Are we failing our rural communities? Motor vehicle injury in British Columbia, Canada, 2001–2007

Are we failing our rural communities? Motor vehicle injury in British Columbia, Canada, 2001–2007

Injury, Int. J. Care Injured 43 (2012) 1888–1891 Contents lists available at ScienceDirect Injury journal homepage: www.elsevier.com/locate/injury ...

176KB Sizes 0 Downloads 13 Views

Injury, Int. J. Care Injured 43 (2012) 1888–1891

Contents lists available at ScienceDirect

Injury journal homepage: www.elsevier.com/locate/injury

Are we failing our rural communities? Motor vehicle injury in British Columbia, Canada, 2001–2007 Nathaniel Bell a,*, Richard K. Simons a,b, Nasira Lakha c, S. Morad Hameed a,b a

Department of Surgery, University of British Columbia, Canada Trauma Services, Vancouver General Hospital, British Columbia, Canada c British Columbia Trauma Registry, Canada b

A R T I C L E I N F O

A B S T R A C T

Article history: Accepted 21 July 2011

In Canada, stratification by geographic area or socio-economic status remains relatively rare in national and provincial reporting and surveillance for injury prevention and trauma care. As injuries are known to affect some populations more than others, a more nuanced understanding of injury risk may in turn inform more effective prevention policy. In this study we assessed rates of hospitalization and death from motor vehicle collisions (MVC) in British Columbia (BC) by socio-economic status (SES) and by rural and urban status between 2001 and 2007. Excess risk in injury morbidity and mortality between different SES groups were assessed using a population attributable fraction (PAF). Over a six-year period rural populations in BC experienced a three-fold increase in relative risk of death and an average of 50% increase in relative risk of hospitalization due to injury. When assessed against SES, relative risk of MVC mortality increased from 2.36 (2.05–2.72) to 4.07 (3.35–4.95) in reference to the least deprived areas, with an estimated 40% of all MVC-related mortality attributable to the relative differences across SES classes. Results from this study challenge current provincial and national reporting practises and emphasize the utility of employing the PAF for assessing variations in injury morbidity and mortality. ß 2011 Elsevier Ltd. All rights reserved.

Keywords: Motor vehicle collisions Rural areas Urban areas Socio-economic status Geographic information system Population attributable fraction

Introduction In Canada, motor vehicle-collisions (MVC) are a leading cause of injury-related morbidity and mortality.1 In British Columbia (BC), motor vehicle-related injuries account for over half of all injuryrelated hospitalizations (55%), ER visits (58%), and partial (54%) and permanent (56%) disability.1 Recent reports, however, suggest that the burden of MVC is improving, showing a 30% decrease in hospitalization, 48% decrease in death, and a 26% decrease in partial and permanent disability from all transport related accidents in BC since 1998.1,2 Despite these substantial improvements, it remains unknown whether these changes are equally distributed across all populations and amongst all geographic regions. In the US, rural crashes have been shown to account for the majority of motor vehicle collisions and deaths.3–6 However, albeit with few exceptions,7 stratification by geographic area in Canadian injury prevention literature remains rare, and is largely void from provincial and national reporting efforts. Current administrative data sources on injury, when linked with secondary administrative data, provide the opportunity to assess

* Corresponding author at: Trauma Services, Vancouver General Hospital, 855 W. 12th Ave., Vancouver, BC V5Z 1M9, Canada. Tel.: +1 604 875 4111x66122; fax: +1 604 875 5348. E-mail address: [email protected] (N. Bell). 0020–1383/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.injury.2011.07.018

whether variations in injury are associated with broader conditions from the social and physical environment. As injuries are known to affect some populations more than others,3,8–10 a more nuanced understanding of injury risk may in turn inform more effective prevention policy. The primary objective of this study was to examine social and geographic variations in motor vehicle-related injury hospitalization and death throughout British Columbia, Canada and identify where and amongst which populations injury morbidity and mortality remain most prevalent. Methods Included in this study were adult (ages >17) patient records from the British Columbia Trauma Registry (BCTR) from January 1, 2001 to March 31, 2007, inclusive. The BCTR is the definitive source of all moderate and severe injury in BC, capturing all persons transported directly or indirectly to one of the province’s designated trauma centres as a result of major trauma.b All injuries with a severity score (ISS) 12 were included in the

b There are currently 13 definitive care/trauma hospitals in BC that regularly contribute data to the BCTR. However, only eight of these facilities have done so since 2001 without interruption, including Vancouver General, Children’s Hospital, Royal Columbian, Lion’s Gate, Royal Jubilee, Victoria General, Kelowna General, and Royal Inland hospital. Data for this study is inclusive to these eight facilities to minimize bias due to yearly decrease/increase in reporting on severe injury.

