Journal of the New Zealand Medical Association, 27-January-2006, Vol 119 No 1228
Pacific Islands Families: First Two Years of Life Study—design and methodology
Janis Paterson, Colin Tukuitonga, Max Abbott, Michael Feehan, Phil Silva, Teuila Percival, Sarnia Carter, Esther Cowley-Malcolm, Jim Borrows, Maynard Williams, Philip Schluter
Following the Second World War, links between Pacific Island groups and the countries around the Pacific Rim intensified and mobility within the region increased substantially.1 Migration, sometimes supported by intergovernmental agreements, was fuelled by the search for employment and a higher standard of living2, resulting in 400,000 people of Pacific Islands ethnicity living in the Rim countries of the Pacific by the mid-1990s.3
Migration to New Zealand has been historically popular since the 1960s, due mostly to its proximity to the Pacific Islands but also due to labour demands resulting from economic growth in the New Zealand economy.4
As one of the fastest growing population subgroups in New Zealand, Pacific peoples (those residents with a Pacific Islands heritage) form an integral part of New Zealand society.5 The Pacific population in New Zealand is now estimated to exceed 284,000,6 comprising 6.8% of the New Zealand population, and Auckland is the preferred region of domicile.7 Samoans constitute the largest group (50%), followed by Cook Island Maori (23%), Tongan (18%), Niuean (9%), Fijian (3%), Tokelauan (3%), and Tuvalu Islanders (1%).7 This ethnic diversity is manifest in differing cultures, languages, strength of acculturation, and corresponding access to (and utilisation of) health and social services.
Despite better employment opportunities in New Zealand, Pacific peoples remain relatively socioeconomically disadvantaged. Their labour force participation and annual median income are lower than in the total population. Pacific people are more likely to be living in poor circumstances with restricted access to higher education, home ownership, and access to functional amenities such as automobiles and telephones. They are also over-represented in multiple adverse health and social statistics.7,8
Furthermore, the health of Pacific infants continues to be an issue of concern with the Pacific infant death rate being higher than the national infant death rate since 1997.7 Pacific infants have high rates of hospitalisation (particularly for respiratory illnesses)8,9 and present at hospital with higher severity of illness than New Zealand children of other ethnicities.10
In an effort to explore and understand both positive and negative facets of Pacific families’ life in New Zealand, the Pacific Islands Families: First Two Years of Life (PIF) Study was instigated. This prospective longitudinal study follows a cohort of Pacific children and their families in order to track the children’s development and wellbeing.
Several theoretical and research perspectives (focussing on individual, family, community, and societal influences on development, health, and wellbeing) underpin this study. These include infant development;11,12 temperament;13–15, attachment;16,17 mutual regulation of infant and caregiver behaviours;18 parenting;19–21 home environment;22,23 social support;24 acculturative stress;25 and the ecological approach emphasising the role of macro-level economic, social, and cultural factors.26
In addition to broader contextual factors, parenting and lifestyle behaviours develop according to different cognitive, linguistic, motivational, and social competencies that are relevant to each culture.27,28 The ‘transactional model’ of human development,29,30 acknowledging the interaction between various influences, will form the basis for the testing of more specific models to account for how the child develops within the context of a Pacific family.
The PIF study’s initial focus is on the key developmental stage of early infancy and the influence of the sociocultural context and family environment on Pacific children at ages 6 weeks, 12 months, and 24 months. The PIF study design is multidisciplined, broad-based and inclusive—capturing information from mothers, fathers, and infants.
General aims of the PIF study are to:
The specific aims of the PIF study (separated by the major health, psychosocial, and behavioural domains) appear in Table 1.
The study will provide ethnic-specific information to a variety of end-users on which to base interventions and inform policy development to help address many negative health and social issues that face Pacific children and families.
Design—Mothers of a cohort of live Pacific Islands infants born at Middlemore Hospital, South Auckland, between 15 March and 17 December 2000 were recruited.
An infant was deemed eligible if at least one parent identified themselves as being of Pacific Islands ethnicity and was a permanent resident of New Zealand. Presentation of study information to mothers and subsequent recruitment was completed within the hospital setting. Maternal home interviews were undertaken at approximately 6-weeks, 12-months, and 24-months postpartum; and paternal home interviews were conducted at approximately 12-months and 24-months postpartum.
