Cohort Profile: The All Our Babies pregnancy cohort (AOB)

Why was the cohort set up?

All Our Babies (AOB) is a community-based, longitudinal pregnancy cohort developed to investigate the relationships between the prenatal and early life periods and outcomes for infants, children and mothers. The design of AOB follows a life course perspective, whereby the influence of early events on long-term health and development of both mothers and children are investigated through examining factors across life stages.¹ AOB spans pregnancy, birth and early postpartum through childhood, and therefore provides the unique opportunity to describe the relations between prenatal events and early life development and to examine key factors that influence child and mother well-being over time.

AOB was originally designed to measure maternal and infant outcomes during the perinatal period, with a particular emphasis on barriers and facilitators to accessing health care services in Calgary, Alberta. Approximately 1 year after recruitment had started, an additional objective, to examine biological and environmental determinants of adverse birth outcomes, specifically spontaneous preterm birth, was added. Recognition of the opportunity to continue to collect relevant life course information on the AOB families, collaborations with content experts and securing additional funding has enabled ongoing follow-up of AOB mother-child dyads. The overall objective was to further investigate risk and protective factors for optimal child development, and to understand the trajectory and impact of poor maternal mental health over time. Mothers have completed questionnaires from pregnancy to 3 years postpartum, and consented to providing the research team with access to their obstetric medical records. Data collection for a 5-year follow-up questionnaire is ongoing. A subgroup within the cohort participated in the ‘prediction of preterm birth’ component and provided blood samples during pregnancy and an umbilical cord blood sample. The continuation of follow-up to 8 years is under way.

Who is in the cohort?

A total of 3387 pregnant women (aged 19 to 47 years), residing in Calgary, Canada, were recruited between May 2008 and May 2011. Women were eligible if they were at less than 25 weeks of gestation at the time of recruitment, at least 18 years old, were accessing prenatal care in Calgary, Canada, and were able to complete written questionnaires in English. The AOB study was approved by the Conjoint Health Research Ethics Board at the University of Calgary (Ethics ID 20821 and 22821). Recruitment methods for the AOB study have been described previously.^2,3 Briefly, a multi-method community-based recruitment strategy was developed involving active and passive recruitment strategies. Active recruitment occurred through a city-wide medical laboratory service where women undergoing prenatal viral serology testing were contacted by telephone by the laboratory staff, to request permission to release their contact information to AOB researchers. Women who consented were contacted to inform them of the study, determine eligibility, answer questions about the study and obtain informed consent. Using another active recruitment strategy, research staff were on site in primary health care waiting rooms to provide women with information about the study, assess eligibility and obtain informed consent. Passive recruitment occurred through posters advertising the study being displayed in places frequented by pregnant women, enabling them to self-identify to the study. Baseline questionnaire data were collected between 2008 and 2011. Table 1 describes the socio-demographic, pregnancy history, psychosocial health, delivery and birth outcomes of AOB participants.

Table 1

Participant characteristics

Category	Characteristics	Na	N (%)
Demographics	Maternal age	3282
	19 years or younger		32 (1.0)
	20–24 years		264 (8.0)
	25–34 years		2344 (71.4)
	35 years or older		642 (19.6)
	Marital status	3353
	Married/common law		3165 (94.4)
	Single/separated/divorced		188 (5.6)
	Maternal education	3355
	High school		367 (10.9)
	Some post-secondary		479 (14.3)
	Graduated post-secondary		1980 (59.0)
	Postgraduate studies		529 (15.8)
	Total household income (before taxes and deductions)	3251
	$39 999 or less		298 (9.2)
	$40 000–$59 999		298 (9.2)
	$60 000–$79 999		419 (12.9)
	$80 000 or more		2236 (68.8)
	Born in Canada	3359
	Yes		2628 (78.2)
	No		731 (21.8)
	Primary language spoken at home	3359
	English		2957 (88.0)
	Other		402 (22.0)
Pregnancy history	Gravidity	3337
	No previous pregnancies		1191 (35.7)
	At least one previous live birth		2146 (64.3)
	Parity	3339
	No previous live births		1636 (49.0)
	At least one previous live birth		1703 (51.0)
Psychosocial health at study intake	Depression, EPDS ≥ 13	3345
	Depression symptoms		270 (8.1)
	Low depression symptoms		3075 (91.9)
	Anxiety, SSAI ≥ 40	3241
	Anxiety symptoms		549 (16.9)
	Low anxiety symptoms		2692 (83.1)
	Stress, PSS ≥ 20	3318
	High stress		601 (18.1)
	Low stress		2717 (81.9)
	Social support, MOS-SSS ≤ 69	3325
	Low social support		459 (13.8)
	High social support		2866 (86.2)
Delivery and birth outcomes	Method of delivery	3055
	Vaginal		2297 (75.2)
	Caesarean section		758 (24.8)
	Number of babies	3057
	Singleton		3021 (98.8)
	Twins		36 (1.2)
	Child sex
	Female	3055b	1467 (47.5)
	Male		1622 (52.5)
	Gestation at birth among singletons	2995
	Term (37 weeks or more)		2784 (93.0)
	Preterm (less than 37 weeks)		211 (7.0)
	Birthweight among singletons	2862
	Not low birthweight (2500 g or more)		2712 (94.8)
	Low birthweight (less than 2500 g)		150 (5.2)