N. Bell et al. / Injury, Int. J. Care Injured 43 (2012) 1888–1891

analysis. In Canada, one of the criteria to qualify for inclusion of patient information into provincial trauma registries is having a single or multiple system ISS score greater than or equal to 12.11–14 This criteria is also used as a working definition for what constitutes submission to the Major Trauma Dataset of the National Trauma Registry (NTR). All injury cases were identified using the International Classification of Diseases 10th revision external cause of injury inclusion codes within the range of V20–29 (motorcycle accident) and V40–79 (vehicular accident). All pre-hospital, in-transport and in-hospital deaths amongst drivers, passengers, and riders where a motorcycle, car, bus, truck, van, or transport vehicle was identified as the cause of death were identified using British Columbia Coroner Service (BCCS) records. BCCS case records were stratified into the same categories as the BCTR data and were inclusive to the entire study period. All incident rates were calculated as age-standardized rates. Population person-years for 2001–2007 were based on BC Stats regional population estimates and projections for provincial Local Health Authority (LHA) geographic boundaries.15 A geographic information system (GIS) was used to spatially link the registry data by postal code and/or city of residence into to the LHA boundary that encapsulated its area. All data linkages were made using the person’s postal code or municipality of residence. LHAs were assigned as ‘urban’ jurisdictions if they contained Census Tract (CT) boundaries and ‘rural’ otherwise. Census CTs exist for all areas in Canada with a core metropolitan population of at least 50,000 people. In BC, these areas include Prince George, Kamloops, Kelowna, Abbottsford, Greater Vancouver, the Capital District of Victoria, and Nanaimo. This schema resulted in 81% of the trauma registry populations being classified as urban and 19% as rural, which was within 1% of the most recent urban and rural population estimates generated by Statistics Canada for BC.16 Area socio-economic status (SES) was assessed using a modified version of the Vancouver Area Neighbourhood Deprivation Index (VANDIX),17 which is a single measure of SES derived from a weighted aggregate of seven variables taken from the census, including in order of weighted importance (weight provided in parentheses): the % of persons without a high school education (0.250), area unemployment rate (0.214), the % of persons with a university degree (0.179), % of lone parent families (0.143), average income (0.089), % of home owners (0.089), and the area’s employment ratio (0.036). The VANDIX was constructed from a 2006 survey of provincial Medical Health Officers (MHOs) asking them to identify which variables they felt best characterized poor health outcomes in BC.17 A portion of the 2006 Census long-form questionnaire has been recalculated for BC LHAs. Five of the seven variables used to construct the VANDIX required no adjustments from the original measure. Two measures, average income and area unemployment rate, required a change from their original calculation. Average income was calculated as the ‘Average Family Economic Income’

1889

and area unemployment was based on the proportion of EI Beneficiaries as a % of population aged 19–64. The number of injuries and population size of each LHA were assigned to one of four SES categories (Low SES, Med-Low SES, Med-High SES, High SES) and stratified by age and population. Age standardized rates were weighted against 2006 national Census counts. Four age groupings were created from the standard population (ages 18–24, 25–44, 45–64, 65+), forcing them to add to 1,000,000 by multiplying the population percentage of each age group into a standard million population. Standard proportions included population counts and percentages of the population aged for 0–17 for building the population weights. With the exception of the aggregated analysis of injuries against SES, the denominator for all yearly age-standardized incidence rates was obtained using the estimated LHA population provided by BC Stats. For the analysis against SES, the total number of person-years in the population denominator was equal to the total population estimate for the 2001 LHA boundaries multiplied by relevant number of years and months of case data from the registries. Age standardized rate ratios were calculated using Poisson log-linear regression. All tests were two sided with a 5% level of significance. We calculated the population attributable fraction (PAF) to assess the relationship between socio-economic inequality and risk of injury morbidity and mortality. Following the method illustrated by Krieger et al., and Hanley, the overall PAF equaled the weighted average of the relevant age-specific PAF, with weights defined by the proportion of cases in each strata.18,19 The reference population was all persons who resided in the least deprived strata. All statistical analyses were generated using SAS software, Version 9.2 for Windows.20 Results Table 1 presents counts of the number of injury hospitalization and death for persons ages 18 and over in BC from 2001 to 2007 by injury mechanism and area SES. A total of 5.1% of the province’s population resided in Low SES strata. Of this 5.1%, 100% of this population was represented by the province’s rural populations. This population group experienced 8.2% of all hospitalizations and 12.1% of all deaths. In contrast, 58.6% of the population was classified into the highest SES areas, with 96.5% of these areas represented by urban populations. Injury frequency increased step-wise across all SES classes relative to the highest SES category. The rate ratios and the population attribute fraction (PAF) are shown in Table 2. Significant stepwise increases in MVC/motorcycle hospitalization and death were observed across SES strata. These trends were most visible amongst injury mortality, with a stepwise socio-economic increase in age-adjusted relative risk ratios from 2.36 (2.05–2.72), 2.75 (2.33–3.25), and 4.07 (3.35–4.95) across all SES strata in reference to the highest (i.e. least deprived) SES areas. The population-adjusted PAF suggests that 40% of MVC fatalities would not have occurred if population rates amongst

