AGINCOURT DEMOGRAPHIC SURVEILLANCE SYSTEM

SOUTH AFRICA

University of the Witwatersrand

Northern Province Department of Health

 

Mark Collinson, Obed Mokoena,  Niko Mgiba, Kathleen Kahn, Stephen Tollman,

Michel Garenne, Kobus Herbst, Elizabeth Malomane and Sheona Shackleton

1.                AGINCOURT DSS SITE DESCRIPTION

 

1.1                      Physical Geography of the DSS Area

 

The Agincourt DSS field-site is situated about 500 kms north-east of Johannesburg in the Agincourt sub-district of Bushbuckridge region, of the Northern Province. Until 1994 the site was located within a ‘homeland’ or bantustan area. The site extends from 24o50’ to 24o56’ south latitude and 31o08’ to 31o25’ east longitude. The altitude range is 400 to 600 metres above mean sea level. The field site with its twenty-one village communities covers 390 sq km, measuring 38 km by 16 km at its widest points.

 

The geo-ecological zone is semi-arid savanna, better suited to game farming and low-density cattle farming than to crop cultivation. Low average rainfall is common, with mean rainfall ranging from 700 mm in the western part of the site to 550 mm in the eastern part. In addition there is a high variability in inter-seasonal rainfall patterns that render the area vulnerable to drought. Some eighty percent of rain falls during the summer months of November to March. The area is affected by drought on average every three and a half years, and even when rainfall is normal it is insufficient to fully supply the domestic or irrigation water needs of the area. The area experiences hot summer and mild winter months. The temperature range is 12-40°C in summer and 5-27° C in winter.

 

1.2     Population Characteristics of the DSS site

           

The total surveillance population is 66,840 people living in 10,500 households, with a population density of 172 persons per square kilometer. The male:female sex ratio for the total population is 0.929. The setting is rural in terms of distance from urban centres and lack of infrastructure. The main ethnic group is Shangaan, although Mozambicans, originally refugees, comprise more than a quarter (29%) of the total population. Both groups are Shangaan-speaking and the Mozambicans are culturally affiliated to the South African host population.  There are mainstream Christian churches, independent African churches and an amalgamation of traditional and Christian beliefs is often practiced. 

 

Unemployment is estimated at 40 – 50%. Formal sector employment involves migrant men who work on the mines and in the manufacturing and service industries of larger towns as well as on nearby game farms, commercial farms and timber plantations. Women make up an increasing proportion of the migrant labour population. Another source of local employment is the public sector. Informal sector activities are widespread, and include food and fruit vending. Pensions are an important source of income for many families. Female headed households constitute 32% of all households (Tollman et al 1995).

 

Almost all villages have at least one primary school and fourteen of the twenty-one villages have a secondary school. Over 40% of adults 25-59 years have received no formal schooling. Six percent have completed secondary school and only 3% have proceeded to some form of post-secondary education. Of those aged 15 –24 years almost all have attended primary school but only 46% have made the transition to secondary school. Although 85% of 10-14 year olds were enrolled in primary school, age of enrollment is frequently delayed.  Adult female literacy (56%) is somewhat lower than adult male literacy (62%) (Tollman et al 1999). 

 

Various types of housing are found, ranging from traditional mud huts to brick dwellings with tin or tiled roofs. Stands are generally too small to support subsistence agriculture and crops grown merely supplement the family diet. Water is pumped, via purification plants in some cases only, to the main reservoirs in the villages. From the reservoirs it is reticulated to the communal taps which serve as the main collection points for village community members. Water is collected manually by women or children usually in 25 litre drums and transported either by wheelbarrow or carried on the head. Water shortage poses a serious problem in most villages. Levels of household sanitation are poor, and pit toilets of varying effectiveness are the norm. All roads are unpaved. Public transport is limited to privately owned mini-bus taxis. Electricity and telephone services, though seriously lacking, are benefiting from recent development initiatives.

 

A health centre with five satellite clinics exist in the field-site, all staffed by nurses. A restricted number of drugs are dispensed from each of these primary care facilities, and the health centre has a small laboratory, able to perform a limited number of diagnostic tests.  An ambulance is based at the health centre. All services are free, and include child health, family planning, ante-natal care, delivery and post-partum care, minor ailments and chronic disease treatments. Although waiting times are long, most of these services are under-utilised. A contributing factor is poor drug supply. Referrals are to two district hospitals, each about twenty-five kilometers from the health centre. The main health problems revealed by verbal autopsy analysis are diarrhoea, kwashiorkor and AIDS in children under five; accidents, violence and AIDS in the 15-49 age group; and chronic degenerative diseases, mainly cardiac, cerebro-vascular, liver and malignant diseases, among those fifty and above (Kahn et al 1999; Tollman et al 1999; Garenne et al 2000).  Seasonal malaria is evident. A high rate of adolescent fertility exists in the midst of escalating HIV sero-prevalence (Garenne et al 2000).