EPDS, Edinburgh Postnatal Depression Scale; SSAI, Spielberger State Anxiety Inventory; PSS, Perceived Stress Scale; MOS-SSS, Medical Outcomes Study Social Support Survey.

^aN indicates the total sample with available data for each variable.

^bN = 3055 AOB participants, 36 of whom had twins, therefore total number of babies was 3091.

Socio-demographic data were unavailable for non-participating women. To assess the representativeness of the AOB sample, the socio-demographic characteristics of AOB participants at the time of recruitment were compared with those of parents of young children in Calgary, Alberta and Canada during the same period.^4–6 A greater proportion of women in the AOB sample (82%) had annual household incomes > $60 000 compared with parenting women in Calgary (65%), Alberta (61%) and Canada (56%) (Figure 1). More AOB participants were married (83%) compared with parenting women in Calgary (73%), Alberta (70%) and Canada (60%) (Figure 1). The proportion of AOB participants who were at least 35 years old, who had completed post-secondary education and who were foreign born were similar to the proportions in Calgary, Alberta and Canada (Figure 1).

Figure 1

Socio-demographic characteristics of women in AOB, Calgary, Alberta and Canada.

Figure 2 shows the numbers of participants accrued and followed up to date. In total, 3387 women enrolled in the study and completed at least one questionnaire. Response rates for the first three questionnaires were 99%, 94% and 90%, respectively. Due to delays in securing funding and ethics approval processes, not all children were in the appropriate age range when each follow-up questionnaire was implemented. Among those eligible, response rates for the follow-up questionnaires were 81% at 1 year, 76% at 2 years and 69% at 3 years (Figure 2).

Figure 2

AOB participant attrition from recruitment through three year follow-up.

Participants who withdrew from the study (n = 938, 28%) included women who actively withdrew from the study, citing a loss of interest, time, comfort with the study or support from their partner, or declining to participate in follow-up research (n = 235, 7%). Participants who passively withdrew from the study included those lost to follow-up, geographical moves and unknown reasons (n = 669, 20%) and women who experienced a miscarriage, pregnancy termination, stillbirth or child death (n = 34, 1%) (Figure 2).

How often have they been followed up?

Between 2009 and 2015, eligible AOB participants were asked to complete questionnaires when their child was 1, 2 and 3 years old (Figure 3). Data collection for the 5-year follow-up questionnaire began in 2013 and is ongoing; the 8-year follow-up is expected to begin in January 2017.

Figure 3

AOB data collection timeline.

The demographic characteristics of participants who have been retained in the AOB cohort compared with those who have discontinued are described in Table 2. Continuing participants were defined as those who completed at least one follow-up questionnaire when their child was 1, 2 or 3 years old. Discontinuing participants included those who actively withdrew from the study or those who were lost to follow-up (excluding pregnancy losses and child deaths). Continuing participants were more likely to be older, be in a stable relationship, have higher educational attainment, have higher family incomes, be born in Canada and primarily speak English in their home (Table 2).

