Home safety education and provision of safety equipment for injury prevention
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The objectives of the review are to:
-
evaluate the effectiveness of home safety education aimed at reducing home injuries in childhood
-
evaluate the effect of providing low‐cost, discounted or free equipment in addition to home safety education
-
evaluate the effect of home safety education and the provision of low‐cost, discounted or free equipment by social group.
Background
In industrialised countries, injuries are now the leading cause of death in childhood, accounting for 40% all child deaths between the ages of 1 and 14 years (UNICEF 2001). The relative contribution of home, road traffic, leisure and other injuries varies by age and by country. In the UK, for children under the age of five, the majority of fatal and non‐fatal medically attended injuries occur in the home (ONS 2001; Morrison 1999). For older children, most fatal injuries occur as a result of a road traffic injury (ONS 2001), but injuries occurring at home account for 40% of medically attended injuries in 5−9 year olds and 25% in 10−14 year olds (Morrison 1999).
Children suffering socio‐economic disadvantage have consistently been found to be at greater risk of unintentional injury than the more advantaged (Roberts 1996; Roberts 1997; Hippisley‐Cox 2002). There is a steep social class gradient in injury mortality. Children from social class V have injury mortality rates five times higher than children from social class I, and the socio‐economic gradient in injury mortality is widening, rather than reducing (Roberts 1996; Roberts 1997). There are also steep socio‐economic gradients in injury morbidity for falls, pedal cycle injuries, pedestrian injuries, poisoning, chemical injuries and burns and scalds (Hippisley‐Cox 2002).
There is evidence that home safety counselling and education, with or without the provision of safety equipment, can be effective in increasing safety equipment possession, increasing safety practices and, in some cases, in reducing injury rates (Bass 1993; DiGuiseppi 2000; DiGuiseppi 2004; Towner 2001). However, there is concern that there may be differential access to (or a differential effect of) educational interventions by social group (Dowswell 2002; van Weeghal 1997; Kendrick 2000). One systematic review specifically addressing social deprivation and the prevention of childhood unintentional injury (Dowswell 2002) concluded there was little evidence relating to interventions aimed at increasing access to injury prevention services or relating to the differential effect of childhood injury prevention interventions by social group. One small randomised controlled trial evaluating the effect of general practitioner safety advice plus low‐cost safety equipment for families on a low income found the intervention increased safety practices and safety equipment possession and use, and that it was equally effective in families receiving means tested benefits as those not receiving benefits (Clamp 1998). Several studies of cycle helmet educational interventions plus helmet provision have assessed the impact of the intervention by social group. These have found mixed results, with one study finding higher helmet wearing rates amongst children from low‐income communities (DiGuissepi 1989) and others finding lower wearing rates in low‐income communities (Farley 1996; Parkin 1993; Towner 1992). Work from other areas suggests preventive interventions may be less effective amongst the socio‐economically disadvantaged. An intervention aimed at increasing uptake of childhood immunisations increased overall rates of immunisation, but widened inequalities in uptake between deprived and affluent populations (Reading 1994). Smoking cessation programmes have also been found to be less effective amongst the disadvantaged (Brown 1999; Richmond 1993; Osler 1998).
Objectives
The objectives of the review are to:
-
evaluate the effectiveness of home safety education aimed at reducing home injuries in childhood
-
evaluate the effect of providing low‐cost, discounted or free equipment in addition to home safety education
-
evaluate the effect of home safety education and the provision of low‐cost, discounted or free equipment by social group.
Methods
Criteria for considering studies for this review
Types of studies
Individual and cluster randomised controlled trials, non‐randomised controlled trials, and controlled before‐and‐after studies will be included.
Types of participants
Children and young people (aged 0−19) and their families. For the third objective (above) studies must either report data on socio‐economic characteristics or have unpublished data on socio‐economic characteristics available for the review. The socio‐economic and demographic characteristics of interest are those previously found to be associated with an increased risk of childhood injury. These include income, receipt of means tested benefits, maternal education, residence in a deprived area, maternal age at birth of first child, single parenthood, ethnicity and non‐owner occupation.
Types of interventions
Home safety education − provided by health or social care professionals, school teachers, lay workers or voluntary or other organisations, to individual children or families, or groups of children or families − the aim of which is to increase home safety practices or use of home safety equipment or reduce home injuries. We shall include interventions offered in primary and secondary healthcare settings (e.g. primary care practices, clinics, out patient departments, accident and emergency departments, hospital wards), schools and the homes of children and families. Interventions involving the provision of free, low‐cost or discounted safety equipment will be included. Community‐based trials with multi‐faceted interventions will only be included if they include individual home safety education or group education.
Types of outcome measures
-
Possession and use of home safety equipment (e.g. stair gates, fireguards, smoke alarms, cupboard locks, window locks, electrical socket covers, curly kettle flexes).
