You are here

Article Text

Article menu

Download PDFPDF

Original article
The economic burden of road traffic injuries: evidence from a provincial general hospital in Vietnam
  1. Ha Nguyen1,2,
  2. Rebecca Q Ivers1,
  3. Stephen Jan1,
  4. Alexandra L C Martiniuk1,3,
  5. Qiang Li1,
  6. Cuong Pham2
  1. 1The George Institute for Global Health, University of Sydney, Sydney, New South Wales, Australia
  2. 2The Center for Injury Policy and Prevention Research, Hanoi School of Public Health, Hanoi, Vietnam
  3. 3Dalla Lana School of Public Health, University of Toronto, Toronto, Canada
  1. Correspondence to Professor Rebecca Q Ivers, Injury Division, The George Institute for Global Health, University of Sydney, PO Box M201, Missenden Rd, NSW 2050, Australia; rivers{at}georgeinstitute.org.au

Abstract

Objective To examine the costs of road traffic injuries (RTIs) in Vietnam and factors associated with increased costs.

Method RTI data were collected in a prospective cohort study on the impact of injuries in Vietnam. Participants were persons admitted to the Thai Binh General Hospital because of RTI. All costs incurred by participants and their family members during hospitalisation were collected, including direct medical costs, direct non-medical costs and indirect costs. Generalised linear models were employed to examine predictors of increased costs including demographic and injury context characteristics.

Results Each RTI hospitalisation costs the patient and family on average US$363 or 6 months of average salary. Income, injury severity, principal region of injury and length of hospital stay were statistically significant predictors of increased costs; age, gender, occupation and road user group were not. After controlling for injury characteristics and income, participants with principal injuries to the lower extremities had a cost 1.28 (95% CI 1.07 to 1.54) times higher than those with principal injuries to the face. Analyses of motorcycle-related RTIs with principal injury to the head also showed increased costs among those without a helmet (1.41 times higher, 95% CI 1.17 to 1.71).

Conclusions RTIs can cause a substantial economic burden to the patient and family. During hospitalisation on average, an RTI would cost approximately 6 months of salary. In addition to interventions to decrease the risk of RTIs, those reducing the severity, such as wearing a motorcycle helmet, should be enforced to minimise the economic and health consequences of injury.

  • Road traffic injuries
  • cost of injuries
  • economic burden
  • cost of hospitalisation
  • Vietnam
  • outcome of injury
  • costs
  • disability
  • health disparities

https://doi.org/10.1136/injuryprev-2011-040293

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Annually, road traffic crashes kill over 1.3 million people and cause over 50 million injuries worldwide.1 ,2 Of these, over 90% of the deaths occur in low and middle income countries.1 The global cost of road traffic crashes has been estimated at US$518 billion.3 In Vietnam, according to the Ministry of Health, in 2008 almost 15 000 lives were cut short due to road traffic injuries (RTIs), accounting for more than 42% of all injury-related mortality and equivalent to a mortality rate of 18.5/100 000.4 Vietnam is among the countries with the highest rate of fatal RTIs in WHO Western Pacific region.5

    Recent decades observed enormous motorisation in Vietnam. In 1990–2009, the number of registered motorcycles had increased by 19 times, from 1.5 million to more than 28.3 million.6 ,7 Such dramatic increases are likely to lead to an increase in RTI fatalities and disability. According to the National Traffic Safety Committee, in 1990–2007, the number of road crashes increased from 5565 to 13 985, the number of fatalities from 2087 to 12 800 and the number of injuries from 4468 to 10 266.6 Efforts have been made to introduce stronger road safety programmes, including the introduction of a motorcycle helmet law in 2007,8 a revised alcohol limit for drivers effective from July 20099 and the revised ordinance of administrative fines in 2010.10 Although it is likely such efforts will result in benefits in the long term, RTIs remain a significant issue.

