Document(s) 10 of 38

Introduction

During the past 30 years, demographic surveillance systems (DSSs) have been established in a number of field research sites in various parts of the developing world where routine vital-registration systems were poorly developed or nonexistent. Although these systems may have been developed differently in terms of their initial rationale, they are all required to track a limited and common set of key variables determining population dynamics and demographic trends. DSSs have similar approaches to defining key variables and their relationships and to developing systems for collection, storage, and analysis of these data. The core concepts presented here draw directly from the ideas and experiences emerging from INDEPTH DSS sites in Africa and Asia. It should be emphasized, however, that even though an effort has been made to standardize the definitions, many DSS sites still define some of the concepts differently.

Demographic surveillance systems

A DSS is a set of field and computing operations to handle the longitudinal follow-up of well-defined entities or primary subjects (individuals, households, and residential units) and all related demographic and health outcomes within a clearly circumscribed geographic area. Unlike a cohort study, a DSS follows up the entire population of such a geographic area.

In such a system, an initial census defines and registers the target population. Regular subsequent rounds of data collection at prescribed intervals make it possible to register all new individuals, households, and residential units and to update key variables and attributes of existing subjects. The core system provides for monitoring of population dynamics through routine collection and processing of information on births, deaths, and migrations — the only demographic events leading to any change in the initial size of the resident population. This core system is often complemented by various other data sets that provide important social and economic correlates of population and health dynamics. These may include information on events such as household formation and dissolution, acquisition and loss of economic assets, and growth or depletion of income.

In many population sites, the DSS may also provide a platform for other studies within the same geographic area. This support varies from one study to another and may include the provision of an initial sampling frame, adjustment for confounding variables, provision of additional explanatory variables, and measurement of the demographic impact of interventions.

Demographic surveillance area

The demographic surveillance area (DSA) is an area with clearly and fairly permanent delineated boundaries, preferably recognizable on the ground (for example, rivers, roads, and clearly demarcated administrative boundaries). The clear delineation of boundaries enables an unambiguous distinction to be made between individuals, households, and residential units to include in the DSS and those to exclude.

The area of a DSS site depends mainly on the size of the population required for demographic surveillance and related research activities (for a typical example, see “Establishing the monitored population” in Chapter 3). The size is also influenced by pragmatic considerations, such as the cost to the research centre and its capacity to manage the associated logistics and human resources. The DSA may expand or shrink over time in response to changing research needs or sources of funding. These changes usually introduce additional complexity, as they alter eligibility criteria and may make it difficult to maintain consistent definitions of internal and external migrations over the period of transition.

Longitudinality

Longitudinal measurement of demographic and health variables is one of the key characteristics of a DSS. This is achieved through repeated visits at more or less regular intervals to all residential units in the DSA to collect a prescribed set of attribute data on registered subjects, who are consistently and uniquely identified. This and recording events affecting these subjects during the interval between visits allow one to construct their history and differentiate DSS data from data collected in multiround surveys and other prospective studies that allow comparison over time only on an aggregated level.

Visits

DSSs collect data during rounds, or cycles, of visits to registered residential units in the DSA. The interval between visits depends on the frequency of the changes in the phenomena under study and on the length of recall intervals for the collected data, and thus on the research focus of each field site. However, like the size of the DSA and observed population, it also depends on funding and logistics. This interval varies from one site to another, ranging from 1 week to 1 year. However, for the majority of DSSs, observations are made at 3- or 4-month intervals. This is widely considered an appropriate interval to ensure comprehensive recording of births, deaths, and migrations, which is the minimum requirement for maintaining the coherence of any DSS .

When intervals between visits are long (a year or more), researchers commonly ignore migration events and instead conduct a full census at each new round. In- and out-migration flows are then inferred through reconciliation of unlinked census records after account is taken of births and deaths between censuses.

Data collected during each fieldwork round are not restricted to key demographic events but may also include the various attributes of the primary subjects. These attributes may be fixed (for example, ethnicity, gender) or changing over time (for example, marital or residential status).