Table 1 Distribution of population, by injury mechanism and area SES, 2001–2007. % of all transport collisionsa

n

Hospitalizations Deaths

Reference population Rural Urban a

2340 1851

77.0 82.0

Injury distribution by VANDIX score (%) Low SES

Med-Low SES

Med-High SES

High SES

8.2 12.1

11.4 19.1

29.5 33.6

50.8 35.0

n

Population geography (%)

Population distribution by VANDIX SES strata (%)

3,178,862 732,801 2,446,061

100.0 23.1 76.9

5.1 100.0 0.0

Includes all pedestrian, cycling, vehicular, and motorcycle injuries.

11.9 67.0 33.0

24.2 32.5 67.5

58.6 3.5 96.5

N. Bell et al. / Injury, Int. J. Care Injured 43 (2012) 1888–1891

1890

Table 2 Age standardized relative risk ratios and population attributable fraction (PAF) by injury mechanism and outcome: British Columbia January 1, 2001–March 31, 2007. Injury mechanism Hospitalizations Vehicle/motorcycle High SES Med-High SES Med-Low SES Low SES Deaths Vehicle/motorcycle High SES Med-High SES Med-Low SES Low SES

Cases

RR (95% CI)

PAF (%)

1189 690 267 194

1.00 1.43 (1.27–1.62) 1.13 (0.95–1.34) 1.90 (1.56–2.31)

649 623 355 224

1.00 2.36 (2.05–2.72) 2.75 (2.33–3.25) 4.07 (3.35–4.95)

14.1%

40.6%

Table 3 Age-standardized prevalence of hospitalization and death amongst drivers and passengers involved in a vehicle/motorcycle collision: British Columbia 2001–2006. Year

Urban Cases

Hospitalization 189 2001 2002 241 2003 304 2004 313 2005 278 2006 337 Death 2001 151 2002 169 2003 180 2004 171 2005 190 2006 163

Rural Rate (95% CI)

Cases

Rate (95% CI)

5.19 6.47 8.10 8.31 7.43 8.27

(4.11–6.27) (5.27–7.67) (6.76–9.44) (6.95–9.67) (6.14–8.72) (7.34–10.10)

69 107 106 107 102 102

6.93 10.56 10.32 10.93 9.79 10.01

(4.51–9.34) (7.56–13.55) (7.36–13.28) (7.81–14.05) (6.90–12.69) (7.04–12.98)

4.14 4.52 4.68 4.52 4.94 4.18

(3.17–5.12) (3.52–5.52) (3.68–5.69) (3.52–5.52) (3.90–5.98) (3.23–5.13)

95 162 120 138 129 139

8.76 15.09 12.01 13.54 12.28 13.32

(6.14–11.39) (11.59–18.58) (8.77–15.25) (10.12–16.97) (9.05–15.52) (9.90–16.73)

persons from the lowest through the highest SES strata were equal. An ‘‘excess’’ rate of 14% for injury hospitalizations was attributable to SES differences between the lowest and highest SES groups. As shown in Table 3, there was little yearly change in prevalence of MVC hospitalization and death across rural and urban jurisdictions. Rural populations consistently experienced a 2–3 fold increase in death following a vehicular or motorcycle collision when contrasted against urban injury mortality. By contrast, disparities between urban and rural populations and rate of hospitalization were reduced, but maintained the same geographic disparity each year. Discussion Growing recognition of the burden of injury over the last two decades in Canada has led to an increased focus on injury prevention and monitoring.21–24 Canadian injury prevention networks such as the Canadian Hospitals Injury Reporting and Prevention (CHIRPP) and SMARTRISK have made important advances in applying data from administrative sources for injury surveillance, assessment of the burden of injury and to inform injury prevention policy. However, these reports largely fail to demonstrate whether the benefits of prevention and trauma care have been equally felt amongst all populations. Motor vehicle injury mortality in rural areas remains substantially higher than amongst urban populations, and that these risks become amplified when also taking into consideration relative variations in socio-economic status. Between 2001 and 2006, relative disparities in injury risk and death between rural and urban populations have remained flat, but with a two- to three-fold increase in relative risk of injury mortality amongst the province’s rural populations. These rates

