Article Text
PDF
Abstract
Objective To investigate the impact of a long-term nurse home visiting programme, embedded within a universal child health system, on the health, development and well-being of the child, mother and family.
Design Randomised controlled trial.
Setting/participants 208 (111 intervention, 97 comparison) eligible at-risk mothers living in a socioeconomically disadvantaged area in Sydney, booking into the local public hospital for confinement.
Intervention A sustained and structured nurse home visiting antenatal and postnatal parenting education and support programme.
Control Usual universal care.
Main outcome measures The quality of the home environment for child development (12–24 months), parent–child interaction and child mental, psychomotor and behavioural development at 18 months.
Results Mothers receiving the intervention were more emotionally and verbally responsive (HOME observation) during the first 2 years of their child's life than comparison group mothers (mean difference 0.5; 95% CI 0.1 to 0.9). Duration of breastfeeding was longer for intervention mothers than comparison mothers (mean difference 7.9 weeks; 95% CI 2.9 to 12.9). There was no significant difference in parent–child interaction between the intervention and comparison groups. There were no significant overall group differences in child mental, psychomotor or behavioural development. Mothers assessed antenatally as having psychosocial distress benefitted from the intervention across a number of areas.
Conclusion This sustained nurse home visiting programme showed trends to enhanced outcomes in many, but not all, areas. Specifically, it resulted in clinically enhanced outcomes in breastfeeding duration and, for some subgroups of mothers, women's experience of motherhood and children's mental development.
Trial registration number ACTRN12608000473369.
Statistics from Altmetric.com
Introduction
Long-term nurse home visiting interventions commencing antenatally and continuing for at least 2 years after birth have been shown in international trials and systematic reviews to be a promising strategy1 for improving maternal and child outcomes in early childhood and later life2,–,4 including child development,5 use of welfare,6 child abuse and neglect, and criminal behaviour.2 7,–,9 Based on this evidence, governments in Australia and internationally10 11 are investing in long-term home visiting for vulnerable children.
What is already known on this topic
▶ Long-term home visitation programmes for parents of young children to promote child development and prevent health and social problems seem to be effective.
▶ The effectiveness of sustained nurse home visiting interventions for vulnerable mothers delivered within a universal child and family heath service has been little studied.
What this study adds
▶ A randomised trial was conducted of an Australian nurse home-visiting intervention to child-age 2 years delivered within a universal maternal child and family health service.
▶ A long-term nurse home visiting programme embedded in a universal child and family health service system can improve family, maternal and child development outcomes.
However, there has been limited research investigating the impact of long-term home visiting programmes in Australia,12 13 and international trials have been conducted primarily in monocultural communities with most home visiting research occurring outside of a universal service delivery system.10
The effectiveness of this model of sustained nurse home visiting for families from immigrant communities, the impact of nurse home visiting when delivered as a component within a comprehensive child and family health and development service system,10 11 and the efficacy for older, multiparous compared with teenage first-time mothers, remains largely unexplored.
Recent findings suggest that, regardless of parity, the group benefitting most from nurse home visiting interventions are mothers living in impoverished areas who have lower psychosocial resources during pregnancy (despite the indices used to measure this).1 14 Trials that have had a broader client base, in terms of parity and risk, for example, the New Zealand Early Start programme,8 15 however, have not yet reported subgroup analyses.
We report the results of the first Australian randomised trial of a sustained nurse home visiting intervention commencing antenatally and continuing to child-age 2 years for families living in a multicultural, socioeconomically disadvantaged urban community. The Miller Early Childhood Sustained Home-visiting (MECSH) programme16 (figure 1) was embedded within the established local comprehensive universal maternal, child and family health services system. We hypothesised that compared with women who received usual universal care, women who received the MECSH programme would report improved family, maternal and child health and developmental outcomes. To determine which groups would benefit most from the intervention, we further hypothesised that the programme may have differential benefits for Australian-born and overseas-born mothers, first-time mothers and multiparous mothers, mothers with one risk factor and those with multiple risk factors, and mothers with psychosocial distress and those without distress at the time of recruitment antenatally. Finally, we hypothesised that the intervention would ameliorate health and developmental disadvantages for vulnerable children compared to the population.
MECSH programme model, goals and structure.