At 12-months and 24-months, a developmental assessment of the child’s adaptive functioning was also undertaken in the family home. Prior to the commencement of the main study, extensive community consultation, staff training, and piloting was undertaken. Formal community liaison commenced in 1996L ethical clearance was received in May 1999; formal consent to recruit mothers at Middlemore was obtained on 10 June 1999; and all 60 mothers giving birth to Pacific infants between 12 July and 24 July 1999 were approached for recruitment into the pilot study.
Participant tracking databases and contact strategies for both mothers and fathers were developed; interview questionnaires were tested and timed; and data entry and management systems established during this piloting phase.
Recruitment—Promotional material was made available through various sources prior to hospital admission—including media, community meetings, antenatal classes and Lead Maternity Carers (LMCS). Eroni Clarke, a former New Zealand All Black rugby player born in Samoa, featured on posters and brochures to raise awareness of the study.
Participant eligibility identification was made through the Birthing Unit in conjunction with the Pacific Islands Cultural Resource Unit that provided a daily printout of Pacific admissions. These were checked for Pacific births and crosschecked with the daily records held in the Birthing Unit.
Under the supervision of Middlemore clinical staff, the Auckland University of Technology (AUT) Pasifika Development Manager (PDM) or the Pacific Liaison Officer (PLO) made the initial approach to identified potential participants; provided brief information about the study; and obtained permission for later contact.
Informed consent was not sought at this point. For potential participants who were not seen at Middlemore Hospital, either because they were discharged home early or transferred to one of the two satellite hospitals, the initial visit was made by the PLO at a location convenient for the mother.
Approximately 6-weeks postpartum, potential participants were allocated to a team of female Pacific interviewers fluent in both English and a Pacific language. In most cases, the interviewers were ethnically matched to the potential participant. The interviewers visited the potential participant in their own homes, fully described the study with the parent(s), and obtained the mother’s informed consent. Once consent was obtained, the interview was carried out in the mother’s preferred language.
When the children reached their first and second birthdays, all maternal participants were re-contacted and revisited by a female Pacific interviewer. Again, consent was obtained before the interview was conducted in the mother’s preferred language. At the time of the interview, mothers were asked to give permission for a male Pacific interviewer to contact and interview the father of the child. If permission and paternal contact details were obtained, then a Pacific male interviewer contacted the father to discuss participation in the study. Once informed consent was obtained from the father, the interview was carried out in the father’s preferred language.
Data collection—Survey instruments are considered relevant and appropriate by both researchers and the Pacific community. Pacific researchers have considerable input into the general measurement framework and specific measures employed to ensure their content acceptability and validity. Where possible, standardised internationally developed measures and scales are used or adapted for the PIF study.
The use of such scales ensures constructs (universal across cultures) are measured and information considered important to child health and development is elicited. Measures employed during the pilot phase that had unsatisfactory reliability and/or validity were removed. Some measures underwent focus group examination to check wording of items and testing procedures.31,32 However, the psychometric properties of most standardised measures used in the pilot were acceptable, and few modifications were required. Table 2 includes the specific dimensions and instruments employed in the main PIF study.
The interview protocol used at the 6-weeks phase was translated into Samoan, Tongan, and Cook Island Maori and checked by fluent speakers of these Pacific languages for clarity of exposition and concept integrity. Only 174 (13%) mothers utilised a non-English version, the majority of these being Tongan. As such, the interview protocols used at the 12-month and 24-month phases were made available in English and Tongan only.
With mother’s consent, supplementary information was obtained from both Middlemore Hospital and Plunket. Maternal obstetric history, delivery details, and infant measures of birth weight for gestational age, head circumference, Apgar scores, and neonatal care were obtained from Middlemore’s Hospital Discharge Summary records. Information about infant’s physical status was obtained from the comprehensive physical assessment carried out by Plunket at 6-weeks and 6-months postpartum.
Data accuracy—Various systems were implemented to ensure data accuracy and consistency. These included: manual coding of each interview protocol to check consistency within the individual interview; accompanying interviewers to check on rapport, informed consent and on the undertaking of standard procedures; and post-interview random phone checks with participants to clarify and confirm specific details of the interviews.
