Background
Child injuries have been identified by the World Health Organization as a growing global public health problem [
1]. There is a need globally to increase awareness of the problem and promote effective ways of reducing the incidence and severity of childhood injuries. The majority of injuries in young children occur in the home, with fire-related injuries being particularly important in terms of resultant disabilities, deaths and costs incurred [
2,
3]. Furthermore, the UK has one of the highest rates for deaths from fire and flames in children aged 0–14 years compared to other high income countries [
4]. In 2011–12 the Fire and Rescue Services in Great Britain attended over 44,300 domestic fires [
5]. Within the same period 11 fatalities were estimated to have happened as a result of accidental fires in the home for the 0–4 age group [
6]. Fires detected by smoke alarms tend to be discovered more rapidly and are associated with a reduced risk of death and property damage [
7‐
9]. Publicity campaigns, such as Fire Kills [
10], have been conducted in the UK in an attempt to increase the number of households which have ‘functioning’
a smoke alarms fitted but few evaluations have been conducted to assess their impact on fatal and non-fatal injuries of young children in terms of their lifetime costs and effects (i.e. quality of life). This is of particular interest because children under the age of 3 years are at the highest risk of burn mortality both with and without smoke inhalation injury [
11].
Four studies [
12‐
15] to date have evaluated the cost-effectiveness (using a decision model-based analyses) of schemes to promote the installation of functioning smoke alarms in the home, with only one of these focusing on the costs and benefits to children [
14]. The economic evaluation by Pitt
et al.[
14], commissioned by the National Institute for Health and Care Excellence (NICE), was primarily based on Ginnelly
et al.[
13] but the analysis was targeted towards reducing unintentional injuries from house fires in children under 15 years of age. This decision model-based analysis found the installation of free smoke alarms to be cost effective. Three determinants were found to be the main drivers of the results obtained by Pitt
et al.; these include the existing prevalence of use of safety devices, the proportion of households that choose to participate in a programme, and the proportion that correctly install or use any devices provided.
The aim of our analysis is to develop a decision analytic model to evaluate the cost-effectiveness of having functioning smoke alarms in households with children less than 5 years of age. We extend the analysis by Pitt
et al.[
14] to include effectiveness data for all previously trialled interventions (i.e. a combination of education, free or low cost equipment giveaway, equipment fitting and/or home safety inspection) to increase uptake of functioning smoke alarms in households and hence, reduce fire-related fatal and non-fatal injuries in children. The cost-effectiveness of all the different interventions is compared.
Discussion
Assessing the cost effectiveness of alternative strategies is important in a public sector system operating under fixed budget constraints. This study evaluated the cost-effectiveness of alternative interventions to increase the household uptake of ‘functioning’ smoke alarms and, consequently, reduce the number and severity of home fire-related injuries in pre-school children. The results of a previous synthesis of evidence on the effectiveness of interventions of interest [
16] were used to populate the model. The authors used a mixed treatment comparisons
d framework to synthesise evidence. This study [
16] indicated that more complex interventions (which include multiple components such as education, equipment and its fitting, and home inspection) have higher probability of increasing the possession of functioning smoke alarms than those less multifaceted. Nevertheless the authors discussed a series of limitations of this analysis, which included: i) the unavoidable existence of some degree of ‘lumping’ of interventions given existent data; ii) the heterogeneous quality of the evidence base; and iii) the existence of some unexplained inconsistency between direct and indirect evidence.
This paper showed that for these interventions to be adopted, decision makers need to be ‘willing to pay’ or displace large amounts of funds. The less complex intervention of Strategy (3) E + FE was identified to have the lowest ICER when compared to usual care (ICER of £34,200 per QALY gained reducing to approx. £4,500 when 1.8 children under the age of 5 assumed per household).
Four studies to date have conducted cost-effectiveness analysis of smoke alarm interventions [
12‐
15]. Two of these studies were UK based [
13,
14] and evaluated the provision and installation of free smoke alarms versus ‘no intervention’. The results from the analysis by Pitt
et al.[
14] informed the NICE public health guidance on the prevention of unintentional injuries among under-15s in the home [
24]. Our study extends the remit of the previous analyses by considering the cost-effectiveness of multiple interventions (i.e. ranging from usual care to more complex interventions comprising a combination of education, free or low cost equipment giveaway, equipment fitting and/or home safety inspection) to increase the installation of functioning smoke alarms in households with young children. This has been achieved by incorporating effectiveness results from a mixed treatment comparison into the cost-effectiveness analysis – which, to the authors’ knowledge, is the first time that this has been done within a public health study. Our analysis also undertakes a number of sensitivity analyses to test the robustness of the findings to assumptions made by the model. These analyses support the finding of our main analysis that more effective but more complex interventions may not necessarily be the most cost effective interventions.