    In addition to their devastating impact in terms of mortality and disability, RTIs also cause a substantial economic burden.11–16 In Bangladesh, the average treatment cost during hospitalisation for an RTI in 2001 was US$86 where the per capita annual income was <US$500.14 Thai estimates in 2004, using a societal perspective, found the average direct and indirect costs per RTI patient were US$2596, exceeding the average annual income of US$2513.16 Such cost information, particularly in resource constrained settings, can help policy makers to assess the resource needs and thus the feasibility of providing and scaling up interventions for RTI prevention, control and treatment. Furthermore, an assessment of such costs highlights the possible economic benefits associated with preventive activities and can provide some of the evidence needed for rational investment decisions.1 ,5 ,17

    To date, few published studies on the costs of RTIs in Vietnam are available and thus the economic burden of RTI is not well known. A study on costs of traumatic brain injury due to motorcycle crashes in Hanoi found that in 2007, average direct costs were US$2365 for a severe traumatic brain injury, US$1390 for a moderate and US$849 for a minor one.18 This study included only 35 participants who had been injured in a motorcycle crash. From prevalence data on RTI data in 2003, the Asian Development Bank estimated that RTI costs to society would account for 0.54% of gross domestic product.19 This study was limited in summary results, which could not go into detail stratification such as by road user or severity level. This relative lack of data on the burden of RTIs in Vietnam, especially economic costs, constrains the ability to identify priorities among RTI categories and prevention efforts. Though the evidence on costs of RTI treatment may not result in policy maker investment into postinjury interventions, as stated in the Global Status Report on Road Safety, studies examining costs and the cost-effectiveness of various interventions are especially needed to inform preventive action and to mobilise support.1 ,20 This study aimed to estimate the costs borne by individuals and their families due to RTI during hospitalisation and examine factors associated with increased costs, such as injury characteristics and the use of helmets.

    Method

    Participants

    A prospective cohort study was conducted to examine the economic impact of injuries in Vietnam. Eligible participants were RTIs admitted to the Thai Binh General Hospital from 1 January 2010 to 31 August 2010, hospitalised for at least 1 day, aged 18 years or older and currently living in the Thai Binh province. Eligible participants and/or their care givers were approached and invited to join in the study voluntarily at hospital admission. Those who consented to join in were asked to provide demographic and injury context information during hospitalisation, and all costs incurred during this period on the day before their discharge. Those who were transferred to central hospital soon after first aid or sent home on request were not included because their short stay did not allow the time to approach and ask for consent. There were 477 participants included in the study. The ethics application for the study was reviewed and approved by the Human Ethics Committee at the University of Sydney, Australia, and the Hanoi School of Public Health Ethics Committee in Vietnam.

    Costing method

    Taking the individual and family perspective, costs reported were direct medical, direct non-medical and indirect costs incurred by participants and their care givers.21 Direct medical costs included out-of-pocket expenditure associated with treatment and care for the injured person such as emergency services, surgery or treatment, paramedical or diagnostic examination tests (eg, x-ray, CT scan), medication (prescribed and over-the-counter drugs), equipment and rehabilitation in the hospital. Direct non-medical costs were those incurred by the injured person and their care givers, such as transportation to hospital, accommodation and meals. Indirect costs refer to lost productivity because of injury treatment and recovery.22 These were estimated by the product of the total days off work over this period and the average daily income of the injured person and their relatives.23 In this study, costs were those incurred during the hospitalisation period.

    Data analysis

    Descriptive statistics were used to present frequencies and percentages for categorical variables (age, gender, occupation, insurance status, types of road user, injury severity by maximum abbreviated injury scale (MAIS) and principal body regions injured). Means were calculated for continuous variables (average monthly income, length of hospital stay and costs). To examine predictors of increased costs during hospitalisation, the outcome variable was the total costs per case, which include direct medical cost, direct non-medical cost and indirect cost incurred by both the participant and their care giver during the hospitalisation period.

    The ordinary least squares of the log transformation and the generalised linear models with log links were initially employed to account for the empirical non-negativity and positively skewed distribution of costing data.24 White's test and the Breusch–Pagan test were used to check for heteroskedasticity of residuals.25 Heteroskedasticity can result in bias and inefficiency in ordinary least squares models. Because of the existence of heteroskedasticity, generalised linear models with log link and γ distribution were selected to examine factors associated with increased cost. The γ distribution was verified by the subsequent modified Park tests.24 Prior to the multivariable modelling, univariable analyses were conducted to examine the association between costs of injuries and individual predictors. Those with a significant association (p value of <0.05) in univariable models and identified to be associated with costs of injuries in prior studies were included in multivariable analyses.

