Health Education Research Advance Access originally published online on March 8, 2005
Health Education Research 2005 20(6):709-718; doi:10.1093/her/cyh021
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
How useful are home safety behaviours for predicting childhood injury? A cohort study
1 Division of Primary Care, University Park, Nottingham NG7 2RD, 2 Queen's Medical Centre, University of Nottingham, Nottingham NG7 2HA and 3 Department of Health Sciences, University of Leicester, Leicester LE1 6TP, UK
4 Correspondence to: D. Kendrick; E-mail: denise.kendrick{at}nottingham.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Little work has examined the utility of home safety behaviours in predicting childhood injury. This study examines the relationship between safety behaviours and child injury using a cohort of 1717 families, with 2357 children aged 0–7 years. Safety behaviours, and sociodemographic and family characteristics were measured using a validated questionnaire, and medically attended injuries were ascertained from medical records. Hospital admission rates were lower amongst children from families with fitted and working smoke alarms [incidence rate ratio (IRR) 0.55, 95% confidence interval (CI) 0.31–0.96], who stored sharp objects safely (IRR 0.44, 95% CI 0.23–0.84) and who had fitted stair gates (IRR 0.57, 95% CI 0.31–1.03). Not having a stair gate and not storing sharp objects safely had high sensitivities and negative predictive values for predicting hospital admission. These findings are unlikely to be explained by reductions in the risk of injuries these items are designed to prevent. Families with a range of safety behaviours may also be ‘safer’ in other ways. Further exploration of factors that may explain lower injury rates in these families is required. Information on safety behaviours may be useful for targeting and monitoring injury prevention activity.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In recent years there has been an increasing emphasis on targeting injury prevention towards those at greatest risk. This has been emphasized in UK government policy documents such as Tackling Health Inequalities (Department of Health, 2003
), and implemented through policies such as Health Action Zones (Department of Health, 2000) and Sure Start (Department of Education and Skills, 2002). The recently published National Service Framework (NSF) for Children's Services states ‘Primary Care Trusts and Local Authorities should develop and monitor injury prevention strategies that target priority areas where there are marked inequalities’ (Department of Health, 2004). Such targeting has most commonly, but not exclusively, been focussed on neighbourhoods, or families, which are socioeconomically disadvantaged.
However, there are some limitations to targeting injury prevention based on socioeconomic disadvantage. First, although significant associations may have been demonstrated between disadvantage and injury occurrence (Roberts and Power, 1996; Roberts, 1997; Hippisley-Cox et al., 2002), the usefulness of disadvantage for the purposes of targeting depends on its sensitivity, specificity and predictive value. Two UK studies have assessed the utility of area-based measures of deprivation [the Townsend score which is a census-based measure constructed from four indicators: housing tenure, unemployment, overcrowding and lack of access to a car (Townsend, 1987)], and individual-level measures of deprivation including receipt of means tested benefits, access to a car, housing tenure, parental unemployment, single parenthood, teenage motherhood and overcrowding (Kendrick and Marsh, 1997, 2001). Both found that although living in a deprived area was one of the best predictors of childhood injury, and was better than most individual-level measures of deprivation, it has a relatively low sensitivity, a higher specificity and a low positive predictive value, especially for hospital admissions for injury. Consequently, targeting based on area measures of deprivation leads to identifying many children as being at risk who would not have an injury and failing to detect a substantial number of those who would have an injury.
One of the reasons why individual or area measures of deprivation may have limited use for targeting injury prevention is that their relationship with injury may be explained by a variety of other factors. One potential explanation for why children in disadvantaged areas have higher injury rates is that such families have poorer access to safety equipment (Kendrick, 1994; Gielen et al., 1995; Roberts, 1996) or for a variety of reasons are less likely to practice safe behaviours (Glik et al., 1993; Hapgood et al., 2000). If such relationships could be demonstrated, safety equipment possession and use, and other safety behaviours may prove to be more useful for targeting injury prevention than measures of deprivation.
Several observational studies have examined the relationship between specific home safety behaviours or safety equipment possession and use, and the injuries that such behaviour or equipment could plausibly prevent. These studies have demonstrated a greater risk of stair ways falls in families without stair gates (Elkington et al., 1999), poisonings in families with unsafe storage practices (Azizi et al., 1994), burns in families with less safe practices related to thermal injuries (van Rijn et al., 1991; Petridou et al., 1998), and fires and fire-related deaths amongst those in homes without smoke alarms (Marshall et al., 1998; Runyan et al., 1992). However, as these studies did not employ a prospective design, they were unable to provide information on sensitivity, specificity and predictive value. One small cohort study by the authors failed to find an association between safety behaviours and injury occurrence, possibly due to a low prevalence of unsafe behaviour, limited power or due to combining injuries of all severities into one outcome measure, potentially obscuring a relationship between safety behaviours and more severe injuries (Hapgood et al., 2001).
