JA KIDS – Jamaica’s second national birth cohort study: aims and methods

ABSTRACT

Birth cohort studies across the world have yielded information that has been used to inform policy and programme decisions that have improved the health and well-being of populations. A few such studies have been conducted in low- and middle-income countries due to funding, methodological and other challenges. This paper briefly reviews the methods of comprehensive birth cohort studies with extensive follow-up of participants through the life course conducted in low- and middle-income countries. It then reviews the first Jamaican birth cohort study of 1986 and discusses the methodological advances in implementing JA KIDS, the second Jamaican birth cohort study conducted in 2011. The aims and methods of JA KIDS are described in detail.

KEYWORDS:

• Birth cohort studies
• low- and middle-income countries
• Jamaica

Introduction

A brief history of birth cohort studies

The British birth cohort study of 1946 was the first of its kind in the world and was conducted to investigate the reason behind falling birth rates (Wadsworth et al., 2006; NSHD1946). Data were collected on 13 687 births occurring in one week in Britain in March of 1946, at each child’s 8-week visit, representing 91% of total births. A follow-up study was subsequently designed to examine the health and progress of one-third of the babies (n = 5362). There have been 25 subsequent contacts with participants, in what is now known as the National Survey of Health and Development (NSHD. At the most recent contact in 2022, participants were 76 years old (NSHD 1946). These studies have guided numerous policy decisions for child and adult health. Birth cohort studies subsequently commenced in Europe (e.g. Northern Finland Birth Cohort of 1966; Danish National Birth Cohort (1997–2003)) and later in the U.S.A. (National Collaborative Perinatal Project of 1959–1966) (Batty et al., 2007).

Canova and Cantarutti identified birth cohort studies as the most appropriate study design to determine the causal relationship between potential biological, social and environmental risk factors during the prenatal or postnatal period and the health status of newborns, children and adulthood (Canova & Cantarutti, 2020). Golding stated that these studies provide more accurate contemporary measurements of exposures of study participants, resulting in information antedating recognition of adverse outcomes and the ability to determine causal mechanism. The addition of biological specimens allows for direct evidence of some exposures (Jones, 2009; Golding et al., 2009). Additionally, most common disorders are caused by a combination of genetic and environmental factors operating at critical points in time. Longitudinal studies allow for multiple factors to be investigated as antecedents, and for multiple outcomes of a single risk factor, such as poverty (Golding et al., 2009).

Disadvantages of high cost and long duration for birth cohort studies have also been identified. In comparison, cross-sectional studies, which provide less rich prevalence and association data, can be conducted in a far shorter time and at a relatively low cost (Golding et al, 2009). However, researchers have described the difference between cross-sectional and cohort studies as the difference between still photographs and a movie (Victora & Barros, 2012).

Birth cohort studies in low- and middle-income countries

Few low- and middle-income countries (LMICs) have conducted national birth cohort studies with extensive follow-up. High costs and methodological challenges not typically experienced in high income countries, such as majority births not occurring in hospital, high rates of internal and external migration, and absence of national databases to facilitate follow-up contacts, have been suggested as reasons preventing their execution (Batty et al., 2007). Transportation and terrain difficulties, as well as limited technical capacity to conduct birth cohort studies, are also considerations in LMIC. Findings that are available from birth cohort studies are important to LMICs, where the majority (84%) of both children and adults live according to the World Bank, and where high levels of childhood mortality and morbidity and chronic diseases in adulthood with origins in childhood persist exist. However, caution in extrapolating findings from birth cohort studies in high-income countries to those in LMICs has been suggested. First, results from high-income countries may not be applicable to LMICs, with differing socio-economic profiles, varying nature and frequencies of exposures and outcomes, and potentially differing confounding factors (Victora & Barros, 2012).