    Collinearity was assessed by examining scatter plots, the correlation matrix and variance inflation factors. No significant collinearity was detected among predictors, including demographic and injury characteristics. Plausible interactions examined were those between length of hospital stay (LOS) and injury severity and principal body region injured. The backward stepwise approach was then employed to remove non-significant predictors and interactions in the full models. STATA V.11 (Stata Corporation, College Station, Texas, US) was employed to perform all statistical analyses.

    Results

    During the 8 months of consecutive recruitment from January 2010 to August 2010, 477 of 542 (88.0%) RTI admissions were recruited. Those not recruited were requested to be sent home (20), transferred or did not give their consent (45). Of those included, there were 48 cases whose consent was sought from their care givers. Motorcyclists represented the largest proportion (75%), followed by bicyclists (14%), pedestrians (10%) and car occupants (1%). Since there were only seven car occupants who sustained injuries, they were not included in most of the detailed analyses.

    Table 1 shows the demographic characteristics of the study sample, along with the costs and LOS by road user type. Overall, the age group that accounted for the largest proportion was also the youngest one (18–29-year old group, the youngest one, 36%). However, among pedestrians or bicyclists, the oldest age group was the one with the largest proportion of injuries. In terms of gender, approximately 75% of the study sample were male subjects. The gender distribution difference was particularly large among motorcyclists, where 82% were male subjects and 18% were female subjects. By occupation, farmers represented the largest proportion of the overall sample (50%) and also farmers represented the largest proportion of each individual road user type (pedestrians: 57%; bicyclists: 43%; and motorcyclists: 46%). Regarding the insurance status, the proportion of coverage ranged from 16% among motorcyclists to 18% among bicyclists.

    View this table:
    Table 1

    Costs (mean and SD) in US$ and LOS (mean and SD) in days by demographic, injury characteristics and road user

    Injury severity and principal injured region along with mean costs and LOS are also presented in table 1. A large proportion of participants experienced moderate (MAIS=2) and serious but not life threatening (MAIS=3) injuries, 43% and 38% respectively. Approximately 8% of injuries were life threatening with probable survival (MAIS=4) and 4% were life threatening with uncertain survival (MAIS=5). Motorcyclists represented the largest number of those injured. Motorcyclists also had the largest proportion of those having an MAIS of at least 3, 52% in comparison with 50% among bicyclists and 36% among pedestrians.

    Table 1 demonstrates that, mean costs increased with injury severity. However, the trend was not completely consistent among pedestrians and bicyclists, where those with MAIS=3 had higher costs than those with MAIS=4. By principal injured body region, those to the lower extremity and head incurred the highest costs, US$434 and US$381 respectively. Injuries to the spine, thorax, abdomen and face regions had the lowest cost. Analyses for pedestrians were limited by the small sample size. With only one pedestrian sustaining an injury to the face, this category was combined with head injury among pedestrians in table 1. Individuals without insurance experienced higher costs than individuals with insurance. This was notable among pedestrians and bicyclists, but not among motorcyclists.

    The major components of costs by road users are shown in figure 1. The mean cost per case, including medical, non-medical and indirect cost during hospitalisation, was US$363 (Table 1, excluding car occupants) or US$367 (including car occupants), more than six times higher than the average monthly income in the Thai Binh province (US$57.926). During this period, most of the costs were direct medical costs, accounting for almost 75% of the total, and costs were relatively uniform across all road users, except for car occupants. Pedestrians had the highest actual direct medical cost (US$306), followed by car occupants (US$296), bicyclists (US$274) and motorcyclists (US$261). For direct non-medical costs, such as expenses for transportation or meals, car occupants experienced the greatest costs (US$61), followed by pedestrians (US$58), motorcyclists (US$44) and bicyclists (US$43). The highest indirect cost was incurred by car occupants almost double that of other road users.

