- Email: [email protected]
The impact of home safety promotion on different social strata in a WHO safe community
Public Health (2006) 120, 427–433
ORIGINAL RESEARCH
The impact of home safety promotion on different social strata in a WHO safe community T. Timpka*, P. Nilsen, K. Lindqvist ¨ping University, Division of Social Medicine and Public Health, Department of Health and Society, Linko ¨ping, Sweden Linko Received 14 April 2005; received in revised form 21 November 2005; accepted 7 December 2005
KEYWORDS Home injuries; Community-based safety promotion; Interventions; Evaluation; Quasi-experimental methods
Summary Objectives: Few studies have investigated the impact of home safety promotion programmes on different social strata. The aim of this study was to investigate the distribution of effects of a community-based home safety programme on home injury rates among families with different connections to the labour market. Methods: A quasi-experimental design was used, with pre- and post-implementation registrations covering the total populations below 65 years of age in the programme implementation area (population 41,000) and in a neighbouring comparison ¨ stergo municipality (population 26,000) in O ¨tland County, Sweden. Results: In the intervention and comparison areas, households in which the adults were not vocationally active displayed the highest rates of home injury. After 6 years of programme activity, the home injury rates for males and females in all social status categories displayed a decreasing trend in the intervention area. The opposite was true for the comparison area, i.e. the incidence of injury increased, with the exception of females in non-vocationally active households. The decline in injury rates in the intervention area was statistically significant for males and females in the employed category and for males in the non-vocationally active category. Changes in injury rates in the comparison area were not statistically significant. Conclusion: The programme was partially successful in that it reduced the injury rate in non-vocationally active households, but it did not influence the injury rate in the employed households. The study design did not allow for conclusions regarding why the post-intervention injury rates remained higher in non-vocationally active households. Further research on the association between the incidence of home injury and socio-economic factors is warranted. Q 2006 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved.
Introduction * Corresponding author. Tel.: C46 13 222383; fax: C46 13 221865. E-mail address: [email protected] (T. Timpka).
After the traffic environment, the home is the most common location of unintentional fatal
0033-3506/$ - see front matter Q 2006 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2005.12.002
428 injuries.1 Studies in Scandinavia have estimated the proportion of home injuries to between 25 and 35% of all injuries requiring treatment in an emergency department.2–4 The home is the most frequent location for fatal and non-fatal injuries among children younger than 15 years of age.5 The most common mechanisms of fatal home injuries among infants and young children are suffocation, fires and burns, and drowning.6 During adolescence, deaths caused by unintentional discharge of firearms account for one-third of all fatalities in the home.5 For the elderly, falls are the main cause of home-related injury events.7 Falls are also a major cause of hospitalization from injury in most countries. About one in three hospital care days for males and more than one-half for females are due to femoral fractures, with the primary cause being falls in the home.8 Home injury prevention and safety promotion programmes have proved to be effective at decreasing the number of injuries that occur in the home.9 However, unintentional injuries within the home environment have not been addressed to the same extent as road traffic or occupational injuries. It may simply be that home injuries are not spectacular enough; when people think about injury, they tend to focus their concern on dramatic, multiple-death injuries that are covered by the media.10 Unfortunately, this obscures the fact that a large number of injuries occur in private homes. Home injuries have been reported to be more common in households with poorer social circumstances.11,12 While evidence supports the position that vulnerable populations living in poor social conditions are at a disproportionate risk of injury,13–16 no studies to date have investigated the impact of home safety promotion programmes on different social strata. The study described in this paper addresses this gap in knowledge. A safety promotion programme addressing home injuries in Motala, Sweden (population 41,000) achieved an 18% overall decrease in the incidence of home injury.17 The objective of this study was to investigate differences in the distribution of this home injury rate reduction among the different social strata in the catchment area. Specifically, the aim was to study, using a quasi-experimental design,18 rates of home injury treated by healthcare organizations among members of households at different levels of labour market integration before and after programme implementation.
T. Timpka et al.
Methods The Motala community is one of the original reference sites for the World Health Organization (WHO) Safe Community accreditation criteria. The Safe Community concept was developed in Sweden in conjunction with the WHO, based on findings from local Swedish injury prevention programmes in the 1970s and 1980s. Scandinavian countries were among the first to implement the Safe Community model in the late 1980s and early 1990s.19 The model emphasizes community participation and multidisciplinary collaboration, recognizing that those most able to solve local injury problems are those people who live in that particular community.20 An assessment of the general structure and process of the Motala programme has been reported previously.21
Study design A quasi-experimental design was used, with preand post-implementation registrations covering the total populations below 65 years of age in the programme implementation area (Motala) and in a neighbouring comparison municipality (Mjo ¨lby) in ¨ stergo O ¨ tland County. The pre-implementation study period covered 52 weeks from 1 October 1983 to 30 September 1984. The post-implementation period covered 52 weeks from 1 January 1989 to 31 December 1989. Changes in the morbidity rates following the intervention were studied using prospective registration of all acute care episodes during the study period.
Data collection Data were collected from two sources. For all patients contacting a healthcare unit located in the intervention area during the study periods, a report form was completed by staff at the care unit with the time of contact and standard personal data. The same form included whether unintentional injury was a possible reason for the contact. For registered patients, specially trained nurses recorded an ICD-8-based diagnostic classification using medical records and discussions with physicians. The routine for data collection was tested in a pilot project. Prior to each study period, the staff at all relevant healthcare units was informed carefully and the routine was practiced for 2 weeks.