-
Safety practices (e.g. storage of medicines, sharp objects and cleaning and other hazardous substances, use of baby walkers, storage of matches and lighters, hot water temperature, holding child whilst holding hot liquid, small objects within reach of child).
-
Self‐reported injury in children and young people aged 0−19 years.
-
Medically attended injury in children and young people aged 0−19 years.
Search methods for identification of studies
A search strategy will be devised to identify randomised and non‐randomised studies (as defined above) using terms to identify injuries, safety equipment, practices and behaviours in the target population. The following sources will be searched.
Bibliographic databases
-
Cochrane controlled trials register
-
Cochrane database of systematic reviews
-
MEDLINE
-
EMBASE
-
CINAHL
-
Database of Abstracts and Reviews of Effectiveness
-
ASSIA
-
PsychINFO
-
Web of Science.
The optimally sensitive search strategy for identifying randomised controlled trials will be used (Dickerson 1994) plus non‐randomised studies will be identified from a search strategy not identifying study type as recommended by the Cochrane reviewer's handbook. All databases will be searched from the date of their inception. Articles in all languages will be retrieved.
The following search strategy will be used to search MEDLINE and adapted as necessary for other databases:
1. randomized controlled trial.pt.
2. randomized controlled trials.sh.
3. randomized controlled trial$.mp. or rRandomized cControlled trials/
4. random allocation.sh.
5. double blind method.sh.
6. single blind method.sh.
7. random allocation/
8. 1 or 2 or 3 or 4 or 5 or 6 or 7
9. clinical trials/ or placebos/
10. controlled clinical trial.pt.
11. comparative stud$.mp.
12. intervention stud$.mp.
13. control group$.mp. or control groups/
14. placebo$.mp. or placebos/
15. evaluation stud$.mp.
16. 9 or 10 or 11 or 12 or 13 or 14 or 15
17. human.sh.
18. exp child/
19. exp infant/
20. exp adolescent/
21. minors/
22. (child$ or adolesc$ or infan$ or young$ or minor$ or toddl$ or bab$).mp.
23. 18 or 19 or 20 or 21 or 22
24. "early intervention (education)"/
25. exp education/
26. exp patient education/ or exp health education/ or exp education/
27. exp public health/ed
28. explode parenting/
29. exp counseling/
30. training.mp.
31. (educat$ or train$ or teach$ or parent$ or counsel$).mp. [mp=title, original title, abstract, name of substance, mesh subject heading]
32. 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31
33. exp accident prevention/
34. safety/
35. exp safety management/
36. safety practice$.mp.
37. exp drug storage/
38. hazardous substances/ae, po [adverse effects, poisoning]
39. 33 or 34 or 35 or 36 or 37 or 38
40. safety equipment.mp. or equipment safety/
41. exp infant equipment/
42. protective devices.mp. or exp protective devices/
43. (fire‐guard$ or fireguard$).mp.
44. (stair$ adj3 gate$).mp.
45. (bab$ adj3 walk$).mp.
46. (protect$ adj3 device$).mp.
47. (kettle$ adj3 (flex$ or cable$ or wire$)).mp.
48. (cook$ adj3 guard$).mp.
49. (smok$ adj3 alarm$ smok$ adj3 detect$).mp.
50. 40 or 41 or 42 or 43 or 44 or 47 or 48 or 49
51. 32 or 50
52. exp accidents/ or exp accidents, home/
53. exp eye burns/ or exp burns, chemical/ or exp burns, inhalation/ or exp burns/ or exp burns, electric/
54. smoke inhalation injury/ or smoke/
55. exp poisoning/
56. carbob monoxide poisoning/ or exp poisoning/
57. exp "wounds and injuries"/
58. (accident$ or burn$ or scald$ or asphyx$ or chok$ or cut$ or suffocat$ or poison$ or fracture$ or wound$ or injur$).mp.
59. exp fractures/
60. suffocation.mp. or asphyxia/
61. exp IPECAC/
62. near drowning/ or exp drowning/ or drowning.mp.
63. 52 or 53 or 54 or 55 or 56 or 57 or 58 or 59 or 60 or 61 or 62
64. 8 and 17 and 23 and (32 or 39 or 50) and 63
65. 16 and 17 and 23 and (32 or 39 or 50) and 63
66. 8 or 16
67. 66 and 17 and 23 and (32 or 39 or 50) and 63.
Websites
-
Injury Prevention Research Centers at the Centers for Disease Control (USA)
-
Health Development Agency (UK)
-
Children's Safety Network (USA)
-
International Society for Child and Adolescent Injury Prevention (International)
-
Child Accident Prevention Trust (UK)
-
Injury Control Resource Information Network (USA)
-
National Injury Surveillance Unit (Australia)
-
Injury Prevention Web and SafetyLit (USA).