Table 2

Comparison of demographic characteristics of continuing and discontinued participants

Characteristic	Continuing participants N = 2384a	Discontinued participants N = 904a	P-value
Maternal age
19 years or younger	13 (0.6)	19 (2.2)	< 0.001
20–24 years	157 (6.6)	106 (12.2)
25–34 years	1735 (72.8)	586 (67.8)
35 years or older	479 (20.1)	154 (17.8)
Marital status
Married/common law	2322 (95.3)	813 (92.1)	0.001
Single/separated/divorced	114 (4.7)	70 (7.9)
Maternal education
High school or some post-secondary studies	547 (22.5)	293 (33.1)	< 0.001
Graduated post-secondary or some postgraduate studies	1889 (77.5)	592 (66.9)
Total household income (before taxes and deductions)
$39 999 or less	157 (6.6)	137 (15.9)	< 0.001
$40 000–$79 999	505 (21.4)	204 (23.8)
$80 000 or more	1698 (72.0)	518 (60.3)
Born in Canada
Yes	1971 (80.8)	631 (71.2)	< 0.001
No	468 (19.2)	255 (28.8)
Primary language spoken at home
English	2186 (89.6)	741 (83.6)	< 0.001
Other	253 (10.4)	145 (16.4)

^aDenominator varies slightly due to missing data for some variables.

What has been measured?

Figure 4 summarizes the broad categories of data collected in the AOB study, spanning pregnancy to 5 years postpartum. In order to understand trajectories of maternal psychosocial health and child development, attention has been paid to assessing the same constructs over time, using the same tools when appropriate. Future analyses, using longitudinal methods such as latent growth class analysis and path analysis, will be used to describe the course of maternal psychosocial health and child development over time.

Figure 4

Constructs measured over time in the AOB study.

Questionnaires

Participants were asked to complete two questionnaires during pregnancy (at < 25 weeks and at 34–36 weeks of gestation) and one questionnaire at 4 months postpartum. Those who agreed to future contact were invited to complete follow-up questionnaires at 1, 2, 3 and 5 years postpartum. AOB questionnaires were developed in collaboration with researchers, health care providers, epidemiologists, decision makers and community programme experts. These questionnaires include standardized measures when available and questions created specifically for the study when standardized measures were not suitable. The first two questionnaires administered during pregnancy collected maternal data only, including socio-demographics, preconception history, pregnancy history, pregnancy experiences, access to prenatal care, health service utilization, psychosocial health, lifestyle and history of mental health disorders, and stressful life events. The third questionnaire, administered at 4 months postpartum, collected maternal and child data including questions about labour and delivery, birth outcomes, breastfeeding, child health, maternal mental health, lifestyle, parenting and the child care environment. The four follow-up questionnaires, administered when the AOB target child was 1, 2, 3 and 5 years old, collected maternal and child data including maternal physical and mental health, health care utilization, parenting, family well-being, childcare environment, community resource use, child health and child development. A detailed list of the data collected, including the measure used if applicable, as part of each questionnaire is provided in Table 3.

