Article Text

Download PDFPDF
Cost and impoverishment 1 year after hospitalisation due to injuries: a cohort study in Thái Bình, Vietnam
  1. Ha Nguyen1,
  2. Rebecca Ivers2,
  3. Stephen Jan2,
  4. Alexandra Martiniuk2,
  5. Leonie Segal1,
  6. Cuong Pham3
  1. 1School of Population Health, University of South Australia, Adelaide, South Australia, Australia
  2. 2The George Institute for Global Health, University of Sydney, Sydney, New South Wales, Australia
  3. 3The Center for Injury Policy and Prevention Research, Hanoi School of Public Health, Hanoi, Vietnam
  1. Correspondence to Ha Nguyen, School of Population Health, University of South Australia, South Australian Health & Medical Research Institute (SAHMRI), North Terrace, Adelaide, SA 5000, Australia; ha.nguyen2{at}unisa.edu.au

Abstract

Background Evidence on the economic impact on individuals and their families following an injury in Vietnam is limited. This study examines the costs and the risk of impoverishment due to hospitalised injuries at 12 months following hospital discharge and associated factors.

Method Employing a prospective cohort design, 892 people hospitalised for injury were recruited from Thái Bình General Hospital in Vietnam in 2010 and followed up for 12 months. All out-of-pocket costs incurred and income lost by injured persons and their caregivers associated with care and treatment of their injuries were reported. To examine associated factors, we used generalised estimating equation models for costs and modified Poisson regression for the risk of impoverishment.

Results The mean total costs by 12 months postdischarge were US$804, nearly 1.2 times the annual average income. Injuries that incurred highest costs were falls (US$950) and road traffic injuries (RTIs) (US$794). At 12-month follow-up, 181 persons (26.9%) became impoverished, with those injured in RTIs and falls at highest risk (26.1% and 35.4%, respectively). Factors associated with higher costs were also those associated with higher risk of impoverishment. These include those injured in RTIs or falls; having higher severity level; principal injured region as upper extremities, lower extremities or head; physical nature of injuries as fracture or concussion injuries; and longer hospitalisation.

Conclusions Injuries impose significant economic burden on injured persons and their families during and beyond hospitalisation. In addition to prevention, there is a need to reform health financing system to protect injured persons from significant out-of-pocket expense for healthcare services.

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

Accounting for 10% of the world's deaths, injuries rank one of the leading causes of death in all regions and countries. Globally, there are 32% more deaths from injuries than deaths from malaria, tuberculosis and HIV/AIDS combined.1 ,2 Although injuries affect people of all ages and income groups, more than 90% of injury deaths are in low and middle income countries (LMICs).1 ,2

According to the most recent statistics on injuries in Vietnam, in 2010, there were more than 39 000 injury deaths, accounting for 11.6% of all mortality and a rate of 43.8/100 000 persons.3 The 10 leading causes of death also include three specific injuries: road traffic injuries (RTIs), drowning and suicide.3 In addition to fatality, injuries lead to millions of non-fatal cases. As reported in the Vietnam National Survey on Injury 2010, an estimated 1.8 million (2092/100 000 persons) were non-fatal injuries. Of these, 36% or 648 000 cases were hospitalised for at least 1 day.4

The large burden of injuries creates significant cost to the national economy, including loss of productivity, costs for treatment and rehabilitation, plus other costs such as administration and property damages. In the USA, Corso et al5 reported in 2006 that for more than 50 million injuries in 2000, the associated lifetime costs were US$406 billion, including US$80 billion for medical treatment and US$326 billion for lost productivity. In Europe, Meerding et al6 reported the lifetime costs for medical treatment for injuries on a cohort of 5755 patients were €1.15 billion, accounting for an estimated 3.7% of total national healthcare expenditure. In Australia, with an estimated 8000 injury deaths (6% of all deaths) and an estimated 400 000 injury hospitalisations each year, the health costs associated with injury in 2004–2005 were estimated at $A3.4 billion, accounting for 7% of national allocated health expenditure.7 In China, the world's most populous country, the annual cost of injury was estimated at US$12.5 billion, almost four times the national public health services budget.8 In a review of 68 studies conducted in 22 LMICs, the median of combined direct and indirect costs of injuries was US$4085 or 97% of gross domestic product per capita.9

