SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Objective

To compare directly the prevalence and risk factors for arthritis and arthritis-attributable activity limitations (AAL) between the US and Canada, and to estimate the population attributable risk percentage (PAR%) for obesity and leisure time physical inactivity.

Methods

We conducted analyses of the 2002–2003 Joint Canada/US Health Survey, which asked about health professional–diagnosed arthritis, and arthritis reported as a cause of disability in specified activities of daily living. We used log-Poisson regression to obtain prevalence ratios for arthritis and AAL, adjusting for education, income, having a regular doctor, physical inactivity, and obesity. PAR% for obesity and physical inactivity were calculated.

Results

The estimated crude prevalence of arthritis and AAL were 18.7% and 9.3%, respectively, in the US and 16.9% and 7.4%, respectively, in Canada. Being American was a significant bivariate predictor of arthritis and AAL, but not after adjustment for obesity and physical inactivity. PAR% for obesity were 14% and 20% for arthritis and AAL, respectively, for Americans and 13% and 17%, respectively, for Canadians, and for physical inactivity were 15% and 21%, respectively, for Americans and 4% and 5%, respectively, for Canadians, with estimates being higher among women.

Conclusion

The higher prevalence of arthritis and AAL in the US may be accounted for by the higher prevalence of obesity and physical inactivity, particularly in women. The high PAR% related to obesity in both countries, and physical inactivity in the US, point to the importance of public health initiatives to reduce obesity and increase physical activity to reduce the prevalence of arthritis and AAL.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Arthritis is one of the most frequently reported disabling chronic conditions in both the US and Canada, and is the leading cause of physical disability, particularly in mid to late life (1–4). In both countries, the prevalence of arthritis and arthritis-associated disability is projected to increase with the aging of the baby boomer population (5–7). The prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitations (AAL) was estimated to be 21.6% and 8.8%, respectively, in the US (based on the 2003 National Health Interview Survey [NHIS]) (5), and 16% and 3.7%, respectively, in Canada (based on the 2003 Canadian Community Health Survey [CCHS]) (8, 9). Generally, the prevalence of both arthritis and AAL has been found to be higher in the US.

Comparison of estimates for the prevalence of arthritis between the two countries is difficult given differences in survey questions and in the case definitions of arthritis and AAL. For instance, the US-based NHIS explores activity limitations due to “arthritis or joint symptoms,” where arthritis includes fibromyalgia; whereas the CCHS asks about “arthritis or rheumatism, excluding fibromyalgia,” and does not mention joint symptoms. It is difficult to know whether differences in prevalence across the two nations are due to differences in methodologies or the prevalence of risk factors. Older age, female sex, and low socioeconomic status (as reflected by low income or low education) are known demographic risk factors for arthritis (10–12). The countries' distributions of income are different, with the gap between the lowest and highest incomes being notably wider in the US than in Canada (13, 14). In addition, obesity is a well-established, lifestyle-related risk factor for arthritis, particularly knee osteoarthritis (12, 15, 16). Being overweight or obese can also worsen the severity of arthritis-associated activity limitations (17). More Americans (29.7%) are obese than Canadians (23.1%) (18), particularly American women (19). The prevalence of obesity is increasing in both the US and Canada (19–23). Physical inactivity is another lifestyle-related risk factor for AAL. Engaging in moderate physical activity is known to improve function (24), and to reduce the risk of disability by nearly 50% (25, 26). Americans also tend to have higher population rates of physical inactivity than Canadians (27).

Differences in access to health care resources might also affect the diagnosis and treatment of arthritis. Canada has a universal health insurance system that provides free access to physician and hospital services (but not medications), whereas in the US health care for those <65 years of age is paid for by private or employment-related insurance, and 11% of Americans are reported to lack such insurance (19). Americans are less likely to have a regular doctor than Canadians and are more likely to have unmet health needs and be dissatisfied with health care services (27).

The availability of the Joint Canada/United States Survey of Health (JCUSH), which was simultaneously carried out in both countries using a common methodology, provides a unique opportunity to directly compare the US and Canada. The objective of this study is to compare the prevalence and risk factors for arthritis and AAL between the US and Canada and to estimate population attributable risk percentages (PAR%) for the contribution of 2 modifiable risk factors, obesity and leisure time physical inactivity, to arthritis and AAL.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Study design and setting.