If it could be demonstrated that home safety behaviours predicted a range of childhood injuries at least as efficiently as individual- or area-level measures of deprivation, this may be useful for public health and injury prevention practitioners for targeting injury prevention programmes and for identifying families who may benefit from items of safety equipment or safety advice. We have therefore undertaken a community-based prospective study, measuring a range of sociodemographic and family characteristics, and using a validated measure of home safety behaviours, to examine the relationship between safety behaviours and childhood injury occurrence in families living in deprived areas of Nottingham, UK.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The objectives of the analyses presented in this paper are to:
- (1) Examine the relationship between safety behaviours and the occurrence of childhood injury.
- (2) Examine the utility of safety behaviours for predicting childhood injury.
- (2) Examine the utility of safety behaviours for predicting childhood injury.
). The trial evaluated the effectiveness of health visitor advice plus access to free or low-cost safety equipment, fitted in the homes of families with children under 5. All families (n = 9909) with children aged under 5 years on the caseloads of 62 health visitors attached to 47 general practices in deprived areas of Nottingham (Townsend score greater than zero) were invited to participate. A Townsend score (see above) greater than zero indicates greater levels of material deprivation.
A total of 3428 families (35%) agreed to participate, and were randomly allocated to the intervention and control arms. The control arm comprised 1717 families with 2357 children. Data on child age and gender were obtained from health visitor caseloads. The exposures of interest were sociodemographic and family characteristics, and safety practices, collected by postal questionnaire at recruitment. Questions on sociodemographic and family characteristics included child age and gender, housing tenure, parental unemployment, overcrowding, receipt of means tested state benefits, lack of access to a car, teenage motherhood, single- or two-parent family, and ethnic group. Questions on safety behaviours included possession and use of fireguards, stair gates, window locks and smoke alarms, and storage of medicines, cleaning products and sharp objects. These safety questions had previously been validated by a home visit (Watson et al., 2003) to compare observed and self-reported practices in a sample of 64 families. The sensitivity of the self-reported questions on safety behaviours for predicting observed safety behaviours ranged from 82 to 100%, whilst the specificity ranged from 60 to 100%. The child poverty index (the percentage of families with children under 16 receiving means tested benefits) at electoral ward level was used as an indicator of area deprivation (Department of Education, Transport and the Regions, 2000). Distance from the centre of the ward of residence to the only hospital with an A & E department in the health district was measured in kilometres.
Three outcome measures were used for this study, which are likely to reflect injuries with a range of severities. These were the primary care injury attendance rate, the A & E attendance rate and the hospital admission rate for unintentional injury during the 2-year follow-up period. Data were collected on medically attended injuries from manually searching primary and secondary care records. Children withdrawing from the study, those placed on the child protection register during the course of the study and those who moved outside Nottingham for whom the date of moving was unknown were excluded from the analysis. Those with a known date of moving were included for the period during which they resided in Nottingham. The number of years at risk of injury was calculated for each child, taking account of date of entry to the study and length of follow-up.
Data analysis
Continuous data have been described using means and SDs where they were normally distributed, and medians and interquartile ranges (IQR) where they were non-normally distributed. Categorical data were explored using frequencies and cross-tabulations, and described using frequencies and percentages. Injury rates with exact Poisson 95% confidence intervals (CIs) were calculated for the three injury outcomes using child years at risk of injury as the denominator.
Poisson regression was used to examine the relationships between sociodemographic characteristics, safety behaviours and injury rates. Where there was significant variation in injury rates between families, a two-level random intercepts model was used, with child at level 1 and family at level 2, and child years at risk of injury as the offset term. Wald tests were used to assess significance for fixed parameter estimates, and the Self and Liang 
2-test for estimates of the variances of random effects (Self and Liang, 1987). Where there was evidence of non-linearity of covariates, these were categorized.
Models were built by adding variables (fixed effects) with P 
0.1 in the univariate analysis into the model, in order of their significance on univariate analysis. They were retained in the model if they were significant at the 5% level. Two-way interactions between covariates that seemed theoretically plausible were examined. Models were checked by examining plots of residual values, leverage and influence.