Methods
Setting and participants
A randomised controlled trial was conducted in a socioeconomically disadvantaged suburb in the greater Sydney area. Mothers were eligible to participate if they did not require the use of an interpreter, and reported one or more of the following risk factors for poor maternal or child outcomes in their responses to routine standardised psychosocial and domestic violence screening conducted by midwives for every mother booking in to the local hospital for confinement: maternal age under 19 years; current probable distress (assessed as an Edinburgh Depression Scale (EDS)17 score of 10 or more) (as a lower cut-off score was used than the antenatal validated cut-off score for depression, the term ‘distress’ is used rather than ‘depression’; use of this cut-off to indicate those distressed approximated the subgroups labelled in other trials as ‘psychologically vulnerable’ or as having ‘low psychological resources’14); lack of emotional and practical support; late antenatal care (after 20 weeks gestation); major stressors in the past 12 months; current substance misuse; current or history of mental health problem or disorder; history of abuse in mother's own childhood; and history of domestic violence. Eligible mothers were randomised to the intervention or comparison group prior to collection of baseline data (see the CONSORT diagram in figure 2). There were no intervention and comparison group differences in the proportion of mothers participating at any data point, nor in mothers' risk profile either between groups or between those participating and those not. The study protocol provides details on eligibility and processes of recruitment.16
CONSORT diagram. *Followed up includes all women who were being actively followed up at that time point, that is, those who completed the previous data point, that data point and those completing any subsequent data point. Participated includes only those who completed that data point.
Intervention
All women received usual antenatal midwifery, obstetric and birthing services. Comparison group women were expected to receive a home visit by a child health nurse within 2 weeks of giving birth, in accordance with standard practice in New South Wales. Women in the intervention group received an average of 16.3 (range 0–52) visits, each of 60–90 min duration, by a child health nurse commencing at on average 26 weeks gestation (range 12–40), and continuing to their child's second birthday (average duration of participation in programme to child-age 57.0 weeks (range 0–122): 82% visited antenatally, 95% visited in the first year postnatally, and 53% visited in the second year).
Measures and procedures
Data were collected antenatally on recruitment, and then at 1, 6, 12, 18 and 24 months postnatally. Primary outcomes measured during the trial period to child-age 2 years were:
▶ The quality of the home environment for child development indexed by aspects of parental care giving, measured by observation and maternal interview in the home at child-age 12 and 24 months using the HOME Inventory.18
▶ Parent–child interaction during free play (sensitive stimulating parenting, detached flat parenting, child engagement) observed in a structured clinic environment at child-age 18 months, measured using the National Institute for Child Health and Development scales of parent–child interaction.19
▶ Child mental, psychomotor and behavioural development at 18 months, measured using the Bayley Scales of Infant Development II.20
Secondary outcomes were:
▶ Perinatal outcomes: low birth weight and preterm birth, extracted from hospital clinical records.
▶ Child respiratory and gastrointestinal illness and immunisation status, measured by maternal self-report at child-age 6, 12, 18 and 24 months, coded dichotomously as the child having experienced at least one episode or illness or not during the first 2 years of the life, and being fully immunised or not at child-age 24 months.
▶ Maternal and household smoking during pregnancy and over the first 2 years of life, measured by maternal self-report when pregnant and at child-age 1, 6, 12, 18 and 24 months, coded dichotomously as mother ever smoking or not, having ever ceased smoking or not (for mothers who smoked), and whether the household was always reported as smoke-free or not during the first 2 years of the child's life.
▶ Breastfeeding over the first 2 years of life, measured by maternal self-report at child-age 1, 6 and 12 months, and collated as total duration (in weeks) of breastfeeding, and age (in weeks) at introduction of solids.
▶ Maternal health (SF-12),21 depression (EDS)17 and social support,22 the number of stressful life events in the previous year (modified Life Events Inventory)23 24 and family functioning (McMaster Family Assessment Device)25 measured by maternal self-report at child-age 1, 12 and 24 months.
▶ Experience of being a mother (Being a Mother),26 assessing women's satisfaction with the role of motherhood, measured at 18 and 24 months.
Maternal self-report and HOME Inventory data were collected by a research assistant who was blind to treatment assignment for baseline and postbaseline assessments, unless inadvertently revealed by the participants. Laboratory parent–child interaction assessments and child development outcomes were undertaken by research staff blinded to group allocation.
Missing data
Cross-sectional data (parent–child interaction, child development) were analysed only for those participants who completed the data point. Child feeding data were collated only for those participants who had reported an age at which they ceased breastfeeding and commenced solids. Smoking data were collated for all participants who had answered questions regarding smoking on at least one data point.
Missing data for longitudinal analysis (HOME Inventory, maternal health, depression, social support, stressful life events, family functioning and experience of being a mother) were dealt with using a three-step procedure to provide a balance between maintaining study power and minimising bias in parameter estimates.27 28 First, participants who had not completed any data points for these outcomes were deleted from analysis. Second, data were imputed for each participant who had completed some but not all of the data points using expectation maximisation (EM) imputation. EM imputation methods used provided a value for missing data points for each individual participant based on both the estimated sample parameters and the actual responses at completed data points, preserving sample size and making use of all available data, while maintaining the variance within the sample. The EM analysis was also used to determine that the data for each outcome were likely to be missing at random. Finally, analyses were conducted and compared using both the imputed data set, and a subset of data containing only participants with complete data, and indicated no bias resulting from imputation. Missing values were then replaced by imputed values in the analysis.
