliable, systematic, standardized, and timely national health information.

Figure 4 shows a selection of national and international surveys that have been conducted in the Region of the Americas to provide information on non-communicable diseases and risk factors.

Source: WHO, World Health Organization, Centers for Disease Control and Prevention (CDC); FCTC, WHO Framework Convention on Tobacco Control; United States Agency for International Development (USAID); United Nations Office on Drugs and Crime (UNODC); Organization of American States (OAS); Inter-American Drug Abuse Control Commission (CICAD); United Nations Children's Fund (UNICEF); human immunodeficiency virus (HIV).