Cost of Injury in Canada Methodology

This report is the fourth of its kind, the previous reports being published in 1998 (using 1995 data), 2009 (using 2004 data) and 2015 (using 2010 data). Due to differences in methodology used in this report, it is not possible to draw direct comparisons with previous reports.

This report presents data at the national level for 2018, accounting for all provinces and territories. The analysis underlying this report was conducted from a societal perspective, using an incidence-costing, human-capital approach. The population of those injured in 2018 was costed over the lifetime of the injured individuals, using an age expectancy of 75 years. The costs, both direct and indirect, were discounted to a present value at 1.5 per cent per annum (CADTH, 2017).

The methodology used in this report expands on the model used to calculate the cost of injury in previous reports. The Electronic Resource Allocation Tool (ERAT) was updated to include more recent disability weights. Physician services, ambulance services, formal and informal caregiving, and direct death costs, which are not accounted for in ERAT, were added to the total cost calculations. Finally, all provinces and territories are accounted for in this report, whereas territorial data were previously unavailable.

Incidence costing

The incidence-costing approach to cost-of-illness studies involves “estimating the lifetime direct and indirect costs of the new cases of a condition or group of conditions which have their onset (incidence) in a given year” (Scitovsky, 1982, p.474). This approach emphasizes that “it is necessary to estimate not only the direct costs of these new cases accruing in the first year, but also the present value of direct costs (the stream of costs associated with the given health problem) which may accrue in the future, until the patient dies” (Scitovsky, 1982, p.474). The stream of future costs is discounted to a present value. Hence, with this approach, the cost of an injury occurrence (i.e., full episode) can be compared to the cost associated with the prevention of that injury.

Under the incidence approach, prevention costs are actually investments, (e.g., $1 invested in bicycle helmets averts $29 in injury costs) and unmanaged injury risks are incremental costs. For example, motorcyclists who do not wear helmets increase the burden of injury on society, while wearing a helmet reduces hospitalization costs by more than $6,000 per patient. The incidence costing approach is useful and accurate to help policy makers assess the benefits of preventing or reducing/ameliorating the incidence of specific health problems.

Direct and indirect costs

Cost-of-illness studies distinguish and measure both direct costs (the value of resources used to treat the persons incurring the illness) and indirect costs (the value lost to society as a result of the illness in question). This report applies the same distinction to cost of injury.

Direct costs are costs to the healthcare system, composed of all the goods and services such as medical supplies, diagnostic imaging, and drugs, used for the diagnosis, treatment, continuing care, rehabilitation, and terminal care of people experiencing an injury. In this costing analysis, these cost categories include expenditures for emergency care, hospitalization, physician services fees, ambulance transportation and rehabilitation.

Under the human capital methodology, indirect costs are societal productivity losses, which account for the injured individual’s inability to perform their major activities. The value of time lost from work due to morbidity, disability, and premature mortality is measured by earnings data. In accordance with the human capital methodology, this includes only foregone earnings calculated as average earnings, adjusted by the participation rate and unemployment rate, over the relevant period within the working life of an individual from ages 15 to 64 years inclusive. A real wage growth rate of one per cent per year was assumed for this study.

It should be noted that the perspective for this report is societal. For example, from a societal perspective, transfer payments such as Canada Pension Plan (CPP), disability, and social assistance are not considered costs since they are a reallocation of resources and the net effect of the transfer to society is zero. As well as these economic costs, there are certain intangible costs associated with injuries, such as pain and suffering, economic dependence, and social isolation which are difficult to quantify in economic terms, and are therefore excluded from the cost calculations. Too many Canadians have their lives and those of their families irrevocably changed forever as a result of injury. The methods in this costing analysis use a conservative approach and therefore the costs are an underestimate.

The Electronic Resource Allocation Tool

It is important to capture the lifetime costs associated with an injury. In order to capture the full episodic costs associated with the various types of injury, this report employed the approach initially developed by SMARTRISK for The Economic Burden of Unintentional Injury in Canada and used in the previous Cost of Injury in Canada Report (2015).

The Electronic Resource Allocation Tool (ERAT) provides a classification and costing framework based on existing provincial/territorial and national injury data and data available from the injury costing literature. In essence, the ERAT combines existing data with variables from the literature in order to model full episodic costs for unintentional and inflicted injuries. The ERAT is a flexible tool that can be updated as new data become available and according to changes in population, injury incidence, and treatment patterns and costs.

The ERAT is held by Parachute and consists of a series of spreadsheets designed to calculate the incidence costs of injury. The ERAT was created to fulfill two major objectives:

• To supply modelling and estimation techniques required to fill critical gaps in the available data
• To serve as a resource tool that can be used by researchers and public health officials at the provincial/territorial and local level to support resource allocation, policy development, and decision-making.

