Home visiting intervention for vulnerable families with newborns: follow-up results of a randomized controlled trial☆
Introduction
EFFORTS TO ADDRESS child abuse and neglect range from prevention and early intervention programs, to out-of-home care services for children removed from dangerous environments. Although treatment for parents and their children remains an important and necessary part of comprehensive service delivery, a broader approach is needed due to the inevitable limitations of interventions that take place after child abuse and neglect have occurred. In comparison, early intervention and prevention strategies seek to assist families in developing skills and knowledge that, over time, may prevent the emergence of dysfunctional parenting behaviors. As such, preventive and early intervention programs are valued not just for their potential to avoid human suffering, but to reduce need for the wide range of services required to address the consequences of child abuse and neglect (James, 1994).
During the past three decades researchers and clinicians alike have attempted to determine appropriate and effective methods for the prevention of child abuse and neglect. The approach used by child protection services is usually to operate at different levels of prevention due to the complexity of personal and social systems factors that predispose families to intrafamilial violence (Browne & Herbert, 1997) and the scarcity of resources available to service providers. Many programs aimed at preventing abuse and neglect (by improving parenting competence, knowledge, skills, and parent-child attachment) have been home-based and controlled trials have demonstrated the value of home-visiting by nurses. Although program diversity has precluded formal meta-analyses of the published literature, substantive and methodological conclusions can be drawn. The more successful preventive interventions tested to date have been conducted with selected, targeted populations MacMillan et al 1994, Olds and Kitzman 1993, Roberts et al 1996, Wekerle and Wolfe 1993. In particular, the home visiting strategy appears to be most effective when targeted to families with a range of characteristics placing them at risk for parenting dysfunction, child abuse, or neglect. Follow-up studies of a successful early trial conducted in the United States (Olds, Henderson, Tatelbaum, & Chamberlin, 1986) examined a diverse range of outcomes from maternal life course, child abuse, and neglect, to children’s criminal and antisocial behavior Olds et al 1997, Olds et al 1998. These longitudinal studies indicated that for disadvantaged families in the United States, intensive home visiting by nurses in the first 2 years of life effectively improved both short and longer term child, parent, and family well-being.
In Australia, program evaluation has been constrained by a number of factors, foremost being the relative low incidence of and difficulty measuring child abuse and neglect and the lack of built-in evaluation of otherwise well-meaning interventions (Vimpani, Frederico, & Barclay, 1996). Furthermore, it has been difficult to control for improved outcomes brought about by changes in wider social family policies (Vimpani et al., 1996). Finally, although the ability to identify and engage appropriate target populations underpins the success of preventive strategies, one of the most consistent themes of the child abuse and neglect literature has been the challenge of recruiting and retaining higher-risk families. Specifically, the problem for community health service providers is that families at risk for child abuse and neglect are frequently among those least likely to access health services. These families are often isolated and marginalized from traditional services due to their personal and/or socioenvironmental characteristics Ferguson et al 1984, Ford et al 1990, Hart 1971. Thus, screening families for potential child abuse and neglect characteristics requires careful assessment based on empirical evidence for risk.
This study sought to address identified limitations within the literature with respect to home visiting. The overall aim of this study was to evaluate the effectiveness of home visiting as a prevention and early intervention strategy using child health nurses, social workers, and parent aides. In accordance with this aim, the hypothesis that the home visiting program would be associated with more favorable adjustment to the parenting role compared with nonintervention, or clinic-based service delivery was tested. Specifically, the following research questions and hypotheses were put:
- 1.
Would a screening instrument used during the immediate postnatal period to identify families with child abuse or neglect risk factors demonstrate community utility?
- 2.
Would a home visiting intervention program beginning in the immediate postnatal period and targeted to families with child abuse or neglect risk factors demonstrate high social validity?
- 3.