Impact of home safety promotion on different social strata
Classification of data The Swedish Socio-economic Index (SEI) was used to classify the individuals in the study areas. The SEI has been employed since the early 1980s to represent social status in most national databases and statistics. The SEI defines social status primarily as being based on occupation. Children and young people are categorized to the SEI group to which their parents’ household belongs. SEI data for all individuals in the intervention and comparison areas were collected from Statistics Sweden (http://www.scb.se). Due to the changes in socio-economic household circumstances concerning retirement from work, individuals older than 65 years of age were excluded from the analyses.22 For the pre-implementation measurement, SEI data originated in the Census survey conducted in 1985. Corresponding data for the postimplementation measurement originated in the 1990 Census survey. Considering that the WHO Safe Community model relies strongly on the existing civic social network, and that occupation is an important determinant for these networks, the detailed SEI categories were used for coding individuals into three secondary categories based on the relation that the household had to the labour market: (1) households in which the vocationally significant member was employed; (2) households in which the vocationally significant member was an entrepreneur or self-employed; and (3) households in which the adults were not vocationally active.
429
the national average (60%). Members of households in which the vocationally significant member was employed constituted the largest share of the population below 65 years of age in both the intervention (82%) and comparison (81%) areas. Members of self-employed households represented 7 and 9%, respectively, and members of households classified as not vocationally active constituted 10% of the population below 65 years of age in both areas.
Statistical methods Injury rates (expressed as per 100 person-years) were calculated by community (intervention and control municipality) for each study period (1983/1984 and 1989), by socio-economic group (employed, selfemployed and not vocationally active), and by gender, as well as for women and men together. Ninety-five percent confidence intervals (CI) were provided for injury rates. For individuals injured more than once, only the first episode during each registration period was included in the calculations. The differences in injury rates between 1989 and 1983/1984 were computed for both areas with 95%CI, and significance tests were conducted. A P-value !0.05 was considered to be statistically significant. Similarly, differences in changes of injury rate between the intervention and control areas were computed using the following expression: Difference in changes of injury rate Z ½post—intervention injury rate in intervention area Kpre—intervention injury rate in intervention areaK
Community characteristics Motala is situated in the western part of the county of ¨ stergo O ¨tland, on the shore of Lake Va ¨ttern. The population was approximately 41,000 during the study period, with 82% living in the central and residential areas and the remainder living in surrounding rural districts. Manufacturing, trade and public administration were the main occupations, with 77% of the gainfully employed working within these fields. The comparison municipality area, Mjo ¨lby (population 26,000), was selected on the basis of socio-economic and demographic similarities to Motala and the availability of injury data. The age and sex mix in both areas were close to the national average and were stable over the registration periods. The average income in both study areas was 93% of the national average. In both areas, the proportion of adults with more than an elementary school education remained 5% below
½post—intervention injury rate in control area Kpre—intervention injury rate in control area All computations were performed using SPSS statistical software (Version 11.5).
Results Quality of injury data During the pre-implementation registration period, which included people of all ages, identity data were missing for less than 1% of the injured patients in the intervention and comparison areas. During the postimplementation registration period, which also included people of all ages, less than 0.5% of the injured patients in the intervention and comparison areas could not be identified in medical records.
6.4 (5.6, 7.2) 4.4 (3.5, 5.2) 5.1 (4.1, 6.1) 3.9 (2.8, 5.0) 3.4 (2.7, 4.2) 2.9 (2.2, 3.6) 4.5 (3.2, 5.8) 3.2 (2.0, 4.5) 3.0 (2.7, 3.3) 1.9 (1.6, 2.2) 4.3 (3.9, 4.6) 2.6 (2.3, 2.9) Intervention area Comparison area
Females
3.6 (3.4, 3.9) 2.2 (2.0, 2.5)
2.8 (2.0, 3.6) 2.7 (1.8, 3.6)
8.0 (6.7, 9.3) 4.9 (3.6, 6.3)
Total Females Males Total Females Males Males
Total
Self-employed Employed
Not vocationally active
T. Timpka et al. Table 1 Rate per 100 person-years (95% confidence interval) of individuals injured at home in 1983/1984 in the intervention and comparison areas, displayed by sex and household relation to labour market.
430
A lower proportion of the injured residents from the intervention area (11/422, 3%, 95%CI 1–5%) compared with the comparison area (28/253, 11%, 95%CI 7–15%) were found to have been directly provided with acute care outside the local healthcare units during the comparison study. Police records did not disclose previously unrecorded injuries.
Pre-intervention injury rates In both the intervention and the comparison areas, households in which the adults were not vocationally active displayed the highest home injury rates (Table 1). In both study areas, males in households in which the vocationally significant member was employed and males in households in which the adults were not vocationally active had a higher incidence of injury than females. The opposite relationship existed for the ‘self-employed’ category, i.e. females displayed higher home injury rates than males in households in which the vocationally significant member was an entrepreneur or selfemployed. The pre-intervention injury rates were noticeably higher in the intervention area than in the comparison area for all three socio-economic groups.
Post-intervention injury rates After 6 years of programme activity, the home injury rates for males and females in all social status categories displayed a decreasing trend in the intervention area (Table 2). The opposite was true for the comparison area, i.e. the injury incidence increased, with the exception of females in non-vocationally active households who demonstrated a minor decline. The decline in injury rates in the intervention area was statistically significant for males and females in the employed category and for males in the non-vocationally active category. Changes in injury rates in the comparison area were not statistically significant. Many of the pre-intervention patterns within the study areas remained the same after the implementation of interventions. Non-vocationally active households had the highest incidence of injury in both study areas, and males sustained injuries more frequently than females in all social status groups except for the self-employed category.