Data sources

Hospitalizations

The Centre of Surveillance and Applied Research within the Public Health Agency of Canada provided the acute hospital separation data from their holdings of the Canadian Institute for Health Information (CIHI) Discharge Abstract Database (DAD). The DAD data did not include hospitalizations for Quebec. Hospitalizations for Quebec were estimated using population proportions.

Records in DAD 2017/18 to 2019/20 with admission year in 2018 were included for extracting injury cases. In diagnoses, if a record had both significant ST diagnostic codes (S00-T79) and external cause codes (V01-Y36) following the ST codes, the record was defined as an injury case.

The DAD included information pertaining to the individual’s age and sex, resource intensity weights (RIWs), length of hospital stay (LOS), ICD-10 CA external cause of injury codes (see Appendix A), and ICD-10-CA nature of injury codes. The data did not break down each individual resource and cost used during the hospital stay, but instead provided the total relative case weight (captured by RIW), associated with different diagnostic characteristics, surgical procedures and resources (drugs and medical supplies) used. As each individual relative case weight and length of stay may vary, an average weight and length of stay was applied to the total number of hospitalizations.

In addition, individual dollar values for diagnostics, surgical procedures and resources were not available and therefore an average national inpatient cost was applied to the average weight and average length of stay in order to obtain the dollar values. The national inpatient cost is a measure that is calculated by CIHI and looks at the cost of a standard hospital stay after adjusting for differences in the types of patients a hospital sees.[1] The cost of a standard hospital stay excluded physician care expenditures. An average value of physician care was applied using information from the National Physician Database at CIHI.

Emergency Department visits

The Centre of Surveillance and Applied Research within the Public Health Agency of Canada also provided the Ontario emergency department (ED) visit data from their holdings of the National Ambulatory Care Reporting System (NACRS). NACRS reporting is complete for Ontario and Alberta only. To obtain an estimate of national ED visits, a ratio of ED visits to hospitalizations for injury in Ontario was applied to the rest of Canada. Ontario NACRS data were used as Ontario is the largest province in Canada by population.

Records in NACRS 2017/18 to 2018/19 with registration year in 2018 and reporting patient in Ontario were included for extracting injury cases. If a record had both diagnostic ST codes (S00-T79) and external cause codes (V01-Y36) following the ST codes, the record was defined as an injury case.

The ERAT uses direct morbidity costs for out-of-hospital treatment related to injuries, using ratios of episodes and related costs of non-hospitalized to hospitalized cases from the United States as Canadian data was unavailable (Miller et al., 1995).

Disability

In order to calculate disability incidence, disability weights were applied to the nature of injuries. Disability weights, available by injury diagnosis codes, for short-term (within the year of injury) and long-term (12 months to lifetime) disability were applied for each injury diagnosis (Gabbe et al., 2006). The incidence of short-term and long-term disability was then calculated by applying the total proportion of injury diagnosis within each external cause, age and sex category. For long-term disability, the percentage requiring long-term treatment was applied (Gabbe et al., 2006).

In ERAT, permanent partial disability is “a condition that results in a permanent disability from which partial recovery is anticipated, along with a return to some form of employment. Complete loss of earning power is expected prior to recovery, after which the worker is expected to return to employment with wages below pre­‐injury wages” (Miller et al., 1995, p.26). Permanent total disability is “a condition equivalent to complete and permanent loss of earning power” (Miller et al., 1995, p.26). To obtain the number of cases with permanent partial and total disability for ERAT, the proportion of those employed with a disability was applied to the disability incidence.[2]

Deaths

Mortality and population data were retrieved from the Canadian Socio-Economic Information Management System (CANSIM), Statistics Canada, for the 2018 calendar year. The external cause of injury for the mortality data was classified according to the International Classification of Diseases, 10th revision (ICD-10).

Direct mortality costs were estimated on a complete episode of events due to an injury-related death. These events range from costs incurred as a result of deaths occurring at the scene, ambulance transportation costs, treatment occurring in the emergency departments and hospitals prior to the death, coroner and autopsy costs and funeral costs.

Lost productivity

Indirect costs were calculated using unemployment rates, labour force participation rates, and average wage rates obtained from Statistics Canada’s CANSIM database and used to estimate the monetary value of the productivity losses resulting from morbidity and premature death. This can include losses related to both paid and unpaid labour. A worker may miss time from paid work (absenteeism), or an injured worker may still show up to work but will not operate at the optimum level (presenteeism) (Public Health Agency of Canada, 2017). The participation rate is the number of labour force participants expressed as a percentage of the population 15 years of age and over. The unemployment rate is the number of unemployed persons expressed as a percentage of the labour force. Unpaid labour includes caregiving, volunteer work, household activities, or any other activity that is outside of the standard labour market. Due to insufficient data on unpaid labour, ERAT only includes estimates of production losses of paid work as a result of morbidity and premature death.