Unique identifiers

Unique identifiers for primary subjects are an indispensable element of DSSs. All systems invariably formulate rules for assigning unique identifiers at the start of the DSS, but their methods for assigning these identifiers to DSS subjects may vary from one site to another. There are two main approaches. One common strategy is to transparently link the subjects in a single residential unit through a hierarchical system of unique numbers. These are built up from a unique number for the residential unit, followed by serial numbers for each of the households within it (where the notion of households applies) and then for each of the enumerated individuals within each household. In this system, the unique number for each individual in the DSS is a composite of the numbers for the residential unit, household, and household member. This may involve creating complex hierarchies, in which the unique number of the residential unit itself is a composite reflecting allocation to regions, areas, and villages (where they exist). This system requires thorough mapping of the DSA before enumeration. It also requires proper training of enumerators to avoid confusion in assigning identifiers. When mapping of the DSA is coupled with georeferencing of residential units, using geographic information system (GIS) technology, global positioning system (GPS) coordinates are assigned as location attributes of the residential units within the database.

The other strategy for assigning identifiers to individuals is to avoid any fixed link to residential units and households. In this system, identifiers for each subject are simply serial numbers incremented each time a new DSS subject is registered. This system requires providing field staff with block allocations of ID numbers with enough latitude to register new subjects. This approach should be coupled with computer generation of the identifiers to safeguard against the assignment of the same ID to multiple subjects on the ground. This strategy helps to preserve people’s anonymity outside their residential units, or when their attribute data are accessed through the database.

Primary DSS subjects

DSSs are typically structured around three main subjects (Figure 1.1) within the DSA. These subjects have both a conceptual and a logistical rationale. From a logistical point of view, it is not feasible to interview all individuals directly, and for this reason individuals are put in groups with physical and social meaning, and information is collected from credible and informed respondents within these groups. The reasons to distinguish between these subjects from a conceptual point of view will be dealt with in greater detail in the following subsections. The three main subjects are (Figure 1.1) as follows:

• Residential units — These are the places where individuals live. They are defined in physical and geographic terms.
• Households — These are the groups to which individual members belong. They are often defined as social subunits of the residential unit.
• Individuals — These are the people who are living in the residential units and households. They are the subject of main interest in any DSS.

Residential units

All DSSs identify residential units as a primary subject of interest, although they vary in the terms they use for these units (for example, compounds or homesteads) and may also differ slightly in their definition of them. Residency, or physical presence within a DSA at a fixed place of abode and for a sufficiently long period, is an essential prerequisite for the enumeration of individuals at risk for demographic events or disease exposure.

In most systems a distinction is made between places of residence and other structures, such as clinics, schools, churches, and stores. Identifying a unifying term for all these structural units may have conceptual merit, and some systems have attempted to do this, as these structural units share many characteristics and this approach simplifies the database hierarchy for handling this concept. In this system an inclusive term such as bounded structure may be used at a higher level and compounds (or homesteads) and facilities at the more specific level.

Households

Households may be variably defined in one or more of the following ways:

• A group of people who consume or make some contribution to food and other shared resources;
• A group of people who have a common allegiance to an acknowledged head of a household;
• A group of people, each of whom is recognized by other members of the household as belonging to a social group; or
• A group of people linked through ties of kinship.

The definition of household and its applicability both as a concept and as a separate DSS subject may vary greatly from one DSS to another. Households may simply be seen as fixed social subunits within residential units. In more complex systems, they may be seen as independent subjects able to change their place of residence while preserving their social identity, and they may have members who are resident elsewhere. In such a system, a clear distinction would be needed between residency, which defines the state of being physically present in a given residential unit for a defined threshold of time, and membership, which defines the state of belonging to a social group irrespective of physical presence. These concepts have a clear overlap with the related concepts of de facto population (persons who are physically present in a place) and de jure population (persons who usually reside in a given place), respectively. The concepts of residency and membership are discussed later in this chapter.

Individuals

The individuals are people of various ages, sex, and other personal characteristics who are residents or members of the DSS residential units or households, respectively. Their personal characteristics may be fixed (sex, date of birth) or change over time (age, marital status). Unless their changes are predictable (like the yearly increment of age), changing characteristics will need to be recorded repeatedly — or their changes will need to be recorded as events — to produce longitudinal trends.

Eligibility

Every DSS is required to define the population under surveillance. As most individuals within any population have places of residence and attachments to social groups, the task of defining the population begins with the identification of the residential units, households (where applicable), and individuals that will be visited and observed. Thereafter, a set of inclusion criteria must be applied to distinguish eligible from ineligible individuals or subjects within each subject category.

As residential units have fixed geographical positions in all DSSs, there are consistent and simple rules for their inclusion: they are included if they are situated in the DSA. In DSSs that deal with households as distinct (and potentially mobile) subjects, these households are eligible if (and while) they are situated in the DSA. This is what is referred to as household residency.