Discussion This study analysed the WHO Safe Community model for safety promotion with regard to associations between pre- and post-intervention home
0.2 (K1.0, 1.5) 0.748 K0.1 (K1.7, 1.5) 0.899 0.5 (K1.5, 2.5) 0.619 1.4 (0.2, 2.6) 0.023 1.5 (K0.5, 3.4) 0.146 1.3 (K0.2, 2.8) 0.078 0.3 (K0.1, 0.7) 0.154 0.4 (K0.1, 0.8) 0.145
0.3 (0.0, 0.7) 0.043
5.0 (4.3, 5.7) K1.4 (K2.5, K0.3) 0.012 4.6 (3.7, 5.4) 4.2 (3.3, 5.1) K0.8 (K2.2, 0.5) 0.222 3.8 (2.7, 4.9) 5.9 (4.8, 7.0) K2.1 (K3.8, K0.4) 0.016 5.4 (4.0, 6.8) 2.6 (1.9, 3.2) K0.9 (K1.9, 0.1) 0.081 4.3 (3.3, 5.2) 3.6 (2.3, 4.9) K0.9 (K2.8, 1.0) 0.356 4.7 (3.1, 6.3) 2.0 (1.2, 2.7) K0.8 (K1.9, 0.3) 0.144 4.0 (2.8, 5.2) 2.7 (2.5, 2.9) K0.9 (K1.2, K0.7) 0.000 2.6 (2.3, 2.8) 2.4 (2.1, 2.6) K0.6 (K1.0, K0.2) 0.002 2.2 (1.9, 2.5)
Females
3.0 (2.7, 3.3) K1.3 (K1.7, K0.9) 0.000 3.0 (2.6, 3.3)
Intervention area Change 1989– 1983 P value Comparison area Change 1989– 1983 P value
Total Females
Not vocationally active
Males Total Females
Self-employed
Males Males
Total Employed
Table 2 Rate per 100 person-years (95% confidence interval) of individuals injured at home in 1989 and change in rates between 1989 and 1983/1984 (95% confidence interval) in the intervention and comparison areas, displayed by sex and household relation to labour market.
Impact of home safety promotion on different social strata
431
injury rates among males and females, and socioeconomic status, as defined by the employment category of the household’s significant member. The observation that no changes in injury rates were recorded in the comparison area suggests that the reduction of home injuries in the intervention area between 1983 and 1989 was likely to be attributable to the safety promotion programme. As expected, those socially disadvantaged as indicated by the SEI categories were at the highest pre-intervention injury risk, indicating that lower socio-economic status is an important risk factor for injury; this is consistent with previous research.16 The present study design did not allow for an investigation into the causes of these differences, although a possible explanation could be the more prevalent use of unsafe domestic products in deprived households. However, this hypothesis is contradicted by a recent study reporting that higher household income and owner-occupied (as opposed to rented) homes are associated with an increased prevalence of fall hazards (e.g. more stairs without railings and greater use of ladders).23 Further research on the association between the incidence of home injury and socio-economic factors is thus warranted. Another finding that requires further study is that females in the self-employed category displayed higher injury rates than males. A possible explanation may be that women in self-employed settings, e.g. farms, mix household activities with ‘enterprise’ work, and thereby introduce additional risks into their homes. With regard to the programme’s aim of equality in safety issues between social groupings, the programme was only partially successful in that it reduced the injury rate in non-vocationally active households but it did not influence the injury rate of the employed households. However, the non-vocationally active households achieved the largest decrease in injury rate of all groups, i.e. the group with the lowest socio-economic status obtained the greatest benefit from the programme. This finding indicates that the programme’s approach of combining a population-based, community-wide strategy with more targeted interventions to groups at increased risk, i.e. children/teenagers and the elderly, was successful in reducing health inequalities. It has been suggested that a ‘pure’ population strategy is only appropriate when risk is widely diffused through the whole population.16 Previous studies have shown that injuries in childhood are related both to poverty at the household level and to living in a deprived neighbourhood, and that these influences are independent.24 This evidence suggests a parallel use of community-wide efforts and more
432 targeted area-based interventions in order to reduce home injuries in children. Another explanation could be substance abuse among adult members of nonvocationally active households, leading to accidents and/or neglect in childminding. While substance abuse has been associated with the occurrence of injury, the association has not been characterized by type of substance and injury type. However, a recent study has reported that alcohol and cocaine use is independently associated with violence-related injuries, whereas opiate use is independently associated with non-violent injuries and burns.25 Screening for substance abuse was not included in the present study, and this hypothesis remains to be addressed in future studies. This study had a number of limitations that are important to consider when interpreting the results. First, it is noteworthy that individuals older than 65 years were excluded from the analyses, although this age category, together with the youngest, is at the highest risk of home injury.26 However, while previous studies have shown that socio-economic factors play an important role in the causation of injuries among young people,27 socio-economic deprivation does not appear to be a risk factor for significant injuries in elderly people.28,29 Instead, lifestyle-related risk factors seem to be more important influences on the occurrence of injury in older age groups. Hence, as the aim of this study was to investigate the impact of home safety interventions in different SEI categories, inclusion of those aged over 65 years would have required separate calculations. In conclusion, the programme was partially successful in that it reduced the injury rate in the non-vocationally active households, but it did not influence the injury rate in the employed households. The study design did not allow for conclusions regarding why the post-intervention injury rates remained higher in the non-vocationally active categories. Further research on the association between the incidence of home injury and socio-economic factors is warranted.