All sociodemographic variables had less than 5% of data missing and such cases were excluded from the analysis. Six percent of data were missing for window locks, and this variable was included in the analysis by creating an indicator variable with categories for missing and non-missing data. The number of safety practices was summed for each family to produce a score. The effect of missing data was examined by creating three scores: one containing only cases with complete data, one assuming missing responses were safe and one assuming they were unsafe. The results were robust to these assumptions, so that the score containing complete data was used in the analyses.
In order to assess how useful safety behaviours are for predicting which families will have an injury, we calculated the sensitivity, specificity, predictive value and positive likelihood ratio of not having specific safety behaviours for predicting hospital admission for injury. For these analyses we focussed on hospital admissions, as these are likely to be the most severe injuries and those whose prevention may be most important. As many injury prevention programmes are targeted at families living in deprived areas, the usefulness of each of the safety behaviours was compared to that of the child poverty index by plotting the sensitivity against 1 – specificity for at least one hospital admission. Analyses were undertaken using Stata (version 7; Stata, College Station, TX) and MLwiN (version 1.1; Institute of Education, University of London).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Baseline data on sociodemographic characteristics, family characteristics and safety behaviours were available for 1642 families (95.6%). Outcome data were available on primary care injury attendances for 2066 children (87.7%), and A & E attendances and hospital admissions for 2273 children (96.4%). The sociodemographic characteristics, family characteristics and safety behaviours of study families are shown in Table I.
|
The frequencies of injury outcomes are shown in Table II. The univariate relationships between safety behaviours and each of the injury outcomes are shown in Table III. None of the safety behaviours were significantly associated with the primary care attendance rate. Children whose families had a fitted and working smoke alarm had a significantly lower A & E attendance rate than those without. Children whose families had a fitted and working smoke alarm, a stair gate and those storing sharp objects safely in the kitchen had significantly lower hospital admission rates than those without these practices. An increase by one in the number of safety practices was associated with a significant reduction in the hospital admission rate.
|
|
The results of the multivariable analyses are shown in Table IV. Children whose families had a fitted and working smoke alarm had an 18% lower A & E attendance rate. There were two possible models for hospital admission rates; this occurred because smoke alarm and stair gate possession confounded the relationship between the other and the hospital admission rate. The smoke alarm and stair gate models explained a similar amount of the variation in the hospital admission rate (16 and 14%, respectively). There was some evidence that children whose families had a fitted stair gate had a lower hospital admission rate (43% lower), and those whose families had a fitted and working smoke alarm had a significantly lower admission rate (45% lower) than those without these practices. In both models, children in families storing sharp objects safely had significantly lower admission rates (55 and 56% in the stair gate and smoke alarm models, respectively) than those in families not storing sharp objects safely. An increase by one in the number of safety practices was associated with a significant reduction (18%) in the hospital admission rate.
|
The sensitivity, specificity and predictive value for not having a smoke alarm, a stair gate and not storing sharp objects safely in the kitchen are shown in Table V. Not having a stair gate and unsafe storage of sharp objects had a high sensitivity and negative predictive value for predicting hospital admission. Table V also shows the positive likelihood ratios for each of the three safety behaviours. Families whose children had a hospital admission for injury were 1.6 times more likely not to have a smoke alarm, 1.3 times more likely not to have a stair gate and 4.8 times more likely not to store sharp objects safely than families whose children did not have an admission. Figure 1 shows the receiver operator curve for the child poverty index (measured on a continuous scale) for at least one hospital admission for injury, with the sensitivity and 1– specificity for not having a smoke alarm, a stair gate and unsafe storage of sharp objects superimposed. Each of the safety practices performed at least as well as the child poverty index in predicting hospital admission.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Principal findings
Having fitted and working smoke alarms and safe storage of sharp objects in the kitchen were associated with significantly lower hospital admission rates for unintentional injuries, and there was some evidence suggesting that having a stair gate may be associated with a lower hospital admission rate. An increasing number of safety practices were associated with a reducing risk of hospital admission. Having fitted and working smoke alarms was also associated with a significantly lower A & E attendance rate. These relationships were independent of sociodemographic and family characteristics and ward level deprivation. Not having a stair gate or unsafe storage of sharp objects identified most of the families whose children were admitted to hospital for an injury, and the majority of families who had a stair gate and stored sharp objects safely did not have an admission. Not having a smoke alarm, a stair gate or unsafe storage of sharp objects performed at least as well as the child poverty index in predicting hospital admission.
