Parts of this chapter have been reproduced from the published papers by Cavallaro et al.1 and Cavallaro et al.2 These are open-access articles distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) licence, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given and an indication of whether changes were made. See: http://creativecommons.org/licenses/by/4.0/.
Chapter outline
This chapter describes the context of adolescent pregnancy and the health visiting services available in England. It briefly describes the implementation of the Family Nurse Partnership (FNP) in England and reviews the literature on its effectiveness. It concludes with a description of the rationale for the present study and study objectives.
Adolescent pregnancy and adverse outcomes
Each year, approximately 3% of babies (~16,000 in 2020) are born to mothers aged < 20 years in England and Wales.3 Adolescent mothers are more likely to experience adversity, be less engaged with education and employment and have rapid repeat pregnancies compared with older mothers.4–7 For their children, young maternal age is associated with a higher incidence of preterm birth, low birthweight8,9 and a greater risk of child maltreatment and associated adverse long-term consequences, including poorer physical health and social, emotional and cognitive outcomes.10–12 These adverse maternal and child outcomes of adolescent pregnancy, due to social adversity, disruption to education and employment and child-rearing practices, are of major importance to public health research and the NHS.13,14 Programmes supporting adolescent mothers, such as the FNP, are therefore likely to remain a priority for the NHS and local authorities (LAs).15
Understanding how best to target services to the most vulnerable mothers is key to improving the health of these mothers and their children. Evidence to help improve targeting is being called for by service providers, who need to understand the value of interventions in the context of their target populations and local services in order to inform commissioning and justify spending.16
The Family Nurse Partnership and early years health visiting in England
Health visiting in England is delivered as part of the Healthy Child Programme. All families with children should receive a minimum of five visits from 28 weeks of pregnancy until the age of two and a half.17,18 These mandated assessments allow health visitors to identify families in need of additional support and offer more intensive support, including additional health visitor contacts and referrals to other, more intensive, programmes. This model of proportionate universalism, an approach combining universal service provision accessible to all with more intensive services proportional to the level of need, has been recommended as key to reducing health inequalities in the UK.19
Several intensive health services aiming to reduce inequalities between adolescent and older mothers have been trialled; of these, FNP, an intensive home visiting programme supporting young first-time mothers, has a strong evidence base from several randomised trials in the USA and is recommended within the UK government’s Healthy Child Programme.19,20 Mothers enrolled in the FNP receive up to 64 home visits by a dedicated Family Nurse from early pregnancy until the child’s second birthday. The FNP aims to improve birth outcomes, child health and development and promote economic self-sufficiency among young mothers.21 Although a randomised trial of the FNP in England found no evidence of benefit on smoking in pregnancy, birthweight, hospital admissions before age 2 or second pregnancy within 2 years, improved cognitive development outcomes were reported, and there remains strong support for the programme locally (see Literature review – evidence of effect of Family Nurse Partnership programmes on maternal and child outcomes).20–23 The Building Blocks 2- to 6-year follow-up reported no evidence of effect on maltreatment outcomes but evidence of improved school readiness, measured by a Good Level of Development in the early years foundation stage profile (EYFSP) and improved educational achievement at Key Stage 1.24 The FNP has been commissioned in > 130 English LAs since 2007 (Figure 1).
FIGURE 1 Family Nurse Partnership site activity status in March 2019 among all FNP sites active during the study period (2010–9), by English LA. (Figure not reproduced here; see the original article.)
Literature review – evidence of effect of Family Nurse Partnership programmes on maternal and child outcomes
Review approach
We conducted a literature review of the effect of FNP programmes on maternal and child outcomes using a combination of PubMed searches for ‘FNP’ or ‘Nurse Family Partnership’ (the original programme name in the USA) and identifying additional references through manual review of reference lists of previously identified papers from the USA, UK and other trials.
The evidence base for the FNP spans multiple countries and includes randomised controlled trials as well as a few non-randomised studies. Findings from this review were synthesised by the country in which the research was conducted, given differences in context and the importance of context (usual care) for the effect of interventions.