The JCUSH was conducted jointly by Statistics Canada and the US National Center for Health Statistics between November 2002 and March 2003. The survey content drew from the NHIS and the CCHS, and was designed to collect information on chronic conditions, functional status, determinants of health, and health care utilization in the two countries. It was a one-time, random, computer-assisted telephone survey, in which the data were collected for 1 adult age ≥18 years per household from noninstitutionalized dwellings sampled from all 50 US states and the District of Columbia and all 10 Canadian provinces. In all selected households, a knowledgeable household member age ≥18 years was first asked to supply basic demographic information on all residents of the household. A household member age ≥18 years was then randomly selected for a more in-depth interview. Individuals living in health care institutions, nursing homes, full-time members of the US or Canadian Armed Forces, and residents of the 3 Canadian territories were excluded.

Variable definitions.

Arthritis was defined as a “yes” answer to the question, “Have you ever been told by a doctor or other health professional that you have arthritis, not including fibromyalgia?” This followed a lead-in question that asked about long-term, doctor-diagnosed conditions, where long-term was defined as a condition that lasted or was expected to last ≥6 months. This question was taken from the CCHS.

AAL.

The survey included a series of questions taken from the section on limitation of activities from the NHIS, all of which were used to determine the presence of AAL for this study. All respondents were first asked to indicate their difficulty level for a series of activities: “By yourself, and without using any special equipment, how difficult is it for you to: to walk a quarter of a mile or half a kilometre; walk up ten steps; stand or be on your feet for about two hours; sit for two hours; stoop, bend or kneel; reach up over your head; use fingers to grasp or handle small objects; lift or carry something as heavy as 10 pounds such as a full bag of groceries; push or pull large objects like a living room chair; go out to things like shopping, movies or sporting events; participate in social activities such as visiting friends, attending clubs and meetings or going to parties; and do things to relax at home or for leisure?” Response choices were: not at all difficult, only a little difficult, somewhat difficult, very difficult, can't do at all, or do not do this activity.

If respondents had any difficulty with any of these tasks, they were then asked to select from a list of 19 conditions, the health problem that caused the difficulty, up to a maximum of 5 health conditions in no particular order. Respondents who indicated no difficulty with any of the tasks were not asked about the causal condition. A dichotomous variable was created to identify those who had limitations due to arthritis/rheumatism versus all others.

Sociodemographic characteristics.

For descriptive analyses, age was dichotomized as <65 years or ≥65 years. However, it was considered continuously for regression analyses. Education was dichotomized, and those with high school or lower education were compared with those educated beyond high school. Quintiles of household income (a derived variable available in the data set) were based on each respondent's household income (from all sources): respondents were assigned to a quintile group such that the weighted count of each quintile group contained approximately one-fifth of the population reporting household income. Those with missing income responses were excluded from the quintiles. Equivalized household income, which was a derived variable in the survey obtained by dividing the household income by the square root of the number of persons residing in the household, was used to create quintiles. The highest income quintile was the reference group in regression. The data set also included a categorical variable for race for the US respondents with the following groups: American Indian/Alaskan Native, Asian, black/African American, white, and other. Race was included in the US models as part of sensitivity analyses, whereby “black/African American” respondents and all “other” respondents were compared with “white” respondents. The Canadian data included only a dichotomous variable differentiating white versus other/multiple races. Sensitivity analyses for race were not carried out for Canada, given the heterogeneity of the nonwhite population: the largest minority group in Canada is Asian, with blacks or Hispanics making up <3% of the population (28).

Regular medical doctor.

Respondents were asked to respond yes or no to the question: “Do you have a regular medical doctor?” They were specifically asked to think about the past 12 months only. Those with a regular medical doctor were compared with those without one. Having a regular medical doctor has been shown to be associated with increased access to health care services (29, 30).

Lifestyle factors.

Body mass index (BMI; based on self-reported height and weight) was classified into 3 categories in accordance with the suggested World Health Organization guidelines: normal/underweight (<25.0 kg/m2), overweight (≥25 kg/m2 to <30 kg/m2), and obese (≥30 kg/m2) (31). For leisure time physical activity, the derived physical activity index (active, moderate, inactive) by Statistics Canada was used. This index is based on the total daily energy expenditure values (kcal/kg/day) for leisure time activities, and takes into account the duration and frequency of the activity. A dichotomous variable was created whereby the inactive were compared with the active or moderately active.