Acknowledgements This study was supported by grants from the Swedish National Rescue Services Board.
References 1. Appy MK. The Home Safety Council and its challenges. Am J Prev Med 2005;28:70–1.
T. Timpka et al. 2. Lindqvist K, Schelp L, Timpka T. Home injuries in a Swedish municipality—consequences and costs. Saf Sci 1999;31:19–29. ¨ konomisk 3. Lauritsen JM, Kidholm K, Skov O, No ¨rga ˚rd L. O daekningsgrad og gennemsnitlige omkostninger ved hospitalsbehandlet tilskadekomst. Ugeskr Laeger 2002;164: 5107–12 [in Danish]. ¨ddningsverket Olyckor i siffror: En rapport om olycksut4. Ra vecklingen i Sverige. Karlstad: Ra ¨ddningsverket; 2004 [in Swedish]. 5. Pollock DA, McGee DL, Rodriguez JG. Deaths due to injury in the home among persons under 15 years of age, 1970–1984. MMWR Morb Mortal Wkly Rep 1988;37:13–20. 6. Widome MD. Pediatric injury prevention for the practitioner. Curr Probl Pediatr 1991;21:428–68. 7. Sjo ¨gren H, Bjo ¨rnstig U. Trauma in the elderly: the impact on the health care system. Scand J Prim Health Care 1991;9: 203–7. ˚. Falko 8. Svanstro ¨m L, Schelp L, Ekman R, Lindstro ¨m A ¨ping, Sweden, ten years after: still a safe community? Int J Consum Saf 1996;3:1–7. 9. Bass JL, Christoffel KK, Widome M, Boyle W, Scheidt P, Stanwick R, et al. Childhood injury prevention in primary care settings: a critical review of the literature. Pediatrics 1993;92:544–50. 10. Christoffel T, Gallagher SS. Injury prevention and public health. Gaithersburg: Aspen Publishers; 1999. 11. Stokes M, Ashby K, Clapperton A. The effect of socio-economic status on injury among Victorians. Victorian injury surveillance & applied research system. Hazard 2002;(49):1–12. 12. Svanstro ¨m L. Evidence-based injury prevention and safety promotion: state-of-the-art. In: Mohan D, Tiwari G, editors. Injury prevention and control. London: Taylor & Francis; 2000. 13. Wilkinson R, Marmot M. Social determinants of health: the solid facts. Copenhagen: WHO Regional Office for Europe; 1998. 14. Laing GJ, Logan S. Patterns of unintentional injury in childhood and their relation to socio-economic factors. Public Health 1999;113:291–4. 15. Cubbin C, LeClerc FB, Smith GS. Socioeconomic status and injury mortality: individual and neighbourhood determinants. J Epidemiol Community Health 2000;54:517–24. 16. Moodie R. Reducing health inequalities: challenges to promoting health and preventing injury. Victorian injury surveillance & applied research system. Hazard 2002; Summer(49):13–20. ˚hlgren M. Evaluation of a 17. Lindqvist K, Timpka T, Schelp L, A home injury prevention program in a WHO Safe Community. Int J Consum Prod Saf 1999;6:25–32. 18. Cook TD, Campell DT. Quasi-experimentation. Boston, MA: Houghton-Mifflin Co.; 1979. 19. Zhao Z, Svanstro ¨m L. Injury status and perspectives on developing community safety promotion in China. Health Promot Int 2003;18:247–53. 20. Bjerre B, Schelp L. The community safety approach in Falun, Sweden—is it possible to characterize the most effective prevention endeavours and how long-lasting are the results? Accid Anal Prev 2000;32:461–70. 21. Lindqvist K, Timpka T, Schelp L. Ten years of experiences from a participatory community-based injury prevention program in Motala, Sweden. Public Health 1996;110: 339–46. 22. Silventoinen K, Lahelma E. Health inequalities by education and age in four Nordic countries, 1986 and 1994. J Epidemiol Community Health 2002;56:253–8.
Impact of home safety promotion on different social strata 23. Marshall SW, Runyan CW, Yang J, Coyne-Beasley T, Waller AE, Johnson RM, et al. Prevalence of selected risk and protective factors for falls in the home. Am J Prev Med 2005;28:95–101. 24. Haynes R, Reading R, Gale S. Household and neighbourhood risks for injury to 5–14 year old children. Soc Sci Med 2003; 57:625–36. 25. Blondell RD, Dodds HN, Looney SW, Lewis CM, Hagan JL, Lukan JK, et al. Toxicology screening results: injury associations among hospitalized trauma patients. J Trauma 2005;58:561–70. 26. Runyan CW, Perkis D, Marshall SW, Johnson RM, CoyneBeasley T, Waller AE, et al. Unintentional injuries in the
433
home in the United States. Part II. Morbidity. Am J Prev Med 2005;28:80–7. 27. Dowswell T, Towner E. Social deprivation and the prevention of unintentional injury in childhood: a systematic review. Health Educ Res 2002;17:221–37. 28. West J, Hippisley-Cox J, Coupland CA, Price GM, Groom LM, Kendrick D, et al. Do rates of hospital admission for falls and hip fracture in elderly people vary by socio-economic status? Public Health 2004;118:576–81. 29. Jones S, Johansen A, Brennan J, Butler J, Lyons RA. The effect of socioeconomic deprivation on fracture incidence in the United Kingdom. Osteoporos Int 2004;15:520–4.