USA randomised trials
Most of the literature on the effectiveness of the FNP stems from three randomised trials of the Nurse Family Partnership (NFP) conducted in the USA by David Olds and colleagues, including over 20 peer-reviewed articles. The trials were conducted in Elmira, New York (enrolment 1978–80), Memphis, Tennessee (enrolment 1990–1) and Denver, Colorado (enrolment 1994–5). A wide range of maternal and child outcomes were examined, with up to 20 years of follow-up.
Overall, the evidence from the three USA trials indicates a positive effect on some – but not all – maternal and child health and educational outcomes. Often, the positive effect was observed in a subsample of particularly vulnerable mothers but not in the entire cohort. The Elmira trial found a one-third reduction in all accident and emergency (A&E) admissions among children up to age 2 and ages 2–4; however, no effect was observed on all hospital admissions ages 2–4.25,26 Although a reduction in mean A&E and hospital admissions for injury/maltreatment of 55% was reported in the second year of life, no such effect was detected for injury/maltreatment in ages 0–1 or 2–4. No effect was observed on this outcome for up to 2 years in the Memphis trial.27 The Elmira trial was the only NFP trial to examine child abuse/maltreatment reports and found a 40% reduction in such reports up to age 15 (although this benefit was not evident before the age of 4 years).26,28
All three USA trials examined the effect of the NFP on child development and educational achievement. Results were mixed in the Elmira trial, with no effect up to age 4 for the whole sample and no effect on intellectual functioning at age 3 or 4 among children identified as maltreated.26,29 There was no difference in intellectual functioning at age 3 or 4 among children of women smoking ≥ 10 cigarettes per day at enrolment; however, they did have higher intelligence quotient (IQ) scores (4.86 points on average) at this age than children of smoking mothers in the control arm.30 The evidence in favour of a positive effect of the FNP was stronger in the Memphis and Denver trials, although these effects were concentrated in the subgroup of mothers with low psychological resources at enrolment (composite score including mental health, sense of mastery and intelligence scores), and even among this group, the positive effects were limited to only a few outcomes by age 6 and/or age 9 in the Denver trial.31–36 Child mortality was only examined in the Memphis trial, which found no evidence of effect up to age 19 for all-cause mortality but a reduction in preventable-cause mortality in the NFP arm (0% vs. 1.6% in one control group, p = 0.04).37
The effect of the NFP on maternal outcomes was weaker than on child outcomes. The Elmira trial found evidence of a 50% reduction in child abuse perpetration reports up to 15 years after giving birth.28,38 There was evidence of 0.5 fewer subsequent births 15 years after the first birth among unmarried women of low socioeconomic status only in the Elmira trial but no evidence at 18 years in the Memphis trial or at 4 years in the Denver trial. However, there was evidence that NFP increased the subsequent birth interval by 5–28 months in all three trials (among the unmarried, low-socioeconomic subgroup in Elmira only).32,34,38,39 There was no evidence of an effect on maternal educational qualifications beyond 6 months31,34,40,41 or on experience of domestic violence in the three trials (weak evidence of a decrease in the Denver trial).28,34,42 Evidence was mixed regarding the effect on drug use or impairment.34,38,42 The effect of NFP on maternal mortality was only examined in the Memphis trial, with no evidence of a difference in all-cause mortality but weak evidence of a 1% reduction for mortality from external causes at 20 years.37
A more recent trial of 5670 Medicaid-eligible nulliparous pregnant mothers recruited between 2016 and 2020 in South Carolina found no evidence of an effect on birth outcomes (preterm birth, low birthweight, small for gestational age and perinatal death).43
England Building Blocks randomised trial
One randomised controlled trial of the FNP (the Building Blocks trial) was conducted in England, enrolling approximately 1600 expectant mothers in 2009–10.44 The Building Blocks trial found no evidence of effect on the four primary outcomes – smoking in late pregnancy, birthweight, second pregnancy within 24 months of first birth and rates of A&E attendance or hospital admissions within 24 months of birth. Some secondary outcomes suggested small positive impacts of the FNP in the first 2 years of life, including maternally reported child cognitive and language development. Safeguarding concerns recorded in primary care records were higher for mothers enrolled in the FNP.