Among the earliest birth cohort studies in LMICs were those in South Delhi, India in 1969–72 (Bhargava et al., 2004), and Guatemala in 1969–77 (Martorell, 2020). The Indian study was focussed on perinatal health and child growth and the Guatemalan study on child nutrition and child growth and development. Later studies reassembled the cohorts and evaluated health and well-being in adulthood. Among LMICs, Brazil has had the longest history of conducting birth cohort studies, with 79 cohort studies commencing prenatally or at birth (Araujo et al., 2020). Of these, only six cohorts had a large number of baseline participants; a comprehensive approach to the health of children, adolescents and adults; and followed participants through different phases of the life cycle: the Ribeirão Preto cohorts of 1978 and 1994 and the São Luís cohort of 1997 and the Pelotas birth cohorts of 1982, 1993, and 2004 (Araujo et al., 2020). There has also been a 2015 Pelotas birth cohort study, following the 11 year cycle of implementation of studies in Brazil (Hallal et al., 2018). The Ribeirao Preto and Sao Luis studies saw children only once, between 7 and 10 years; the 1978 study also saw participants for adult contacts up to 37–39 years (Barbieri et al., 2022). The three studies allowed for comparative time trend analysis of the same city, as well as comparisons of two socio-demographically populations in different cities.

The 1982 Pelotas birth cohort study is considered one of the largest and longest-running birth cohorts in LMICs. Following enrolment of 5,914 children born in Pelotas, a relatively affluent city of Brazil, children were seen at 1,2,4,13,15, 18, 19, 23, and 30 years (Horta et al., 2015). The 1993 and 2004 studies utilised the same methodology of the 1982 study and were designed to allow for temporal analyses (Victora et al., 2008; Santos et al., 2014; Goncalves et al, 2018). Additionally, the later studies had more contacts in the first year of life. For example, the 1993 study saw participants at 1,3, and 6 months and 1,4, 6, 9, 11, 12–13, 15, 18 and 22 years. The 2015 study was unique in that data which were collected in pregnancy.

Other studies in LMICs with similar characteristics to the Brazilian studies included the Birth to Twenty cohort study in South Africa (Bt20+) and the Cebu Longitudinal Health and Nutrition Survey (CLHNS). The Bt20+ is Africa’s longest running longitudinal study and its only intergenerational study (Richter et al., 2007). The study enrolled 3273 singleton children born to women resident in Soweto-Johannesburg during a 7-week period in 1990. Some antenatal interviews were conducted in late 1989. There has been regular follow up of participants, as well as their offspring up to the third generation, over almost three decades.

The CLHNS commenced as an interdisciplinary study of infant-feeding patterns (Adair et al., 2011), but expanded to include a wide range of maternal and child health and demographic issues. Conducted over a one year period in 1983–1984, a sample of 3080 women were enrolled. Surveys were conducted in the last trimesters of pregnancy, after birth, every 2 months for 24 months, and when participants were 8, 11, and 15 years. Tracking surveys of participants were completed in 2002, 2005, 2007 and 2009.

The 1986 Jamaican birth cohort study

The first Jamaican birth cohort study began as the Jamaican Perinatal Mortality and Morbidity Survey (JPMMS). The aims of this study were to determine more accurately the perinatal mortality and morbidity rates and to identify the maternal, social and environmental factors predictive of fetal and early infant death (A. McCaw-Binns et al., 2011), in order to inform health policy and practice to reduce mortality rates (A. McCaw-Binns et al., 2011).

Patterned on the British birth cohort studies, 10 500 pregnant women and children were enrolled at delivery in the main study, representing 94% of all births in the country during the 3-month period 1 September to 31 October 1986 (A. M. McCaw-Binns et al., 1996). Morbidity and mortality components were also included, the details of which are described elsewhere (A. M. McCaw-Binns et al., 1996).

Similar to the British birth cohort studies, the JPMMS transitioned to the Jamaican Birth Cohort Study (A. McCaw-Binns et al., 2011). Researchers had intended to obtain data on children and families at important transition points in children’s lives: at the end of pre-school (6–7 years), primary (11–12 years) and secondary education (15–16 years). Inability to obtain funding prevented the 6–7-year follow-up. However, at 11–12 years and 15–16 years, factors impacting children’s cognition, educational attainment and behaviour were determined in an urban sub-sample (n = 1760; n = 1656). Limitation of funding precluded national samples at these contacts. The first contact in adulthood occurred when participants were 18–20 years and focussed on early determinants of chronic diseases (Ferguson et al., 2020). Adult participants were last seen in 2015–2017 when participants were 28–30 years old.

