VICTORA, C. G. et al.
Cad. Saúde Pública, Rio de Janeiro, 19(5):1241-1256, set-out, 2003
Birth cohort studies from developed countrieshave contributed enormously to the under-standing of the long-term consequences of con-ditions in early life. The first prospective na-tional birth cohort study was started in theUnited Kingdom in 1946 (Wadsworth, 1991),and new cohorts were started in 1958, 1970, and2000 (Smith & Joshi, 2002). In the United States,a large birth cohort will be started in 2004 (seehttp://www.nichd.nih.gov/despr/cohort).However, large birth cohort studies are rareoutside developed countries. Harpham et al.(in press) have recently reviewed longitudinalstudies from developing countries with samplesof 1,000 or more children. Of these, only threebirth cohorts were followed for 10 years ormore. In chronological order, these include the1982 Pelotas (Brazil) cohort of approximately 6,000 births; the 1983-1984 Cebu (Philippines)study with approximately 3,000 newborns(CHLNS, 2002); and the 1990 Johannesburg (South Africa)
Birth to Ten Study
of some 4,000births (Richter et al., 1995). With the exceptionof the Cebu study – terminated in 1999 – theother two cohorts are still being followed. A smaller study is being conducted in Guatema-la, where about 400 subjects born from 1969 to1977 are still being followed (Stein et al., 2002).In this paper we provide an update on the
Pelotas and Birth Cohort Study
, with emphasison the design and methods used in recentphases of the study. The importance of birthcohort studies and their main methodologicalchallenges are highlighted. Special attention isgiven to the issue of losses to follow-up, possi-ble strategies for improving response rates incohort studies, and integration of longitudinalepidemiological and ethnographic methods.
The study was carried out in the Brazilian city of Pelotas, Rio Grande do Sul State. The urbanpopulation grew from 214,000 in 1982 to 320,000in 2000. This is a relatively affluent area of Brazil, near the Southern border with Uruguay and Argentina, with a per capita gross domes-tic product of US$ 2,700 (Klering, 1992) in the1990s. The population descends mostly fromPortuguese and Spanish settlers, Amerindians,and Africans brought as slaves; German immi-grants are also represented. Inter-racial mar-riages are common. The main economic activi-ties are agriculture (mostly rice farming andcattle-raising), commerce, and education (thecity is home to two large universities). Over90% of the households are connected to thepublic water supply and half to the sewage dis-posal network. When the study began, the in-fant mortality rate in the city was some 40deaths per thousand live births. There is onephysician for each 260 inhabitants, health carebeing provided free of charge at 32 basic healthfacilities, five specialized health centers, andfive general hospitals.
Early phases of the study: 1982-1986
The methods used in the early phases of thecohort have been described elsewhere (Barroset al., 1990, 2001; Victora et al., 1992). The study began as a perinatal health survey (Barros,1986) including all 6,011 infants born in threematernity hospitals (accounting for 99.2% of allbirths in the city). The 5,914 live-born infants were weighed with regularly calibrated pedi-atric scales (Filizolla, Brazil) to the nearest 10g.Birth length was not recorded. Mothers were weighed and measured and answered a shortquestionnaire on socioeconomic, demograph-ic, and health-related variables.The cohort children were followed up atseveral points in time(Table 1). Initially, thoseborn from January to April 1982 were targetedin the
1983 Follow-up Study
through the ad-dresses obtained during the hospital interview, when they were aged 8-16 months (mean age11.3 months).To calculate the proportion of children lo-cated in each follow-up visit, those known tohave died were added to those examined. In1983, 66 of the 1,919 children born from Janu-ary to April 1982 were know to have died; addedto the 1,457 whose mothers or caretakers wereinterviewed, these accounted for 79.3% of thechildren who were targeted. Address errors were the main reason for non-response. Non-response rates quoted in the present paper areslightly different from those presented in earli-er publications (Barros et al., 1990), since they did not fully account for children who haddied.The subsequent phases included the
(January-April 1984), whenthe mean age was 19.4 months (range 12-29),and the
1986 Follow-up Study
(December 1985-May 1986) (mean age 43.1 months; range 35-53). To minimize losses to follow-up, in eachround the approximately 70,000 urban house-holds were visited in search of children born in1982. After the census was completed, children who still had not been located were searchedfor at their last known address. This approach