Where uncertainty over adopting a particular intervention based on existing information exists, the expected consequences of this uncertainty can be quantified. This informs the decision maker of the consequences for the public sector (in £s) of the possibility of making the wrong decision, and informs the maximum value of conducting further research to reduce and improve decision making. In our analysis this was quantified to be approx. £49,900 at a cost effectiveness threshold of £30,000
e[
18]. The decision maker should consider conducting new research only if the costs of the research are lower than this value.
At the basis of the analyses conducted in this study there are a range of limitations. These include, firstly, the difficulty in categorising some of the interventions reporting in the effectiveness studies due to inadequate descriptions of the interventions; for example, education in the different studies may have been of varying intensity. Secondly, although the impact on the results of changing many of the assumptions made in the modelling were investigated in the sensitivity analyses undertaken, not all assumptions were able to be investigated; for example, there is some evidence that a child admitted to hospital with a burn is more likely to be admitted in the future with another burn than with another injury [
25]. Thirdly, we know social inequalities exist in the possession of ‘functioning’ smoke alarms in families with children under 5 in the UK and therefore future research may investigate whether more complex interventions may be more cost effective in some social groups [
26]. Finally, data on burn treatment costs is country specific; therefore, the results from this analysis (based on UK data) may not necessarily be generalizable to other countries of different healthcare systems.
While economic evaluation has been widely used in the past two decades to support decision making in the health care setting, its use has only recently been applied within public health [
20,
27‐
29]. Methodological challenges specific to public health include: (i) the attribution of effects (both intended and unintended) of the policy on the targeted population and problem; (ii) the costs and consequences which should be analysed, considering the feasibility of the programme; (iii) the acceptability of the policy by the relevant stakeholders, which often involves subjective judgements, beliefs, values and interests of the actors concerned, and iv) obtaining an equilibrium between an efficient and an equitable allocation of resources [
30,
31]. In our analysis we chose a Public Sector perspective, however, if we restricted the analysis to the NHS and PSS (i.e. focusing on healthcare related costs and omitting law enforcement, and Fire and Rescue costs) the ICER for Strategy (3) E + FE, marginally increased from £34,200 to £35,561 per QALY. If we expand the perspective to include property damage, cost of fatality (i.e. coroners, autopsy) and cost of equipment incurred by individual households but not lost productivity costs, then the ICER for Strategy (3) E + FE substantially increased to approx. £74,000 per QALY.
In this paper important findings were made about the cost effectiveness of interventions in promoting the uptake of ‘functioning’ smoke alarms and consequently, in reducing child injuries at home. However, there continues to be insufficient evidence to inform and support public health policy/decision making. This state of affairs can be changed, but it will require strong direction to ensure the priorities for economic evaluation evidence become organised and coordinated at local, regional and national levels.
Endnotes
a‘Functioning’ implies that the safety device is fully operational.
bA severe fire-related injury was defined as one that requires inpatient stay greater than five days in an intensive care unit. It was assumed that any child suffering a severe injury (particularly burns) would suffer some form of disability and would carry that impairment for the rest of its life. A child experiencing this event would therefore suffer a decrement in (health related) quality of life and would be subject to additional health costs for the rest of its lifetime. A minor or moderate fire-related injury is assumed not to have any significant decrement in children’s quality of life or any additional on-going health costs.
cNote that a predefined threshold does not exist outside of the health sector and therefore the £20,000 to £30,000 range of values is occasionally used throughout to support the interpretation of results.
dMixed treatment comparisons (also known as network meta-analysis) [
32‐
35] are an extension of standard (pairwise) meta-analysis that enable the simultaneous comparison of all evaluated interventions within a single coherent analysis.
eIntervention time horizon is assumed to be of 10 years and the annual effective population (i.e. expected number of single child households under 5 per year in the UK) considered is approx. 31,000 (ONS 2010).
Acknowledgements
We would like to thank Professor Jon Nichol for providing data from the HALO study. We would also like to thank members of the Keeping Children Safe study for their help with suggesting sources of evidence to inform parameter values.
Funding
This study presents independent research commissioned by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research funding scheme (RP-PG-0407-10231). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.
Pedro Saramago was funded through a Medical Research Council Capacity Building Grant (grant number G0800139) and the Portuguese Fundacao para a Ciencia e a Tecnologia (grant number SFRH/BD/61448/2009).
Andrea Manca contribution was made under the terms of a Career Development Fellowship issued by the NIHR (grant CDF-2009-02-21).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
PS designed and implemented the decision analytic model, conducted the cost-effectiveness analysis and prepared the final manuscript. NJC designed and implemented the decision analytic model, oversaw and advised on all elements of the cost effectiveness analysis and prepared the final manuscript. AJS and DK conceived of the study, helped to develop and populate the decision analytic model, assisted with interpretation of the results and reviewed the final manuscript. MH and KD advised on the model structure, supplied data to inform the model parameters and reviewed the final manuscript. AM advised on elements of the cost effectiveness analysis and reviewed the final manuscript. All authors read and approved the final manuscript.