Chapter 3
The Burden of Arthritis in Canada

Chapter 3 presents three overall indications of the impact of arthritis on the population: mortality*, life expectancy and health-adjusted life expectancy (HALE), and economic costs.

Mortality

Marie DesMeules, Claudia Lagacé, J. Denise Power 

Introduction

Since arthritis and related conditions are not seen as life threatening, little research has focused on the impact of arthritis on mortality; of the research that there is, most relates solely to the mortality of individuals with rheumatoid arthritis.^1-8 Rheumatoid arthritis is associated with a significant increase in mortality because of both the disease itself and other co-morbid or co-existing conditions such as cardiovascular disease, infections and renal disease.^5,7,9 Although deaths for which arthritis is indicated as an underlying cause are relatively rare, arthritis actually causes more deaths than many other conditions that are traditionally viewed as greater health threats.

Methods

Mortality rates were calculated using the Canadian annual mortality data for the period of 1985-1998, with the 1991 Canadian population as a standard. All deaths for which the underlying cause was recorded as arthritis and related conditions, as described in Table 3A-1 in the Methodological Appendix at the end of the chapter, were included in the analyses.

Results

The results that follow represent only a fraction of deaths related to arthritis, since only deaths with arthritis as an underlying cause could be included. The mortality database does not currently provide data on secondary, or contributing, causes of death. Arthritis is more commonly seen as a contributing cause of death, either as a result of complications from treatment (such as gastrointestinal [GI] bleeding related to anti-inflammatory drug use) or as one of its long-term complications (such as cardiovascular disease).^5-7,9-11 To more fully describe the impact of arthritis on mortality, deaths due to GI bleeding were also examined.

Mortality from Arthritis by Age and Sex

In 1998, 497 men and 257 women in Canada died with arthritis and related conditions identified as the underlying cause. Mortality rates increased with age for both sexes. Among men, rates ranged from 2 deaths for every million men aged 54 years or less to 219 deaths per million in those 75 years and older. Among women in the same age groups, rates ranged from 5 deaths per million to 267 per million. Overall, mortality rates for arthritis and related conditions were higher among women than men in every age group, with approximately 4 female deaths for every 3 male deaths (Figure 3-1).

Figure 3-1 Number of deaths and mortality rates (deaths per 100,000) attributed to arthritis and related conditions, by age and sex, Canada, 1998

Source: Canadian Mortality Database 1998, Statistics Canada

Mortality by Type of Arthritis

Between 1985 and 1998, the risk of mortality from arthritis varied greatly by type of arthritis: risk of death from connective tissue diseases (such as lupus) was approximately 3 times higher than from osteoarthritis (Figure 3-2). The number of deaths with rheumatoid arthritis as an underlying cause was twice the number with osteoarthritis as an underlying cause. Mortality rates ranged from approximately 2 deaths per million for rheumatism to approximately 12 deaths per million for connective tissue diseases.

Figure 3-2 Standardized mortality rates (deaths per 100,000) for all ages, by type of arthritis, Canada, 1985-1998

Source: Canadian Mortality Database 1985-1998, Statistics Canada

Trends in Mortality Over Time and Provincial/territorial Comparisons

Mortality rates for arthritis and related conditions for each age group remained relatively stable from 1985 to 1998 (Figure 3-3).

Overall, age- and sex-standardized mortality rates by province/territory for arthritis and related conditions reflected the national rate of 2.4 per 100,000 (Figure 3-4). The Yukon recorded the highest mortality rate (5.1 deaths per 100,000). New Brunswick had the second highest (2.9 per 100,000). The Yukon's relatively high rate was due to higher mortality rates for osteoarthritis (2.3 per 100,000) and connective tissue diseases (1.7 per 100,000). Mortality rates in Newfoundland and Labrador and Prince Edward Island were below the national rate, at 2.0 per 100,000 and 1.7 per 100,000 respectively.

Figure 3-3 Mortality rates (deaths per 100,000) for arthritis and related
conditions, by year and age group, 1985-1998, Canada

Source: Canadian Mortality Database 1985-1998, Statistics Canada

Figure 3-4 Age- and sex-standardized mortality rates (ASMR) (deaths per 100,000) for arthritis and related conditions, by province/territory, Canada, 1985-1998

Source: Canadian Mortality Database 1985-1998, Statistics Canada

Comparisons with Mortality from Other Causes

Table 3-1 compares the number of deaths and the mortality rate for arthritis and related conditions with other conditions commonly presumed to be more serious and life threatening. In 1998, arthritis was a more common underlying cause of death in Canada than melanoma, asthma or HIV/AIDS, especially among women.