The value of lost production was estimated for the working age population comprising individuals aged 15 to 64 years. For this study, it is assumed that all individuals contribute to society. Caregiving, defined as care received by the injured individual within the year of injury and the main condition for which the injured individual sought help, can be classified as either a direct cost or an indirect cost depending on whether a formal, or direct, payment was made (Public Health Agency of Canada, 2018). Informal caregiving provided by family, friends and neighbours are considered as indirect costs, and formal caregiving provided by paid workers and organizations are considered as direct costs. Proportion of formal and informal caregiving for injury was applied to the direct and indirect costs (Public Health Agency of Canada, 2018).

Population denominators

Estimates of populations by age and sex for 2018 were obtained from Statistics Canada. See Reference Population.

Data limitations

The information in this report can be considered a very conservative estimate given data on doctor visits, clinic visits and the toll on individuals are not accounted for. In addition, costs toward injury prevention programming have not been included.

There may be variations in data between this report and provincial/territorial reports. This is due to the date when data are collected, as well as the use of ratios to produce national estimates for certain data points. Ratios were used for obtaining estimates for national hospitalizations as hospitalizations for Quebec were not included in the DAD data. Ratios were also used to estimate the national ED visits using Ontario NACRS data as the NACRS data holding was incomplete at the national level.

The data sources used in this analysis are largely reliant on existing administrative data. The quality of the data is reliant on the expertise of professional data coders, who must interpret written descriptive information into ICD-10-CA codes. Some ICD-10-CA codes may be too broad and, in some cases, lack specific detail about the injury. The data retrieved are from administrative datasets that use a standardized coding system, thus information pertaining to circumstances of the incident was not available.

Direct costs for diagnostic and surgical treatment, drug expenditures, clinical treatment, therapy, and rehabilitation are not readily available for each patient. Comprehensive searches through hospital health records, medical clinic records, and insurance systems would be needed in order to extract all of the required information. Average national costs of relevant health services were used to estimate individual patient costs in dollar values.

Drug expenditures outside of hospital care are excluded as prescription data was not available.

Indirect costs did not include transfer payments made by government or social services. Additionally, the productivity losses of unpaid work activities, such as searching for jobs, household services, raising children, caring for aging parents, shopping and volunteering, and school attendance were excluded due to data availability.

This analysis used an accepted standard methodology to estimate the value of lost production for the working age population comprising individuals aged 15 to 64 years. This approach is limited in that it does not account for the growing number of Canadians aged 65 and over remaining active in the workforce. As a result, total indirect costs based on lost productivity presented in this report are an underestimate. This methodology will be revisited when developing future reports.

Intangible costs associated with injuries, such as pain and suffering, economic dependence, and social isolation, which are difficult to quantify in economic terms, were excluded from the cost calculations.

Some numbers and percentages may not add up due to rounding. There may be variations in data between this report and provincial or territorial reports.

References

CADTH. (2017). Guidelines for the Economic Evaluation of Health Technologies: Canada (4th ed.). https://www.cadth.ca/sites/default/files/pdf/guidelines_for_the_economic_evaluation_of_health_technologies_canada_4th_ed.pdf

Gabbe, B.J., Lyons, R.A., Simpson, P.M., Rivara, F.P., Ameratunga, S., Polinder, S., Derrett, S., & Harrison, J.E. (2016). Disability weights based on patient-reported data from a multinational injury cohort. Bulletin of the World Health Organization, 94(11), 806–816C. http://dx.doi.org/10.2471/BLT.16.172155

Miller, T., Pindus, N., Douglass, J., and Rossman. S. (1995). Databook on Nonfatal Injury Incidence, Costs and Consequences. Washington: The Urban Institute Press.

Public Health Agency of Canada. (2017). Economic burden of illness in Canada, 2010. https://www.canada.ca/en/public-health/services/publications/science-research-data/economic-burden-illness-canada-2010.html

Scitovsky, A.A. (1982). Estimating the direct costs of illness. Millbank Memorial Fund Quarterly: Health and Society, 60(3), 463-491. https://doi.org/10.2307/3349802

---

[1] For more information, see CIHI, Cost of a Standard Hospital Stay

[2] Employment rates among Canadians with disabilities are available from Statistics Canada, based on the Canadian Survey on Disability (CSD). Findings from the 2017 CSD were used for this report: https://www150.statcan.gc.ca/n1/pub/89-654-x/89-654-x2018002-eng.htm