Statistical analyses.

Log-Poisson regression, with the robust error variance option, was used to estimate prevalence ratios as predictors of our 2 outcomes, arthritis and AAL, using no arthritis or no AAL as the reference categories. Prevalence ratios can be interpreted as being similar to relative risks. In a series of 3 separate models for each outcome, country was added first, followed by the 3 sociodemographic characteristics (age, education, and income) along with regular medical doctor, and finally the 2 lifestyle factors (obesity and leisure time physical inactivity) were added third. An interaction between obesity and physical inactivity was tested. Models were also stratified to obtain estimates by country and sex. In separate sensitivity analyses, data were reanalyzed to account for individuals who did not report income. Also, race was included in regressions for the US, to check whether this changed prevalence ratios for BMI and leisure time physical activity. Prevalence ratios were used to calculate PAR%, to take into account differences in the prevalence of the risk factors to provide a directly comparable estimate for the 2 countries. PAR% were calculated for obesity and physical inactivity, using the following formula: [P(PR−1)] / [P(PR−1) + 1] × 100, where P is the weighted proportion of the population with exposure to the risk factor and PR is the prevalence ratio.

The SAS computer statistical package, version 9.1 (SAS, Cary, NC), was used to perform all analyses. Data were weighted to reflect the sample design and are representative of the Canadian and American household populations age ≥18 years in 2002/2003. Weights were adjusted for nonresponse, both at the household and person level. All descriptive data were presented as population-weighted proportions, with bootstrapped chi-square tests used for bivariate comparisons between the 2 countries. Because log-Poisson regression is an interactive procedure, it was not possible to calculate bootstrap weights; therefore, a rescaled weight was used with incorporation of a design factor to allow for clustering in the sample.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The overall response rate to the JCUSH survey was 65.5% in Canada and 50.2% in the US. Data were obtained for 3,505 Canadians and 5,183 Americans. The study population was weighted so as to be representative of 206 million American adults and 24 million Canadian adults residing in households during 2002/2003. The response rates were >95% for each of the main study variables (arthritis, AAL, BMI, and physical inactivity) used in the analysis.

The estimated prevalence of arthritis and AAL were higher in the US (18.7% [95% confidence interval (95% CI) 17.6–19.8] and 9.6% [95% CI 8.8–10.5], respectively) than the equivalent estimates in Canada (16.8% [95% CI 15.5–18.2] and 7.7% [95% CI 6.8–8.6], respectively). The prevalence of arthritis was particularly higher for American women (23.3% [95% CI 21.7–25.0]) compared with Canadian women (19.6% [95% CI 17.7–21.5]), with the prevalence of AAL being 13.0% (95% CI 11.7–14.3) and 9.2% (95% CI 7.9–10.6) for US and Canadian women, respectively. The overall arthritis prevalence was similar for US men (13.6% [95% CI 12.1–15.1]) and Canadian men (14.0% [95% CI 12.2–15.8]). The prevalence of AAL was also similar, being 6.0% (95% CI 5.0–7.0) and 6.1% (95% CI 4.8–7.3) for US and Canadian men, respectively. Overall, 44.7% of Americans and 39.3% of Canadians with arthritis reported AAL. The proportion of those with arthritis and AAL with selected sociodemographic, health care access, and lifestyle characteristics by country are presented in Table 1. In both countries, the proportion of those with arthritis or AAL who were age >65 years, had low income, or had low education were similar. As might be expected, the percentage who had access to a regular medical doctor was slightly higher in Canada, but not significantly different from the US. The most notable difference was that the percentage of those with arthritis or AAL who were obese or physically inactive was much higher in the US than in Canada.

Table 1. Distribution of demographic characteristics and lifestyle factors among those with arthritis and AAL in Canada and the US, 2002–2003*
AttributeArthritisAAL
USCanadaPUSCanadaP
  • *

    Values are the percentage (95% confidence interval). AAL = arthritis-attributable activity limitations.

  • Significance of testing the difference in the proportion of a variable between the US and Canada, by chi-square.