The JPMMS has produced six published technical reports, 11 books or book chapters, and 57 published papers to date. Additionally, the data have been used for capacity building; four PhD theses and four Masters have utilised the data. Probably the most important outcome is the use of the findings to guide evidence-based policy decisions. Four of the 30 papers and a number of the technical reports and book chapters report on policy implications identified by the study and include policy and programme recommendations such as strategies to reduce maternal and perinatal mortality, address violence prevention among youth, and improve children’s health and well-being (Ashley & McCaw-Binns, 1998; Samms-Vaughan, 2008; Ashley & McCaw-Binns, 2008; M-Caw Binns, 2009).

Origin of the JA KIDS study

Twenty-five years after the first, a second birth cohort study was considered necessary. Apart from collection of health data at birth, the JPMMS had not collected any data during the early childhood period, the importance of which had become more widely known in the two and a half decades since the first study. Even if the 6–7 year study had been implemented, our current knowledge is that the years prior to the age of 6 years are the most critical for brain development. As such, the overall aim of the JA KIDS study was to identify those factors that impacted children’s early health and development, with the intention to guide policy decisions to improve child outcomes. The first phase of the study was therefore designed to collect data multiple times within the first 2 years of life when children’s brains are known to be developing rapidly.

The study design and implementation was influenced by the knowledge and experience gained from cohort studies conducted between 1986 and 2011. this influenced study design. The importance of collecting data, on maternal health and well-being, including biological samples, in the antenatal period had been documented. Biological samples were important to measure environmental exposures that would not otherwise be detected or quantified, validate exposure data collected in other ways; and measure outcome traits or intermediate phenotypes (Jones, 2009). The JA KIDS study design included collection of data and biological samples in the antenatal period.

Early birth cohort studies, such as the JPMMS, had focussed primarily on maternal health and well- being at birth; fathers were never interviewed. At subsequent contacts, data were primarily obtained from parents on children’s well-being with limited attention to maternal and paternal well-being, and their parenting and relationship styles. The JA KIDS study design included collection of data directly from mothers and fathers about their physical and mental well-being. In particular, fathers were invited to attend and participate in interviews at all in-person contacts. There had also been technological advances in the collection and management of data. Computer-assisted telephone interviewing was reported to be the mechanism of choice for public health surveys based on cost-effectiveness and interviewing and data entry efficiencies, despite there being some methodological concerns of bias towards those with telephones and those at home and the absence of paper-based data to address typing errors (Choi, 2004).

Materials and methods

Study phases

The JA KIDS study was conducted in two phases. Phase 1 saw the establishment of the birth cohort and collection of data from families and children in the first two years of life. Phase 1 aimed to determine the impact of the prenatal, socio-emotional, physical environment and genetic factors on children’s health growth and development to collect data on young children that would form the basis for a longitudinal study to determine the impact of the varied early years environments on long-term child and adolescent development; and using the information obtained to make evidence-based recommendations to the Government of Jamaica to inform early childhood policies and programmes and thereby improve the outcomes of children. Phase 2, conducted when children were 4 years old, had a focus on school readiness and aimed to determine the association between the quality of the various early childhood environments prior to pre-school entry and child development status and school readiness, in order to provide policy recommendations on improving the quality of early childhood development to the government of Jamaica.

Target population

The JA KIDS study participants include children born in all fourteen parishes across Jamaica from 1 July to 30 September 2011, and their parents or primary caregivers.

Exclusions

The JA KIDS cohort excludes any child born outside of Jamaica and outside of the cohort period even if parents were pre-enrolled during the antenatal period.

Pre-enrolment

Women with estimated dates of delivery between 1 July and 30 September 2011 were eligible for pre-enrolment. JA KIDS field interviewers recruited mothers from all antenatal clinics in the public health care system and from private antenatal care clinics operated by obstetricians and general practitioners. In the public health system, clinic registers were reviewed and eligible women were identified and invited to participate. Enrolment in the private sector was more challenging due to a large number and variety of antenatal service care providers (general practitioners, obstetricians), an absence of a comprehensive listing of such providers, and an absence of incentive for participation.