Strengths and weaknesses of the study
This is the first UK study to prospectively examine the relationship between sociodemographic characteristics, safety behaviours and childhood injury. The use of a questionnaire validated by home observations (Watson et al., 2003) provides greater confidence that self-reported practices reflect actual practice. The study had sufficient power to demonstrate relationships between sociodemographic characteristics, safety behaviours and injury rates. This was achieved as a result of high follow-up rates, especially for A & E attendances and hospital admissions.
The study was not designed to examine the relationships between safety behaviours and the specific injuries they could potentially prevent. This would require data on injury mechanisms, which is frequently incompletely recorded in medical records, especially those in primary care (Agass et al., 1990; Marsh et al., 1995). In addition, as some injury mechanisms are rare, such as falls down stairs or house fires, a much larger sample size would have been required. Although we collected data on a range of outcomes including primary and secondary care attendances and hospital admissions for injury, without an objective measure of injury severity we cannot conclude that certain safety behaviours are associated with more or less severe injuries.
Changes in safety behaviours over time, either because of changing safety needs as children grow older or increased awareness or availability of equipment, may have resulted in misclassification of the safety behaviours of some families. This will tend to underestimate the effect of safety behaviours on the risk of injury. Hence, failure to find an association between some behaviours and injury occurrence may have resulted either from using a single measurement of behaviours at baseline rather then repeated measurements over time or from using a relatively long follow-up period when use of the safety equipment and behaviours may be dependent on the age of the children within a family.
It is possible that families have over-reported safety behaviours (Scott, 1997). Only safety behaviours that were validated by home observations (Watson et al., 2003), and had high sensitivities and specificities for predicting observed behaviours, were used in an attempt to minimize this. However, families agreeing to the home visit may have been less likely to over-report safety than those not agreeing, in which case some over-reporting may still have occurred. Such over-reporting would tend to bias the estimated effect of safety behaviours towards unity, so underestimating the true effect of safety behaviours on injury.
Finally, study participants were those that agreed to participate in a randomized controlled trial evaluating a child injury prevention intervention. Families agreeing to participate in a trial may well differ from those that do not agree to participate. It is likely that families who were more safety conscious and more interested in injury prevention were more likely to participate in the trial. As some non-participants (n = 629) also returned the baseline questionnaire, we have been able to compare the safety behaviours of participants and non-participants. Participants differed from non-participants for only one of the safety behaviours; they were significantly less likely to store sharp objects safely in the kitchen than non-participants (39 versus 43%, 
= 4.16, P = 0.04). We therefore consider it unlikely that differences between participants and non-participants will limit the generalizability of our findings to a large degree.
Comparisons with previous work and explanation of the study findings
There is little research examining the relationship between safety behaviours in the home and injury outcomes. Our previous study found no association between the safety score and whether a child in the family had a medically attended injury (Hapgood et al., 2001). The positive findings from this study may have resulted from greater power, a higher prevalence of unsafe behaviour or through using separate injury outcomes that may reflect injuries of differing severities.
Although several case-control studies have found associations between safety behaviours and the specific injuries that they could plausibly prevent (van Rijn et al., 1991; Runyan et al., 1992; Azizi et al., 1994; Marshall et al., 1998; Petridou et al., 1998), it is important to highlight that this is unlikely to explain our findings. As the number of admissions for injuries resulting from house fires, stair way falls and cuts from sharp objects stored in the kitchen will have been small in our study, it is unlikely that lower hospital admission rates amongst families with smoke alarms, stair gates and who stored sharp objects safely reflect reductions in these injuries. This suggests that families who are ‘safe’ in these ways are also ‘safer’ in other ways than families without these behaviours. This may reflect differences in beliefs about the preventability of childhood injury (Girasek, 2001), perceptions of the risk of or susceptibility to injury (Glik et al., 1993; Wortel et al., 1995), beliefs about the necessity and success of undertaking safety measures (Wortel et al., 1995; Russell et al., 1996), perceived social norms (Sellstrom and Bremberg, 1996) or in child supervisory styles, practices and the effective use of ‘safety rules’ (Sparks et al., 1994; Morrongiello et al., 2001, 2004a,b; Soubhi, 2004). Further work is required to explore the relationship between these factors, deprivation, safety behaviours and childhood injuries.