Statistical analysis
Analysis was intention to treat. Differences between the intervention and comparison groups were assessed using independent t tests for cross-sectional parametric analyses, repeated measures analysis of variance (ANOVA) for longitudinal parametric analyses, and contingency tables for non-parametric analyses. Four a priori analyses were undertaken comparing outcomes for the hypothesised subgroups using 2×2 factorial ANOVA to assess main (intervention vs comparison group) effects and interaction (intervention by subgroup effects for mothers who were Australian-born vs overseas-born, first-time vs multiparous, had one vs multiple risk factors, and scored EDS ≥10 (antenatal psychosocial distress) vs EDS <10), while maintaining the power of the whole sample. Child development and perinatal outcomes were compared with population norms using one-sample t test and binomial analyses.
For all analyses two-tailed tests were undertaken: findings with α<0.05 were determined to be statistically significant; findings with α<0.1 are reported as trends. Effect sizes (ESs) were calculated for all trends and statistically significant findings; (ES≈0.5 (d)) was considered clinically meaningful. All main (group) effects are reported, together with significant findings and trends in interaction (subgroup) effects.
Ethics
The trial received approval from both the Sydney South West Area Health Service (Western Zone) and the University of NSW Human Ethics Committees.
Results
There were 208 participants in the study. Maternal risk and demographic characteristics are outlined in table 1. There were no significant demographic or risk profile differences between the groups at any data collection point. First-time mothers in the MECSH trial had a mean age of 24.6 years (SD 5.8) compared with multiparous mothers' mean age of 29.1 years (SD 6.0). Only 25 of 65 first-time mothers were teenagers.
Baseline maternal demographic and risk characteristics of the trial groups
Home environment
Overall, mothers who received the intervention were more emotionally and verbally responsive when observed at 12 and 24 months than mothers in the comparison group (table 2). Intervention group mothers who had been psychosocially distressed antenatally provided a significantly more emotionally and verbally responsive environment than psychosocially distressed comparison group mothers. First-time mothers, overseas-born mothers and mothers with more than one risk factor in the intervention group also were more emotionally and verbally responsive than their comparison group counterparts; however, the effect sizes were small.
Quality of the home environment from a child development perspective, as measured by the HOME Inventory18 (12–24 months)
There were no significant main intervention effects for other components of the quality of the home environment. There were, however, significant interaction effects indicating differences between some subgroups of intervention and comparison mothers in provision of appropriate play materials, organisation of the physical and temporal environments, and provision of opportunities for variety in daily stimulation (table 2). There were no group or subgroup differences in avoidance of restriction and punishment.
Parent–child interaction and child development
Overall, there were no significant differences in laboratory assessments of sensitive stimulating parenting, detached flat parenting or child engagement between the intervention and comparison groups, or subgroups, at 18 months (table 3). There were no significant main effects for the intervention in child mental, psychomotor or behavioural development. There was, however, a clinically meaningful difference in the mental development of children of mothers who were psychosocially distressed antenatally in the comparison group, who had poorer mental development compared with their intervention group counterparts. Children of the psychosocially distressed and non-distressed mothers in the intervention group, and the non-distressed comparison group mothers, were not significantly different to the general population norm of 100 (SD 15).20 The mean score for children of the antenatally psychosocially distressed mothers in the comparison group, however, was statistically different to the population norm (t=3.522, p=0.003, power 94%, ES (d)=0.92), with 47.4% of these children being delayed (ie, Mental Developmental Index <85).
Parent–child interaction and child development outcomes at 18 months (n=108)
Child health, maternal and family outcomes
Tables 4 and 5 show child health, maternal and family outcomes. Only nine infants in the study (4.5%) were of low birth weight (<2500 g), with no significant group difference. The proportion of low birthweight infants in this sample was lower than reported nationally (6.4%). There was also no significant difference in the proportion of infants born preterm, which was consistent with the proportion reported nationally (8.2%).
Child, maternal and family outcomes
Child feeding, maternal and family outcomes
Children in the intervention group were breastfed for significantly longer than children in the comparison group. This difference was attributable to overseas-born mothers in the intervention group feeding for significantly longer than overseas-born mothers in the comparison group. Australian-born mothers in both the intervention and comparison groups breastfed for an average of 10.3 (SD 11.1) and 5.5 (SD 5.0) weeks, respectively. Both groups commenced children on solids at an average age of 5 months.
Overall, there was no difference between the intervention and comparison group participants' experience of being a mother. Mothers who were psychosocially distressed antenatally, first-time mothers and mothers born overseas who received intervention, however, were more likely to report a more positive experience of being a mother than those same subgroups of mothers in the comparison group. There were no significant group or subgroup differences in maternal health, social support or family outcomes.