Hand searching
-
Abstracts from the 1st to 7th World Conferences on Injury Prevention and Control
-
Injury Prevention
-
Reference lists of articles included in the review will be searched.
Identifying unpublished research
-
Search of the National Research Register and Current Controlled Trials
-
Register of the Medical Editors Trial Amnesty
-
Authors of papers included in the review will be asked for details of unpublished research
-
Sure Start (UK), Home Start (UK), Royal Society for the Prevention of Accidents (UK), National Children's Homes (NCH), the Fire and Rescue Service (UK), the Community Practitioners and Health Visitors Association (CPHVA) and Local Authorities (UK) will be contacted to provide details of any home safety programmes that may meet the inclusion criteria
-
Dissertations.
Data collection and analysis
Study selection
A two‐stage screening process will be used. Titles and abstracts of articles will be scanned independently by two reviewers to identify which articles to retrieve in full. Where there is disagreement between reviewers, the full article will be retrieved. Full articles will be independently reviewed by two researchers, using a standard form listing inclusion criteria. Disagreement between researchers will be dealt with by consensus‐forming discussions including a third member of the research team.
Data extraction
We will attempt to obtain individual patient data (IPD) for all eligible studies. Where IPD are obtained, datasets will be formatted across studies to ensure uniformity. If IPD are not available or authors are unwilling to provide them, they will be asked to supply grouped outcome data by socio‐economic characteristics allowing the use of consistent sub‐groupings of patients in the analysis.
A standard data extraction form will be designed and used to extract data on participants, socio‐economic characteristics, interventions and outcomes for extracting data from publications. Data will be extracted by two researchers independently and compared. Any discrepancies will be identified and investigated by referral back to the original article by a senior member of the research team with prior experience of data extraction for systematic reviews. Data extraction will be undertaken blind to the author and institution of the article. Special care will be taken to avoid the inclusion of multiple reports pertaining to the same individuals, for example in trials reporting outcomes over multiple time periods.
Assessing the methodological quality of studies included in the review
Where information regarding study quality is not available from publications, or requires clarification, authors will be contacted to provide this. Quality will be assessed independently by two reviewers. The degree of agreement will be assessed using kappa coefficients. Assessment of quality will be undertaken by researchers blind to the results of each study. For randomised controlled trials allocation concealment, blinding of outcome assessment and completeness of follow‐up will be used as the three markers of trial quality. For non‐randomised controlled trials outcome assessment and completeness of follow‐up will be markers of quality, plus assessment of the distribution of confounders. We will examine the influence of individual aspects of study quality and their influence on effect sizes will be explored in a sensitivity analysis.
Defining outcomes to be used in the analysis
All the outcome measures that will be used in the analysis are naturally dichotomous. These include:
-
possession and use of home safety equipment (e.g. stair gates, fireguards, smoke alarms, cupboard locks, window locks, electrical socket covers, curly kettle flexes)
-
safety practices (e.g. storage of medicines, sharp objects and cleaning and other hazardous substances, use of baby walkers, storage of matches and lighters, hot water temperature, holding child whilst holding hot liquid, small objects within reach of child)
-
one or more self‐reported injuries in children and young people aged 0−19 years
-
one or more medically attended injuries in children and young people aged 0−19 years.
Assessing heterogeneity
Heterogeneity between the results of studies included in the review will be explored using forest and Galbraith plots and chi‐squared tests. Heterogeneity relating to social group will be explored using sub‐group and meta‐regression analyses as described below. The reasons for any other sources of heterogeneity will be explored using sub‐group analyses and meta‐regression where appropriate. If individual level data are available, this will enable a more powerful examination of heterogeneity, which will be undertaken using both a classical approach, and a Bayesian multi‐level approach described below. Where significant heterogeneity exists, the results from studies will not be pooled, and explanations will be sought for the sources of heterogeneity such as differences in study populations, in settings, in baseline levels of risk or in interventions.
Combining treatment effects
Pooled results will be presented as odds ratios and 95% CI for the binary outcomes of possession of items of safety equipment, safety practices and occurrence of at least one self‐reported or medically attended injury. Random effect models will be used predominantly to allow for and quantify the degree of statistical heterogeneity present between individual studies. Where cluster randomised trials have been reported without appropriate adjustment for clustering, we will approach the authors for the individual patient level data and adjust for clustering in the analysis. If individual level data are not available, we will adjust the reported treatment effect for clustering using the intra‐class correlation coefficient (ICC) of the study if available. Otherwise, we will use the ICC of similar cluster randomised trials as part of a sensitivity analysis. The number needed to treat (NNT) to prevent one injury, or to facilitate one family possessing safety equipment or undertaking a safety practice will be calculated.