Data entry and security—Two researchers developed codes based on the first 50 cases for the open-ended questions and interviewer observations. All data are double entered into an electronic database (SPSS Data Entry Builder 2.0) that employs comprehensive data validation and checking rules. This database includes no personal identification and is held in secure password protected storage under the responsibility of the PIF study’s Co-Directors in accordance with the requirements of the New Zealand Privacy Act (1993) and the Health Information Privacy Code (1994).
Participant labelling in this database is made by reference number only. Identification information is stored on a separate computer in a password-protected file. All interviews are confidential, and only PIF study staff authorised by the Co-Directors have access to computerised data. All information is treated as sensitive data. Participants own their data and reserve the right to withdraw it at any time.
Sample size—Balancing the competing demands for increased statistical power in longitudinal studies33 against conducting a feasible, efficient, and cost-effective study, a final cohort of approximately 1000 mothers of Pacific infants at 24-months was targeted. To compensate for likely attrition (assumed to be 10% per measurement wave), it was determined that approximately 1400 eligible mothers needed to be recruited. This sized cohort would generally have adequate power for inter-ethnic comparisons of major groups such as Samoan, Tongan, and Cook Islands Maori. However, it is recognised that power will not always be adequate for analyses involving smaller ethnic groups; for the detection of small differences between groups; or for more complex analyses involving a greater number of categories (e.g. different feeding methods in Islands-born versus NZ-born Tongans).
Statistical analyses—In terms of the more general analytic approach, cross-sectional analyses at each assessment time will be undertaken, allowing the prevalence of key outcomes to be estimated, as well as the association between risk factors and key outcome variables. Longitudinal analysis techniques will be used in the modelling of developmental pathways including structural equation modelling, mixed linear models, and generalised estimating equations. The gathering of data from mothers, fathers, and secondary sources allows the concordance between reports of child outcomes and parental functioning to be assessed.
Fisher’s exact test was employed to determine whether differential attrition existed over time for a selection of basic sociodemographic variables. Differential attrition would have occurred if both 12-month and 24-month distributions of participation and non-participation were significantly different to the 6-week distributions, using a significance level α=0.05 to define statistical significance.
Ethics—Careful consideration is continually given to the ethical aspects of this longitudinal study with Pacific peoples. Ethical approval for the pilot and main PIF study was obtained from the Auckland Branch of the National Ethics Committee, the Royal New Zealand Plunket Society, and the South Auckland Health Clinical Board.
Of the mothers with Pacific infants who had been born at Middlemore Hospital between 15 March and 17 December 2000, 1590 gave consent to the 6-week postpartum home visit. The recruitment and participation rate at the 6-weeks, 12-months, and 24-months interview schedule is depicted in Figure 1.
From the 1376 primary respondents of the cohort at 6-weeks, there were 1368 biological mothers, 1 foster mother, 6 adoptive mothers, and 1 grandmother. For ease of exposition, we shall refer to this group collectively as ‘mothers’ hereafter. Twenty-three of the 1376 mothers gave birth to twins, but as one twin member was stillborn the overall number of children forming this cohort totalled 1398. Of these, 374 (27%) infants were first-born and 1339 (96%) infants were discharged with their mothers from hospital. Middlemore hospital records revealed that the mean birth weight was 3584 grams (SD: 613 grams) with a range of 650 to 5390 grams.
The mean age of all mothers was 27.9 years (SD: 6.2 years), the range was 14 to 57 years, and 8% of mothers were younger than 20 years (see Table 3). Most (1107 [80%]) mothers were living together in married or de facto partnerships. 454 (33%) mothers were New Zealand-born. In terms of schooling, 535 (39%) mothers had no formal educational qualifications, 464 (34%) had secondary school qualifications only, and 377 (27%) had post-school qualifications.