Discussion
This randomised trial of the MECSH programme demonstrated several beneficial outcomes for children and their mothers; however, it was limited by the size of the sample. As it was an intervention programme embedded in an existing service, the number of participating nurses, and hence mothers, was limited by the funding, infrastructure and management capacity within the service. Funding was provided for three nurses, who could each manage a caseload of 25 families at any given time. This small sample size has limited the power of some analyses, and also limited the capacity to assess the effectiveness of the MECSH programme for women with more complex risks such as domestic violence or drug and alcohol issues. The analyses were also limited by low participation at some data collection points.
Nevertheless, the MECSH trial showed some significant results and some trends that require replication in larger samples of mothers drawn from a similarly widely defined at-risk group, including older, multiparous mothers, and mothers with higher levels of education than have been reported in other trials.1 14 Mothers of infants and toddlers in the intervention group provided a home environment that was statistically significantly more supportive of their child's development through more verbal and emotional responsivity; however, the effect size was small. Intervention group children were breastfed longer, particularly those of overseas-born mothers, and the subgroup of children of mothers who had been psychosocially distressed antenatally had clinically better mental development scores than their counterparts from the comparison group.
Mothers assessed antenatally as having psychosocial distress showed benefit from the MECSH programme across a number of areas, including child development, their experience of being a mother, and small effects in a number of domains of the quality of the environment from a child development perspective; emotional and verbal responsivity, organisation of the environment and provision of appropriate play materials. These results are similar to those found in other sustained nurse home visiting studies,1 14 although the intervention impacted on a broader range of domains of the home environment for this subgroup of women than has been reported previously.1 An increasing body of evidence from both animal and human studies suggests that stress in pregnancy has significant impacts on developmental and behavioural outcomes for children.29 While the mental development of children of mothers who were not distressed antenatally in both the intervention and comparison groups was comparable with the general population, children's development was particularly poor in the distressed subgroup in the absence of the MECSH intervention, suggesting that sustained nurse home visiting may be particularly effective in ameliorating some adverse developmental impacts for children of mothers with antenatal distress. However, this needs replication in a larger sample.
Overseas-born mothers showed benefit in the duration of breastfeeding, their experience of being a mother, and small effects for emotional and verbal responsivity, although benefits were greater for Australian-born mothers in the provision of appropriate learning materials. Benefit accrued particularly for first-time mothers in their experience of being a mother, and in the two HOME subscales of provision of appropriate learning materials and emotional and verbal responsivity. The small effects in emotional and verbal responsivity reflect the findings for studies focused on teenage, first-time mothers only, albeit in this case with an older group of primiparous women. The outcomes for higher risk (two or more) compared with lower risk (one risk only) mothers showed small benefits in responsivity, organisation of the environment and provision of appropriate play materials. In these domains, the intervention improved the outcomes for higher risk women, who otherwise would do more poorly compared to lower risk women.
Conclusion
These findings are important in the context of the broader implementation of sustained nurse home visiting programmes. The MECSH programme had many features similar to other successful programmes1 15: it was delivered by professional nurses, was research based, and paid high attention to fidelity, quality and supervision. There were, however, three key differences between MECSH and other research programmes. First, the MECSH programme was embedded in and delivered, managed and supervised by the established clinical service, within a comprehensive universal child and family health services system. The findings reported here demonstrate that outcomes of a magnitude similar to stand-alone research trials are possible in the context of established services, and hence are likely to be replicable in widespread implementation of programmes that also employ rigorous monitoring processes.
Second, the MECSH programme showed that in many outcome areas there were no differences in outcomes for first-time and multiparous mothers. This challenges the assumptions of many nurse home visiting programmes that their benefits are confined to first-time mothers.1
Finally, the MECSH programme, unlike other trials, identified mothers with psychosocial distress during pregnancy using a population-based screening tool (EDS) rather than identifying them post hoc using specific research-based measures.1 14 An issue to be faced in widespread implementation is how mothers with lower psychosocial resources can be identified in the population.1 14 The MECSH trial demonstrates that this group can be easily identified and enrolled in effective intervention programmes embedded with comprehensive services at the population level.
Acknowledgments
The authors would like to acknowledge the contribution of the child and family health nurses who delivered the intervention to the families, research assistants involved in the data collection and entry (Roberta Chavez, Sheryl Scharkie, Astrid Toscan and Amanda Kemp). The authors also thank Fiona Byrne for assistance in the preparation of this paper for publication, and thank participating families.
References
Footnotes
-
Funding The trial was funded by the Australian Research Council (LP0560285), Sydney South West Area Health Service, the NSW Department of Community Services and the NSW Department of Health.
-
Competing interests None.
-
Ethics approval The trial received approval from both the Sydney South West Area Health Service (Western Zone) and the University of NSW Human Ethics Committees.
-
Provenance and peer review Not commissioned; externally peer reviewed.