Implications for practice and further research
Although area-based measures of deprivation are routinely collected, and hence may appeal as an easy and cheap method of identifying families at high risk, there are several reasons why data on safety behaviours may be particularly useful. First, not only does such data identify families at higher risk of injury, but it also identifies potential reasons why a family is at higher risk. Second, we have demonstrated that families lacking certain safety behaviours are at higher risk of a range of injuries, not just those that could be prevented by that safety behaviour. Consequently, high-risk families identified using this method should then be offered injury prevention appropriate to preventing a range of childhood injuries. Third, if data on safety behaviours is collected within the child health promotion programme in routine contacts such as the 6–8 week check or immunizations, these contacts will also provide opportunities to provide safety advice and help families to access safety equipment. Finally, targeting can take place within such contacts as the health care provider can identify those at higher risk during the contact. Such data collection and targeting need not be confined to health care providers. The recent NSF for children's services has charged home visitors with providing advice on preventing a range of injuries (Department of Health, 2004) and enquiring about safety behaviours is a prerequisite to providing appropriate advice.
If data on safety behaviours were routinely recorded in the Personal Child Health Record and recorded electronically, they could act as positive health indicators to inform decisions about health promotion priorities, to raise awareness within communities and amongst policy makers regarding injury risk, and to monitor injury prevention activity (Catford, 1983).
Although we have demonstrated that certain safety behaviours are associated with a reduced risk of injury, a larger study is required to explore the relationship between safety behaviours and injury severity, using a measure of injury severity that is independent of health service utilization. As the prevalence of some safety behaviours, e.g. smoke alarm ownership, is increasing (Roberts, 1996), further work will be required to examine whether this changes the relationship between safety behaviours and risk of injury.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Agass, M., Mant, D., Fuller, A., Coulter, A. and Jones, L. (1990) Childhood accidents: a practice survey using general practitioners' records and parental reports. British Journal of General Practice, 40, 202–205.
Azizi, B.H., Zulkifli, H.I. and Kassim, M.S. (1994) Circumstances surrounding accidental poisoning in children. Medical Journal of Malaysia, 49, 132–137.
Catford, J.C. (1983) Positive health indicators—towards a new information base for health promotion. Community Medicine, 5, 125–132.[Medline]
Department for Education and Skills (2002) Making a Difference for Children and Families. Sure Start. Department for Education and Skills, London.
Department of Health (2000) The NHS Plan. The Stationery Office, London.
Department of Health (2003) Tackling Health Inequalities. A Programme for Action. Department of Health, London.
Department of Health (2004) National Service Framework for Children, Young People and Maternity Services. Core Standards. Department of Health, London.
Department of the Environment Transport and Regions (2000) Indices of Deprivation 2000. London: Department of the Environment, Transport and Regions, London.
Elkington, J., Blogg, S., Kelly, J. and Carey, V. (1999) Head injuries in infants: a closer look at baby walkers, stairs and nursery furniture. NSW Public Health Bulletin, 10, 82–83.
Gielen, A.C., Wilson, M.E., Faden, R.R., Wissow, L. and Harvilchuck, J.D. (1995) In-home injury prevention practices for infants and toddlers: the role of parental beliefs, barriers and housing quality. Health Education Quarterly, 22, 85–95.[Web of Science][Medline]
Girasek, D.C. (2001) Public beliefs about the preventability of unintentional injury deaths. Accident Analysis and Prevention, 33, 455–465.
Glik, D., Kronenfeld, J. and Jackson, K. (1993) Safety behaviors among parents of preschoolers. Health Values, 17, 18–27.
Hapgood, R., Kendrick, D. and Marsh, P. (2000) How well do socio-demographic characteristics explain variation in childhood safety practices? Journal of Public Health Medicine, 22, 307–311.
Hapgood, R., Kendrick, D. and Marsh, P. (2001) Do self reported safety behaviours predict childhood unintentional injuries? Injury Prevention, 7, 14–17.
Hippisley-Cox, J., Groom, L., Kendrick, D., Coupland, C., Webber, E. and Savelyich, B. (2002) Cross sectional survey of socioeconomic variations in severity and mechanism of childhood injuries in Trent 1992–7. British Medical Journal, 324, 1132–1786.
Kendrick, D. (1994) Children's safety in the home: parents' possession and perceptions of the importance of safety equipment. Public Health, 108, 21–25.[CrossRef][Web of Science][Medline]
Kendrick, D. and Marsh, P. (1997) Injury prevention programmes in primary care: a high risk group or a whole population approach? Injury Prevention, 3, 170–175.
Kendrick, D. and Marsh, P. (2001) How useful are sociodemographic characteristics in identifying children at risk of unintentional injury? Public Health, 115, 103–107.[CrossRef][Web of Science][Medline]
Marsh, P., Kendrick, D. and Williams, E.I. (1995) The impact of audit on the recording of childhood accidental injury information by primary health care teams. Audit Trends, 3, 47–52.