Ethnicity of mothers was self-identified and also coded according to the sub-groupings of the 1996 Census. Of the cohort mothers, 650 (47%) self identified their major ethnic group as Samoan; 289 (21%) as Tongan; 232 (17%) as Cook Islands Maori; 59 (4%) as Niuean; 47 (3%) as Other Pacific (this includes mothers identifying equally with two or more Pacific groups, equally with Pacific and Non-Pacific groups, or with Pacific groups other than Samoan, Tongan, Cook Island, or Niuean); and 99 (7%) as Non-Pacific. These frequencies are broadly similar to those seen in the New Zealand Pacific Island population.7
No important differential attrition was observed for any of the sociodemographic variables investigated in Table 3. While the distribution of participants was statistically different to non-participants at 12-months for ethnicity (p=0.001) and years lived in New Zealand (p=0.007), no such difference existed at 24-months (ethnicity, p=0.20; years lived in New Zealand, p=0.15).
999 of the mothers interviewed at 12-months had partners who met eligibility criteria to act as collateral respondents, of whom 825 (83%) were interviewed. Most (820 [99%]) fathers interviewed at 1 year were the biological fathers of the children, with five adoptive or stepfathers. Again, for ease of exposition, we shall refer to this group collectively as ‘fathers’ hereafter. Most (786 [95%]) fathers were living with the biological mother of the child in a married (77%) or de facto (18%) relationship. The mean age was 32.1 years (SD: 7.3 years), the range was 17 to 65 years, and 1% of the fathers were younger than 20 years.
Approximately one-quarter (203 [25%]) of the fathers were born in New Zealand. 440 fathers (53%) self-identified their major ethnic group as Samoan; 199 (24%) as Tongan; 73 (9%) as Cook Islands Maori; 26 (3%) as Niuean; 28 (3%) as Other Pacific (this includes fathers identifying equally with two or more Pacific groups, equally with Pacific and Non-Pacific groups, or with Pacific groups other than Samoan, Tongan, Cook Island or Niuean); and 59 (7%) as Non-Pacific.
In terms of education, 481 (58%) fathers had no formal educational qualifications, 220 (27%) had a secondary school qualification only, and 122 (15%) had a post-school qualification.
Of the 757 secondary respondents interviewed at 24-months, almost all (754 [99%]) were the biological fathers of the children in the cohort with three adoptive or stepfathers. There were no significant differences in age, martial status, smoking status, highest educational qualification, English fluency, and household income demographics of this group compared to that interviewed at 12-months. However, there were fewer Tongan fathers interviewed at 24-months (20% vs 24%, p<0.001), and more who had lived in New Zealand for over a decade (42% vs 40%, p=0.002).
Birth mothers served as respondents for all child assessments at 12-months and for 97% of 1,064 child assessments at 24-months postpartum. Mean height and weight of the children at 24-months was 89.0 cm (SD: 4.4 cm) and 14.4 kg (SD: 2.2 kg) respectively.
The PIF study is a large, scientifically and culturally robust longitudinal study that has achieved respectable participation rates over the first 24-months of its life. Both maternal and paternal figures have the opportunity to participate, something that has been rarely undertaken in previous longitudinal studies of this nature. Both parental figures have been highly responsive and receptive to our invitation to participate in each of the measurement waves despite the many demands upon their time.
We believe this stems from several important design attributes embedded within the PIF study, including: the integration of Pacific people in all aspects of the study; the single-based hospital design; and having home-visits conducted by ethnic and gender-matched interviewers.
The integral involvement with Pacific people in the consultation, design, development of the interview protocols and instruments, recruitment, elicitation, and interpretation of information and governance of the study is critical. Such input gives credibility, cultural appropriateness, and Pacific community involvement and ownership of the study. These features coupled with the perceived importance and relevance of the study’s contents by the participants themselves should ensure that response rates remain respectable throughout.
The PIF study directors are guided by the Pacific Peoples’ Advisory Board who monitored the general direction of the study. The Board’s primary responsibility is to protect and enhance the study to maximise benefits for Pacific communities. In addition, through the sharing between Pacific and non-Pacific researchers, the PIF study serves as a useful platform for the advancement and training of Pacific researchers and field staff.
A single hospital-based recruitment catchment enabled the target population to be easily identified and accessed; streamlines and enhances the cost-effectiveness of procedures for the approach and recruitment; and increases the ‘health’ credibility of the study for the participants. Also, this institutional sampling frame lends itself to a ‘study champion’; a person or persons with considerable passion and vigour, who understands the local system and ensures smooth running of study. Both the PDM and PLO performed well in this capacity for the PIF study.