Mortality from Treatment Complications - Gastrointestinal (GI) Bleeding

Patients with arthritis are among the most frequent users of non-steroidal anti-inflammatory drugs (NSAIDs),¹² although these drugs are also used for other painful and inflammatory disorders. GI complications are the most common type of adverse drug reaction that may occur with the use of NSAID therapy.¹² Approximately 107,000 hospitalizations and 16,500 deaths occur each year in the United States as a result of NSAID use.¹³ The estimated mortality rate due to GI toxicity from NSAID use by arthritis patients is about 2 deaths per 1,000 people with arthritis per year.^14,15

In 1998, 1,322 Canadians died from GI bleeding (Table 3-2). The number of deaths and the mortality rate from GI bleeding increased with age, and each was higher among men than women. Since data on contributing (secondary) causes of death for the whole country are unavailable, GI bleeding mortality rates specifically due to the treatment of arthritis cannot be determined. However, since people with arthritis are the most frequent users of NSAIDS, these data indicate that mortality from arthritis presented earlier in this chapter has likely been underestimated.

Discussion

Although relatively rare, arthritis is a more common underlying cause of death in Canada than melanoma, HIV/AIDS or asthma. This chapter has underestimated the mortality burden of arthritis in Canada because the data do not include deaths for which arthritis was a contributing cause as a result of complications from treatment (such as GI bleeding from NSAID use). Data on contributing causes of death for the whole country are currently unavailable. Statistics Canada plans to provide this information by 2005. The introduction of new families of anti-inflammatory drugs, such as COX-2 inhibitors, which are believed to lower the risk of adverse effects on the GI tract, is expected to lead to a decrease in mortality associated with arthritis.

Life Expectancy and Health-Adjusted
Life Expectancy (HALE) 

Doug Manuel, Claudia Lagacé, Marie DesMeules, Robert Cho, J. Denise Power 

Introduction

Mortality and life expectancy are often used to describe the health status of a population, according to the assumption that greater life expectancy implies better health.¹⁶ Although arthritis is usually not a fatal condition, it causes more deaths than many other well-known diseases, such as melanoma. As one of the most prevalent chronic conditions in Canada, arthritis is also a leading cause of disability. As a result, when conditions such as arthritis are examined, measures of both mortality and morbidity (overall health status) need to be considered. These two measures can provide contrasting views of a disease or condition.

In the effort to provide measures of population health that take into account both mortality and morbidity, summary measures, such as health-adjusted life expectancy (HALE), have been developed.¹⁷ HALE adjusts overall life expectancy, or life years lived, according to the amount of time spent in less-than-perfect health or with disability.¹⁷ It sheds more meaning on longer life by determining whether an increase in the average lifespan is accompanied by better quality of life.¹⁶

This section considers both life expectancy and HALE in describing the influence of arthritis on the quality of life of Canadians. (Details regarding the calculation of these measures can be found in Table 3A-2 in the Methodological Appendix at the end of this chapter.) Data from the CCHS and Canadian annual mortality data were used to calculate these measures.

The Impact of Eliminating Arthritis on Life Expectancy and HALE

Currently, life expectancies for Canadian women and men at birth are 81.2 and 75.6 years respectively (Table 3-3). If arthritis were eliminated, overall average life expectancy would increase by 0.35 years for all females and 0.16 years for all males in the population.

HALE is estimated to be 69.8 years for women and 66.5 years for men (Table 3-4). If arthritis were eliminated, Canadian females would gain 1.5 years in HALE and males would gain almost 1 year. Therefore, eliminating arthritis would result in a gain of more than 1 year of good health for females and close to 1 year for males, combined with a small overall gain in life expectancy.

Discussion

Disease-specific life expectancy has no direct policy implications without consideration of the prevalence of the condition in the population, its adverse consequences and the potential for eliminating either the disease or its consequences.^16,18 Success in the battle against arthritis, one of the leading chronic health problems in Canada, could considerably increase HALE within the population, particularly in the case of women. Eliminating this rarely fatal disease, however, would contribute less to extending average life expectancy. Most people with a diagnosis of arthritis will be recommended for treatment and monitoring. Clearly, improvements in arthritis treatment hold great potential for increasing the number of “healthy” years lived by Canadians.