Enrolment at birth

Trained field research officers were stationed at all twenty-six (26) public and private hospitals. Information on eligible mothers who delivered in the previous 24 hours, as well as babies who were unwell or who had died were obtained from delivery registers. Mothers were then approached, sensitized about the study, consented and interviewed. Consent at enrolment included consent for further contacts, but also included withdrawal from the study at any time without any consequences. Recruitment of fathers was opportunistic; all fathers who visited the hospital were invited to participate and be interviewed.

Follow-up contacts

There have been three main contacts subsequent to the antenatal and birth contacts to date (Figure 1). When study children were 9–12 months old, attempts were made to contact all study participants by telephone; a structured questionnaire was completed by telephone interview. At 18–22 months of age, there were two types of contacts. A 10% random sample of families (mothers, fathers and children) was invited for face-to-face interviews and developmental assessments of their children at one of six locations over a 5-month period, and a 25% random sample was targeted for telephone interviews. At 48–54 months, a targeted sample that included families who participated in three or four previous contacts was invited to participate through in-person interviews and developmental assessments of their children.

Figure 1. Flow Chart of JA KIDS Participants.

Data collection procedures

Questionnaires

At each contact with families, questionnaires were completed that provided socio-demographic data, age-specific information on children’s health development, well-being and parent–child interactions, as well as information on parenting, parental health and well-being, family functioning and social support systems. Questionnaires completed by mothers in the antenatal period and at delivery also included information on past and current reproductive and pregnancy health and environmental exposures. Questionnaires completed by fathers at delivery and at the 18–22 month face-to-face contact additionally enquired of paternal perspectives on pregnancy, parenting and child rearing. Data on neonatal admissions were extracted from hospital records throughout the country and neonatal admission questionnaires completed.

Standardised assessments

At the 18–22 month contact, parents completed standardised developmental screening questionnaires for their children, and children who attended in person additionally completed diagnostic developmental and behavioural assessments using standardised instruments. Parents completed literacy and verbal comprehension tests, and anthropometric data (weight and height) were completed for parents and children. Head circumference measurements were also obtained for children.

At the 4–5 year old contact, children completed standardised developmental, behavioural academic, verbal comprehension and executive function tests. Weight and height measurements were also obtained.

Biological samples

Blood samples were obtained from primarily from mothers who consented in the antenatal period; a few samples from mothers who consented but who were missed were collected in the immediate postpartum period. Umbilical cord blood samples were obtained at birth from children of parents who consented and whose delivery took place at the University Hospital of the West Indies, located adjacent to the research site.

Research partnerships and sub-studies

Research partnerships were critical to study implementation. Some studies that were not part of the original study were easily added, many because of existing research partnerships. Questions on sexuality of Jamaican fathers were incorporated in the fatherhood questionnaire at birth and salivary samples for testosterone were obtained from a small sample of fathers for analysis through a partnership with the University of Las Vegas, Nevada (Gray et al., 2015, 2017). Parenting Partners Caribbean provided funding for additional information to be obtained on the social supports available to parents of cohort children that were admitted to either a Neonatal Intensive Care Unit or Special Care Nursery. The main funders of the study, the Inter-American Development Bank provided an additional grant to conduct follow-up research with the firstborn children in each parish using a variety of observational methods, developmental tasks and interviews conducted in children’s homes (Firstborns Project). These mixed data collection methods were used to enhance study findings and provide a richer perspective on the lives and experiences of Jamaican children and their families. In addition, the Firstborns Project provided valuable information on the physical, socio-emotional and cognitive environments of the children. A short documentary series was produced from this research. During the early months after the COVID-19 pandemic, a grant from the World Bank allowed for the impact of the pandemic on children and families to be assessed, utilising existing contacts with families from the 4–5 year old study contact (Samms-Vaughan et al., In Press).