 

Forced relocation of communities under the apartheid regime, in the 1940-1960 period, followed by the formation of ethnically divided homelands in the 1970’s, has had a significant impact on the social, economic and demographic profile of the population. The government of the time sought to massively exploit the labour potential of rural South Africa, while simultaneously restricting development. Homeland governments were given hegemonic powers with dubious results. Homeland health services, although poorly organised and managed, rested on a network of mission hospitals and clinics. They were thus less fragmented than their urban counterparts, leading to a more comfortable fit with South Africa’s new decentralized, district-based health system. Densely settled rural villages, with cash-based economies were the resultant pattern of settlement, with males aged 20 and 59 years largely absent from the permanent population. Recent changes in government have affected movement patterns. With more freedom of movement people are tending to move to rural towns. These towns are becoming development nodes along paved roads through these rural areas.

 

2.                AGINCOURT DSS PROCEDURES

2.1       Introduction to the Agincourt DSS Site

 

The original objectives of the Agincourt study were (Tollman 1999):

 

a.       To provide essential information on the demography, health status and fertility status of the Agincourt community as a basis for the improved formulation, implementation and assessment of district-level programmes;

 

b.       To serve as a sentinel field site providing accurate information on the population dynamics of rural communities in South Africa to inform the evolution of rural health and development policy;

 

 

c.       To provide the capacity and a database to support more advanced community based studies and field trials in the future.

 

The current primary objective relates directly to c. above, namely to provide a research infrastructure and longitudinal database for a range of community-based studies relating to burden of disease, health systems interventions and social/household/community dynamics. This to inform decentralized health and social policy.

 

The Agincourt baseline census was conducted in 1992. The original population under surveillance was 57,509 persons in 8,896 households. By 1999 this had increased to 66,840 in 10,500 households. Verbal autopsies and maternity histories were introduced in 1993. A partnership between the Agincourt DSS, the study communities and the local health services was established and is carefully nurtured (Tollman 1995).

 

The Agincourt DSS data is updated every 12 months. Residents are defined as either ‘permanent’ (resident in the study site for six months or more in the preceding year) or ‘migrant’ (resident in the study area for less than six months, but nevertheless regarding the Agincourt area as “home”). A verbal autopsy is conducted on all deaths by a trained lay fieldworker in the vernacular, and assessed by medical practitioners (Kahn et al in press).  The software system consists of a relational database constructed in Microsoft Access 2000. The main demographic, health and socio-economic variables measured routinely by the DSS include: births, deaths, migrations in- and out-, household relationships, resident status, refugee status, education, antenatal and delivery health-seeking practices.  In the 1999 DSS update, information on chronic cough was collected for a study on active case-finding for tuberculosis.  In  2000, information on labour force participation in both the formal and informal sectors is being collected; during 2001 a module describing the burden of disabilities is planned.

 

The Agincourt DSS is the foundation for the Agincourt Health and Population Programme (AHPP), a research initiative of the University of the Witwatersrand. It is housed within the Health Systems Development Unit of the Faculty of Health Sciences. The AHPP has strong ties to the Northern Province Department of Health and the Bushbuckridge Regional and District Health Services.  The core management team consists of the AHPP leader, senior researcher, field research manager and site manager. The DSS field team consists of four supervisors, twenty fieldworkers, one VA supervisor and four VA fieldworkers, all employed on a contract basis for the duration of data collection. In 2000 a part-time data-form checker has been employed. The data-capture team consists of a supervisor and two data typists.

 

Work underway within the AHPP addresses a portfolio of clinical, public health, population and social challenges including: (a) changing patterns of mortality, fertility and migration; (b) increasing circulatory disease and interpersonal violence among adults; (c) estimating the burden of respiratory syncytial virus, and investigating persisting kwashiorkor among children under-5; (d) links between the health and population status of former refugees (Mozambicans) and their livelihood strategies; (e) household and community dynamics, in particular resource flows, adaptive strategies, and the consequences of aging and adult death on the household; (f) labour migration and its impact on men’s sexual behaviour and use of contraception; and (g) evaluation of public service and adolescent-oriented  interventions against HIV/AIDS and TB.

