Evaluation of the Family Nurse Partnership (FNP) Scotland: A Natural Experiment Approach

This study was designed to evaluate the effectiveness of Family Nurse Partnership (FNP) in Scotland – by linking routine administrative data.

Background

The Nurse Family Partnership (NFP) is a licensed intensive home visiting programme for first time young mothers and their children developed in the United States and introduced into Scotland in 2010 following initial adaptation, at which point it became known as the FNP in Scotland. The programme is provided from early pregnancy until the child’s second birthday and follows a schedule of home visits to promote better pregnancy and birth outcomes, improve child health and development, and improve maternal life course. Following an evaluability assessment which explored options for an impact evaluation, a natural experiment using linked routine administrative data was commissioned.

The research investigated whether young mothers and their children participating in the FNP programme (Clients) had better outcomes than those who were not participating in FNP (Controls). Mothers and their children who had already completed the FNP programme in Scotland were followed up, anonymously, using their health, education and social care information. The Electronic Data Research and Innovation Service (eDRIS) safe haven was used to securely hold all the anonymised data which was made available to the project team via a remote portal.

The findings of the study will contribute to the developing evidence for determining the effectiveness of FNP programme in Scotland and will inform policy for expansion and continuation of the programme.

Methods

Cardiff University's evaluation of the Family Nurse Partnership specialist home visiting programme has been summarised in a methods paper published by the Scottish Government. This document describes the detailed methodological approach taken by the study team which is based on a natural experiment.

The cohort included 8,118 young mothers: 3,203 FNP Clients and 4,915 Controls. FNP Clients were women enrolled onto the FNP programme between January 2010 and March 2016 in one of ten Scottish Health Boards offering FNP at that time and their first-born child(ren). Controls were mothers and their first-born child(ren) eligible for FNP but with an antenatal booking date outside of FNP recruitment periods, between January 2009 and March 2016.

Results

Key findings 

Maternal tested outcomes

When comparing families in receipt of specialist home visiting and those who did not:

• There were no differences in rates of maternal use of alcohol or drugs during pregnancy, in the time between first and subsequent pregnancy or birth, nor rates of mothers with a subsequent birth within 2 years of having her first child.

Child tested outcomes

• Children from families visited by a Family Nurse were more likely to have been breastfed at 10-14 days post-partum and at 6-8 weeks post-partum. There were no statistically significant differences on breastfeeding initiation nor duration of breastfeeding, among a small cohort of women with dates recorded for ceasing breastfeeding.
• There was a statistically significant reduction in the child’s exposure to second hand smoke over time in the FNP group compared to the Control group and statistically significant more children in the FNP group were registered with a dentist by aged 2 years.
• A significantly higher proportion of children in the Controls had any newly suspected child development concerns recorded at 27-30 months, with no other differences in any other child development outcomes.
• There were no statistically significant differences on the following outcomes: rates of pre-term delivery before 37 weeks gestation, time to first dental admission, dental attendance by 27-30 months, being of a healthy weight at 27-30 months or at the Primary 1 (P1, 5-6 year olds) review, being admitted to hospital with an unintentional injury in the home by age 2 and by age 5 years, time to first A&E attendance with an accidental injury, or time to first hospital admission for a serious injury.
• There was no evidence of a statistically significant difference between groups in children achieving their age relevant attainment level (P1, 5-6 year olds and P4, 8-9 year olds) for Reading, Writing, Listening and Talking, Literacy, and for Numeracy. There is some evidence of impact at a health board level.
• There were no statistically significant differences observed between groups in the age of a child’s first child protection investigation, the number of child protection investigations per child up to the age of five, nor in the average number of days a child spent on the child protection register.

Conclusions

We found statistically significant differences between FNP Clients and Controls on five of the 39 tested outcomes assessed. Formally adjusting to account for the large number of comparative tests reduces these to two programme effects: exposure to second hand smoking and breastfeeding at 10-14 days. As the study used only routine data, other outcomes of relevance to FNP could not be assessed but remain of importance.