Age ≥65 years40.9 (37.8–44.0)39.2 (35.2–43.1)0.6846.2 (41.7–50.7)45.5 (39.4–51.5)0.91
Education ≤ high school59.4 (56.4–62.4)63.3 (59.2–67.5)0.3563.1 (58.6–67.6)64.4 (58.3–70.4)0.83
Lowest income quintile31.3 (27.8–34.7)30.9 (26.7–35.1)0.9338.3 (32.9–43.6)34.4 (28.1–40.7)0.57
Has a regular doctor91.5 (89.6–93.5)95.0 (93.1–96.9)0.1490.4 (87.5–93.4)96.1 (93.4–98.7)0.12
Obese body mass index32.3 (29.3–35.4)23.1 (19.4–26.8)0.0235.0 (30.6–39.5)27.4 (21.9–32.8)0.20
Physically inactive70.7 (67.7–73.7)55.8 (51.6–60.1)0.0077.2 (73.4–81.1)58.8 (52.6–65.0)0.00

The results of log-Poisson analyses with arthritis as the outcome are presented in Table 2. Three stepwise models are presented, showing the effect of sequentially adding the country variable, demographic variables and the regular medical doctor variable, and lifestyle variables. Country on its own was a significant predictor of arthritis (model 1), with a prevalence ratio of 1.11. Although country became a nonsignificant predictor with adjustment for sociodemographic factor (model 2), the prevalence ratio was similar to that in model 1. However, the ratio is reduced to close to 1 with the addition of the overweight, obesity, and physical inactivity variables in model 3. In the final model, older age, being female, low education, low income, having a regular medical doctor, being obese or overweight, and leisure time physical inactivity were all associated with increased risk of reporting arthritis.

Table 2. Predictors of arthritis in the US and Canada, 2002–2003: prevalence ratio estimates from log-Poisson regression models*
Covariates (reference category)Model 1Model 2Model 3
  • *

    Values are the relative risk (95% confidence interval).

  • P > 0.05.

Country: US (Canada)1.11 (1.00–1.22)1.10 (1.00–1.23)1.01 (0.91–1.13)
Age, continuous 1.04 (1.04–1.04)1.04 (1.04–1.05)
Female sex (males) 1.44 (1.26–1.63)1.45 (1.27–1.66)
Education ≤ high school (> high school) 1.26 (1.10–1.44)1.20 (1.05–1.37)
Income quintile (highest quintile, 5)   
 4 1.00 (0.80–1.26)0.97 (0.77–1.22)
 3 1.00 (0.79–1.26)0.96 (0.76–1.21)
 2 1.28 (1.03–1.59)1.18 (0.96–1.47)
 1 1.46 (1.17–1.83)1.33 (1.07–1.66)
Has regular doctor (no regular doctor) 1.73 (1.35–2.22)1.61 (1.26–2.06)
Body mass index (normal/underweight)   
 Overweight  1.21 (1.04–1.40)
 Obese  1.85 (1.60–2.13)
Physical inactivity (active)  1.28 (1.12–1.46)

The results of the analyses, stratified by country and looking at predictors of arthritis among men and women, are presented in Table 3. For the most part, the prevalence ratios for arthritis in the stratified models were similar to those for Table 2, model 3, with the exception of income and physical inactivity. Low income was found to be a significant predictor of arthritis in the US, but not in Canada, with higher prevalence ratios for men than women. Leisure time physical inactivity was a significant predictor of arthritis in the US but not Canada. In sensitivity analyses, accounting for those with missing values for income did not substantially change the results. Race was not found to be a significant predictor in the US (data not shown); comparable data for Canada were not available.

Table 3. Predictors of arthritis in the US and Canada, 2002–2003: prevalence ratio estimates from country- and sex-stratified log-poisson regression models*
Covariates (reference category)USCanada
AllMen onlyWomen onlyAllMen onlyWomen only
  • *

    Values are the relative risk (95% confidence interval). NA = not applicable; BMI = body mass index.

  • P > 0.05.