ORIGINAL RESEARCH
The impact of home safety promotion on different social strata in a WHO safe community T. Timpka*, P. Nilsen, K. Lindqvist ¨ping University, Division of Social Medicine and Public Health, Department of Health and Society, Linko ¨ping, Sweden Linko Received 14 April 2005; received in revised form 21 November 2005; accepted 7 December 2005
KEYWORDS Home injuries; Community-based safety promotion; Interventions; Evaluation; Quasi-experimental methods
Summary Objectives: Few studies have investigated the impact of home safety promotion programmes on different social strata. The aim of this study was to investigate the distribution of effects of a community-based home safety programme on home injury rates among families with different connections to the labour market. Methods: A quasi-experimental design was used, with pre- and post-implementation registrations covering the total populations below 65 years of age in the programme implementation area (population 41,000) and in a neighbouring comparison ¨ stergo municipality (population 26,000) in O ¨tland County, Sweden. Results: In the intervention and comparison areas, households in which the adults were not vocationally active displayed the highest rates of home injury. After 6 years of programme activity, the home injury rates for males and females in all social status categories displayed a decreasing trend in the intervention area. The opposite was true for the comparison area, i.e. the incidence of injury increased, with the exception of females in non-vocationally active households. The decline in injury rates in the intervention area was statistically significant for males and females in the employed category and for males in the non-vocationally active category. Changes in injury rates in the comparison area were not statistically significant. Conclusion: The programme was partially successful in that it reduced the injury rate in non-vocationally active households, but it did not influence the injury rate in the employed households. The study design did not allow for conclusions regarding why the post-intervention injury rates remained higher in non-vocationally active households. Further research on the association between the incidence of home injury and socio-economic factors is warranted. Q 2006 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved.
Introduction * Corresponding author. Tel.: C46 13 222383; fax: C46 13 221865. E-mail address: [email protected] (T. Timpka).
After the traffic environment, the home is the most common location of unintentional fatal
0033-3506/$ - see front matter Q 2006 The Royal Institute of Public Health. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.puhe.2005.12.002
428 injuries.1 Studies in Scandinavia have estimated the proportion of home injuries to between 25 and 35% of all injuries requiring treatment in an emergency department.2–4 The home is the most frequent location for fatal and non-fatal injuries among children younger than 15 years of age.5 The most common mechanisms of fatal home injuries among infants and young children are suffocation, fires and burns, and drowning.6 During adolescence, deaths caused by unintentional discharge of firearms account for one-third of all fatalities in the home.5 For the elderly, falls are the main cause of home-related injury events.7 Falls are also a major cause of hospitalization from injury in most countries. About one in three hospital care days for males and more than one-half for females are due to femoral fractures, with the primary cause being falls in the home.8 Home injury prevention and safety promotion programmes have proved to be effective at decreasing the number of injuries that occur in the home.9 However, unintentional injuries within the home environment have not been addressed to the same extent as road traffic or occupational injuries. It may simply be that home injuries are not spectacular enough; when people think about injury, they tend to focus their concern on dramatic, multiple-death injuries that are covered by the media.10 Unfortunately, this obscures the fact that a large number of injuries occur in private homes. Home injuries have been reported to be more common in households with poorer social circumstances.11,12 While evidence supports the position that vulnerable populations living in poor social conditions are at a disproportionate risk of injury,13–16 no studies to date have investigated the impact of home safety promotion programmes on different social strata. The study described in this paper addresses this gap in knowledge. A safety promotion programme addressing home injuries in Motala, Sweden (population 41,000) achieved an 18% overall decrease in the incidence of home injury.17 The objective of this study was to investigate differences in the distribution of this home injury rate reduction among the different social strata in the catchment area. Specifically, the aim was to study, using a quasi-experimental design,18 rates of home injury treated by healthcare organizations among members of households at different levels of labour market integration before and after programme implementation.
T. Timpka et al.
Methods The Motala community is one of the original reference sites for the World Health Organization (WHO) Safe Community accreditation criteria. The Safe Community concept was developed in Sweden in conjunction with the WHO, based on findings from local Swedish injury prevention programmes in the 1970s and 1980s. Scandinavian countries were among the first to implement the Safe Community model in the late 1980s and early 1990s.19 The model emphasizes community participation and multidisciplinary collaboration, recognizing that those most able to solve local injury problems are those people who live in that particular community.20 An assessment of the general structure and process of the Motala programme has been reported previously.21
Study design A quasi-experimental design was used, with preand post-implementation registrations covering the total populations below 65 years of age in the programme implementation area (Motala) and in a neighbouring comparison municipality (Mjo ¨lby) in ¨ stergo O ¨ tland County. The pre-implementation study period covered 52 weeks from 1 October 1983 to 30 September 1984. The post-implementation period covered 52 weeks from 1 January 1989 to 31 December 1989. Changes in the morbidity rates following the intervention were studied using prospective registration of all acute care episodes during the study period.
Data collection Data were collected from two sources. For all patients contacting a healthcare unit located in the intervention area during the study periods, a report form was completed by staff at the care unit with the time of contact and standard personal data. The same form included whether unintentional injury was a possible reason for the contact. For registered patients, specially trained nurses recorded an ICD-8-based diagnostic classification using medical records and discussions with physicians. The routine for data collection was tested in a pilot project. Prior to each study period, the staff at all relevant healthcare units was informed carefully and the routine was practiced for 2 weeks.