Table 3

Variables collected in all questionnaires

Variables	Q1 < 25 weeks of gestation	Q2 34–36 weeks of gestation	Q3  4 months postpartum	Q4  1 year	Q5  2 years	Q6  3 years	Q7  5 years
Socio-demographics
Household income	X			X		X	X
Maternal education	X					X
Marital status	X			X	X	X	X
Maternal age	X
Background	X
Primary language	X
Paternal age	X
Family background	X
Family status					X	X
Neighbourhood safety		X		X		X
Maternal health/pregnancy history/perinatal outcomes/health behaviours
Overall health	SF-12	SF-12	SF-12	SF-12	SF-12	SF-12	SF-12
				1-item	1-item	1-item	1-item
Maternal health service utilization	X	X	X	X
Pregnancy history	X
Pre-pregnancy experiences	X
Pregnancy experiences	X	X	X
Labour and delivery experiences			X
Post-birth health				X
Breastfeeding			X	X
Feeding difficulties			MCH Feeding Scale	MCH Feeding Scale
BMI/height and weight	X	X
Exercise	X	X				X	X
Sleep						X
Food group intake	X	X
Food security	X	X	X
Substance use	X	X	X	X		X	X
Physical activity					X	X	GLTE
Housing security			X			X
Subsequent birth control					X
Subsequent pregnancies					X	X
Subsequent live births					X	X
Maternal psychosocial health
Maternal history of mental health	X
Maternal history of stressful life events	X				RAINE Stressful Life Events Scale
Maternal history of abuse		X
Overall emotional health	X	X	X	X	X	X
Depression	EPDS	EPDS	EPDS	EPDS	CES-D	CES-D	CES-D
Anxiety	SSAI	SSAI	SSAI	SSAI & STAI	SSAI	SSAI	SSAI 6-items
Stress	PSS	PSS	PSS	PSS		PSS	PSS
Social support	MOS SSS	MOS SSS	MOS SSS	NLSCY SSS	NLSCY SSS	NLSCY SSS	NLSCY SSS
Optimism		LOT-R
Abuse			X
Balancing responsibilities				X	X	X
Relationship happiness				X		X
Relationship tension and conflict	X					X	X
Relationship experiences							ECR (short form)
Adverse childhood experiences						ACEs
Birth outcomes
Number of babies			X
Child sex			X
Gestation at birth			X
Birthweight			X
Child health
Overall child health				X	X	X	X
Diagnosed long-term health condition				X	X	X	X
Ear infections				X	X	X
BMI/height and weight				X	X	X	X
Child health service utilization			X	X	X	X
Child vaccinations			X	X	X	X	X
Child sleep			X	X	X		X
Child oral health					X	X
Child nutrition						NutriSTEP	X
Child development
Overall development				ASQ	ASQ	ASQ	ASQ
Language abilities				CDI: Words and Gestures	CDI: Words and Gestures	CDI: Words and Sentences	CCC-2
Temperament						CBQ-SF
Behaviour					BITSEA	NLSCY CBCL	BASC
Autism screen					M-CHAT
Family history of developmental delay or learning disability				X		X
Parenting
Parenting morale			PMI
Parent-child interaction				BPCIS	BPCIS
Parenting style			PACOTIS Parental hostile- reactive behaviours subscale			NLSCY parenting scales	NLSCY parenting scales
			4-items
Parenting self-efficacy and competence				PACOTIS Parental self-efficacy subscale	PACOTIS Parental self-efficacy subscale
				5-items	5-items
Maternal separation anxiety				MSAS
Parenting alliance				DAS
				1-item
Reading and literacy beliefs						PRBI
Family environment
Community resource use			X	X	X	X	X
Mother work status		X	X	X	X	X	X
Maternity leave			X	X
Father work status					X
Paternity leave				X
Paid leave/benefits			X	X
Child care arrangement			X	X	X	X	X
Child exposure to a second language				X	X	X	X
Division of household jobs					X
Father contribution to caregiving						X
Parent-child interaction				X	X	X
Home literacy environment						X	X
Child screen time					X	X	X
Child physical activity					X	X	GLTE
Child enrolment in preschool or kindergarten							X
Child injury							X

X refers to investigator-derived variables. The full versions of standardized measures were used, unless the number of items used for a modified version of a measure is noted.

SF-12, Short Form Health Survey-12; MCH Feeding Scale, Montreal Children’s Hospital Feeding Scale; GLTE, Godin Leisure-Time Exercise Questionnaire; EPD, Edinburgh Postnatal Depression Scale; CES-D, Center for Epidemiologic Studies Depression Scale; SSAI, Spielberger State Anxiety Inventory; STAI, Spielberger Trait Anxiety Inventory; PSS, Perceived Stress Scale; MOS SSS, Medical Outcomes Study Social Support Survey; NLSCY SSS, National Longitudinal Survey of Children and Youth Social Support Scale; LOT-R, Life Orientation Test – Revised; ECR (short form), Experiences in Close Relationships Scale (short form); ACEs, Adverse Childhood Experiences Checklist; NutriSTEP, Nutrition Screening Tool for Every Preschooler; ASQ, Ages and Stages Questionnaire; CDI Words and Gestures, MacArthur-Bates Communicative Development Inventories: Words and Gestures; CDI Words and Sentences, MacArthur-Bates Communicative Development Inventories: Words and Sentences; CCC-2, Child Communication Checklist-2; CBQ-SF, Rothbart Child Temperament Scale; BITSEA, Brief Infant Toddler Social Emotional Assessment; NLSCY CBCL, National Longitudinal Survey of Children and Youth Child Behaviour Checklist; BASC, Behaviour Assessment System for Children; M-CHAT, Modified Checklist for Autism in Toddlers; PMI, Parenting Morale Index; BPCIS, Brigance Parent Child Interaction Scale; PACOTIS, Parental Cognitions and Conduct Toward the Infant Scale - Parental Hostile-Reactive Behaviors subscale; NLSCY Parenting Scales, National Longitudinal Survey of Children and Youth Parenting Scales; PACOTIS Parental Self-Efficacy subscale, Parental Cognitions and Conduct Toward the Infant Scales - Parental Self-Efficacy subscale; MSAS, Maternal Separation Anxiety Scale; DAS, Dyadic Adjustment Scale; PRBI, Parent Reading Belief Inventory.