Economic Burden

Julie Stokes, Sylvie Desjardins, Anthony Perruccio 

Introduction

Establishing the costs associated with arthritis from any single source presents a dual challenge. First, different sources present different cost components related to arthritis and often under the banner of musculoskeletal conditions. Second, different sources use slightly different definitions of arthritis and rheumatism: some include particular sub-types of arthritis and related conditions whereas others do not.

This chapter presents the most recent (1998) cost values available from the Public Health Agency of Canada of Health Canada.¹⁹ All values presented are in 1998 dollars.

Total costs associated with arthritis include both direct and indirect costs:

Direct costs are defined as the value of goods and services for which payment was made and resources used in treatment, care and rehabilitation.¹⁹ These include hospital care expenditures, drug expenditures, physician care expenditures and additional direct health care expenditures.

Indirect costs refer to the dollar value of lost production due to illness, injury, disability or premature death. Disability measures the value of activity days lost due to short-term and long-term disability (morbidity due to short-term and long-term disability), and premature death measures the value of years of life lost due to premature death (mortality costs).

The Cost of Musculoskeletal Diseases

In 1986, the economic burden of musculoskeletal diseases (ICD-9 710-739) in Canada was estimated to be $11.4 billion,²⁰ which made it the fourth most costly disease group. Seven years later, estimates ranked this group second, at $19.0 billion.²¹ This ranking was maintained in 1998, when the total economic burden was estimated at $16.4 billion. Indirect costs accounted for more than 5 times the direct costs ($13.7 billion and $2.6 billion respectively).¹⁹

Hospital care expenditures accounted for more than one-half of the direct costs of musculoskeletal disease ($1.4 billion) in 1998, and drug and physician care expenditures were estimated to be 23% ($614.3 million) and 22% ($578.2 million) of direct costs respectively. Long-term disability ($12.6 billion) represented over 90% of indirect costs.

Musculoskeletal diseases represented the most costly disease group for women in Canada in 1998 ($8.2 billion) and the third most costly disease group for men ($8.1 billion). All direct cost components were slightly higher for women than for men. Among indirect cost components, however, men's costs for morbidity due to long- and short-term disability were higher than women's.

Costs Attributed to Arthritis

In 1998, estimates placed the economic burden of arthritis (ICD-9 714-716, 721) in Canada at approximately $4.4 billion (Table 3-5), representing just over one-quarter of the total cost of musculoskeletal diseases. Arthritis accounts for nearly one-third of hospital care expenditures for musculoskeletal disease, over 40% of drug expenditures, and more than one-quarter of both musculoskeletal mortality costs and morbidity due to long-term disability.

Of the total arthritis expenditures in 1998, $908.9 million (20%) were direct costs and $3.5 billion were indirect costs (80%). Figure 3-5 shows the relative magnitude of the cost components for arthritis. Morbidity costs due to long-term disability accounted for 76.3% of arthritis costs, by far the largest cost component of the arthritis burden at nearly $3.4 billion. The largest direct costs were hospital care expenditures at $458 million and drug expenditures at $263 million, representing 10.3% and 5.9% of total costs respectively.

Figure 3-5 Economic burden of arthritis, by cost component, Canada, 1998

Source: Economic Burden of Illness in Canada, 1998 and custom tabulation by the Economic Research Analysis
Section, Strategic Policy Directorate, Public Health Agency of Canada, Health Canada

In terms of breakdown by sex, women incurred greater costs related to arthritis than men. They accounted for approximately 60% of hospital care expenditures, prescription drug expenditures, and mortality costs, and one-half of morbidity costs due to long-term disability.

Seniors (aged 65 years and over) accounted for most of the direct costs associated with arthritis: 70% of hospital care expenditures and nearly one-half of total expenditures on prescription drugs. They accounted for less than one-quarter of the arthritis morbidity costs due to long-term disability. Nearly 70% of this cost was incurred by the 35-64 year age group.

The economic burden of musculoskeletal conditions in Canada accounted for 10.3% of the total economic burden of all illnesses but only 1.3% of health science research.

Discussion

In constant dollars, the economic burden of musculoskeletal diseases appears to have decreased in Canada since 1993.¹⁹ The majority of the decrease is due to a reduction in disability costs: in 1993, morbidity costs due to disability totalled $16.3 billion (in 1998 $), and in 1998 disability costs were $13.6 billion. Decreases in both long-term and short-term disability costs have also been noted for other chronic diseases, such as cardiovascular diseases, respiratory diseases and nervous system/sense organ diseases.