Age, continuous1.04 (1.04–1.05)1.04 (1.03–1.05)1.04 (1.04–1.05)1.04 (1.03–1.04)1.04 (1.03–1.05)1.04 (1.03–1.04)
Female sex (males)1.47 (1.26–1.70)NANA1.33 (1.12–1.58)NANA
Education ≤ high school (> high school)1.19 (1.02–1.38)1.41 (1.08–1.85)1.09 (0.92–1.30)1.24 (1.03–1.51)1.28 (0.93–1.75)1.22 (0.96–1.55)
Income quintile (highest quintile, 5)      
 40.98 (0.76–1.27)1.44 (0.97–2.16)0.73 (0.53–1.01)0.91 (0.66–1.23)1.07 (0.66–1.73)0.76 (0.51–1.13)
 30.97 (0.75–1.26)1.09 (0.69–1.73)0.89 (0.65–1.20)0.90 (0.66–1.22)0.95 (0.59–1.53)0.81 (0.55–1.21)
 21.22 (0.96–1.56)1.66 (1.11–2.49)1.01 (0.75–1.35)0.92 (0.68–1.24)1.10 (0.68–1.78)0.76 (0.53–1.09)
 11.36 (1.07–1.74)2.02 (1.31–3.12)1.09 (0.82–1.46)1.16 (0.86–1.55)1.57 (0.99–2.50)0.88 (0.61–1.26)
Has doctor (no doctor)1.60 (1.23–2.08)1.85 (1.21–2.82)1.39 (1.01–1.91)1.83 (1.27–2.63)1.58 (0.95–2.62)2.22 (1.32–3.74)
BMI (normal/ underweight)      
 Overweight1.19 (1.00–1.40)1.01 (0.76–1.33)1.36 (1.11–1.67)1.30 (1.07–1.58)1.12 (0.82–1.52)1.47 (1.15–1.86)
 Obese1.82 (1.55–2.13)1.53 (1.13–2.06)2.08 (1.73–2.51)2.01 (1.64–2.47)1.68 (1.19–2.38)2.27 (1.77–2.91)
Inactive (active)1.31 (1.12–1.52)1.35 (1.06–1.71)1.27 (1.05–1.54)1.09 (0.93–1.28)1.06 (0.82–1.37)1.13 (0.91–1.40)

Prevalence ratio estimates for predictors of AAL are presented in Table 4. Again, 3 sequential models are presented. Country was a significant predictor in model 1, with the US having a higher risk of AAL than Canada. When controlling for age, sex, education, income, and health care in model 2, country remained a significant predictor. However, after adjusting for obesity and physical inactivity, country became a nonsignificant predictor (model 3) with a prevalence ratio of 1.15. Being female, older age, lower education, lower income, being obese or overweight, and leisure time physical inactivity were associated with AAL.

Table 4. Predictors of arthritis-attributable activity limitations in the US and Canada, 2002–2003: prevalence ratio estimates from log-Poisson regression models*
Covariates (reference category)Model 1Model 2Model 3
  • *

    Values are the relative risk (95% confidence interval).

  • P > 0.05.

Country: US (Canada)1.25 (1.08–1.45)1.31 (1.11–1.54)1.15 (0.97–1.35)
Age, continuous 1.05 (1.04–1.05)1.05 (1.04–1.06)
Female sex (males) 1.76 (1.44–2.14)1.78 (1.46–2.18)
Education ≤ high school (> high school) 1.18 (0.96–1.44)1.10 (0.90–1.35)
Income quintile (highest quintile, 5)   
 4 0.88 (0.60–1.28)0.84 (0.57–1.23)
 3 1.18 (0.82–1.69)1.10 (0.77–1.57)
 2 1.53 (1.09–2.15)1.35 (0.96–1.90)
 1 2.03 (1.43–2.87)1.76 (1.26–2.48)
Has regular doctor (no regular doctor) 1.64 (1.12–2.39)1.48 (1.02–2.14)
Body mass index (normal/underweight)   
 Overweight  1.36 (1.09–1.71)
 Obese  2.29 (1.84–2.85)
Physical inactivity (active)  1.42 (1.15–1.75)

The results of stratified analyses, looking at predictors of AAL by country and sex, are presented in Table 5. The overall findings were similar by country. Race was also not a significant predictor of AAL in the US (data not shown).

Table 5. Predictors of arthritis-attributable activity limitations in the US and Canada, 2002–2003: prevalence ratio estimates from country- and sex-stratified log-Poisson regression analyses*
Covariates (reference category)USCanada
AllMen onlyWomen onlyAllMen onlyWomen only
  • *

    Values are the relative risk (95% confidence interval). NA = not applicable; BMI = body mass index.