Impact of home safety promotion on different social strata
Classification of data The Swedish Socio-economic Index (SEI) was used to classify the individuals in the study areas. The SEI has been employed since the early 1980s to represent social status in most national databases and statistics. The SEI defines social status primarily as being based on occupation. Children and young people are categorized to the SEI group to which their parents’ household belongs. SEI data for all individuals in the intervention and comparison areas were collected from Statistics Sweden (http://www.scb.se). Due to the changes in socio-economic household circumstances concerning retirement from work, individuals older than 65 years of age were excluded from the analyses.22 For the pre-implementation measurement, SEI data originated in the Census survey conducted in 1985. Corresponding data for the postimplementation measurement originated in the 1990 Census survey. Considering that the WHO Safe Community model relies strongly on the existing civic social network, and that occupation is an important determinant for these networks, the detailed SEI categories were used for coding individuals into three secondary categories based on the relation that the household had to the labour market: (1) households in which the vocationally significant member was employed; (2) households in which the vocationally significant member was an entrepreneur or self-employed; and (3) households in which the adults were not vocationally active.
429
the national average (60%). Members of households in which the vocationally significant member was employed constituted the largest share of the population below 65 years of age in both the intervention (82%) and comparison (81%) areas. Members of self-employed households represented 7 and 9%, respectively, and members of households classified as not vocationally active constituted 10% of the population below 65 years of age in both areas.
Statistical methods Injury rates (expressed as per 100 person-years) were calculated by community (intervention and control municipality) for each study period (1983/1984 and 1989), by socio-economic group (employed, selfemployed and not vocationally active), and by gender, as well as for women and men together. Ninety-five percent confidence intervals (CI) were provided for injury rates. For individuals injured more than once, only the first episode during each registration period was included in the calculations. The differences in injury rates between 1989 and 1983/1984 were computed for both areas with 95%CI, and significance tests were conducted. A P-value !0.05 was considered to be statistically significant. Similarly, differences in changes of injury rate between the intervention and control areas were computed using the following expression: Difference in changes of injury rate Z ½post—intervention injury rate in intervention area Kpre—intervention injury rate in intervention areaK
Community characteristics Motala is situated in the western part of the county of ¨ stergo O ¨tland, on the shore of Lake Va ¨ttern. The population was approximately 41,000 during the study period, with 82% living in the central and residential areas and the remainder living in surrounding rural districts. Manufacturing, trade and public administration were the main occupations, with 77% of the gainfully employed working within these fields. The comparison municipality area, Mjo ¨lby (population 26,000), was selected on the basis of socio-economic and demographic similarities to Motala and the availability of injury data. The age and sex mix in both areas were close to the national average and were stable over the registration periods. The average income in both study areas was 93% of the national average. In both areas, the proportion of adults with more than an elementary school education remained 5% below
½post—intervention injury rate in control area Kpre—intervention injury rate in control area All computations were performed using SPSS statistical software (Version 11.5).
Results Quality of injury data During the pre-implementation registration period, which included people of all ages, identity data were missing for less than 1% of the injured patients in the intervention and comparison areas. During the postimplementation registration period, which also included people of all ages, less than 0.5% of the injured patients in the intervention and comparison areas could not be identified in medical records.
6.4 (5.6, 7.2) 4.4 (3.5, 5.2) 5.1 (4.1, 6.1) 3.9 (2.8, 5.0) 3.4 (2.7, 4.2) 2.9 (2.2, 3.6) 4.5 (3.2, 5.8) 3.2 (2.0, 4.5) 3.0 (2.7, 3.3) 1.9 (1.6, 2.2) 4.3 (3.9, 4.6) 2.6 (2.3, 2.9) Intervention area Comparison area
Females
3.6 (3.4, 3.9) 2.2 (2.0, 2.5)
2.8 (2.0, 3.6) 2.7 (1.8, 3.6)
8.0 (6.7, 9.3) 4.9 (3.6, 6.3)
Total Females Males Total Females Males Males
Total
Self-employed Employed
Not vocationally active
T. Timpka et al. Table 1 Rate per 100 person-years (95% confidence interval) of individuals injured at home in 1983/1984 in the intervention and comparison areas, displayed by sex and household relation to labour market.
430
A lower proportion of the injured residents from the intervention area (11/422, 3%, 95%CI 1–5%) compared with the comparison area (28/253, 11%, 95%CI 7–15%) were found to have been directly provided with acute care outside the local healthcare units during the comparison study. Police records did not disclose previously unrecorded injuries.
Pre-intervention injury rates In both the intervention and the comparison areas, households in which the adults were not vocationally active displayed the highest home injury rates (Table 1). In both study areas, males in households in which the vocationally significant member was employed and males in households in which the adults were not vocationally active had a higher incidence of injury than females. The opposite relationship existed for the ‘self-employed’ category, i.e. females displayed higher home injury rates than males in households in which the vocationally significant member was an entrepreneur or selfemployed. The pre-intervention injury rates were noticeably higher in the intervention area than in the comparison area for all three socio-economic groups.
Post-intervention injury rates After 6 years of programme activity, the home injury rates for males and females in all social status categories displayed a decreasing trend in the intervention area (Table 2). The opposite was true for the comparison area, i.e. the injury incidence increased, with the exception of females in non-vocationally active households who demonstrated a minor decline. The decline in injury rates in the intervention area was statistically significant for males and females in the employed category and for males in the non-vocationally active category. Changes in injury rates in the comparison area were not statistically significant. Many of the pre-intervention patterns within the study areas remained the same after the implementation of interventions. Non-vocationally active households had the highest incidence of injury in both study areas, and males sustained injuries more frequently than females in all social status groups except for the self-employed category.