Results for follow-up to age 6 showed no difference for most maltreatment outcomes between the trial arms, including referrals to social services, children referred as Children in Need (CiN), duration for which children were assessed as in Need, children with a child protection plan (CPP) or who were Looked After.24 However, children in the FNP arm spent on average 2 months less time in care than children in the usual care arm. There was no evidence of a difference in children not attending a hospital outpatient appointment, attending A&E for injury or ingestion or being admitted to hospital for the same causes.
Nonetheless, there was evidence of FNP’s impact on some – but not all – child development and educational outcomes. There was no difference in Special Educational Needs (SEN) provision up to age 6 or educational attendance for ages 2–4. Children of FNP mothers were more likely to achieve a Good Level of Development at school entry (age 5) than in the usual care arm, with a greater beneficial impact on total point score (across 17 learning goals) observed for children of younger mothers. At Key Stage 1, children in the FNP arm were more likely to reach the expected level for reading; no other differences were observed for Key Stage 1 outcomes. The beneficial effects of the FNP were stronger among boys than girls (reading and writing), among younger mothers at enrolment (mathematics and writing) and among mothers not in employment, education or training at the time of enrolment (writing).
Other randomised trials in Germany, the Netherlands and Canada
Several other trials have been conducted in high-income countries. The VoorZorg trial in the Netherlands, enrolling in 2007–9, reported a reduction in child abuse/maltreatment reports by age 3 in the FNP arm, as well as a reduction in some types of interpersonal violence at 32 weeks of pregnancy and 24 months after birth.45,46 A trial of a FNP-based model in Germany (Pro Kind) reported improved child development among high-risk women only, but no evidence of a difference in subsequent births within 2 years.47 The follow-up trial evaluating outcomes at age 7 reported fewer behavioural problems in children, less child-abusive parenting, fewer maternal mental health problems and higher maternal life satisfaction in the intervention arm.48 Some positive effects on mother–daughter interactions were also reported for a small subsample who agreed to participate in video recording.49
One randomised controlled trial in Canada has not yet published results on primary outcomes but has reported preliminary findings on a number of secondary outcomes, with a reduction in prenatal cannabis use and a modest reduction in cigarette use in smokers associated with the intervention but no reduction in rates of prenatal cigarette and alcohol use.50
Non-randomised studies in Australia, Scotland and the USA
Non-randomised studies in the USA and Australia have reported reduced preterm births, child maltreatment, infant death and subsequent births among FNP participants compared with controls, as well as higher high school attainment and different patterns of A&E attendance.51–56 These studies adjusted for confounders through propensity score matching, frequency matching and entropy balancing, although due to limited maternal characteristics, the potential for residual confounding remains. The Australian study compared participants to eligible women who were not referred to and never participated in the programme, thereby also being subject to likely residual confounding. One prospective cohort study in Scotland has not yet reported results.57
Other evidence
A randomised controlled trial of group FNP in England, administering 44 FNP sessions to groups of 8–12 expectant mothers, found no evidence of effect on parenting or maternal sensitivity or on secondary outcomes [except for a higher proportion of mothers breastfeeding at 6 months, odds ratio (OR) 3.2; 95% confidence interval (CI) 0.99 to 10.6].58
Synthesis
The available evidence on the effectiveness of the FNP is summarised in Appendix 1.
