Evaluation of the Safe At Home safety equipment scheme

During 2009-2011, the Department for Education funded an £18 million scheme in England to provide safety education and equipment to the poorest families with children under 5. The scheme (Safe At Home) was overseen by the Royal Society for the Prevention of Accidents (RoSPA) and led by local authorities. It included the provision of equipment such as stair gates, fire guards and cupboard locks to over 66,000 families. To date, the effectiveness of this scheme in preventing child injuries from occurring has not been analysed on a national scale.

The aim of the study is to determine if the Safe at Home scheme was associated with the following in children under age 5:

1) a reduction in hospital admissions due to unintentional injuries of any type likely to have occurred within the home

2) a reduction in hospital admissions due to unintentional injuries likely to have occurred within the home that could plausibly be prevented by the equipment provided by the scheme (e.g. falls down stairs, fire burns, poisoning)

The study will also determine if the scheme:

1) was cost-effective as delivered (targeted at the most deprived areas) and

2) would be cost-effective if implemented in ‘lower risk’ populations e.g. those with lower baseline injury incidence or residing in more affluent areas

The HES data was transferred in full, by NHS Digital, to the SAIL databank/UK SeRP at the Swansea University, where data was processed and analysed. The HES data has been merged with home safety equipment supply data using Lower Super output area fields in both datasets. PEDW data from Wales, available in SAIL, has also been used to provide some of the control population. Access to the HES/PEDW/equipment supply data by researchers at the University of Nottingham will be via the secure platform connection to SAIL databank/UK SeRP.

Results will only contain aggregated data with small numbers suppressed in line with the HES analysis guide. The analysis was undertaken at Lower Super Output Area (LSOA) level and results presented at LSOA, Clinical Commissioning Group (CCG) and Local Authority level. This is because CCGs and Local Authorities (Public Health and Social Care) commission health and social care interventions and so the results will need to be presented in a way that is relevant to them.

Outputs will be reported in peer-reviewed academic journals and then disseminated further through professional publications, conferences and public dissemination routes.

The number of years of data requested was from 2003/4 to 2015/16. This number of years will provide adequate study power and achieve a sample size of 72 time points recorded on each lower super-output area; that is 6 times per year for the 5.25 years prior to the intervention, 2 years during the intervention delivery and 4.75 years after the intervention between January 2004 to December 2015. This sample provides at least 85% power to detect an effect size of 0.5 for auto correlations up to values of 0.5.

The data fields requested are restricted to those that are specific to the analysis that will be undertaken, with the primary outcome of injury admission rates in the intervention versus the control areas.

The data requested was only the episode where age at start of the episode is under the age of five.

The HES data from NHS Digital will enhance what is known about the effectiveness of this type of scheme. Until now effectiveness of home safety equipment provision has only been assessed at the local level (e.g. local evaluations) and as such, there is no robust effectiveness and cost effectiveness data. The National Institute for Health and Care Excellence, NICE, currently recommend that local areas commission home safety equipment provision schemes for families with young children and on the lowest income levels, based on local evaluation data. The results may therefore inform future NICE guidance, depending on the outcome of the analyses.

None of the work will take place outside of England and Wales. These data will not be used for commercial purposes. Data will not be provided, at record level, to any third parties and will not be used for direct marketing.

HES data from NHS digital will only be used for the purposes stated within this agreement.

Article 6 justification of processing the data.

The legal basis for processing these data is to undertake a public task: i.e. the data processing is necessary for the data controllers to perform research in the public interest which is to evaluate the effectiveness and cost-effectiveness of public sector spend on health activities. Depending on the results of the study, the data controllers will either show that the intervention being studies is good use of public money and should therefore continue to be provided for the most vulnerable in society, or is not good value for money in which case it should not continue. This information could only otherwise be ascertained if there was a sufficiently large randomised trial, but this would be costly and unethical without having first evaluated the existing data here that are available.

Article 9 - Special category data

In this study the data controllers are processing health data and the justification for this under Article 9 is for reasons of public interest in the area of public health, to ensure high standards of quality and safety of health care and the data controllers have put in place the necessary information security processes.

RoSPA are only acting in an advisory capacity (e.g. providing context to help interpret the data) and do not have any control over how the data are processed.

Pseudonymised, non-sensitive record level HES data was transferred from NHS Digital to the SAIL databank/UK Secure e-Research Platform (UKSeRP) for secure storage, processing and statistical/health economic analysis. The legal basis for this is Section 261(1) of the Health and Social Care Act and necessary in order to undertake a public task under Article 6 of GDPR.

The SAIL Databank is now powered by the UK Secure e-Research Platform (UKSeRP), developed by the Health Informatics Group at Swansea University, with support from the Farr Institute of Health Informatics Research funded by Medical Research Council. UKSeRP, a is high powered data management and sharing technology, is infinitely scalable to suit a range of use cases including imaging, genomics and analysis of free text. It benefits from carefully designed Information Governance to ensure person-based data with high privacy risk is managed to the highest standards, and is ISO 27001 certified.