 

Community feedback and dialogue are integral to the Agincourt research process. Information from the DSS and related research initiatives is communicated to the study communities through printed “village fact sheets” and by ad hoc community meetings. This facilitates community involvement in local health action and related development activities. Information is regularly discussed with district and regional health service managers, as well as senior officials of the Northern Province and National Department of Health.

 

2.2            Agincourt DSS Data Collection and processing

 

Several factors influenced the choice of the Agincourt field site, in particular its location, some distance from any tar road or township settlement; the presence of a health centre with satellite clinics and unrealised potential to function as a referral network; the need to develop rational referral patterns, de-linked from constraints imposed by homeland boundaries; and the presence of large numbers of Mozambicans displaced by the recent civil war.

 

2.2.1        Field Procedures

 

a)      Mapping

 

Hand-drawn maps of each village were made for the initial census in 1992. These included roads, dwellings and other reference landmarks such as railway lines, power-lines, shops, churches and soccer fields. Since then, village maps are updated each year by both specific fieldwork exercises as well as by routine correction and updating of maps during census fieldwork. The maps make it possible for any member of the team to return to a particular household without risk of ambiguity

 

b)      Initial Census and regular update rounds

 

Five census rounds have been completed to date, (1992 ‘Baseline’, 1993/1994, 1995, 1997, 1999), with a sixth round underway in 2000. Rounds are conducted in the dry season, viz. July to November. A fieldworker interviews the best respondent available at the time of visit. Individual information is checked for every household member. All events are recorded that have occurred since the previous census and any status observations updated. Where possible questions are directed to particular household members, for example maternity history or pregnancy outcome information is asked directly from the woman involved. If appropriate respondents are not available the fieldworkers undertake revisits, usually during evenings and on weekends, with a limit placed at two revisits per household.

 

In 1999, with the aim of increasing the speed of data collection, the field team was expanded to four supervisors, twenty fieldworkers, one VA supervisor and four VA fieldworkers. The team operates out of five field-offices, provided by clinics or community members at no charge.   

 

c)      Continuous surveillance of events

 

The demographic surveillance system involves a continuous recording of vital events in the population namely births, deaths, in and out-migrations. From the second round onwards, a verbal autopsy (VA) has been conducted on every death to determine the most probable cause. VA’s are conducted concurrently with the census, but with a separate team of fieldworkers dedicated to VA interviews only.  

 

d)      Supervision and Quality Control

 

To ensure data quality supervised visits and random duplicate visits are conducted. For supervised visits, the supervisor goes into the field with the fieldworker and observes some interviews. After each interview constructive feedback is given to the fieldworker with the aim of improving interview technique. Random duplicate visits are conducted by the supervisor on 2% of the population. After a careful explanation is given the whole interview conducted again, differences between the first and second interviews identified and possible reasons for these determined. From this data quality can be assessed and error rates computed.

 

Furthermore, form checking occurs in a structured system at four levels of the field organisation. The checks become more detailed as the form progresses through the system. An error is returned to the fieldworker for correction and where necessary a revisit is done. Supervisors keep track of forms using printed checklists.

 

2.2.2        Data Management and Analysis

 

a)      Data handling

 

Existing details of each household are printed onto the census forms. The fieldworker checks this information and, in addition, status fields are updated for each household member. Separate event forms exist for pregnancy outcomes, deaths, migrations and maternity histories and are only completed if one of these events occurred in the inter-censal period. Death forms are completed in duplicate so that one copy can be passed on to the verbal autopsy team. The set of forms resulting from each interview is stapled together in a predetermined order.  Supervision checklist forms allow for the monitoring of data collection for each dwelling.

 

b)      Data processing

 

When a form has left the field and passed all quality checks it is captured onto a software system. Currently, data is captured using simultaneous data entry with three computers on a network writing to a database on a server. The software system is a relational database currently held in Microsoft Access 2000. A custom-made data entry programme has been developed which sits on top of the Access database and, by mirroring the format of the data forms, provides an easy-to-use interface between the user and the database.  The database consists of related tables which store different aspects of the data. The main table is the ‘Individual’ table which stores key information on all individuals encountered. The ‘Residence’ table provides information on individual residence episodes, indicating how and when a person entered and exited a particular location in the field site. A ‘Memberships’ table records information on how and when an individual entered and exited a particular household  (i.e. a social grouping defined as people ‘eating from the same pot’). A table exists for each event category, viz. ‘Births’, ‘Deaths’, ‘Migrations’ and ‘Maternity histories” as well as an ‘Observations’ table, which records information about each interview. In addition, a range of status observation tables record information about individuals that are updated at different frequencies during census rounds. These include ‘Residence Status’ (updated in rounds 1 through 6), ‘Education Status’ (rounds 1 and 4), ‘Cough status’ (round 5) and ‘Labour status’ (round 6).  