Rules for individuals, particularly in highly mobile populations, are more complex. The most typical approach is to simply base their eligibility on residence, that is, physical presence. Individuals are eligible if (and while) they are resident at eligible residential units. This is what is referred to as individual residency. Another approach,

based on social linkages, rules that individuals are eligible if (and while) they are members of eligible households. This requires careful and consistent definitions of household and membership and can allow individuals who are not resident to remain as members of the household and therefore to qualify for observation.

Residency and membership

Clear geographical boundaries for the DSA and well-defined physical boundaries for residential units are minimal prerequisites for following up DSS subjects consistently and arriving at numerators and denominators for rate calculations. In systems where residential units and households are separate subjects and there is a separate relationship between individuals and each of those subjects — expressed as residency and membership, respectively — these concepts become substantially more complex.

Observing an individual’s presence in, or absence from, a specific residential unit requires clear rules for residency status. The physical presence of an individual for a very short time may not be taken into account when the amount of time spent in the residential unit is computed. Conversely, the noncontinuous presence of an individual, with short periods of absence, may be considered continuous residency if he or she meets a threshold for inclusion.

Residency and membership statuses are assigned at the start of the DSS, based on prescribed eligibility rules. Thereafter, new residency episodes may commence as a result of births or in-migrations exceeding a prescribed threshold of duration, and current residency may end because of deaths or out-migrations, again exceeding a prescribed threshold of duration. New membership episodes may commence as a result of events that initiate a social relationship with a household, such as birth, marriage, adoption, or household formation, and may be terminated by events that end such a relationship, such as death, divorce, or household dissolution.

Core DSS events

To know the size of the registered resident population at any time, a DSS collects information about three core events that alter this size, namely, births, deaths, and migrations. These events are described by the following fundamental demographic equation:

[1.1]

where P is the population; B is the number of births; D is the number of deaths; I is the number of in-migrants; O is the number of out-migrants; and t₀, t₁ is the time interval of their occurrence.

An underlying principle for recording events in a DSS is that of a population at risk. Mortality, fertility, and migration rates are calculated by counting the number of deaths, births, or migrations occurring within a registered population exposed to the risk. For example, an individual who is not resident within the DSA is not considered at risk of dying within the area. Consequently, most DSSs do not observe nonresident individuals or households and do not record their events.

Births and fertility

Pregnancies and their outcomes for all women registered in the DSS are recorded regardless of the place of occurrence of such events. The recording of births has two purposes: for estimating fertility and for identifying a criterion for registering an individual. To estimate fertility, a DSS should record all pregnancy outcomes, including miscarriages (<28 weeks), induced abortions, stillbirths (=28 weeks), and live births. All live births are then registered as individual members of the DSS, independent of subsequent survival. In some DSSs, fieldworkers take note of live births to visitors to the DSA to alert the data collector in the next round to register the mother (if she becomes eligible) and her child. This procedure is very helpful, as it greatly improves the accuracy of dates of birth of newly born babies and increases reporting of births from eligible mothers with frequent in- and out-migration.

Although most DSSs will report their estimates of the fertility of a specific age group of women, usually 15–49 years, they should also record births to women outside this age group.

The underreporting of pregnancies and their outcomes is a major problem across all DSSs. Some DSSs have used the recording of pregnancies during routine update visits to improve birth coverage. Pregnancy observation has also been used to increase the reporting of other pregnancy outcomes, particularly miscarriages, induced abortions, and stillbirths. However, this requires an update-visit interval of <5 months so that a notification of pregnancy can be obtained in one round, followed by the recording of the pregnancy outcome in the next visit.

Deaths and mortality

Deaths of all registered and eligible individuals are recorded, regardless of the place of death. It may be impossible to record the deaths of previously eligible individuals who then out-migrated. In this case, observation of their survival is censored at the time of migration. Information about the death of visitors to the DSA is sometimes collected, but it is only used in mortality estimates if a de facto population estimate is available for each day.

Underreporting of deaths is typically less of a problem than that of births, because a death is widely known and remembered. Exceptions are the deaths of young (and yet unregistered) infants, particularly perinatal deaths, if cultural beliefs or grief hinders reporting.

Some DSSs collect more detailed information about deaths to establish the cause of death, generally through the so-called verbal autopsies (VAs).

Migrations and mobility

Two types of migration events occur:

• External migration — where residence changes between a residential unit in the DSA and one outside it; and
• Internal migration — where residence changes from one residential unit to another in the same DSA.