Medical records

Obstetric and birth data were obtained through data linkages to mothers’ and children’s prenatal and birth medical records from the Alberta Health electronic database, and have been previously described.^2,3 The medical records contributed obstetric and birth data not captured in the questionnaires (e.g. pregnancy complications, antepartum risk score, Apgar score etc.) and some data elements that were also collected in the 4-month postpartum questionnaire (type of delivery, gestational age, birthweight etc.). A validation study comparing the two sources found maternal recall of infant characteristics and labour and delivery outcomes was valid (sensitivity and specificity > 85%).⁷ Both versions of these common data elements have been retained in the dataset, providing the option to utilize either source depending on the research question and analysis being conducted.

Biological specimens

Maternal blood at 17–23 and 27–33 weeks of gestation and umbilical cord blood, when retrievable, was collected at birth for participants in the ‘prediction of preterm birth’ group (n = 1871). Maternal blood samples were collected into PAXgene^™ tubes (PreAnalytix/BD Canada, Mississauga, ON, Canada) for RNA extraction, EDTA tubes for DNA extraction, a heparin tube to isolate plasma and a serum collection tube. Umbilical cord blood (3–5 ml) is routinely collected at all hospital births to establish red blood cell antigens. The research team accessed the unused portion of these umbilical cord blood samples from consenting participants and transferred them to study freezers. All samples have been stored at −80°C. The biological specimen collection and storage provide whole blood, plasma and serum samples from which lymphocytes, cytokines and proteins may be isolated and RNA and DNA can be extracted for microarray analysis. Currently, total RNA has been extracted and microarray has been performed on 326 maternal blood samples (165 women) by the Centre for Applied Genomics (TCAG; Hospital for Sick Children, Toronto, ON, Canada). Data were deposited into the National Center for Biotechnology Information Gene Expression Omnibus (accession number: GSE59491).

Utilization of H1N1 and seasonal influenza vaccine sub-study

A subgroup of AOB participants (n = 509) who received the second questionnaire between 18 November 2009 and 31 March 2010 were asked about their utilization of the H1N1 vaccine and seasonal influenza vaccine during the 2009–10 flu season. During this year, the World Health Organization had declared the spread of the H1N1 virus a pandemic. A seven-item questionnaire, created specifically for this study, was mailed to eligible participants with the second AOB questionnaire. Questions were asked about women’s utilization of the H1N1 and seasonal influenza vaccines, timing of receipt of both vaccines, contraction of influenza, timing of contraction of influenza and whether they had taken time off work due to influenza symptoms during their pregnancy.

Impact of local flood sub-study

A sub-study was conducted to understand the impact of flooding that occurred in Calgary, Alberta, in June 2013 on families and their children. Torrential rain and flooding led to a local state of emergency being declared, evacuation of over 100 000 residents and power outages across the province. All AOB participants who were living in Calgary at the time and who agreed to be contacted for additional research were mailed a flood impact questionnaire in the autumn of 2013. The questionnaire asked women about their experiences during the flood and damages incurred, help/aid provided and received, stress responses such as post-traumatic stress, psychological distress, social cohesion and relationship dynamics. Participants who did not return a questionnaire were asked to answer three brief questions from the flood impact questionnaire by telephone or e-mail. In total, 1923 women participated in the flood impact study, of whom 1442 participants completed the flood questionnaire and 471 participants responded to the three flood questions by telephone or e-mail.

