According to the Coalition for Evidence-Based Policy:
The primary evidence for the Nurse-Family Partnership (NFP) model's effectiveness consists of three randomized controlled trials (RCTs), in which low-income mothers were (or were not) assigned nurses based on lottery, and then researchers followed both those who did and those did not receive nurse visits, collecting a broad range of data. We find the evidence from these studies to provide a compelling case that the model in question improves life outcomes for children and mothers, across a variety of regions and population profiles.
A thorough literature review of home visitation programs points to NFP as a standout among such programs in terms of the methodological strength and encouraging implications of its studies.2 In addition, NFP is one of the seven social programs that meet the Coalition for Evidence-Based Policy's Top Tier evidence standard.3
In our discussion below, we only report results that are statistically significant at the .05 level.
The first RCT of the NFP model, in Elmira County, NY, followed families for 19 years after birth and found that youth who had participated in the NFP program were about half as likely to ever have been arrested or convicted as youth in the control group.5 The follow up 15 years after birth (but not the one 19 years after) measured child abuse and neglect and found that the families who received nurse-visits had about half as many reports of child abuse and neglect as the control group.6
As of August 2010, this evaluation has published follow ups from 3, 6, and 9 years after birth. The 6-year follow up found that children of visited mothers had better vocabulary and intellectual functioning, and were reported by their mothers to have fewer behavioral problems compared to the control group.7
The follow up at age 9 found no statistically significant effects on children for the whole sample,8 but it did find effects for mothers in the treatment group. Compared to the control group, nurse-visited mothers had 13% fewer live births, had been with their current partner 18% longer, and had spent 12% less time on welfare and 10% less time on food stamps during the nine years after their child's birth.9
As of August 2010, the Denver study had followed children and mothers for 4 years post-birth. The study found no effects on the full sample of children.10 For mothers, the study found that about half as many nurse-visited mothers reported domestic violence from partners in the 6-month period prior to the interview as did mothers in compared to the control group. It also found that nurse-visited mothers had greater intervals between the births of their first and second children, and reported enrolling their children less frequently in preschool.11
We are in general concerned about researchers' conflicts of interests, which could bias results or lead to publication bias (the possibility that studies are more likely to be published when they show positive results).12
In the case of the NFP program, David Olds is the co-author of all publications from the three main studies of NFP discussed above, and also initiated development of the program.13
According to the NFP National Service Office, the cost to a community/agency averages $4,500 per family per year served in the NFP program, and ranges from $2,914 to $6,463 per family per year.14 These figures imply a typical cost of $10,800 per family served in the NFP program.15 We do not know the details of what costs were included in these figures.
Coalition for Evidence-Based Policy, "Nurse-Family Partnership."
“The following are the main conclusions of this review:
”¢ The popularity of home visiting has been driven by the results of a few studies of programs such as the Nurse-Family Partnership that demonstrate long-term benefits for parents and children.
”¢ However, there are very few long-term studies of home visiting programs.” Gomby 2005, Pg 2.
“The popularity of home visiting has been propelled by the findings of large and long-term benefits in a few studies (most notably, the studies of the Nurse-Family Partnership).” Gomby 2005, Pg 9.
“A hallmark of this program [Nurse-Family Partnership] is its use of research to determine program effectiveness and to improve services. Program implementation is monitored carefully in each site, using a management information system that is integral to the program. Longitudinal follow ups are being conducted in Elmira, Memphis, and Denver.” Gomby 2005, Pg 65.
“The Nurse-Family Partnership meets the Top Tier Evidence standard, defined per recent Congressional legislation to include: Interventions shown in well-designed and implemented randomized controlled trials, preferably conducted in typical community settings, to produce sizeable, sustained benefits to participants and/or society.” Coalition for Evidence-Based Policy, “Nurse-Family Partnership.”
As of October 5, 2010, there were 7 interventions labeled "Top Tier" on the web site Coalition for Evidence-Based Policy, "Social Programs that Work."