Injuries can also lead to significant financial hardship to the injured and their families, particularly in countries where out-of-pocket payment for healthcare service is common. However, evidence on this issue is scarce in Vietnam. Based on 35 traumatic brain injuries due to motorcycle incidents, Hoang et al10 reported the total direct costs at US$2365, US$1390 and US$849 for severe, moderate and minor cases, respectively; 84% of households experienced catastrophic expenditure because of the treatment costs after 1 year. From a sample of 892 hospitalised injuries, Nguyen et al estimated the average direct costs during hospitalisation were US$317.2. At discharge, 27% of samples faced catastrophic expenditure.11 While these studies provided important evidence of the economic burden of injuries, they are either limited in small sample or short study period. Continuing care and treatment required beyond hospitalisation result in further costs, which could push the injured and their families into poverty. In a study of injuries in 1999–2001 in Vietnam, Thanh et al12 found that the risk of impoverishment of household with injury was 1.21 times (p=0.08) the risk of those without injuries. To better understand the longer term economic impact of injuries, our study estimates the costs by 1 year after hospital discharge and associated factors. We also examine the impact of injuries on impoverishment (ie, the risk of descending into poverty because of the healthcare payment for their injuries) and associated factors.

Methods

Study setting and participants

A prospective cohort study was conducted in Thái Bình, a province 100 km south of the capital of Vietnam. Between 1 January 2010 and 31 August 2010, 892 study participants were recruited consecutively from Thái Bình General Hospital (TBGH), the largest public trauma hospital in the province. According to the Provincial Department of Health statistics on injuries in the first 8 months of 2010 in the province, our sample accounted for almost 88.0% of all hospitalised injuries for people aged 18 years old and over.13

Study participants were individuals aged 18 years or older who were admitted to TBGH for at least 1 day due to an injury, had a current residential address within the Thái Bình province area and consented to participate in the study. Participants were approached by research assistants (RAs) only after their transfer from emergency department to inpatient ward, and their condition had been stabilised. Due to the cultural aspect that people prefer to die at home,14 fatal or near-fatal cases not admitted to inpatient ward and in the hospital were not included. For those participating in the study, we collected data during their stay in the hospital, and at 1, 2, 4 and 12 months after discharge.

During their hospitalisation, RAs approached participants and their caregivers to identify their eligibility, get their consent and collect data on basic sociodemographic characteristics (eg, age, gender, occupation, monthly income), injury characteristics (eg, external cause, severity measured by the MAIS, the principal body region injured) and expenses related to their injury care and treatment. MAIS was the highest score measured by the Abbreviated Injury Scale, which is an anatomical scoring system created by the Association for the Advancement of Automotive Medicine to represent the ‘threat to life’ associated with an injury. The scale ranges from 1 to 6 representing: minor injury (1); moderate (2); serious (3); severe, life threatening (4); critical, survival uncertain (5) and unsurvivable (6).15 During the follow-up (at 1, 2, 4 and 12 months after discharge), trained community health workers (community RAs) approached participants at their homes to continue collecting data on ongoing expenses related to their injury care and treatment.

Costing method

We collected data on all out-of-pocket costs incurred by participants and their caregivers associated with treatment and recovery for their injuries during hospitalisation and over the 12 months following discharge. Costs were categorised as direct medical, direct non-medical and indirect costs (box 1).16 All cost data were collected by means of extracting from hospital bills and face-to-face interviews with participants and their caregivers.

Box 1

Calculation of total injury costs

Embedded Image

*Average daily income was the respondent's self-reported income or the average per capita income in the province in 2010 if the respondent was not in paid employment.