Language sub-study

A subgroup of AOB participants (250 mother-child dyads) have been recruited to participate in a study investigating early influences of phonological awareness. Phonological awareness describes a child’s ability to detect and manipulate the sound structure of spoken words and is one of the most reliable and robust predictors of later reading ability.⁸ In this sub-study, 5-year-olds were tested in person during a laboratory visit on measures of phonological awareness using the Comprehensive Test of Phonological Processing 2nd Edition,⁹ and on linguistic and reasoning abilities using the Clinical Evaluation of Language Fundamentals 5¹⁰ and the Wechsler Preschool and Primary Scale of Intelligence: Block Design subtest.¹¹ Pure-tone hearing screens were administered according to the American Speech-Language-Hearing Association 1997 Standards to detect hearing loss. Data were collected from mothers on maternal/family and environmental characteristics, home literacy practices and maternal literacy beliefs and attitudes, as part of the AOB questionnaires. Assessments are ongoing.

What has it found?

A complete list of publications from AOB can be found at [allourfamiliesstudy.com]. Key findings to date have been summarized by outcome (Box 1

Box 1 Key findings

• Whole blood gene expressions and clinical factors were associated with spontaneous preterm birth (SPTB) in asymptomatic women. Integrating clinical data and gene expression data in a multivariate model resulted in the prediction of SPTB with 65% sensitivity and 88% specificity in asymptomatic women.

• Anxiety during pregnancy coupled with previous psychosocial stress increased women’s risk of late preterm birth (34–36 weeks of gestation) when controlling for demographic, obstetrical, medical and lifestyle factors.

• A comparison of data collected from maternal report on 4-month postpartum questionnaires and medical records revealed that maternal self-report of labour, delivery and birth outcome is highly valid (sensitivity and specificity > 85%).

• A total of 17% of children experienced delayed development at 1 year of age. Prenatal depression, preterm birth, low postpartum community engagement and less frequent parent-child interaction were identified as risk factors for delayed development at age 1 year.

• In all, 13% of children were classified with a late talking language delay at age 2 years. Risk factors included male sex and a family history of language delay. Protective factors included providing informal play opportunities, reading and sharing books daily and receiving care at a child care centre.

).

Pregnancy outcomes

Depression and anxiety during pregnancy

The prevalence of depression during pregnancy in the second or third trimester was 11%, whereas 23% of women experienced anxiety on at least one occasion during pregnancy.¹² Chronic antenatal depression, defined as experiencing depression during both the second and third trimesters, occurred for less than 3% of women. Chronic antenatal anxiety was experienced by 9% of women. Risk factors for chronic antenatal depressive symptoms included history of poor mental health, poor physical health, lower optimism, high perceived stress, low social support, unplanned pregnancy and infertility treatment.¹² Risk factors for chronic antenatal anxiety symptoms included history of poor mental health, lower optimism, high perceived stress, low social support and partner tension.¹² Understanding which women will experience long-standing challenges with anxiety and depression can facilitate early identification and targeting of scarce mental health resources.

Characteristics of women who consume alcohol during pregnancy

Almost half (49%) of women reported consuming some alcohol during pregnancy, including before they knew they were pregnant.¹³ Prior to pregnancy recognition, 13% of women reported at least one episode of binge drinking, defined as consuming five or more drinks on one occasion. After pregnancy recognition, 43% of women drank some alcohol, almost all at low to moderate levels. Risk factors for binge drinking in early pregnancy included lower educational attainment, nulliparity, unintended pregnancy, smoking in the year preceding pregnancy, binge drinking in the year preceding pregnancy and low dispositional optimism [which is defined as a global expectation that more good (desirable) things than bad (undesirable) things will happen in the future].^13,14 Independent predictors of alcohol consumption after pregnancy recognition included unintended pregnancy, pre-pregnancy body mass index < 25.0 kg/m², smoking in the year before pregnancy and binge drinking in the year before pregnancy.¹³ Common risk factors identified were unintended pregnancy and substance use behaviours before pregnancy and aligns with previous research.^15,16

Birth outcomes

Prediction of preterm birth

Whole blood gene expressions and clinical factors were associated with spontaneous preterm birth (SPTB) in asymptomatic women.¹⁷ Specifically, gene sets and pathways associated with inflammation were upregulated in women with SPTB compared with women who had term deliveries. In contrast, women with SPTB had lower RNA metabolism, RNA processing and T cell activation compared with women who had term deliveries. Significant clinical factors associated with SPTB included alcohol consumption during pregnancy, history of preterm birth, history of abortion, urinary tract infection during pregnancy and anaemia during pregnancy. A multivariate model integrating clinical data and gene expression data was found to predict SPTB with 65% sensitivity and 88% specificity in asymptomatic women.¹⁷