Partnerships also facilitated collection, storage and analysis of biological samples. Antenatal blood samples were analysed through a partnership with the University of Michigan (Manandhar et al; In Press (this journal); cord blood samples were analysed for heavy metals through a partnership with the University of Texas at Houston (Rahbar et al., 2015). Some partnerships, such as those with the World Bank, UNICEF, and Jamaican government agencies, such as the CHASE Fund and the National Health Fund facilitated data analysis and production of specific technical reports.

Steering committee

The JA KIDS study was guided by an inter-institutional Steering Committee co-chaired by the University of the West Indies and the Planning Institute of Jamaica, and included representatives from the Jamaican Ministries of Education, Health and Labour and Social Security, the Early Childhood Commission, the Statistical Institute of Jamaica, the National Environment and Planning Agency and the National Parenting Support Commission. The Steering Committee provided technical and logistical support to the study and was structured to ensure that policy implications were identified early and transmitted to the relevant Government of Jamaica Ministries and agencies to impact national development goals, programmes and plans.

Results

Jamaica’s Registrar General’s Department advised that in 2011, there were 39 673 births with 11 124 births during the July–September cohort period. In total, 9,742 mothers and 3,410 fathers were enrolled in the study at delivery, representing 87.6% of mothers who delivered during this period and 30.7% of fathers. Five per cent of mothers refused to participate while 7.5% of mothers were not interviewed, either because mothers were discharged shortly after delivery or were too ill to participate.

Of the 5,204 mothers who were pre-enrolled in the antenatal period, 4576 (87.9%) were matched to mothers delivering during the cohort period. Reasons for mothers not matched to those delivering in the cohort period included the mother delivering outside of the cohort period due to pre-term or post-term birth, the mother delivering at a non-participating birthing centre (one was identified during the study), miscarriage, abortion, maternal death or the mother was missed or refused participation in the study.

Figure 1 shows that the 9–12 month contact included 7609 participants, representing 78.1% of the JA KIDS enrolled and 68.4% of the total births. At 18–20 months, a selected 3,485 participants were seeing, representing a targeted 35.8% of the JA KIDS population. At 4–5 years, there were 1311 participants, representing 69% of the targeted group of participants who had attended most or all previous contacts.

Not shown in Figure 1 is that 1186 maternal blood samples and 144 cord blood samples were obtained.

Prior socio-demographic analysis of the cohort sub-samples (Reece et al, In Press, this journal) has indicated that apart from the sample of women enrolled in pregnancy that were biased towards the public sector, the other samples were similar in maternal age, and socio-economic status as measured by a composite indicator that included maternal education, but differed by educational level as a single variable.

Discussion

A second national birth cohort study was able to be successfully implemented in Jamaica, an LMIC, in 2011, twenty-five years after the first study. Apart from Brazil where birth cohort studies have been conducted with regularity every 11 years in recent times, Jamaica is, to the authors’ knowledge, the only other LMIC that has conducted more than one birth cohort study. Existence of a prior study was important as the JA KIDS study design and implementation benefitted from the experience gained during the execution of the first birth cohort study, in addition to advances in research in birth cohort studies across the world since 1986. Collection of data in pregnancy, frequent contacts during the critical early childhood period, including fathers as important providers of information, collecting biological samples and collecting data using computer assisted telephone interviewing (CATI) were all advances when compared with the 1986 study. CATI was particularly useful in reducing study costs occasioned by transportation, and facilitating parent participation without long periods of travel to study sites.

In comparison to the other studies in LMICs, particularly the later studies in Pelotas, the JA KIDS study similarly collected data in the antenatal period and had more frequent contacts in the early childhood period. However, the JA KIDS study collected biological samples in the antenatal period and at birth as well as at other contacts. There was also a specific focus on collecting data from fathers’ perspectives. This was culturally important as the public perception of fatherhood was often negative, and fathers had not had the opportunity for their voices to be heard.

Apart from providing critical information in the early childhood period, the existence of a birth cohort study framework was beneficial in other ways. First, there was a capacity to study other areas of child health and development in depth at relatively low cost, by incorporating a number of sub-studies. Second, population-based data that were required at short notice, such as that on the impact of the COVID-19 pandemic on infants and children, were able to be obtained in a relatively short time, because of the existence of the cohort study framework.