Similar to other studies on cost of injuries,5 ,6 ,17 direct medical items included emergency service, surgery, paramedical services, diagnostic tests (such as X-ray, CT scan), medication (prescribed and over-the-counter drugs), equipment and rehabilitation. For the purpose of our study, only costs that were paid out-of-pocket by patients were abstracted from hospital bills. Information on non-medical costs incurred by the participants and their caregivers such as for transportation to the hospital and clinics, and accommodation and meals were also collected. To minimise recall bias during the follow-up, each participant and/or caretaker of the participant was provided with an expenditure diary in which to record expenses incurred. Community RAs checked the suitability of expenses and recorded them into the questionnaire. At each follow-up, they also collected details on participants’ strategies to cope with direct expenses for their treatment, such as using savings, borrowing from others or selling assets. Participants were followed for only 12 months after discharge; therefore, any strategy incurred beyond this period was not captured.

Indirect costs were loss of income due to time spent in injury treatment and rehabilitation.5 ,6 ,17 Using the human capital approach, these were estimated by multiplying the number of days off work by their reported average daily income, which were based on their reported number of working days per month and monthly salary before their injuries. For those not in paid employment, we used the average wage to represent the value of the production they could have contributed had they not been hospitalised or not taking care of the injured person. According to the National Household Living Standards survey, the average monthly per capita income in Thái Bình province in 2010 was equivalent to approximately US$57.90.18 The calculation of indirect costs was based on regular income, and did not include sources such as borrowing, selling of asset or withdrawal from savings.

Impoverishment

Households can be pushed into poverty because of excessive out-of-pocket payment for healthcare services and loss of income. In measuring impoverishment due to healthcare payment for injuries, the national poverty line was used as the threshold. It was Vietnamese dong (VND) 400 000 (US$20) per person per month (Decision 09/2011/QĐ-TTg).19 If their preinjury income was above the poverty line, an injured person's household was considered impoverished if their remaining income per person after all direct expense and income loss was below the poverty line (box 2).20

Box 2

Calculation of impoverishment

Embedded Image

Data analysis

Descriptive statistics were used to describe study participants, including age, occupation and external cause of injuries. Costs of injury by external cause were computed in terms of mean and SD. To explore factors associated with higher costs over time, generalised estimating equation (GEE) models with log link and gamma distribution were used. The GEE estimator is almost as efficient as the maximum likelihood estimator, and GEE models can produce robust variance estimates even if the within-subject associations among the repeated measures are mis-specified.21–23 In examining factors associated with impoverishment at 12 months postdischarge, we used modified Poisson regressions to estimate the RRs. Modified Poisson regression is a Poisson regression applied to binomial data using a robust error variance, which can estimate RRs consistently and efficiently.24 ,25 A p value of <0.05 was considered as statistically significant. All analyses were performed using STATA V.12.0 (Stata, USA).

Results

In this study, we followed 892 participants for 12 months after their hospitalisation. There were 160 participants (17.9%) who lost to follow, with 126 lost at 1-month, 3 at 2-month, 10 at 4-month and 21 at 12-month follow-ups. Of these, 41.3% (66/160) were with unidentified addresses, 10.0% (16) refused to continue, 26.3% (42) moved to other provinces and 22.5% (36) died. In general, the distribution of study participants by major demographic and injury characteristics remained stable (figure 1). For instance, at hospitalisation and all follow-ups, participants aged 18–29 years always accounted for the largest proportions, 27.8% at hospitalisation, 27% at 1-month, 2-month and 4-month follow-ups and 26.6% at 12-month follow-up. Participants hospitalised due to RTIs were always the group with largest proportions (between 53.1% and 54.0%), followed by fall (between 29.3% and 30.4%). The majority of injuries were those with MAIS of 2 and 3, accounting for between 40.1% and 42.6%, and between 40.7% and 41.4%, respectively. However, compared with the remaining, the lost-to-follow-up sample had a significantly larger proportion of injuries with MAIS of 4 and 5.