The three USA trials of FNP showed mixed but overall positive impacts on child health and development outcomes and some maternal outcomes, similar to the more recent Netherlands trial. These results contrast with the Building Blocks trial in England, which showed no evidence of impact of FNP on most child outcomes, with the exception of some cognitive outcomes. There are two main contributing explanations for the difference in results observed in England compared with the USA and the Netherlands: first, there are likely important differences in usual care contexts between different countries. The social safety net is likely to be stronger in England than in the USA, with better access to services for adolescent mothers not enrolled in FNP (including the minimum five mandated health visiting contacts, universal health care free at the point of care, services provided through free Children’s Centres, etc.), which may explain the lack of association for most outcomes in England. For example, the mothers in the control arm of the Building Blocks trial received an average 16 health visiting contacts up to the child’s second birthday. Nonetheless, the usual care group in the Netherlands trial probably had access to similar or better levels of care than in England, with 9–11 home visits before the child’s second birthday, as well as support from child welfare and mental health organisations,59 indicating that there are important factors (beyond access to services) shaping the effect of FNP.
Second, there are notable differences in the eligibility criteria for FNP in England compared with other countries. The main eligibility criterion for enrolment in FNP in England is maternal age: adolescents who are aged up to 19 at last menstrual period (LMP), and who are first-time mothers, are eligible for enrolment as long as they live in a LA with a FNP site and are enrolled before 28 weeks of pregnancy. From November 2016, a few FNP sites extended their eligibility criteria to enable enrolment after 28 weeks gestation and among mothers aged 20–24 at LMP with other markers of vulnerability, recognising that mothers in these groups could also benefit.60 Eligibility criteria for the Building Blocks trial (nulliparous, age ≤ 19, living in the catchment area of a local FNP team, < 25 weeks of pregnancy)44 are therefore aligned with the practice in English FNP sites at that time. Young age is the main eligibility criteria for FNP in England, based on the ease of identifying the youngest adolescent mothers, associations between adolescent motherhood and social adversity, disrupted education and employment13,61 and other factors contributing to poor birth and health outcomes among their children.5,6,62
In contrast, additional socioeconomic criteria such as unemployment, low educational level or low income are used in combination with maternal age in other countries,27,35,63 based on logic models of how the original NFP was expected to provide benefits.64 As a result, the population of young mothers enrolled in trials in other countries are a more selected and vulnerable group than in England, who may stand to benefit more from the FNP (as evidenced by greater effectiveness in socioeconomically deprived groups demonstrated in the USA trials).32,36,38
Rationale for the present study
Usual care available to adolescent mothers is likely to have declined between the Building Blocks trial study period and after the introduction of austerity measures in England – in particular, health visitor budgets have decreased since responsibility for commissioning health visiting services shifted from the NHS to local government in England in 2015.65 Furthermore, evidence from the USA trials suggests that the youngest, most disadvantaged mothers are likely to benefit most from FNP.22 Adequately powered subgroup analyses are needed to examine whether some groups of families benefit from FNP more than others. In addition, constrained conditions under which trials are conducted often do not match the complexity of real-world implementation of programmes.66
Linkage of existing administrative records provides a cost-efficient means of evaluating services as they are implemented in the real world by bringing together data from different sectors on a range of outcomes. They also allow for a sufficiently large sample size for subgroup analyses. Our population-based study aims to use longitudinal linked observational data between the health, education and social care sectors to evaluate the effects of FNP on outcomes of eligible mothers and their children up to age 7 to generate evidence on the factors that may influence effectiveness and programme engagement (including participant characteristics, setting, provider and programme delivery).67,68 Evaluating outcomes for approximately 30,000 FNP families and up to 1 million controls built on the results of the Building Blocks trial24,44 will provide increased statistical power to detect smaller differences, differences in rarer outcomes and subgroup differences for which the Building Blocks trial was underpowered. Use of these data for the real-world evaluation of FNP is important and necessary to inform the targeting and commissioning of services by generating evidence on which groups of mothers and their children benefit from the real-world implementation of FNP in England.
Research aims and objectives
We aimed to evaluate the real-world, ongoing implementation of FNP in England on the outcomes of mothers participating in FNP and their children. Specifically, our objectives were to:
- determine the rate of and characteristics associated with enrolment in FNP among young mothers across LAs in England
- determine the effect of FNP on maternal and child outcomes, including identifying which families benefit the most from FNP
- identify contextual and programme factors that might influence the effects of FNP.