Stress and preterm birth

Cumulative psychosocial stress [defined as the combined effects of state anxiety during pregnancy (≥ 40 on the Speilberger State Anxiety Scale) and at least one of: (i) history of mental health disorder; (ii) history of abuse; or (iii) negative feelings toward their pregnancy] was a statistically significant independent risk factor for late preterm birth (34–36 weeks of gestation), controlling for demographic, obstetric, medical and lifestyle factors.¹⁸ This finding is consistent with previous research¹⁹ and aligns with the theory of allostatic load.^20,21 Cumulative psychosocial stress was an independent risk factor for preterm birth (< 37 weeks of gestation) among women with low levels of either social support or optimism, but not among women with higher levels of these factors.¹⁸ These findings suggest that a history of psychosocial vulnerability combined with current anxiety symptoms in pregnancy increases the risk for preterm birth.

Child development outcomes

Developmental delay

The proportion of children with delayed development at 1 year of age, defined as scoring in the ‘monitoring zone’ on at least two of the five developmental domains of the Ages and Stages Questionnaire (ASQ), was 17%.²² Key risk factors for delayed development at 1 year included maternal prenatal depression, preterm birth, low community engagement and non-daily parent-child interaction.²² Child development can be adversely influenced by poor maternal mental health²³ and socioeconomic status of the family.^24,25 Among children with these influences, our results suggest protective factors for developmental delay include higher parenting self-efficacy, relationship happiness, higher social support, community engagement and daily parent child interaction.²² These findings suggest that strategies such as those aimed at supporting mother’s mental health, increasing mother’s social support networks, parenting confidence, encouraging use of community resources and encouraging parents to engage in daily reading and play with their child, may positively influence child development at 1 year.

Late preterm (34 to 36 weeks of gestation) infants were more likely to demonstrate risk of developmental delay in communication and gross motor domains on the ASQ at 12 months compared with term infants (≥ 37 weeks of gestation).²⁶ Corrected age was used for preterm infants. Both late preterm and term infants who required neonatal intensive care admission were at greater risk of communication delay compared with infants who were not admitted, after adjusting for maternal and infant factors.²⁶ These findings suggest late preterm status and admission to neonatal intensive care can be used to identify infants at risk of developmental delay at 1 year. Early identification of risk provides an opportunity for developmental assessment and referral to early intervention programming.

Language delay

The prevalence of late talking (defined as scoring < 10th percentile on the MacArthur-Bates Communicative Development Inventories: Words and Sentences) in toddlers between 24 and 30 months of age was 12.6%.²⁷ Risk factors for late talking included male sex and a family history of late talking and/or diagnosed speech or language delay, whereas protective factors included providing informal play opportunities, reading and sharing books daily and receiving care at a child care centre, as compared with other forms of care which were not significant.²⁷ The protective factors identified in this study describe opportunities for language-based social interaction with a variety of communication partners, which may contribute to decreased risk for delayed early vocabulary.

What are the main strengths and weaknesses?

Key strengths of the All Our Babies (AOB) study include the recruitment and retention of a large community-based pregnancy cohort with comprehensive psychosocial, lifestyle and biological data spanning pregnancy to early childhood. The AOB study used standardized validated measures or scales for the majority of constructs, including those related to demographics, mental health and child development. Prospective data collection on social, psychological and contextual variables in addition to health-related data provides the opportunity to consider, from an ecological framework, the role of numerous factors on child development and maternal mental health. The longitudinal nature, with frequent early follow-up, enables the identification of sensitive windows to inform screening and intervention. The representativeness of the sample in terms of race and ethnicity is a strength. Further, the AOB pregnancy cohort is aligned with the Developmental Origins of Health and Disease (DOHaD), a field of research that examines how environmental factors interact with genotypic variation to alter the capacity of an individual human being to cope and adapt to its environment across the life span. The collection of biological specimens and environmental data from women during the prenatal period and early postpartum period provides the opportunity to undertake innovative longitudinal research that considers early gene-environment interactions as influencers on lifetime risk for disease and health. AOB took advantage of time-sensitive events, including the H1N1 pandemic and a local natural disaster, providing real-time data that could be linked to earlier exposures. To our knowledge, the AOB study is the only contemporary Canadian pregnancy cohort of its size with such extensive prenatal data related to demographics, lifestyle, maternal mental health, health service utilization, community resources, child care and biology coupled with its comprehensive collection of repeated child development measures over time.