Figure 1

Distribution of study participants over time by (A) age, (B) occupation, (C) external cause of injury and (D) injury severity (MAIS) Maximum Abbreviated Injury Scale.

Mean total costs per participant, including direct and indirect components, during hospitalisation and each follow-up are presented in table 1. Highest direct costs were incurred during hospitalisation, and decreased over time postdischarge. There were also reduction trends for indirect costs. For instance, for the interval of 2 months between 2-month and 4-month follow-ups, the average indirect costs per participant per month were US$12 (US$24÷2). Then, for the 8 months’ interval between 4-month and 12-month follow-ups, it was US$5.90 (US$47÷8). From initial hospitalisation to the last follow-up, the mean total costs per participant was US$804, including US$625 direct and US$179 indirect costs. Of the total costs, more than 60% were incurred during the follow-up period. Across injury causes, the mean total costs per participant were highest for fall injuries (US$950), followed by RTIs (US$794) and burns (US$742). Of the total costs, on average, the direct component accounted for 77.8%. It was highest in sharp-object-related injuries (84.8%) and lowest in assault-related injuries (72.9%).

Table 1

Direct and indirect costs (US$) per participant by external cause, total and at each follow-up

Analyses of mean total costs per participant over time by injury characteristics controlling for demographic characteristics are summarised in table 2. Groups with a relative difference of >1 indicates that for the same time point (hospitalisation or same follow-up), the mean total costs were larger than the referent group. For instance, the mean costs per RTI patient were 1.49 times (95% CI 1.29 to 1.72) the mean costs per patient of injury other than RTIs or falls. Patients who were transferred from other hospitals (indirect admission) incurred costs 17% (95% CI 4% to 32%) higher than those admitted directly to TBGH. Other factors associated with higher costs include higher ISS, principal injured regions other than face, fracture and concussion injuries and longer hospital stays. For health insurance, the mean costs appear to be smaller for those who actually used their insurance than those who did not (relative difference of 0.93). However, they were not statistically significant (p=0.339).

Table 2

Relative difference in mean total costs over time by injury characteristics controlling for age, gender, occupation and income—multivariable generalised estimating equation models

Impoverishment due to healthcare payments for injuries and income loss among participants whose preinjury income was above the poverty line, broken down by external cause and severity level, are summarised in table 3. At the 12-month follow-up, there were 181 (26.9%) patients who were impoverished. The number of those impoverished was highest among RTI patients (95) and lowest among assaults (0). Fall patients experienced the highest proportion of impoverishment (35.4%), followed by RTIs (26.1%) and burns (25.0%). In terms of injury severity, the proportion of those impoverished was higher for injuries with higher MAIS, from 5.6% for those with MAIS of 1 to 54.5% for those with MAIS of 5.

Table 3

Cumulative numbers and rates of impoverishment over time by external cause and severity

Table 4 shows the multivariable analyses of factors associated with increased risk of impoverishment controlling for demographic factors. RTI and fall patients were 1.80 times (95% CI 1.18 to 2.75) and 1.88 times (95% CI 1.19 to 2.97), respectively, more likely to be impoverished than those who experienced other injuries. Indirectly admitted patients had higher risk of impoverishment (RR 1.33, 95% CI 1.05 to 1.68) than directly admitted ones. Other factors also statistically significantly associated with higher risk of impoverishment were higher MAIS, having principal injured region to the lower extremities or to the head, with injuries as fracture or concussion, and longer hospital stay.