The hypothesis was tested that more favorable adjustment to the parenting role would be associated with a home visiting intervention program compared with nonintervention or clinic based service provision only and would be evidenced by: (a) lower rates of maternal depression, parenting stress, and child abuse potential; (b) higher scores for home environment and parent-child interactions observed in the home; and (c) higher levels of parental knowledge and practice of child safety and preventive health care.
- 4.
The hypothesis that maternal, family and environmental factors identified in the immediate postnatal period would be predictive of adjustment to the parenting role at follow-up was tested.
Section snippets
Participants
One hundred and eighty one women with newborns were initially enrolled to participate in the randomized-controlled trial designed to test these hypotheses. Women were recruited from a major urban teaching and tertiary referral hospital in Queensland between 11 January 1996 and 9 September 1996. Women were selected for inclusion to the study on the basis of two criteria sets. Women in the immediate postnatal period were approached personally to consent to participate if they: (a) had at least
Home environment
The 45-item semi-structured observation/interview infant version of the Home Observation for Measurement of the Environment (HOME) Inventory was used to measure level of developmentally simulating experiences available for the child and overall quality of the home environment from a developmental perspective at infant ages 6 weeks and 12 months (Caldwell & Bradley, 1984). The HOME Inventory is a standardised measure of quality of the home environment using binary choice items clustered into six
Child health
A 28-item, 12-month child health questionnaire was developed to examine parental knowledge and practice of preventive childcare. Items measuring parental distress related to childhood feeding and sleeping variants were included in this questionnaire. Parental perception of their own child’s eating and feeding behavior was measured using a 10-point scale from “no problem” to “major problem.” Similarly, participants were asked to rate perception of their own child’s sleeping pattern using a
Utility of the screening tool
Of the 636 (63.09%) self-report questionnaires returned by women in the immediate postnatal period during recruitment, 614 (96.5%) completed the question “Are you a sole parent?” Of these, 15.0% (n = 92) reported sole parent status. Overall, 4.2% (n = 27) of respondents reported having sought termination of their pregnancy. Five hundred and ninety four (93.4%) respondents completed questions related to domestic violence. Of these, 7.2% (n = 43) reported being the victim of at least one form of
Consent to participate in the immediate postnatal period
Of the 636 (63.09%) women approached in the immediate postnatal period who returned completed inclusion questionnaires, 463 (72.8%) gave written consent to participate in the longitudinal trial. Characteristics reported by women who gave consent to participate in the project longitudinally were compared with characteristics reported by women who wished only to complete the inclusion questionnaire. Data were analysed using Statistical Package for the Social Sciences (SPSS) for Windows Release
Parenting stress index
For PSI data, a repeated measures MANOVA (SPSS) design was used to test for differences over time between groups. Parent domain scores indicated improvement over time for parenting stress associated with symptoms of postnatal depression, parent-infant attachment, parent’s sense of competence, and parent health. Specifically, results of the repeated measures MANOVA showed a significant overall main effect for time for: (a) symptoms of postnatal depression, F(2, 260) = 3.09, p < .05; (b)
Home environment
At 6 weeks, the HOME Inventory subscale that measured provision of appropriate play materials became a dichotomous variable when redundant items were removed (to make it age appropriate) and was thus tested using the chi-square statistic (SPSS). Each subscale of the HOME Inventory and the total HOME scale was tested for interaction with parity and tested separately for interaction with baseline CAPI abuse scale score. No significant interaction effects between group and parity, or group and
Child health
At 6 weeks, no statistically significant differences were found between the groups for utilisation of other community services (Armstrong, Fraser, Dadds, & Morris, 1999). At 12 months, means for perception of child’s feeding behaviors for intervention group and comparison group were 2.28 and 1.78 (SDs = 2.76 and 2.29, respectively). Means for perception of child’s sleeping patterns for intervention group and comparison group were 2.61 and 2.10 (SDs = 3.03 and 2.51, respectively). The higher