    Figure 1
    Figure 1

    Mean costs reported by road users in US dollars.

    Table 2 presents the mean costs and LOS among motorcyclists by principal injured region and helmet use. Overall, helmet wearing was statistically significantly associated with lower costs (US$337 vs US$432, p<0.01) and lower LOS, but differences in LOS were not statistically significant (7.4 days vs 8.0 days, p=0.18). The difference is statistically significant among motorcyclists having the head as principal injured region (US$322 vs US$507, p<0.001 and 6.8 days vs 8.2 days, p<0.05), but not among motorcyclists whose principal injured region was other than the head.

    View this table:
    Table 2

    Mean costs (US$) and LOS (days) among motorcyclists by principal injured region and helmet use

    Factors associated with increased costs among all road users were examined and are presented in table 3. In univariable analyses, demographic characteristics including age, gender, occupation, insurance status and type of road users were not statistically significant factors. However, higher MAIS, having principal injury to extremities or the head and longer hospital stay were statistically significantly associated with increased cost. For instance, costs of RTIs with MAIS=2 were 1.62 times higher (95% CI 1.21 to 2.18) than those incurred by injuries with MAIS=1. Comparing costs incurred by individuals whose principal injuries were to the face and those whose principal injuries were to the spine, thorax or abdomen region, there was no statistically significant difference (95% CI 0.68 to 1.33). However, costs paid by those with principal injuries to the head region were 1.54 (95% CI 1.19 to 1.99) times higher than principal injuries to the face. Regarding LOS, each additional day in hospital was associated with 11% increase in cost (95% CI 10% to 12%).

    View this table:
    Table 3

    Relative difference in mean costs by demographic and injury characteristics: univariable analyses of all 470 participants (excluding car occupants)

    Despite statistically significant associations identified in prior studies, age, gender, occupation and type of road user were not statistically significantly associated with costs and therefore were not included in the multivariable analyses.13 ,14 ,16 Backward stepwise approach found that holding other variables constant, monthly income (p=0.02), injury severity MAIS, principal injury regions, LOS (p<0.001) and a first order interaction between injury severity and LOS (p=0.01) were statistically significant in predicting costs (table 4). For instance, in participants with the same injury characteristics, every US$10 increase in monthly income incurred 1% higher costs (95% CI 0% to 2%). Participants with principal injuries to the lower extremity region had a total cost 1.28 (95% CI 1.07 to 1.54) times higher than that of those with principal injuries to the face. Interaction between injury severity and LOS indicates that the change in cost by LOS would be different for each level of MAIS. The relative difference of 0.99 indicates that increase in costs by LOS would become smaller among individuals with higher MAIS. Those with insurance seemed to incur lower costs than those without insurance (0.89 times). However, the insurance status was not statistically significant (p=0.11).

    View this table:
    Table 4

    Relative difference in mean costs by demographic and injury characteristics: multivariable analyses* of all 470 participants (excluding car occupants)

    The effect of helmet use on cost due to head injuries among motorcyclists is presented in table 5. In the multivariable regression analyses for this specific group, no interaction was found to be statistically significant. Controlling for demographic characteristics, not wearing a helmet and longer hospital stay were found to be statistically significant factors associated with increased costs. Those without helmets had a cost 1.41 times higher than those with helmet (95% CI: 1.17 to 1.71) after adjusting for demographic characteristics and length of stay. In terms of LOS, each additional day staying in the hospital was associated with a 13% (95% CI 11% to 15%) increase in costs.

    View this table:
    Table 5

    Relative difference in mean costs and associated factors among RTIs involved with motorcycle and having head as the principal injured region (335 motorcyclists with reported helmet use)

    Discussion

    This study contributes to our understanding of the costs borne by individuals and families due to RTIs during hospitalisation and associated factors in Thai Binh, Vietnam. From the individual and family perspective with the human capital approach, the average costs of an RTI, including direct medical, direct non-medical and indirect costs, were US$363, more than 6 months of average salary. By using the human capital approach, indirect costs can be readily estimated from the individual current market earning27 and also minimise the varying values of safety among individuals in their answers as in the willingness to pay approach.28