The HES data will be merged in SAIL with home safety equipment supply data using Lower Super output area fields in both datasets. PEDW data from Wales, available in SAIL, will be used also to provide some of the control population. The home safety equipment dataset contains no individually identifiable level data. It is non-sensitive scheme activity data supplied by RoSPA to the SAIL databank. There will be no record level linkage and the request is not asking NHS Digital to do the aggregated linkage of RoSPA LSOA level data and HES data.

There will be no attempts by the study to identify any individual.

The primary analysis is a controlled interrupted time series analysis with injury rate data points at Lower Super Output Area (LSOA) level. Because injuries are more common in deprived areas, differ between boys and girls and by child age, control areas will be matched on how rural or urban the area is (using Office for National Statistics classification at LSOA level), deprivation, sex and age profiles. As the scheme was implemented in England only and in households with the lowest income and in areas with the highest injury rates, the study will need to supplement the pool of potential control areas with areas of similar levels of deprivation and injury rates in Wales using the PEDW data (the scheme was not implemented in Wales).

Intervention (England) areas will therefore be matched at LSOA level to control areas (England and Wales) on rurality (ONS classification), deprivation, sex and age profiles.

The rates of injury-related hospital admission will be compared in the intervention and control areas before, during and after the scheme’s implementation using an interrupted time series analysis. The applicant will undertake a cost-effectiveness analysis, estimating the incremental cost per hospital admission averted and determine if the cost-effectiveness of the scheme varies according to the level of deprivation of the areas it is implemented in.

The hypothesis is that provision of home safety equipment prevents injuries in children that require hospitalisation and that the provision is cost effective to the public sector.

Data will not be accessible by any parties other than substantive employees of the Swansea University (SAIL databank/UKSeRP) and substantive employees of the University of Nottingham (accessed via remote secure platform connection).

The project described in this agreement will lead to at least two published papers (a 'clinical' effectiveness paper and a cost effectiveness paper) in peer reviewed journals, plus dissemination at the National Society for Academic Primary Care and Safety 2021 world conferences and other professional networks (e.g. via RoSPA). It is anticipated that the papers will be submitted for publication by summer 2021.

The main analysis took longer than expected and study team are now completing the economic analysis, which could take another 2-3 months. then study team will need to submit the papers for peer review publication. The peer review process can take 6-9 months, and depending on the comments that reviewers make, study team may need to go back to the data to re-do analyses. This is why a further extension request has been requested.

Papers will be submitted to Journal of epidemiology and community health and the Journal of Public Health. There is funding for at least one open access publication which would be free for anyone to read.

Research summaries/briefings will be specifically targeted at commissioners within CCGs and Local Authorities who are the most likely to commission population-level safety equipment schemes and work with Public Health England and the National Institute for Health and Care Excellence to disseminate their findings in evidence reviews that they produce. A lay summary of the research will also be produced and, via the communications teams at the universities of Nottingham and Swansea disseminate this to the media and to RoSPA for their dissemination to parents.

Outputs will contain only aggregate level data with small numbers suppressed in line with the HES analysis guide. Results will be presented at Local Authority and Clinical Commissioning Group level as these are the footprints for current commissioning arrangements for health (NHS), Public Health and Social Care services. It would be aimed at Public Health Consultants, GPs, Commissioners and third sector organisations who may provide equipment on a charitable basis.

The study team will work with patient public representatives and the 3rd sector (e.g. The Royal Society for the Prevention of Accidents (RoSPA) and Child Accident Prevention Trust (CAPT) to develop dissemination materials for local authorities and CCGs to help them ensure that commissioned services are designed to maximise injury prevention.

The dataset itself used in the analysis will not be made publicly available and no analysis tools will be generated.

The National Institute for Health and Care Excellence (NICE) have developed public health guidance on home injury prevention in children: PH29 (unintentional injuries: prevention strategies for under 15s) and PH30 (Unintentional injuries in the home: interventions for under 15s). In these publications NICE recommend that commissioners of health and social care services provide local safety equipment schemes for families most in need (on low incomes and with children under 5). The results from this analysis will inform future updates of this guidance and will provide additional health economic / cost effectiveness data that does not currently exist. This will be achieved by members of the study team sharing the results of the study with the National Accident Prevention Strategy Group which they sit on with Public Health England who are reviewing the child accident prevention guidance, in conjunction with NICE. Members of the study team also sit on the advisory group for this work. This is in addition to usual publication in peer reviewed journals and media engagement when the studies are published.

The results will also inform the commissioning of home safety equipment schemes via the direct dissemination methods to CCGs and Local Authorities. This dissemination will happen via the Faculty of Public Health (individuals from the study team are members and will share the findings via special interest groups in child health and injury prevention) and via the Royal Society for the Prevention of Accidents who are part of the study group and have access to a large practitioner audience nationally and internationally.

The results may indicate that the scheme is not likely to be effective, in which case NHS and Public Health funding should be used to commission alternative injury prevention programmes that are shown to be effective.

If however the study indicates that provision is effective and cost effective it would inform a strong case for local areas to commission schemes in the future. After national funding ended, most areas stopped providing equipment at that time.

The study team have been contacted by NICE to ask on the progress of the research and they are now awaiting the final outputs to be able to update their guidance (PH29 and 30).

The other outputs will be produced once these have been through the peer review process.

The study team have presented provisional results to an international conference and received useful feedback regarding this analysis.