 

c)      Data Quality Assurance and links back to field operations

 

The software system incorporates built-in validation checks. Data that are not plausible. (e.g. a date of death occurring before a date of birth) are prevented from entering the database. When these errors occur the form is put to one side, reviewed by the data manager and when necessary returned to the relevant team supervisor for resolution. Data that are unusual, though possible, (eg. a delivery to a woman older than 50) are also flagged by the system and reviewed by the data manager. As data are entered computer checks are done to look for invalid codes, missing values, consistency within records, consistency between records, correct spelling of place names and duplicate entries. A useful data quality check after a census is a comparison of the village of origin of internal in-migrants with the village of destination of internal out-migrants, as well as a review of demographic trends. 

 

d)      Data analysis and dissemination

 

Basic analyses are done to produce village fact sheets, community feedback information, sampling frames and denominator information. Further data cleaning and demographic analyses are conducted to produce reliable population information. A monograph of the ‘baseline’ findings was produced in 1994 and all findings of specific scientific or policy interest are published in local and international peer-reviewed journals. Presentations are made to policy-makers at sub-district, district, regional, provincial and national levels. 

 

3.     AGINCOURT DSS BASIC OUTPUTS

 

3.1 Demographic Indicators generated by the Agincourt DSS Site

 

The total (de jure) population was 66,840 in 1999. Of these, the permanent population (resident in the site for more than six months of the preceding year) was 56,566. The sex ratio (male/female) in the total population was 0.929, falling to 0.712 in the permanent population aged 15-49 years. The age structure in the total population at end-1999 was as follows: 2.3% aged < 1year, 12.0% aged 0-4 years, 27.6% aged 5-14 years, 55.9% aged 15-64years, and 4.5% age 65+ years. The total fertility rate was 2.72; the percentage of female-headed households was 32%; the dependency ratio was 0.79; the infant mortality rate was 43.0/1000 among males and 45.1/1000 among girls; mean household size was 6.4; the adult literacy rate* in females was 56% and in males 62%.

 

Migration surveillance is conducted by recording data on individuals moving into or out of a household during the period between census rounds. Moves are classified as internal if they have their origin and destination within the field-site villages, otherwise they are classified as external moves. The place of origin or destination, date of the move, and reason for its occurrence, are captured for each move.

 

3.2 Demographic trends in the Agincourt DSS data

There has been a profound decline in total fertility rate (TFR) from around 6 births per woman in 1970-74 to 2.72 in 1999 (Garenne at al 2000). Mortality had been declining for some time until about 1993. Since 1994 an increase in mortality in three age groups has been documented: young adults aged 20-49 years (both sexes), children aged 0-4 years (both sexes) and older adult woman aged 50-64 (male mortality continues to decline in this age group). Migration trends show a net population loss due to an excess of external out-migration over external in-migration (about 1 % of the population per year). The main focus of departures has been nearby towns, in particular Mkhuhlu. This move towards Mkhuhlu was particularly strong in the 1994-1995 period (Collinson et al 2000).

 

 
 

 

 

 

 

 

 

 

 

 

 

 

Figure 1: Population pramid for the Agincourt DSS Site

Table 1.  Age and Sex Specific Mortality in the Agincourt DSS Site

 

 

 

Age

 

 

 

 

 

 

 

Deaths (nDx)

 

Person Years (NPYx)

 

 

 

 

 

 

 

Male

Female

 

Male

Female

 

 

 

 

 

 

 

0

 

59

65

 

3,877

3,866

1-4

 

76

75

 

17,147

17,093

5-9

 

14

19

 

23,175

23,002

10-14

 

14

12

 

20,119

19,943

15-19

 

16

18

 

17,741

17,494

20-24

 

32

30

 

14,014

15,098

25-29

 

46

30

 

11,122

12,356

30-34

 

53

36

 

9,027

10,365

35-39

 

61

37

 

7,198

8,572

40-44

 

44

36

 

5,634

7,025

45-49

 

69

20

 

4,559

5,111

50-54

 

48

22

 

3,322

3,572

55-59

 

56

19

 

2,697

3,285

60-64

 

41

48

 

1,980

3,132

65-69

 

56

63

 

1,733

3,351

70-74

 

58

66

 

1,352

2,086

75-79

 

70

65

 

1,021

1,583

80-84

 

29

33

 

415

507

85+

 

27

40

 

292

479