Having a second national birth cohort, twenty-five years after the first, allows for analysis of temporal data, and potentially the impact of policy and programme interventions implemented in the time period between the studies. However, temporal changes also impacted study implementation. For example, the period of enrolment was 3 months for JA KIDS compared to two months for the JPMMS, in order to obtain a reasonably similar sample size (10,500 vs 9742). This is reflective of a fall in Jamaica’s birth rate between 1986 and 2011.

A main strength of this study, similar to the 1986 study, is its national reach across all fourteen parishes of Jamaica, thereby reducing the likelihood of bias, particularly geographic, at enrolment. Another strength is the large sample size; among the long running birth cohorts in LMICs, the Jamaican birth cohort studies have had the largest enrolment at birth. However, this also presents a challenge. Limited funding precluded subsequent follow up of the entire sample for the previous birth cohort study and of the early studies of this cohort.

There were a few limitations in establishing the birth cohort. Antenatal data collection in the private sector was challenging, as was mentioned earlier, because of large number of providers and more limited collation of information on pregnant women in this sector when compared with the public sector. Rapid discharge of healthy mothers and babies meant that some families were missed. The refusal rate was higher in the 2011 study than in the 1986 study in which there was data collection for 94% of all births.

Researchers have argued that although birth cohort studies in LMICs are smaller in size, and less mature, they offer similarly detailed and valuable data (Batty et al., 2007). This data would be relevant not only to the country of study, in this case Jamaica, but potentially to other socio-demographically and culturally similar countries, such as Caribbean and African countries. The expertise gained from the conduct of birth cohort studies in LMICs with similar methodological challenges can be used to facilitate and empower researchers from similar countries to conduct their own birth cohort studies and/or to critically evaluate the relevance of the findings for their own countries. Either of these options would contribute to capacity building for birth cohort studies in LMICs.

Like other similar cohort studies around the world, a rich database from which to analyse causal factors affecting children’s early development and later adult outcomes has been established. Other birth cohort studies, including the 1986 Jamaican Birth Cohort Study, have in later years been used to investigate the early origins of adult physical, social and behaviour strengths and disorders. Many of the long running studies in LMICs, such as the South Africa and Pelotas studies, have also been evaluating inter-generational trends.

The 1986 Jamaican birth cohort study has demonstrated its usefulness in providing information to guide policy development. The 2011 study also has huge potential for critical policy development in early and later childhood and adulthood. The implementation of the 2011 study benefitted from a number of partnerships that provided additional funding to that provided by the main funding agency, but limited funding still impacted implementation. As in other birth cohort studies, extension of the JA KIDS study to include further contacts with participants and further data analysis to achieve this potential will be dependent on the recognition of the importance of birth cohort studies for this purpose and the subsequent availability of adequate funding.

Ethical approval

Ethical approval was granted by the Ethics Committee for the Ministry of Health – Jamaica (approval #198; 2011) and the University of the West Indies (approval # ECP 122 10/11). All participants provided written informed consent.

Acknowledgment

We are grateful to all the families who took part in this study, the staff in health centres and hospitals throughout Jamaica for their help during recruitment, and the JA KIDS team, of interviewers, computer and laboratory technicians, clerical and administrative workers, research scientists, volunteers and managers at the University of the West Indies (Mona). The Inter-American Development Bank (Grant ref: ATN/JF-12312-JA; ATN/OC-14535-JA) and the University of the West Indies, Mona Campus provided core support for JA KIDS. Secondary data analysis at the Centre for Maternal and Newborn Health in the UK was funded through a Global Health Grant (Project number OPP1033805) from Bill and Melinda Gates Foundation and WHO. Additional support was provided by the World Bank, UNICEF, the CHASE Fund, the National Health Fund, Parenting Partners Caribbean, the University of Nevada – Las Vegas, the University of Texas Health Science Centre at Houston and Michigan State University and its partners. This publication is the work of the authors listed, who will serve as guarantors for the contents of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Inter-American Development Bank [ATN/JF-12312-JAATN/OC-14535-JA].