The estimates presented here, as well as those for arthritis, are based on principal diagnosis only; secondary and subsequent diagnoses were not captured. As a result, the cost estimates are considered to be conservative. Musculoskeletal diseases are often a contributing cause of cardiovascular or digestive disease and are not captured in the estimates.^22-26

The costs for arthritis presented here are less than the costs estimated by Coyte,²⁷ at $6.2 billion (baseline estimate, converted to 1998 $), assuming that expenditure values remained unchanged since 1994. The subset of arthritis conditions (ICD-9 714-716, 721) used by Health Canada in their analyses was a different and more restricted set than that employed by Coyte. Coyte's definition of arthritis (ICD-9 098.5, 099.3, 274, 696.0, 710-720, 725-729, v78.4, v43.6) closely mirrored the definition of arthritis and related conditions used in the other chapters of this publication. From the frequency of these diagnoses, it is assumed that the definition used by Health Canada represents nearly 60% of the cases in the broader definition. If this is so, then inflating the Health Canada figure to include the broader definition of arthritis narrows the gap between the estimates. Nonetheless, both sources demonstrated a similar proportional breakdown of direct and indirect costs.

The costs presented in this chapter exclude expenditures for care in institutions other than hospitals, costs related to health professionals other than physicians (such as rehabilitation professionals) and direct health expenditures (such as for over-the-counter medications, assistive devices and informal care giving). As well, the value of time lost from work and leisure activities by family members or friends who care for the patient are not included. As a result, these data likely underestimate the total cost of arthritis. In addition, the drug expenditures presented here pre-date the availability of new arthritis medications such as COX-2 inhibitors and biologic disease-modifying anti-rheumatic drugs (DMARDs), which are costly.

While arthritis affects predominantly women and older people, Canadians between the ages of 35 and 64 years incur nearly 70% of long-term disability costs due to arthritis. Using earnings to establish the value of lost production places more emphasis on diseases prevalent among people with high incomes, many of whom are men, than on diseases suffered by the poor, the elderly and women.²⁷ Therefore, the estimate of $4.4 billion should be viewed as the lower end of the range of the true costs of arthritis and related conditions. Furthermore, no economic analyses can calculate the intangible personal costs such as arthritis-related pain, suffering and loss of opportunity.

Even though the cost estimates for musculoskeletal diseases, including arthritis, should be interpreted in the context of the methods, assumptions and limitations from which they were calculated,¹⁹ they still provide a sense of the magnitude of the economic burden of this disease group in Canada. Arthritis represents an important economic burden, especially for women and those in the 35-64 year age group. The cost component that contributes the most to this burden is morbidity due to long-term disability.

Implications

The impact of arthritis is greater in terms of health and disability than in terms of mortality. Arthritis control approaches need to focus on improving health and reducing disability.

Reducing arthritis-related disability has the potential to reduce indirect costs and increase HALE for the population as a whole.

Projections indicate that people aged 55 years and over will account for the greater part of the increase in the number of people affected with arthritis. Research also indicates that a greater proportion of people with arthritis than people with other chronic conditions are not in the labour force. As a result, long-term disability expenditures for arthritis and related conditions are expected to increase substantially in the near future.

Future cost studies of arthritis could adopt a more inclusive definition of arthritis and aim to use the full range of available data, such as those presented in this publication. As well, initiating new partnerships among those involved with arthritis and building on existing relationships will be necessary to clarify what information is currently available and what is missing.

An imbalance between the proportion of expenditures in health science research directed towards musculoskeletal diseases and the proportion of their contribution to the total economic burden of disease has been noted.

With the advent of new treatments, the surveillance of changes in direct costs in relation to indirect costs is essential. By helping to establish the best courses of action when making decisions about the treatment of arthritis, surveillance has the potential to reduce morbidity and decrease costs in the long run.

New treatments for arthritis and related conditions also require that surveillance for this condition include monitoring of changes in mortality and HALE. Making available contributing causes of death data would lead to a more accurate description of the full impact of arthritis on mortality.

*    Mortality data were provided to Health Canada from the Canadian Vital Statistics databases at Statistics Canada. The cooperation of the provincial and territorial vital statistics registries that supply the data to Statistics Canada is gratefully acknowledged.

[Previous] [Table of Contents] [Next]