Marshall, S.W., Runyan, C.W., Bangdiwala, S.I., Linzer, M.A., Sacks, J.J. and Butts, J.D. (1998) Fatal residential fires: who dies and who survives? Journal of the American Medical Association, 279, 1633–1637.
Morrongiello, B.A., Midgett, C. and Shields, R. (2001) Don't run with scissors: young children's knowledge of home safety rules. Journal of Pediatric Psychology, 26, 105–115.
Morrongiello, B.A., Ondejko, L. and Littlejohn, A. (2004a) Understanding toddlers' in-home injuires: I. Context, correlates and determinants. Journal of Pediatric Psychology, 29, 415–431.
Morrongiello, B.A., Ondejko, L. and Littlejohn, A. (2004b) Understanding toddlers' in-home injuries: II. Examining parental strategies and their efficacy, for managing child injury risk. Journal of Pediatric Psychology, 29, 433–446.
Petridou, E., Trichopoulos, D., Mera, E., Papadatos, Y., Papazoglou, K., Marantos, A. and Skondras, C. (1998) Risk factors for childhood burn injuries: a case-control study from Greece. Burns, 24, 123–128.[CrossRef][Web of Science][Medline]
Roberts, I. (1996) Smoke alarm use: prevalence and household predictors. Injury Prevention, 2, 263–265.
Roberts, I. (1997) Cause specific social class mortality differentials for child injury and poisoning in England and Wales. Journal of Epidemiology and Community Health, 51, 334–335.
Roberts, I. and Power, C. (1996) Does the decline in child injury mortality vary by social class? A comparison of class specific mortality in 1981 and 1991. British Medical Journal, 313, 784–786.
Runyan, C.W., Bangdiwala, S.I., Linzer, M.A., Sacks, J.J. and Butts, J. (1992) Risk factors for fatal residential fires. New England Journal of Medicine, 327, 859–863.[Abstract]
Russell, K.M. and Champion, V.L. (1996) Health beliefs and social influence in home safety practices of mothers with preschool children. Image—Journal of Nursing Scholarship, 28, 59–64.
Self, G.S. and Liang, K.Y. (1987) Asymptotic properties of maximum likelihood estimators and likelihood ratio tests under nonstandard conditions. Journal of the American Statistical Association, 82, 605–610.[CrossRef][Web of Science]
Scott, I. (1997) You can't believe all that you're told: the issue of unvalidated questionnaires. Injury Prevention, 3, 5–6.
Sellstrom, E. and Bremberg, S. (1996) Perceived social norms as crucial determinants of mother's injury-preventive behaviour. Acta Paediatrica, 85, 702–707.[Web of Science][Medline]
Soubhi, H. (2004) The social context of childhood injury in Canada: integration of the NLSCY findings. American Journal of Health Behavior, 28(Suppl. 1), S38–S50.
Sparks, G., Craven, M.A. and Worth, C. (1994) Understanding differences between high and low childhood accident rate areas: the importance of qualitative data. Journal of Public Health Medicine, 16, 439–786.
Townsend, P. (1987) Deprivation. Journal Social Policy, 16, 125–146.
van Rijn, O.J., Bouter, L.M., Kester, A.D., Knipschild, P.G. and Meertens, R.M. (1991) Aetiology of burn injuries among children aged 0–4 years: results of a case-control study. Burns, 17, 213–219.[CrossRef][Web of Science][Medline]
Watson, M., Woods, A. and Kendrick, D. (2002) Injury prevention: working together on an RCT. Community Practitioner, 75, 172–175.
Watson, M., Kendrick, D. and Coupland, C. (2003) Validation of a home safety questionnaire used in a randomised controlled trial. Injury Prevention, 9, 180–183.
Wortel, E., de Geus, G.H. and Kok, G. (1995) Behavioral determinants of mothers' safety measures to prevent injuries of pre-school children. Scandinavian Journal of Psychology, 36, 306–322.[Web of Science][Medline]
Received on August 25, 2004; accepted on September 17, 2004
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. C. LeBlanc, I. B. Pless, W. J. King, H. Bawden, A.-C. Bernard-Bonnin, T. Klassen, and M. Tenenbein Home safety measures and the risk of unintentional injury among young children: a multicentre case-control study Can. Med. Assoc. J., October 10, 2006; 175(8): 883 - 887. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||