Discussion This study analysed the WHO Safe Community model for safety promotion with regard to associations between pre- and post-intervention home
0.2 (K1.0, 1.5) 0.748 K0.1 (K1.7, 1.5) 0.899 0.5 (K1.5, 2.5) 0.619 1.4 (0.2, 2.6) 0.023 1.5 (K0.5, 3.4) 0.146 1.3 (K0.2, 2.8) 0.078 0.3 (K0.1, 0.7) 0.154 0.4 (K0.1, 0.8) 0.145
0.3 (0.0, 0.7) 0.043
5.0 (4.3, 5.7) K1.4 (K2.5, K0.3) 0.012 4.6 (3.7, 5.4) 4.2 (3.3, 5.1) K0.8 (K2.2, 0.5) 0.222 3.8 (2.7, 4.9) 5.9 (4.8, 7.0) K2.1 (K3.8, K0.4) 0.016 5.4 (4.0, 6.8) 2.6 (1.9, 3.2) K0.9 (K1.9, 0.1) 0.081 4.3 (3.3, 5.2) 3.6 (2.3, 4.9) K0.9 (K2.8, 1.0) 0.356 4.7 (3.1, 6.3) 2.0 (1.2, 2.7) K0.8 (K1.9, 0.3) 0.144 4.0 (2.8, 5.2) 2.7 (2.5, 2.9) K0.9 (K1.2, K0.7) 0.000 2.6 (2.3, 2.8) 2.4 (2.1, 2.6) K0.6 (K1.0, K0.2) 0.002 2.2 (1.9, 2.5)
Females
3.0 (2.7, 3.3) K1.3 (K1.7, K0.9) 0.000 3.0 (2.6, 3.3)
Intervention area Change 1989– 1983 P value Comparison area Change 1989– 1983 P value
Total Females
Not vocationally active
Males Total Females
Self-employed
Males Males
Total Employed
Table 2 Rate per 100 person-years (95% confidence interval) of individuals injured at home in 1989 and change in rates between 1989 and 1983/1984 (95% confidence interval) in the intervention and comparison areas, displayed by sex and household relation to labour market.
Impact of home safety promotion on different social strata
431
injury rates among males and females, and socioeconomic status, as defined by the employment category of the household’s significant member. The observation that no changes in injury rates were recorded in the comparison area suggests that the reduction of home injuries in the intervention area between 1983 and 1989 was likely to be attributable to the safety promotion programme. As expected, those socially disadvantaged as indicated by the SEI categories were at the highest pre-intervention injury risk, indicating that lower socio-economic status is an important risk factor for injury; this is consistent with previous research.16 The present study design did not allow for an investigation into the causes of these differences, although a possible explanation could be the more prevalent use of unsafe domestic products in deprived households. However, this hypothesis is contradicted by a recent study reporting that higher household income and owner-occupied (as opposed to rented) homes are associated with an increased prevalence of fall hazards (e.g. more stairs without railings and greater use of ladders).23 Further research on the association between the incidence of home injury and socio-economic factors is thus warranted. Another finding that requires further study is that females in the self-employed category displayed higher injury rates than males. A possible explanation may be that women in self-employed settings, e.g. farms, mix household activities with ‘enterprise’ work, and thereby introduce additional risks into their homes. With regard to the programme’s aim of equality in safety issues between social groupings, the programme was only partially successful in that it reduced the injury rate in non-vocationally active households but it did not influence the injury rate of the employed households. However, the non-vocationally active households achieved the largest decrease in injury rate of all groups, i.e. the group with the lowest socio-economic status obtained the greatest benefit from the programme. This finding indicates that the programme’s approach of combining a population-based, community-wide strategy with more targeted interventions to groups at increased risk, i.e. children/teenagers and the elderly, was successful in reducing health inequalities. It has been suggested that a ‘pure’ population strategy is only appropriate when risk is widely diffused through the whole population.16 Previous studies have shown that injuries in childhood are related both to poverty at the household level and to living in a deprived neighbourhood, and that these influences are independent.24 This evidence suggests a parallel use of community-wide efforts and more
432 targeted area-based interventions in order to reduce home injuries in children. Another explanation could be substance abuse among adult members of nonvocationally active households, leading to accidents and/or neglect in childminding. While substance abuse has been associated with the occurrence of injury, the association has not been characterized by type of substance and injury type. However, a recent study has reported that alcohol and cocaine use is independently associated with violence-related injuries, whereas opiate use is independently associated with non-violent injuries and burns.25 Screening for substance abuse was not included in the present study, and this hypothesis remains to be addressed in future studies. This study had a number of limitations that are important to consider when interpreting the results. First, it is noteworthy that individuals older than 65 years were excluded from the analyses, although this age category, together with the youngest, is at the highest risk of home injury.26 However, while previous studies have shown that socio-economic factors play an important role in the causation of injuries among young people,27 socio-economic deprivation does not appear to be a risk factor for significant injuries in elderly people.28,29 Instead, lifestyle-related risk factors seem to be more important influences on the occurrence of injury in older age groups. Hence, as the aim of this study was to investigate the impact of home safety interventions in different SEI categories, inclusion of those aged over 65 years would have required separate calculations. In conclusion, the programme was partially successful in that it reduced the injury rate in the non-vocationally active households, but it did not influence the injury rate in the employed households. The study design did not allow for conclusions regarding why the post-intervention injury rates remained higher in the non-vocationally active categories. Further research on the association between the incidence of home injury and socio-economic factors is warranted.