In addition to the analyses outlined above, a Bayesian multi‐level modeling approach will be used to synthesise IPD from individual and cluster randomised studies and to explore both individual and study level covariates. However, as it is perceived that IPD will not be available for all studies, this framework will also permit summary data to be included when only this is available.
Handling missing data
Data can be missing at different levels when carrying out a meta‐analysis. Most extreme, is when whole studies are missing. The impact of such missing data is assessed using tools to address publication bias as described below. Less extreme, but nevertheless important, is the situation when only partial data is available from a study. We hope to minimise such missing data by requesting the IPD from the original study researchers. Where such attempts are unsuccessful and data cannot be obtained at all, or estimates have to be derived by making further assumptions, the robustness of the overall findings will be assessed through sensitivity analyses.
Sub‐group analyses
Studies may either have been undertaken in a specific socio‐economic group (e.g. low‐income families), or have been undertaken across socio‐economic groups, in which case IPD on socio‐economic variables may be available. Both these types of studies will be included in the review. Where IPD are available, the relationship between treatment effect and socio‐economic characteristics will be assessed by: (a) sub‐group analysis using categorical IPD, or (b) meta‐regression where IPD on continuous measures of socio‐economic status are available. If authors are unable to provide individual patient level data, meta‐regression will be undertaken using summary continuous measures of socio‐economic circumstances such as deprivation scores, income, age at birth of first child or years of maternal education as explanatory variables to explore the relationship between treatment effect and socio‐economic characteristics. For binary measures of socio‐economic characteristics, such as single parenthood or non‐owner occupation, analyses will be undertaken to determine the treatment effect in each sub‐group. Dowswell and Towner have drawn attention to the wide range of socio‐economic variables that have been measured in studies of childhood injury prevention programmes (Dowswell 2002) and it is possible that the relationship between treatment effect and socio‐economic variables will vary depending on the variables used. We will, therefore, undertake sub‐group analyses combining studies that have used comparable socio‐economic variables. If sufficient data are available to undertake sub‐group analyses for several socio‐economic variables, the problems associated with multiple significance testing will be addressed in the analysis and interpretation of the results.
Where there are sufficient studies, pre specified sub‐group analyses will be undertaken examining the effect of factors that may influence outcome including:
-
duration of follow‐up
-
setting of intervention
-
intensity of intervention
-
whether equipment was provided or not
-
individual versus group education.
In addition, where sufficient data are available, sub‐group analyses will be undertaken to examine the relationship between baseline risk and treatment effect with respect to socio‐demographic characteristics that have previously been shown to be associated with injury risk including:
-
ethnicity
-
previous injury
-
family type
-
child age
-
child gender.
Assessing publication bias
The possibility of publication bias will be examined using funnel plots and Egger's test. If there is evidence that publication bias does exist, its likely impact on the results will be assessed using the trim‐and‐fill method, which estimates and adjusts for the numbers and outcomes of studies estimated as being missing and provides a sensitivity analysis to assess the robustness of the results to the likely degree of publication bias in the literature.
Sensitivity analyses
The individual contribution of each study to the pooled result will be assessed graphically. Sensitivity analyses will be undertaken to assess the effect of removing the most influential study from the analysis. The robustness of the findings with respect to study quality will be assessed by comparing treatment effects between:
-
randomised and non‐randomised controlled trials
-
randomised controlled trials with adequate allocation concealment, blinded outcome assessment and at least 80% follow‐up versus those without all three quality markers.
Dissemination of the results of the review
Whilst the review is in progress, it will be included in the UK National Research Register, the title will be registered with the Cochrane Injuries Group and the protocol will be published by the Cochrane Injuries Group. The review will form part of the Cochrane Library and will, therefore, be available electronically to subscribers' worldwide. We will obtain advice from the Cochrane Users Group and the Cochrane Collaboration's Consumer Network regarding dissemination. The results will be published in peer‐reviewed journals, presented at International and National conferences and made available to the UK National Health Service's Centre for Reviews and Dissemination and to the UK Health Development Agency for its "Evidence Base". The results, and recommendations for policy and practice, will be made available to the UK Department of Health, the International Society for Child and Adolescent Injury Prevention, the Child Accident Prevention Trust (UK), the Royal Society for the Prevention of Accidents (UK) and National Children's Homes (UK). Relevant professional bodies will be informed of the results including, in the UK, the Royal College of Paediatrics, the Royal College of General Practitioners, the Faculty of Public Heath Medicine, and the Community Practitioners and Health Visitors. The results will also be made available, in the UK, to the Children's Taskforce for informing the National Service Framework for Children and the External Working Group for healthy children and young people, whose remit is to develop standards for promoting health and wellbeing in childhood.
Involving users in the review
We will obtain advice from the Cochrane Collaboration's Consumer Network regarding user involvement.