  • P > 0.05.

Age, continuous1.05 (1.04–1.06)1.06 (1.05–1.07)1.05 (1.04–1.05)1.05 (1.04–1.05)1.04 (1.02–1.06)1.05 (1.04–1.06)
Female sex (males)1.82 (1.46–2.28)NANA1.47 (1.10–1.96)NANA
Education ≤ high school (>high school)1.09 (0.88–1.36)1.09 (0.73–1.64)1.09 (0.84–1.41)1.19 (0.86–1.66)1.12 (0.63–1.98)1.24 (0.84–1.82)
Income quintile (highest quintile, 5)      
 40.85 (0.56–1.29)1.11 (0.50–2.43)0.71 (0.44–1.17)0.80 (0.45–1.42)1.00 (0.37–2.72)0.67 (0.35–1.31)
 31.11 (0.75–1.66)1.33 (0.61–2.88)0.98 (0.62–1.54)1.06 (0.62–1.82)1.46 (0.57–3.74)0.81 (0.42–1.55)
 21.40 (0.96–2.04)2.38 (1.17–4.82)1.06 (0.68–1.63)1.03 (0.60–1.74)1.53 (0.61–3.84)0.73 (0.40–1.36)
 11.83 (1.26–2.66)3.64 (1.76–7.53)1.27 (0.83–1.95)1.36 (0.80–2.30)2.11 (0.85–5.19)0.95 (0.51–1.77)
Has doctor (no doctor)1.41 (0.96–2.09)1.63 (0.84–3.14)1.27 (0.80–2.04)2.81 (1.33–5.96)2.93 (0.98–8.78)2.66 (0.94–7.52)
BMI (normal/ underweight)      
 Overweight1.35 (1.05–1.74)1.38 (0.86–2.21)1.44 (1.07–1.94)1.27 (0.92–1.75)0.97 (0.56–1.69)1.53 (1.05–2.23)
 Obese2.27 (1.79–2.86)2.47 (1.54–3.96)2.27 (1.73–2.98)2.44 (1.75–3.39)2.09 (1.18–3.72)2.60 (1.77–3.81)
Inactive (active)1.46 (1.15–1.86)1.41 (0.94–2.10)1.49 (1.11–2.00)1.12 (0.86–1.46)1.13 (0.74–1.73)1.11 (0.79–1.57)

An interaction between obesity and leisure time physical inactivity was found to be nonsignificant in the overall model, as well as in those stratified by country for both arthritis and AAL.

The PAR% for obesity contributing to arthritis was similar between the 2 countries, and slightly higher for the US (20%) than Canada (17%) for AAL (Table 6). Sex-stratified analyses showed higher PAR% for Americans, except for obesity contributing to arthritis, for which the PAR% was higher for Canadian men. Large differences were evident in the PAR% between Americans and Canadians for physical inactivity for both arthritis and AAL.

Table 6. Prevalence and population attributable risk percentages (PAR%) of obesity and physical inactivity in the US and Canada for arthritis and arthritis-attributable activity limitations (AAL), by sex, 2002–2003
Type of estimateUSCanada
TotalMenWomenTotalMenWomen
  • *

    Corresponding prevalence ratio used to calculate PAR% was significant at P > 0.05.

Obesity      
 Population prevalence19.8319.6320.0214.7617.7811.85
 PAR%      
  Arthritis13.94*9.42*17.78*12.98*10.79*13.08*
  AAL20.18*22.41*20.24*17.48*16.26*15.90*
Physical inactivity      
 Population prevalence57.5152.3862.2747.5642.7552.21
 PAR%      
  Arthritis15.01*15.49*14.39*4.172.506.36
  AAL20.88*17.5023.42*5.305.365.55

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

To our knowledge, this is the first study to provide a direct comparison of US and Canadian data in search of between-country differences that might be associated with differences in the prevalence of arthritis and AAL. The prevalence estimates for both countries were of a similar order of magnitude, suggesting that differences reported previously are largely attributable to differences in survey methods. Nevertheless, the prevalence of both arthritis and AAL was higher in the US, and the findings suggest that this is largely accounted for by the higher prevalence of obesity and physical inactivity in the US than in Canada. This is reflected in the PAR% estimates, particularly for physical inactivity.