Table 4

Factors associated with impoverishment—multivariable modified Poisson models, controlling for age, gender, occupation and income

Discussion

This study has estimated the total economic costs by 12 months postdischarge per person. The costs of US$804 exceed the average annual income in the province, US$695 in 2010.18 Because of the higher proportions of more severe injuries in the lost-to-follow-up sample (17.9%) and the higher costs associated with more severe injuries, the actual average costs would be higher than the estimated if there was no loss to follow-up. The injured and their families sustained significant economic costs beyond hospitalisation. The total costs and the proportions of impoverishment were higher for those experiencing RTIs or falls, injuries with principal injured region to upper extremities, lower extremities or to the head. These were mainly because they were associated with higher severity levels. They needed more medical care and had longer period off from work, and consequently, they incurred higher costs, both direct and indirect, during hospitalisation and follow-up. These results are similar to other studies, which found that costs of injuries extend far beyond the acute period of the event.21 ,26 The trend of higher costs for indirect-admission patients than direct ones was also because the indirect admissions were associated with more severe injuries.

Similar to studies estimating costs around 1 year postinjury,10 ,27 we found that direct costs accounted for the majority of the total costs. For example, studying costs of major trauma in Queensland, Australia, Rowell et al reported direct costs at $A78 577, more than three times the indirect costs.27 However, when estimating in terms of lifetime costs, indirect component accounts for a larger proportion.5 ,28 ,29 In Corso et al's study, indirect costs represented about 80% of the total. The difference in distribution of the direct and indirect components in the total costs found in our study was mainly because the costs estimated were limited to the 12-month period after hospital discharge. In addition, since we did not include costs associated with fatal injuries, which would incur significant indirect costs due to the loss of lifetime earnings, the average proportion of indirect costs in the total was lower.

At 12-month follow-up, 26.9% of the sample descended into poverty. This is 10 times higher than the average rate of impoverishment due to health payment in Vietnam in 2010 (2.5%).30 The high risk of impoverishment found in this study was likely due to the high costs of treatment and out-of-pocket payment as the main mechanism to pay for healthcare services. In Vietnam, the rolling-out of financial autonomy in public hospital since 2000s has provided hospitals with opportunities to generate additional resources, improve their financial sustainability and provide wider range of services. However, such reform has resulted in increased costs of services and financial burden to patients.31 ,32 Health insurance was introduced in 1990s as a mechanism to protect patients from financial burden of healthcare services. However, more than 20 years after the implementation and continuous amendments, studies indicate that the benefits of health insurance are inconsistent.30 ,33 ,34 Using data from the National Living Standard Survey 2002–2010, Van Minh et al found statistically significantly lower rates of impoverishment in household with health insurance enrolees only in 2004 and 2010. In our study, participants who reported using their insurance seem to have lower risk of impoverishment; however, it was only borderline statistically significant (p=0.054). In terms of total costs, it was far from statistically significant (p=0.339).

Vietnam has been well on track to achieving Millennium Development Goal 1, which aims to halve the proportion of people with an income of less than US$1 a day by 2015.35 However, as a country with a very high incidence of injuries, the economic impact associated with injuries would impose challenges to current efforts in poverty reduction in Vietnam. This highlights the priority of injury prevention to reduce the economic hardship for the injured and to achieve the Millennium Development Goal. There is also a need for a strong and urgent reform of the health financing system. The reform needs to first address the issue of provider-induced supply of unnecessary care, including overuse of diagnosis test/examination and the overprescription of drugs,31 ,32 major drivers of cost escalation. Second, the system needs to resolve the problem of a large proportion of out-of-pocket spending for patients by facilitating the process of usage of health insurance. While the government investment has significantly improved health insurance coverage in the population, data on usage show that the proportion of people admitted to hospital and used their health insurance dropped from 64% in 2006 to 52% in 2008. Even in the poorest group in the population, the usage of insurance was just 62%.36 Differential treatment in almost all public hospitals of insurance enrolees, including longer wait time and lower quality of care at overcrowded facilities, are crucial factors leading to insured patients forfeiting all insurance benefits and paying out-of-pocket to receive better services.37 ,38

