Statistics from Altmetric.com
Randomised controlled trial with 15 years follow up.
Public antenatal clinic and private obstetrics offices in a semirural area of central New York State, USA.
400 women (11% African-American) who were <25 weeks gestation, had no previous live births, and had ≥1 sociodemographic risk factor (<19 y, unmarried, or low socioeconomic status). Most were unmarried (62%), ≥19 years (52%), and had low socioeconomic status (59%). Follow up was 81%.
Stratified by maternal race, marital status, and geographic region, women were allocated to 1 of 4 treatment groups: (1) sensory and developmental screening for children at 12 and 24 months (n=94); (2) treatment 1 plus free transportation to prenatal and child health care until the child's second birthday (n=90); (3) treatment 2 plus prenatal nurse home visits (mean 9 visits) (n=100); and (4) treatment 3 plus home visits (mean 23 visits) for 2 years after delivery (n=116).
Main outcome measures
Substantiated reports of child abuse or neglect, subsequent births, duration of welfare benefits, and mothers' substance abuse, arrests, and convictions.
Analysis was by intention to treat, and adjusted for sociodemographic variables. Women who received home visits until 2 years after delivery (group 4) were compared with women who did not (groups 1 and 2). Nurse home visits reduced state verified cases of child abuse and neglect (log incidence difference 0.77, p<0.001). No differences were found for subsequent births; months that women received welfare; behavioural impairment because of substance abuse; arrests; or convictions. Subgroup analysis of women who were unmarried and from low socioeconomic status households (40%) showed that home visits reduced child abuse and neglect (log incidence difference 1.61, p<0.001), subsequent births (mean difference 0.5 births, p=0.02), months that women received welfare (mean difference 29.9 mo, p=0.005), behavioural impairment because of substance abuse (log incidence difference 0.58, p=0.005), arrests (log incidence difference 1.74, p<0.001), and convictions (log incidence difference 1.65, p<0.001).
Prenatal and postnatal home visits by nurses reduced the incidence of child abuse and neglect over a 15 year period. Home visits to unmarried women from low socioeconomic households also reduced subsequent births, use of welfare, substance abuse, arrests, and convictions.
These 2 excellent trials by Kitzman et al and Olds et al add strength to the large, often poor quality, studies on home visiting with pregnant women and young families. The replication of Olds' 1970s home visiting trial by Kitzman et al shows that such programmes are successful with disadvantaged groups and can be successfully administered in a public health setting. The 15 year follow up by Olds et al provides new information about the effects on women, families, and society.
Kitzman et al confirm that prenatal home visits by nurses can improve the health of pregnant women. The failure to improve birth weight is consistent with previous studies,1 as is the lack of change in preterm births, which is probably explained by racial differences in risk profiles and greater surveillance of intervention groups.2 Nevertheless, the improvements in maternal hypertension and kidney infection are important, independent of effects on fetal health.
The study by Kitzman et al confirms the effects of postnatal visits on child health and parenting,3 in particular, the contribution of home visits to childhood injury prevention.4, 5 The study by Olds et al fills a gap in earlier research by including outcomes for women as well as children. A change in outcomes such as verified cases of child abuse and criminal behaviour is noteworthy. However, Kitzman et al did not show short term effects on maternal education and employment. These outcomes are difficult to change, and long term follow up will be important.
These 2 studies overcame many problems common to community research. They had high participation and follow up rates, concealed and random allocation, well designed and blinded outcome measurement, and clear theoretical foundations. The subgroup analysis by Olds et al of low income, unmarried women was not initially part of the design and should be interpreted with care. However, Kitzman et al improved on the initial trial by increasing the sample size, adding 3 important factors to stratification and hypothesising a greater effect on women with few psychological resources. These 2 studies strengthen the conclusion that teaching, counselling, family support, and referral services provided by nurses to first time mothers in their homes change serious short and long term outcomes. Programmes which offer a similar integrated service, with adequate contact time by well trained nurses, who individualise the intervention and address the full context of women's lives, are likely to change the prospects of disadvantaged young women and children. These results, however, cannot be generalised to other programme designs.
Managers need to scrutinise home visiting programmes in light of those proved to be effective, looking at theoretical linkages, risk factors, and the specific quality and quantity of service. Nurses providing home visits must examine the content of their interventions, incorporate the full range of educational and supportive elements, and continue visits throughout infancy.
This research reinforces the importance of adequate funding for home visiting for mothers with few resources. The public health significance of the changes in the intervention group in the study by Olds et al should not be ignored: every third woman chose to limit her family by 1 child and every family received welfare for 1 year less. Although the changes were not statistically significant for the total sample, at a population level such changes would be enormous.
Sources of funding: National Institute of Mental Health and US Department of Health and Human Services.
For article reprint: Dr D L Olds, University of Colorado Health Sciences Center, 1825 Main Street, Denver, CO 80218, USA. Fax +1 303 864 5236.
A modified version of this abstract appears in Evidence-Based Mental Health 1998 May and Evidence-Based Medicine 1998 May-Jun.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.