    Point estimates of mean cost across road users show the highest costs to be among pedestrians and the lowest among motorcyclists. However, the difference was not statistically significant, likely because they are all vulnerable road users.1 Regarding insurance, costs among insurance holders were lower than for people without insurance. This difference was small and non-significant suggesting that in Vietnam insurance currently provides very limited protection from the costs associated with road injury. In addition, recent evidence also shows low utilisation of health insurance even among the poor (52% on average and 62% among the poorest quintile).29

    Costs of more severe injuries were significantly higher than those of the least severe ones, similar to the findings of Hoang et al. In their study, the average costs incurred by severe, moderate and minor brain injury patients were US$2365, US$1390 and US$849, respectively.18 A small sample size limited their analyses to descriptive statistics. Our study went further by investigating all injury types related to motorcycle crashes, and employed multivariable analyses to control for potential confounders including demographic and injury characteristics. The multivariable analysis shows that higher income was statistically significantly associated with higher costs, which seems to contradict the crude estimate in table 1. It can be seen that retiree and unemployed individuals, who had a lower income than white- and blue-collar workers, had the highest average costs. In fact, RTIs among retiree and unemployed individuals were likely to be more severe and result in a longer hospital stay than among workers. These resulted in higher crude costs among retiree and unemployed individuals.

    This study also indicated that costs of injury were found to be positively associated with severity and associated with the principal injury region and LOS. Furthermore, the statistically significant interaction between injury severity and LOS showed that the relative increase in costs associated with longer LOS became smaller at higher injury severity. This is because more severe RTIs were associated with higher costs. Therefore, additional costs associated with longer LOS among more severe RTIs would contribute less to the relative increase than among less severe RTIs.

    Examining costs by principal injury region provides an insight into potential interventions to minimise costs. Costs incurred by individuals with the head as principal injured region were 54% higher than costs among those with the face as principal injured region. Among motorcyclists in particular, those not using a helmet incurred costs 41% higher than those using a helmet. Prior research has identified that protection of the head by using helmet among motorcyclists can reduce the risk of injury by 72%.30 In addition, helmet was found to be cost-effective.31 ,32 In the USA, for instance, a benefit to cost ratio ranges from 1.33 to 5.07.31 Thus, wearing a helmet to protect head injuries would reduce mortality and reduce costs associated with hospitalisation.30 ,33 ,34 Although not a trial, our cost analyses provide additional supportive evidence of the beneficial effects of wearing a helmet.

    A number of limitations exist in our study. First, costs, both direct and indirect, were captured only during the hospitalisation period, and additional costs are likely to accrue following participants' hospital discharge. Second, our cost estimates include those incurred by individuals and their families, but exclude property damage, pain and stress. Therefore, the current figures would underestimate the actual costs. Another limitation which may have resulted in underestimation is the exclusion of cases transferring to higher level hospital or being sent home on request. These were usually more severe cases and thus would incur higher costs. Future study designs should consider investigating longer term costs, including a wider range of road users, transferred cases and also account for a wider scope of costs such as intangible costs.

    In conclusion, the study provides the evidence of significant costs during hospitalisation due to RTIs. By increasing prevention efforts, such costs can be saved for other living needs such as education, housing or other healthcare needs. With severity, LOS and principal injured regions such as head and extremities as factors associated with increased costs, interventions to reduce severity, injuries to the head, extremities and LOS would be highly desirable in addition to prevention of RTIs. With further evidence of beneficial effects of helmet use, recent mandatory helmet law needs to be closely monitored and enforced.

    What is already known on the subject

    • Despite efforts which have been made to improve road safety, road traffic injuries (RTIs) remain the leading cause of injury-related death and hospitalisation in Vietnam.

    • To date, few analyses on costs of RTIs in Vietnam are available and the household economic burden of RTI is not well understood.

    What this study adds

    • On average, the cost of a road traffic injury (RTI) in 2010 in Thai Binh was US$363, more than 6 months of average salary.

    • Injury severity, principal region of injury and length of hospital stay were statistically significant predictors of increased costs.

    • Helmet use was found to be significantly associated with decreased costs among motorcycle-related RTIs with principal injury to head.