This study has a number of limitations. First, we were unable to include fatal cases, and therefore, limited our estimates to injury survivors. We noted the number of patients transferred to the hospital and requested to be sent home (28 patients). These would likely be fatal or near-fatal cases. However, because we did not collect any data from them, we were not able to provide detailed analyses for these cases. In general, the economic costs incurred by these patients would be large due to complete loss of productivity, and therefore, our calculation would underestimate the actual costs. Another important aspect that also leads to underestimation is the range of cost items collected. Including other costs such as property damage, administration, health system or those from a societal perspective would provide a more comprehensive estimate.39 The estimate would be higher, indicating even larger economic costs and impact of injuries. This would provide even stronger evidence for advocating for investing in injury prevention. A further source of underestimation is the method used to calculate productivity losses, which does not take into account intangible costs, including reduced quality of life, and the pain and suffering experienced by injured persons and their families because of the injuries.40 ,41 Finally, since this was a hospital-based cohort, we were unable to capture cases that were not hospitalised, either because of financial barriers or injuries requiring no hospitalisation. These limitations provide implications for future research to study a more comprehensive picture of the burden of injuries in Vietnam. Though limited to injury survivors and the economic burden to their families, our study demonstrates that the financial costs of hospitalised injuries are substantial, and the risk of impoverishment is high.

Conclusion

This study is one of very few cohort studies examining the economic burden of injuries in Vietnam. It demonstrated the high economic costs of injuries. The size of the study enabled more detailed quantitative examinations of demographics and injury characteristics associated with higher costs and risk of impoverishment. RTIs and falls were identified as injuries accounting for the largest proportions of all hospitalised injuries, incurring the highest costs and also associated with the greatest risk of impoverishment. While health insurance is considered as an important mechanism to protect patients from high costs of healthcare services and their impact, our study revealed only a borderline statistically significant effect of health insurance in that role. This result provides important implications for policy-makers to continue efforts in injury prevention as well as to reform mechanisms to facilitate usage of health insurance and to control regulations for health service costs in Vietnam.

What is already known on the subject?

  • Injury is an important cause leading to significant mortality and burden of disease.

  • The majority of the global burden of injury is in low and middle income countries.

  • To date, few studies investigate the economic impact on individuals and their families following the injury.

What this study adds?

  • Injuries lead to significant economic burden to individuals and their families, well beyond the hospitalisation period.

  • The mean costs 12 months after hospital discharge were US$804, with 60% incurred during the period after hospitalisation.

  • Road traffic injuries and falls accounted for the largest proportions of hospitalised injuries, incurring the largest costs and the highest risk of impoverishment.

  • Health insurance had a borderline statistically significant effect in lowering the costs and protecting individuals and their families from impoverishment.

Acknowledgments

The researchers would like to acknowledge the Atlantic Philanthropies for financial support for the prospective cohort study on health and economic consequences of hospitalised injuries in Thái Bình province, and Vietnam, which provided data for this study.

References

Footnotes

  • Contributors HN, RI, SJ, AM and CP conceived the study, developed and refined the study design, instrument and analysis plan. HN analysed the data and wrote the draft. RI, AM, SJ, LS and CP provided feedback regarding analyses and provided critical revision of draft manuscripts.

  • Funding The study received funding from The Atlantic Philanthropies in a project to build capacity on injury prevention research for the Hanoi School of Public Health, Vietnam (grant number AP1533). The funder has no involvement in the study design; collection, analysis and interpretation of data; the writing of the manuscript; the decision to submit the manuscript for publication.

  • Competing interests None declared.

  • Ethics approval The ethics application for this study was reviewed and approved by the Human Ethics Committee in University of Sydney, Australia; the Hanoi School of Public Health Ethics Committee and Thái Bình General Hospital in Vietnam. All participants were provided with an information statement, which specifies that their participation in the study is entirely voluntary. They are not obliged to participate and can withdraw at any time. Whatever may their decision be, it will not affect their relationship with the research staff or the treatment they receive from the hospital or other health services.

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

  • Data sharing statement Data for this study were from the prospective cohort study on health and economic consequences of hospitalised injuries in Thái Bình province, Vietnam. Parties interested in obtaining the data are referred to the Center for Injury Policy and Prevention Research, Hanoi School of Public Health, Vietnam.