are in contrast to provincial reporting on MVC rates, which illustrate relative decreases in hospitalization and death. However, provincial reporting does not stratify by urban or rural status or by socio-economic variables. In this study, both rurality and SES were measured using previously defined standards and indicators and, as such, represent a reliable perspective for monitoring injury trends across socio-economic and geographic lines. Our results also suggest that motor vehicle-related mortality in British Columbia follows a social gradient. Considering that the lowest SES scores completely overlapped the rural LHAs, the data suggest that MVC-related injury amongst the province’s rural populations is getting worse, not better. This trend was further reflected by the PAF, which suggested that underlining 40% of all MVC-related deaths are population differences attributable to SES. That these disparities in injury risk continue despite more frequent educational, outreach, and legislative efforts remains problematic and deserving of greater attention. No study that attempts to measure the geographic and social determinants of health is free from error. One of the limitations of this study is that we report on SES using area-based measures and not individual-based measures. Challenges of the ecological fallacy are endemic in injury prevention research due to the lack of non-clinical individual-level patient information. However, previous studies have shown that the relationship between SES and injury persists when measured at the individual-level.9,25,26 This study also reports on rural/urban injury using data aggregated from place of residence, not place of injury. However, we found that 81% of the coroner (BCCS) case records occurred in the person’s municipality of residence. A sensitivity analysis was not performed on BCTR data as approximately 50% of case records have missing incident location information. This is a potentially important limitation, though previous studies have shown that in over 90% of injury cases, the individual’s area of residence and geographic area of injury have been coincident.27,28 The data presented in this study were derived from trauma centres that continually provided data to the BCTR over the study period, which may have resulted in under-reporting of the burden injury amongst the province’s rural populations. In addition, the classification of BC populations into the SES strata demonstrated the potentially important confounding relationship between socio-economic status and rurality. In this analysis, the most deprived SES strata were completely represented by the province’s rural populations. The relationship between deprivation, geography, and injury mortality is complex. This analysis raises important questions about the culture of rurality in BC and the importance of a shared dimension between our socio-economic and physical environments. Large inequalities in injury risk should be considered with designing further injury surveillance and public health care policies to reduce its consistent socio-economic and geographic pattern. Data reported in this analysis also represent injuries that are traditionally classified as moderate or severe and therefore cannot be directly compared against results produced by SMARTRISK, which reports on all-cause injury morbidity and mortality across BC and Canada. There is a need to expand on this analysis and incorporate lower acuity information from both BC and across Canada. Current data availability and linkage capabilities increase opportunity to measure and report on factors that contribute to injury prevalence and add to a growing area of research on injury and the environment. Conclusion The primary focus of injury prevention and control is to reduce hospitalization and loss of life from what has been shown

N. Bell et al. / Injury, Int. J. Care Injured 43 (2012) 1888–1891

to be largely preventable and predictable causes of morbidity and mortality. Our study suggests that there are opportunities to strengthen our commitment to reducing the burden of injury, particularly amongst rural populations. Findings from the PAF in relation to SES suggest that 40% of all deaths from vehicular collisions would not have occurred if population rates between the lowest and highest SES strata were equalized. Moreover, the disparity in injury risk across socio-economic and geographic lines should not be undervalued as the province’s rural populations represented the entire low SES category. The PAF method does well to combine the known facts of the determinants of injury and should be considered in future policy initiatives, interventions, and reporting on injury trends. The distinctiveness of injury and its impact across geographic and socio-economic lines may in fact suggest that in many instances injury is reducible to an urban and rural form and that new public health perspectives and policies need to emerge in order to reduce the growing geographic and socio-economic disparities in injury. Conflict of interest statement The authors have no conflict of interest. Acknowledgement The authors thank the British Columbia Trauma Registry that generously provided data for this study. Nathaniel Bell was supported by a Canadian Institutes of Health Research (CIHR) postdoctoral fellowship grant. References 1. SMARTRISK. The economic burden of injury in Canada. Toronto, ON; 2009. 2. Angus DE, Cloutier E, Albert T, Che´nard D, Shariatmadar A. The economic burden of unintentional injury in Canada. Toronto, ON: SMARTRISK; 1998. 3. Muelleman RL, Wadman MC, Tran TP, Ullrich F, Anderson JR. Rural motor vehicle crash risk of death is higher after controlling for injury severity. Journal of Trauma-Injury Infection and Critical Care 2007;62(1):221–5. 4. Muelleman RL, Walker RA, Edney JA. Motor-vehicle deaths – a rural epidemic. Journal of Trauma-Injury Infection and Critical Care 1993;35(5):717–9. 5. Maio RF, Green PE, Becker MP, Burney RE, Compton C. Rural motor vehicle crash mortality: the role of crash severity and medical resources. Accident Analysis and Prevention 1992;24:631–42. 6. Baker SP, Waller A, Langlois J. Motor-vehicle deaths in children – geographic variations. Accident Analysis and Prevention 1991;23(1):19–28.