    Acknowledgments

    The researchers would like to acknowledge the George Institute for Global Health, The University of Sydney and Atlantic Philanthropies for financial support and student scholarship.

    References

    1. ↵
      World Health Organization. Global Status Report on Road Safety: Time for Action. Geneva: World Health Organization, 2009.
    2. ↵
      World Health Organization. Global Status Report on Road Safety. 2011. http://www.who.int/violence_injury_prevention/road_safety_status/en/index.html (accessed 29 Sep 2011).
    3. ↵
      World Health Organization. Violence and Injuries: The Facts. Geneva: World Health Organization, Department of Violence and injury prevention and disability (VIP), 2010.
    4. ↵
      Administration of Preventive Medicine MoH. Injury Mortality Statistics in 2008. Hanoi: Ministry of Health, 2009.
    5. ↵
      World Health Organization. Road Safety in the Western Pacific Region: Call for Action. Manila: The WHO regional office for the Western Pacific, 2009.
    6. ↵
      National Transportation Safety Committee. Statistics on Road Traffic Accident 2007. Vol. 2008. Hanoi, 2007.
    7. ↵
      National Transportation Safety Committee. Report on Traffic Safety in July and Seven Months in 2009. 2009. http://www1.mt.gov.vn/ykienatgt/default.asp?Param=category&catid=14&ArticleId=3497 (accessed 25 Mar 2010).
    8. ↵
      Government of the Socialist Republic of VietNam, Resolution 32/2007/NQ-CP, A number of urgent measures to reduce traffic accidents and jams, dated March 29th 2007, Government, Editor. Hanoi, Vietnam, 2007.
    9. ↵
      National Assembly. Law 23/2008/QH12: Road Traffic Law. Hanoi: National Assembly, Hanoi, Vietnam, 2008.
    10. ↵
      Government of the Socialist Republic of VietNam, Decree 34/2010/ND-CP, Ordinance of Administrative Fine related to road traffic, dated April 2nd 2010, Government, Editor. Hanoi, Vietnam, 2010.
    11. ↵
      1. Bastida JL,
      2. Aguilar PS,
      3. Gonzalez BD
      . The economic costs of traffic accidents in Spain. J Trauma 2004;56:883–8; discussion 888–9.
      OpenUrlPubMedWeb of Science
    12. ↵
      1. Connelly LB,
      2. Supangan R
      . The economic costs of road traffic crashes: Australia, states and territories. Accid Anal Prev 2006;38:1087–93.
      OpenUrlCrossRefPubMed
    13. ↵
      1. Hijar M,
      2. Arredondo A,
      3. Carrillo C,
      4. et al
      . Road traffic injuries in an urban area in Mexico. An epidemiological and cost analysis. Accid Anal Prev 2004;36:37–42.
      OpenUrlCrossRefPubMed
    14. ↵
      1. Mashreky SR,
      2. Rahman A,
      3. Khan TF,
      4. et al
      . Hospital burden of road traffic injury: major concern in primary and secondary level hospitals in Bangladesh. Public Health 2010;124:185–9.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Naci H,
      2. Baker TD
      . Productivity losses from road traffic deaths in Turkey. Int J Inj Contr Saf Promot 2008;15:19–24.
      OpenUrlCrossRefPubMed
    16. ↵
      1. Riewpaiboon A,
      2. Piyauthakit P,
      3. Chaikledkaew U
      . Economic burden of road traffic injuries: a micro-costing approach. Southeast Asian J Trop Med Public Health 2008;39:1139–49.
      OpenUrlPubMed
    17. ↵
      1. Rice DP
      . Cost of illness studies: what is good about them? Inj Prev 2000;6:177–9.
      OpenUrlFREE Full Text
    18. ↵
      1. Hoang TM,
      2. Pham TL,
      3. Vo TTN,
      4. et al
      . The costs of traumatic brain injury due to motorcycle accidents in Hanoi. Vietnam: Cost Effectiveness and Resource Allocation, 2008:6;17–23.
    19. ↵
      Asian Development Bank-Association of Southeast Asian Nations Regional Road Safety Program. Vietnam: Accident Costing Report AC 10, 2005.
    20. ↵
      Department of Health Systems Financing - Health Systems and Services. WHO Guide to Identifying the Economic Consequences of Disease and Injury. Geneva, Switzerland, 2009.
    21. ↵