Service satisfaction
Statistically significant group differences were found for each scale used to measure satisfaction with community child health services. Greater satisfaction for the home visiting program group compared with comparison group participants accessing standard clinic-based services was found. ANOVA detected significant group differences for communication, F(1, 148) = 12.08, p < .05; convenience, F(1, 144) = 16.13, p < .05; interpersonal manner, F(1, 146) = 20.42, p < .05; general satisfaction, F(1,
Prediction for high parenting stress at 12-month follow-up
Table 9 presents results of the linear regressions using scores for total PSI and EPDS at 12-month follow-up and CAPI at 18-month follow-up as dependent variables. First, for total PSI at 12-month follow-up, block 1 resulted in no significant prediction. However, the addition of block 2 variables resulted in a significant equation [F (4,64) = 22.62, p < .001] and accounted for 61% of the variance. Addition of the block 3 variables improved prediction to 63% of the variance [F (1,63) = 3.30, p
Discussion
The self-report questionnaire was given to women at a time when they were steeped in adapting to their role as a parent and at a time when a plethora of information was being presented to them. The high return rate of this questionnaire indicated that risk for serious response bias was negligible. However, caution must be applied when attempting to draw conclusions beyond the context of recruiting for inclusion to a research project. The ability to identify vulnerable families in order to
Conclusions
The ability to identify depression levels, stress levels, and coping skills of all new parents would offer a possibility for preventing some cases of child abuse and neglect in this community. For community child health nurses undertaking routine home and family appraisals in the postnatal period, these aspects of assessment are important. This study indicates that community practitioners would be best advised to assess these variables rather than make intuitive assumptions of risk. Future
Acknowledgements
The authors wish to acknowledge the valuable contribution made by participating families and the following individuals: Joan Mee, Kathryn Scroope, and Lois Nielson (child health nurses; Cathy Huston and Bronwyn Herbert (social workers); Kathy Whitefield and Rebecca Hoge (research assistants); and the Royal Women’s Hospital, Brisbane. The research for this paper is attributable to the Community Child Health Service, Royal Children’s Hospital and District Health Service, Queensland, Australia;
References (42)
- et al.
Stressed parents with infantsReassessing physical abuse risk factors
Child Abuse & Neglect
(1999) - et al.
Validation of the Edinburgh Postnatal Depression Scale (EPDS) in non-postnatal women
Journal of Affective Disorders
(1996) Assessing physical child abuse riskThe child abuse potential inventory
Clinical Psychology Review
(1994)- et al.
Defining and measuring patient satisfaction with medical care
Evaluation and Programme Planning
(1983) - et al.
Prevention of child physical abuse and neglectPromising new directions
Clinical Psychology Review
(1993) Parenting Stress Index (3rd ed.)
(1990)- et al.
A randomized, controlled trial of nurse home visiting to vulnerable families with newborns
Journal of Paediatrics and Child Health
(1999) - Australian Bureau of Statistics. (1992). 1989–1990 National health survey: Children’s immunisation, Australia....
- et al.
The Edinburgh Postnatal Depression ScaleValidation for an Australian sample
Australian and New Zealand Journal of Psychiatry
(1993) - et al.
Assessment of the home environment
Preventing family violence
Home observation for measurement of the environment
Quantitative methods in psychologyA power primer
Psychological Bulletin
Detection of postnatal depressionDevelopment of the 10-item Edinburgh postnatal depression scale
British Journal of Psychiatry
Australia’s mothers and babies 1995
Programs that mitigate the effects of poverty on children
Future Child
Evaluation of Hawaii’s Healthy Start Program
Future Child
The utilization of preschool health and education services
Social Science and Medicine
Plunket contacts in the first year of life
New Zealand Medical Journal
Links between early post-partum mood and post-partum depression
British Journal of Psychiatry
The inverse care law
Lancet
Cited by (0)
- ☆
Funding bodies for this research include the Community Child Health, Royal Children’s Hospital and District Health Service, The Abused Child Trust (Queensland), The Creswick Foundation (Victoria), and the National Health and Medical Research Council (Australia).