GiveWell, "Summary of Nurse-Family Partnership Study Results."
The publication for the follow up when the children are 19 years old is Eckenrode et al. (2010). When considering the whole sample, it finds the following statistically significant (at 5% significance level) effects of NFP on the 19-year olds:
The publication from follow up when the children are 15 years old is Olds et al. (1997). When considering the full sample, it finds only one statistically significant (at 5% significance level) effects on mothers and child abuse and neglect of NFP: the families with who got nurse-visits had fewer substantiated reports of child abuse and neglect. The average incidence was 0.29 in the treatment group and 0.54 in the control group (P
The publication for the follow up when the children were 6 years old is Olds et al. (2004a). When considering the whole sample, it finds the following statistically significant (at 5% significance level) effects of NFP on the 6-year olds:
"Nurse-visited children had higher scores on tests of intellectual functioning and receptive language (92.34 vs 90.24, P = .03, and 84.32 vs 82.13, P = .04, respectively) and were reported by their mothers to have fewer problems in the borderline or clinical range of the CBCL Total Problems scale (1.8% vs 5.4%, P = .04)." Olds et al. 2004a, Pgs 1554, 1556.
"Effects on the first-born children of nurse-visited women at age 9 (versus the control group):
Coalition for Evidence-Based Policy, "Nurse-Family Partnership."
For the full sample, the 9-year follow up found (at least) the following statistically significant (at 5% significance level) effects on mothers:
"12% less time on welfare during the nine years (5.2 months per year for the nurse-visited women vs. 5.9 months per year for control group women). 10% less time on food stamps during the nine years (7.0 months per year vs. 7.8 months per year). 13% fewer subsequent live births (an average of 0.81 births vs. 0.93).... 18% more time with their current partner (an average of 61.6 months vs. 52.4 months)." Coalition for Evidence-Based Policy, "Nurse-Family Partnership."
Olds et al. 2004b, Table 4, Pg 1565.
The publication for the follow up when children were 4 years old is Olds et al. (2004b). When considering the full sample, it found the following statistically significant (at 5% significance level) effects on mothers:
Olds et al. 2004b, Pgs 1563 and 1565.
For more, see our blog post on publication bias.
"Nurse-Family Partnership is founded on the pioneering work of David Olds, professor of pediatrics, psychiatry, and preventive medicine at the University of Colorado Denver. While working in an inner-city day care center in the early 1970s, Olds was struck by the endemic risks and difficulties in the lives of low-income children. He realized the children needed help much earlier—at home, with their mothers, when they were infants, and even before they were born. Olds' determination to help young children and families get a better start in life led to the development of a nurse home visitation program for first-time, low-income moms and their children.
Over the next 30 years, he tested the program in randomized controlled trials with three different populations: Elmira, New York, in 1977; Memphis, Tennessee, in 1988; and Denver, Colorado, in 1994."
Nurse Family Partnership National Service Office, "Program History."
"Built upon the pioneering work of Professor David Olds, Nurse-Family Partnership's model is based on more than 30 years of evidence from randomized, controlled trials that prove it works. Beginning in the early 1970s, Dr. Olds initiated the development of a nurse home visitation program that targeted first-time mothers and their children. Over the next three decades, he and his colleagues continued to test the program in three separate, randomized, controlled trials with three different populations in Elmira, N.Y., Memphis, Tenn., and Denver, Colo." Nurse Family Partnership National Service Office, "A Cornerstone of Nurse-Family Partnership."
"Following is information about the average cost of a family in the NFP program to a community/agency and is approved for public dissemination:
Nurse-Family Partnership typically costs $4,500 per family per year with a range throughout the country of $2,914 - $6,463 per family per year based primarily on variations in costs of living." Michelle Stapleton, email to GiveWell, November 16, 2010.
$4,500 * 2.4 = $10,800
"The program costs approximately $11,200 per woman over the three years of visits (in 2008 dollars)." Coalition for Evidence-Based Policy, "Nurse-Family Partnership."