Where nonresident household members are ignored, only external migration affects the size of the population, resulting in either the registration of a new in-migrant or the termination of follow-up of an out-migrant. However, recording internal migration is very important to ensure the accuracy and validity of DSS data. The DSS needs to identify internal migrations and migrants and collect supporting information to avoid double counting of individuals and to ensure that their exposure to the social and physical environment is correctly apportioned. Migrations influence the registration of births and deaths; for example, a death would not be recorded for an individual who out-migrated before his or her death.

Defining the circumstances under which a migration is acknowledged to have occurred is notoriously difficult, not only for DSSs, but even for vital-registration systems and censuses. Different DSSs have different criteria. One approach, generally known as the “50% rule,” considers individuals resident if they have spent most of the time between two data-collection visits within the DSA. Any former resident who has not spent at least 50% of the time in the DSA would be recorded as having out-migrated.

However, many rural communities have individuals who regularly and predictably change residence for seasonal work, employment, or educational opportunities. The terms circular and pendular migration are often used. In the Hlabisa DSS, a newly established system in an area of very high population mobility, individual residency has been replaced with household residency as a registration criterion. Consequently, although out-migrations are recorded, the fieldworkers do not automatically terminate follow-up observations.

Migration is a repeatable event — an individual may make several migrations over time, both internally and externally. To maintain longitudinal integrity of data concerning individuals, a DSS should establish whether an external in-migrant has previously been registered in the DSS. The individual’s current and previous records should be matched so that he or she is not handled as a new individual in the system but as an individual under observation for several periods.

Episodes

Episodes are a logical complement to events. They are meaningful and identifiable segments of time started and ended by events. The life of an individual, for instance, can be understood as an episode that started with the individual’s birth and ended with his or her death. In the same way, residential units or households can be said to be episodes that start when they are formed and end when they are dissolved.

The usefulness of the concept of episodes is not limited to primary subjects. It applies equally to associations between them and therefore provides a useful framework for handling residency, membership, marital status, and many other concepts. Episodes also make it much easier to formulate and implement validation rules regarding events.

Other events

In addition to births, deaths, and migrations, other events are of interest for our understanding of demographic, health, and social dynamics. One event on which data are commonly collected relates to nuptiality or marital status. Most DSSs collect information about events such as marriage, defined as an event that starts a marital relationship, and divorce, that is, an event that ends a marital union. Other events recorded by DSSs depend on their complexity and research interests but may include the change of a head of household, a household’s formation or dissolution, or the construction or destruction of building structures.

Nuptiality and conjugal relationships

DSSs collect data on nuptiality primarily because of the important influence of marital patterns on fertility. Marriage as a start of an episode is easily identified, although a period of sexual union may have preceded marriage. The ending of a conjugal relationship can be less clearly marked, because it may not always be the death of one of the partners or a divorce, but a period of separation. In DSAs where the nonmarital fertility rate is high, other conjugal relationships become important, and the systems record informal relationships as well as formal marriages. However, in taking on this broader approach to sexual relationships, the DSSs must overcome two hurdles:

• The difficulty of establishing the starting and ending events of conjugal relationships that are not marked by official ceremonies; and
• The difficulty of establishing the link between two or more partners (in polygamous relationships, for example). For nonmarital conjugal relationships, where the partners often do not cohabit, greater efforts are needed to establish this link in a database than is the case for marital unions.

Construction and disintegration of residential units

At any given time, new residential units may be under construction and other residential units may be at various stages of disrepair following natural disasters or abandonment. The physical state may be distinct from the functionality of the residential unit; that is, it is possible that a residential unit is physically intact but long abandoned, and apparently broken-down units may still have households and individuals living in them. It is also possible that broken-down or destroyed units may subsequently be rebuilt, when the owner returns.

As the state of the residential unit is often — if not always — a good indication of its functionality, a DSS should make provision to track both its physical state and function.

Events occurring in households

Similarly, households can go through important changes affecting their composition and socioeconomic and health conditions. New households may form within an existing residential unit when, for example, a son takes a wife and establishes a family of his own or when a polygynous man takes another wife. Separate households may merge to form a new household, or a complete household may move to settle at another residential unit. Households may lose one or more members over time and decrease in size, or they may completely dissolve through a process of slow attrition or a major environmental or social disaster.

In environments with substantial social flux and instability, it is important to keep track of these events and their effects on the formation and dissolution of households. This is essential if DSSs have conceptualized households as subjects in their own right. Because they also influence patterns of individual presence at a residential unit, these household changes have important implications for the composition of the residential unit as a whole.

Document(s) 10 of 38