A limitation of this study is the use of participant-completed measures to operationalize psychosocial health and child development outcomes as an alternative to clinical diagnoses or reports from other sources. Therefore, the outcomes are based on standardized screening tools and participants’ responses on mental health and child development measures. Consequently, the possibility for misclassification must be considered when interpreting the findings. Although rates of mental health challenges and past adversity are in alignment with published findings, the issue of generalizability is important. AOB includes and has retained urban women and those with higher incomes and who are married. Consequently, generalizability to those most vulnerable (young maternal age, food and housing insecurity) should be done with caution. An additional limitation was the collection of maternal and child data only. Ideally data would have been collected from fathers, to examine paternal contributions to child development outcomes as well. The participant questionnaires were detailed and comprehensive, to permit the collection of socio-demographic, psychosocial, health, contextual and child data. During pilot testing, on average participants completed the questionnaires in 30 min. Given the emphasis on child development in later questionnaires, time for completion increased in subsequent follow-up waves: as such, it is possible that participant fatigue during questionnaire completion may have contributed to attrition across time for some women. Finally, although significant efforts have been made to retain participants over time, we acknowledge that the longitudinal design and logistics related to timing of funding and development of follow-up questionnaires has impacted on study attrition over time and may be a source of selection bias, particularly for the follow-up questionnaires at 1, 2 and 3 years. The differences observed in the demographic characteristics of continuing participants (older maternal age, stable relationship, higher education, higher income, Canadian-born) compared with discontinued participants should be taken into consideration when interpreting the findings.

Can I get hold of the data? Where can I find out more?

The All Our Babies questionnaire and medical record data are stored at Secondary Analysis for Generating Evidence (SAGE), a secure data repository managed by the Alberta Centre for Child, Family and Community Research. Requests for data and collaborations are welcomed. Those interested in collaborations can contact Nikki Stephenson at the University of Calgary at []. For further information, please access our study website at [allourfamiliesstudy.com].

Profile in a nutshell

• AOB is a community-based longitudinal pregnancy cohort of mother-child dyads investigating maternal, birth and child development outcomes.

• A total of 3387 pregnant women (aged 19 to 47 years), residing in Calgary, Canada, enrolled and completed at least one questionnaire between May 2008 and May 2011. The majority were married, well educated, had annual household incomes > $60 000 and were born in Canada.

• Mothers have completed six questionnaires spanning pregnancy to 3 years postpartum and provided access to their medical records. A subgroup of participants provided two maternal blood samples during pregnancy and an umbilical cord blood sample at birth; 2449 mother-child dyads remain eligible for future follow-up.

• The AOB dataset currently includes comprehensive psychosocial data spanning pregnancy to 3 years postpartum, linkage to medical records and maternal and umbilical cord blood samples for a subgroup of the cohort. When complete, 5-year follow-up questionnaire data will be added and an 8-year follow-up is planned.

• Requests for data and collaborations are welcomed. Those interested in collaborations can contact Nikki Stephenson at the University of Calgary at [].

Funding

Initial funding to investigate maternal and infant outcomes during the perinatal period was provided by Three Cheers for the Early Years, Alberta Health Services. Expansion of the cohort, to include collection of biological specimens to predict determinants of adverse birth outcomes, was funded by Alberta Innovates Health Solutions [Interdisciplinary Team Grant #200700595] and the Global Alliance to Prevent Prematurity and Stillbirth [GAPPS award #12006]. Subsequent collaborations focusing on the developmental trajectories of AOB children have been supported by numerous grants, including the Social Sciences and Humanities Research Council, Alberta Health Services, the Alberta Children’s Hospital Foundation, the Alberta Centre for Child, Family and Community Research, Upstart the United Way of Calgary, Women and Children’s Health Research Institute, the Canadian Foundation for Fetal Alcohol Research and the Max Bell Foundation.

Acknowledgements

The authors acknowledge the contribution and support of AOB team members and our participants. We are extremely grateful to all the families who took part in this study and the whole All Our Babies team. We are extremely grateful to the investigators, coordinators, research assistants, graduate and undergraduate students, volunteers, clerical staff and managers.

Conflict of interest: None declared.