Acknowledgements This study was supported by grants from the Swedish National Rescue Services Board.
References 1. Appy MK. The Home Safety Council and its challenges. Am J Prev Med 2005;28:70–1.
T. Timpka et al. 2. Lindqvist K, Schelp L, Timpka T. Home injuries in a Swedish municipality—consequences and costs. Saf Sci 1999;31:19–29. ¨ konomisk 3. Lauritsen JM, Kidholm K, Skov O, No ¨rga ˚rd L. O daekningsgrad og gennemsnitlige omkostninger ved hospitalsbehandlet tilskadekomst. Ugeskr Laeger 2002;164: 5107–12 [in Danish]. ¨ddningsverket Olyckor i siffror: En rapport om olycksut4. Ra vecklingen i Sverige. Karlstad: Ra ¨ddningsverket; 2004 [in Swedish]. 5. Pollock DA, McGee DL, Rodriguez JG. Deaths due to injury in the home among persons under 15 years of age, 1970–1984. MMWR Morb Mortal Wkly Rep 1988;37:13–20. 6. Widome MD. Pediatric injury prevention for the practitioner. Curr Probl Pediatr 1991;21:428–68. 7. Sjo ¨gren H, Bjo ¨rnstig U. Trauma in the elderly: the impact on the health care system. Scand J Prim Health Care 1991;9: 203–7. ˚. Falko 8. Svanstro ¨m L, Schelp L, Ekman R, Lindstro ¨m A ¨ping, Sweden, ten years after: still a safe community? Int J Consum Saf 1996;3:1–7. 9. Bass JL, Christoffel KK, Widome M, Boyle W, Scheidt P, Stanwick R, et al. Childhood injury prevention in primary care settings: a critical review of the literature. Pediatrics 1993;92:544–50. 10. Christoffel T, Gallagher SS. Injury prevention and public health. Gaithersburg: Aspen Publishers; 1999. 11. Stokes M, Ashby K, Clapperton A. The effect of socio-economic status on injury among Victorians. Victorian injury surveillance & applied research system. Hazard 2002;(49):1–12. 12. Svanstro ¨m L. Evidence-based injury prevention and safety promotion: state-of-the-art. In: Mohan D, Tiwari G, editors. Injury prevention and control. London: Taylor & Francis; 2000. 13. Wilkinson R, Marmot M. Social determinants of health: the solid facts. Copenhagen: WHO Regional Office for Europe; 1998. 14. Laing GJ, Logan S. Patterns of unintentional injury in childhood and their relation to socio-economic factors. Public Health 1999;113:291–4. 15. Cubbin C, LeClerc FB, Smith GS. Socioeconomic status and injury mortality: individual and neighbourhood determinants. J Epidemiol Community Health 2000;54:517–24. 16. Moodie R. Reducing health inequalities: challenges to promoting health and preventing injury. Victorian injury surveillance & applied research system. Hazard 2002; Summer(49):13–20. ˚hlgren M. Evaluation of a 17. Lindqvist K, Timpka T, Schelp L, A home injury prevention program in a WHO Safe Community. Int J Consum Prod Saf 1999;6:25–32. 18. Cook TD, Campell DT. Quasi-experimentation. Boston, MA: Houghton-Mifflin Co.; 1979. 19. Zhao Z, Svanstro ¨m L. Injury status and perspectives on developing community safety promotion in China. Health Promot Int 2003;18:247–53. 20. Bjerre B, Schelp L. The community safety approach in Falun, Sweden—is it possible to characterize the most effective prevention endeavours and how long-lasting are the results? Accid Anal Prev 2000;32:461–70. 21. Lindqvist K, Timpka T, Schelp L. Ten years of experiences from a participatory community-based injury prevention program in Motala, Sweden. Public Health 1996;110: 339–46. 22. Silventoinen K, Lahelma E. Health inequalities by education and age in four Nordic countries, 1986 and 1994. J Epidemiol Community Health 2002;56:253–8.
Impact of home safety promotion on different social strata 23. Marshall SW, Runyan CW, Yang J, Coyne-Beasley T, Waller AE, Johnson RM, et al. Prevalence of selected risk and protective factors for falls in the home. Am J Prev Med 2005;28:95–101. 24. Haynes R, Reading R, Gale S. Household and neighbourhood risks for injury to 5–14 year old children. Soc Sci Med 2003; 57:625–36. 25. Blondell RD, Dodds HN, Looney SW, Lewis CM, Hagan JL, Lukan JK, et al. Toxicology screening results: injury associations among hospitalized trauma patients. J Trauma 2005;58:561–70. 26. Runyan CW, Perkis D, Marshall SW, Johnson RM, CoyneBeasley T, Waller AE, et al. Unintentional injuries in the
433
home in the United States. Part II. Morbidity. Am J Prev Med 2005;28:80–7. 27. Dowswell T, Towner E. Social deprivation and the prevention of unintentional injury in childhood: a systematic review. Health Educ Res 2002;17:221–37. 28. West J, Hippisley-Cox J, Coupland CA, Price GM, Groom LM, Kendrick D, et al. Do rates of hospital admission for falls and hip fracture in elderly people vary by socio-economic status? Public Health 2004;118:576–81. 29. Jones S, Johansen A, Brennan J, Butler J, Lyons RA. The effect of socioeconomic deprivation on fracture incidence in the United Kingdom. Osteoporos Int 2004;15:520–4.