1891

7. Kmet L, Macarthur C. Urban–rural differences in motor vehicle crash fatality and hospitalization rates among children and youth. Accident Analysis and Prevention 2006;38(1):122–7. 8. Haynes R, Reading R, Gale S. Household and neighbourhood risks for injury to 5–14 year old children. Social Science & Medicine 2003;57(4):625–36. 9. Cubbin C, LeClere FB, Smith GS. Socioeconomic status and injury mortality: individual and neighbourhood determinants. Journal of Epidemiology and Community Health 2000;54(7):517–24. 10. Clark DE. Motor vehicle crash fatalities in the elderly: rural versus urban. Journal of Trauma-Injury Infection and Critical Care 2001;51(5):896–900. 11. Sibley AK, Tallon JM. Major injury associated with all-terrain vehicle use in Nova Scotia: a 5-year review. Canadian Journal of Emergency Medicine 2002;4 (4):263–7. 12. Karmali S, Laupland K, Harrop AR, Findlay C, Kirkpatrick AW, Winston B, et al. Epidemiology of severe trauma among status Aboriginal Canadians: a population-based study. Canadian Medical Association Journal 2005;172(8):1007–11. 13. Bell N, Schuurman N, Hameed SM. A small-area population analysis of socioeconomic status and prevalence of severe burn/fire-related injury in British Columbia, Canada. Burns 2009;35(8):1133–41. 14. Girotti M. Report of the trauma expert panel. Available from: www.ontla.on.ca/ library/repository/mon/17000/272640.pdf; 2006 [accessed 29.06.11]. 15. BC Stats. Regional population estimates and projections: population projections to 2036 for BC. Available from: http://www.bcstats.gov.bc.ca/data/popstart.asp; 2010 [accessed 18.10.10]. 16. Statistics Canada. Population estimates and projections. Available from: http:// www40.statcan.gc.ca/101/cst01/demo62a-eng.htm; 2006, 2009-09-22 [accessed 22.04.10]. 17. Bell N, Schuurman N, Oliver L, Hayes MV. Towards the construction of placespecific measures of deprivation: a case study from the Vancouver metropolitan area. The Canadian Geographer 2007;51(4):444–61. 18. Hanley JA. A heuristic approach to the formulas for population attributable fraction. Journal of Epidemiology and Community Health 2001;55:508–14. 19. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Painting a truer picture of US socioeconomic and racial/ethnic health inequalities: the public health disparities geocoding project. American Journal of Public Health 2005;95(2):312–23. 20. SAS Institute Inc.. SAS/STAT1 version 9.2 of the SAS system for Windows: Copyrightß 2008. Cary, NC, USA: SAS Institute Inc.; 2009. 21. Mackenzie SG, Pless IB. CHIRPP: Canada’s principal injury surveillance program. Injury prevention 1999;5(3):208–13. 22. CHIRPP. Leading cause of death and hospitalization in Canada. Available from: http://www.phac-aspc.gc.ca/injury-bles/chirpp/index.html; 2006. 23. CIHI. National trauma registry report: hospital injury admissions 1998/1999. Toronto: Canadian Institute for Health Information; 2001. 24. PHAC. Leading causes of death, Canada, males and females combined. Ottawa. www.publichealth.gc.ca; 2005. 25. Cubbin C, Smith GS. Socioeconomic inequalities in injury: critical issues in design and analysis. Annual Review of Public Health 2002;23:349–75. 26. Simpson K, Janssen I, Craig WM, Pickett W. Multilevel analysis of associations between socioeconomic status and injury among Canadian adolescents. Journal of Epidemiology and Community Health 2005;59(12):1072–7. 27. Joly MF, Foggin PM, Pless IB. Geographical and socioecological variations of traffic accidents among children. Social Science & Medicine 1991;33(7):765–9. 28. Simons R, Brasher P, Taulu T, Lakha N, Molnar N, Caron N, et al. A populationbased analysis of injury-related deaths and access to trauma care in ruralremote northwest british columbia. Journal of Trauma 2010;69(1):11–9.