      1. Akobundu E,
      2. Ju J,
      3. Blatt L,
      4. et al
      . Cost-of-illness studies: a review of current methods. Pharmacoeconomics 2006;24:869–90.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      Rice DP, MacKenzie EJ, and Associates. Cost of Injury in the United States: A report to Congress. San Francisco, CA: Institute for Health and Aging, University of California and Injury Prevention Center, The Johns Hopkins University, 1989.
    23. ↵
      1. Drummond MF,
      2. Sculpher MJ,
      3. Torrance GW,
      4. et al
      . Methods for the Economic Evaluation of Health Care Programmes. 3rd edn. Oxford: Oxford University Press, 2005.
    24. ↵
      1. Manning WG,
      2. Mullahy J
      . Estimating log models: to transform or not to transform? J Health Econ 2001;20:461–94.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      Stata Web Books—Regression with Stata—Chapter 2-Regression Diagnostics. UCLA: Academic Technology Services, Statistical Consulting Group. http://www.ats.ucla.edu/stat/stata/webbooks/reg/chapter2/statareg2.htm
    26. ↵
      General Statistics Office of Vietnam. Results of the Survey on Household living standards 2010, General Statistics Office of Vietnam. Hanoi, Vietnam, 2011.
    27. ↵
      1. Butchart A,
      2. Brown D,
      3. Huynh DK,
      4. et al
      . Manual for Estimating the Economic Costs of Injuries Due to Interpersonal and Self-directed Violence. Geneva: World Health Organization Press, 2008.
    28. ↵
      1. Segel JE
      . Cost of illness studies—A primer: RTI International, RTI-UNC Center of Excellence in Health Promotion Economics. North Carolina, US, 2006.
    29. ↵
      Ministry of Health and Health Partnership Group. Joint annual health review 2010-Vietnam's health System on the Threshold of the Five-year Plan 2011-2015. Hanoi, Vietnam: Ministry of Health, 2011.
    30. ↵
      1. Liu BC,
      2. Ivers R,
      3. Norton R,
      4. et al
      . Helmets for preventing injury in motorcycle riders. Cochrane Database Syst Rev 2008;(1):CD004333.
    31. ↵
      1. Hyder AA,
      2. Waters H,
      3. Phillips T,
      4. et al
      . Exploring the economics of motorcycle helmet laws–implications for low and middle-income countries. Asia Pac J Public Health 2007;19:16–22.
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Waters HR,
      2. Hyder AA,
      3. Phillips TL
      . Economic evaluation of interventions to reduce road traffic injuries–a review of the literature with applications to low and middle-income countries. Asia Pac J Public Health 2004;16:23–31.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Brandt MM,
      2. Ahrns KS,
      3. Corpron CA,
      4. et al
      . Hospital cost is reduced by motorcycle helmet use. J Trauma 2002;53:469–71.
      OpenUrlPubMedWeb of Science
    34. ↵
      1. Eastridge BJ,
      2. Shafi S,
      3. Minei JP,
      4. et al
      . Economic impact of motorcycle helmets: from impact to discharge. J Trauma 2006;60:978–3; discussion 983–4.
      OpenUrlPubMedWeb of Science

    Footnotes

    • Funding The study was funded by Atlantic Philanthropies, and HN was supported on an international student scholarship by the University of Sydney. RQI and SJ were funded by the National Health and Medical Research Council of Australia. ALCM was funded in part by an unrestricted educational Fellowship from Merck Inc in 2011. CP was funded in part by Atlantic Philanthropies.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval Ethics approval was provided by the Human Ethics Committee in University of Sydney, Australia, and the Hanoi School of Public Health Ethics Committee in Vietnam.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement Data used in this study are a component of a larger cohort study which includes hospitalisation and 1-year follow-up data on costs and quality of life of participants. Further analyses and publications will be developed by authors.