Instrumental in the PIF study’s recruitment and retention success to date are the research base for fieldwork staff and the home visits by ethnic and gender matched bilingual or multilingual Pacific interviewers. The research base was a readily identifiable ‘home’ for the field workers—where they engaged, learnt, and developed with fellow colleagues and the study investigators. In terms of home visitations, participants find them convenient and comfortable, particularly for those without private transportation. Moreover, they find the experience less disruptive for the family especially if they have other children who need attention.
The ethnic and gender-matched bilingual or multilingual Pacific interviewers facilitates the development of a rapport between themselves and the participants, and language barriers or difficulties are circumvented. The availability of survey instruments in English, Samoan, Tongan, and Cook Island Maori, in the first instance, probably contributed to the strong participant response.
Other design efficiencies included the utilisation of data from alternative available sources, such as Middlemore Hospital Discharge Summary records and Plunket’s records including the comprehensive physical assessment. The capture and use of these data decreases participant burden and thereby is likely to increase participation rates. Additionally, use of the data ‘value-adds’ to the records collected by each of these organisations themselves.
While also a strength, a weakness of the PIF study is that it only included infants born at Middlemore Hospital in the sampling frame. It could be opined that this may impinge on the representiveness and generalisability of the study’s findings. However, more Pacific infants are born at Middlemore than any other New Zealand hospital and we demonstrate that the proportions of Pacific ethnic subgroups enrolled were broadly similar to national proportions which suggest that any inherent biases due to this recruitment strategy are likely to be small.
Although not a weakness in the PIF study itself, the lack of equivalent contemporaneous data for other non-Pacific ethnic groups means that interethnic group comparisons cannot be readily made. Recognising this deficiency, the Ministry of Social Development and Heath Research Council of New Zealand recently called for proposals for a longitudinal study of New Zealand Children and Families.34
Because of the success of the PIF study, the Pacific Islands Families Study: Transition to School (PIF: TTS) was conceived and launched in 2004. This successor study follows the PIF cohort of Pacific children and their families, and aims to assess the children’s development and wellbeing at ages 4 and 6 years. In particular, this study will focus on the cohort and their families as they negotiate critical developmental transitions, notably the transition to school.
The PIF study has been designed to advance scientific knowledge in a number of disciplines and provide public benefits through the provision of good quality information. It is envisaged that findings from this study will inform policy development and assist programme implementation for a variety of stakeholders working towards maximising the potential of Pacific families and communities within broader New Zealand society. With the large, scientifically and culturally robust study that has achieved respectable participation rates, we believe these aims will be met.
Author information: Janis Paterson, Associate Professor and Co-Director, Pacific Islands Families: First Two Years of Life Study1; Colin Tukuitonga, former Co-Director, Pacific Islands Families: First Two Years of Life Study1 and Pacific Health Research Centre, Department of Maori & Pacific Island Health2; Max Abbott, Pro Vice-Chancellor and Dean1; Michael Feehan, Adjunct Associate Professor1 and Co-President3; Phil Silva, Adjunct Professor1 and Founding Director, Dunedin Multidisciplinary Health & Development Research Unit4; Teuila Percival, Paediatrician5; Sarnia Carter, Research Fellow, Pacific Islands Families: First Two Years of Life Study1; Esther Cowley-Malcolm, Manager, Pacific Research Workforce Development, Pacific Islands Families: First Two Years of Life Study1; Jim Borrows, Research Manager1; Maynard Williams, Senior Research Fellow and Statistician1; Philip Schluter, Professor of Biostatistics1.
Acknowledgements: The Pacific Islands Families: First Two Years of Life (PIF) Study is supported by grants awarded from the Foundation for Science, Research and Technology; the Health Research Council of New Zealand; and the Maurice and Phyllis Paykel Trust. The authors also gratefully acknowledge the families who participated in the study; the Pacific Peoples Advisory Board; and the other members of the PIF research team.
Correspondence: Professor Philip Schluter, Faculty of Health & Environmental Sciences, Auckland University of Technology, Private Bag 92006, Auckland 1020. Fax: (09) 921 9877; email: firstname.lastname@example.org
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals