Table 1

Key components of the economic evaluation of the SOSA programme

Decision problemDoes the SOSA programme offer value for money for improving home safety, reducing childhood injuries and improving health?
Type of evaluationCost-effectiveness and cost–utility analyses alongside a non-randomised controlled before and after study
PopulationChildren aged between 2 and 7 months of age with parents aged 18 years or older, residing in within one of eight electoral wards of Nottingham City. Cohort 1 was recruited in September 2017, cohort 2 in March 2018 and cohort 3 in September 2018.
Setting and perspectiveEight electoral wards in Nottingham city, NHS and local authority perspective28
Time horizonTwo years for each cohort
InterventionThe SOSA programme involved using evidence-based home safety promotion delivered to families within the four SOSA wards. Service providers (health visiting team members, family mentors (peer family support workers) and children’s centres staff) were specifically trained to deliver the SOSA programme. The SOSA programme was tailored to a family’s needs, and included referral/signposting to partner organisations for additional risk assessments (eg, home fire risk assessments provided by the fire service), and referral to charities for safety equipment (if available).
The components delivered by each practitioner group were:
Health visiting teams:

  1. Provided home safety advice through use of age-specific checklists at infant (9–12 months) and toddler (2–2.5 years) child health reviews.

  2. Used the checklists to guide advice to families following attendance at the emergency department for an injury as required.

  3. Discussed with or distributed monthly safety messages and safety week resources to parents in ad-hoc contacts or clinics.


Family mentors:

  1. Undertook home safety activities from the family mentor manual.

  2. Discussed with or distributed monthly safety messages to parents at home visits.


Children’s centres:

  1. Ran four safety weeks per year.

  2. Discussed with or distributed monthly safety messages to parents.

ComparatorUsual care: Four wards received home safety promotion from health visiting teams and children’s centres that was already provided as part of routine care and did not have access to the SOSA programme. Family mentors were not available in control wards.
CostsNational currency (£) at 2019/2020 prices
OutcomesPrimary outcome: Number of homes which adopted the three key safety practices (one fitted and working smoke alarm, a safety gate on stairs and storing poisons out of reach)
Secondary outcomes: Number of injuries avoided, Quality-Adjusted Life Years (QALYs) gained
Discounting3.5% per annum
Analytical strategy

  1. For each cohort, we estimated the total healthcare cost, number of homes with the three key safety practices, number of injuries and QALYs

  2. Incremental healthcare costs, number of homes with the three key safety practices, number of injuries, and QALYs were estimated using regression modelling with variables specified a-priori using multi-level mixed regression analysis controlling for29 :

  • Matched ward.

  • Mother’s age at birth of first child.

  • Number of children (aged under 16 years) at home.

  • Single-parent family.

  • Index of Multiple Deprivation 2019 score.

  • Whether the house had the three key safety practices at baseline (excluded from primary analysis as that analysis took account of change from baseline).

(3) The primary measure of cost-effectiveness was the incremental cost-effectiveness ratio (ICER) per additional house with the three key safety practices Embedded Image defined as
Embedded Image
Secondary measures of cost-effectiveness were
ICER per injury avoided Embedded Image
Embedded Image
And ICER per QALY gained Embedded Image
Embedded Image
One-way sensitivity analysesVarying SOSA programme costs, incremental healthcare costs, incremental homes with three key safety practices, incremental injuries avoided, and incremental QALYs gained were varied between 0.5 and 2 times their initial value. Results were plotted using a Tornado Plot.
Probabilistic sensitivity analysisBootstrapping with replication, sampling 10 000 times, generating pairwise incremental costs and outcomes, allowing estimation of 95% CIs on ICERs, and generation of scatterplots of incremental cost versus incremental outcomes and cost-effectiveness acceptability curves.30

  • NHS, National Health Service; SOSA, Stay One Step Ahead.