Overall, the risk factors for reporting arthritis and AAL, namely, being female, having lower education or income, and obesity, were similar to those found in other studies (28, 32, 33). In all models, low income was found to be a stronger predictor of arthritis and AAL in the US than in Canada, and in men than in women. This is consistent with the overall findings from the JCUSH, which showed that Americans in the lowest income quintile reported worse health, obesity, and severe mobility limitations more frequently than their Canadian counterparts; this differential did not exist at the other end of the income spectrum (19). Low income among men may increase the risk of arthritis through more physically demanding jobs. Although the overall prevalence of arthritis and AAL in men was similar between countries, the relative magnitude of the risk factors was somewhat different, particularly for AAL, with higher contributions from low income for American men, and a higher contribution from obesity for Canadian men.

Despite differences in the health care systems, in both countries the proportion of those with arthritis reporting having a regular medical doctor was high, and this was significantly associated with arthritis and AAL. The survey asked about arthritis that had been diagnosed by a health professional, and people with arthritis of sufficient severity to cause activity limitation might well need to be in contact with medical care. The slightly higher estimates for Canadians could be a reflection of universal access to the health care system.

Our finding that obesity was a strong risk factor for arthritis and particularly of AAL confirms previous findings (12, 15–17, 34, 35). The estimates of the PAR% for arthritis and AAL from this study underline the importance of reducing the population prevalence of obesity. The PAR% estimates for obesity were somewhat higher than the estimate for (unspecified) arthritis from an Australian study (36).

Physical inactivity has been less well studied as a risk factor for arthritis and AAL. In stratified analysis we found that leisure time physical inactivity was only a risk factor for Americans. Likewise, the PAR% for physical inactivity contributing to arthritis and AAL was much larger for the US than for Canada. The higher population prevalence of physical inactivity in the US likely contributes to this difference; not only were Americans with arthritis or AAL more likely to be inactive, but this was also true of the nonarthritis American population. These findings point to the public health importance of increasing levels of physical activity, particularly in the US. Nevertheless, with almost half of the population reporting physical inactivity, Canadians can also not afford to be complacent.

The prevalence ratios and associated PAR% for obesity and physical inactivity were generally higher when predicting for AAL than for arthritis alone. In addition to contributing to arthritis, obesity and physical inactivity may make further independent contributions to AAL. These results confirm the potential role of reduced obesity and increased physical activity in preventing activity limitations for individuals with arthritis.

Despite the mounting evidence that physical activity can prevent disability and improve function (37), the proportion of the US and Canadian populations with arthritis or AAL who choose to be physically inactive remains high. Female sex, older age, black race, being overweight or obese, poorer self-reported health, and limitations resulting from joint symptoms have been associated with not engaging in recent exercise/physical activity in the US (38). In Canada, physical inactivity has been associated with being female, older (age ≥75 years), underweight or overweight, having worse self-rated health, having functional limitations, having severe pain, or not having prescription drug insurance coverage (39). In other words, groups who are most likely to be physically inactive are those that are more likely to have arthritis or AAL.

This study had several limitations that are common to secondary databases. This was a cross-sectional study, limiting the possibility of causal inferences. Although the study used a common methodology to obtain the data, there may be differences in reporting between the US and Canada. Because these countries differ in the prevalence and distribution of racial minorities, race could not be included in our models. Similarly, lack of comparable data meant we were unable to adjust for insurance coverage in our models. Insurance coverage is an important determinant of health status and outcomes, particularly in a country like the US where universal health coverage does not exist and insufficient financial resources may act as a deterrent to obtaining health care. Nevertheless, studies have found that Canadians tend to be similar to insured Americans on factors such as regular access to a medical doctor (19). The definition of arthritis was general, we were not able to distinguish between different types of arthritis conditions, and the reliability of the survey question is unknown. We grouped underweight and normal weight together. Being underweight might pose a different risk of developing arthritis-attributable limitations; because this group comprised a small percentage of the sample (∼2.5%), it is unlikely that it had a noticeable impact on the final estimates. Furthermore, it is possible that the nature and types of co-occurring conditions might differ between the two countries. Finally, the variable for physical activity was based on leisure time activities only, so important components of work- and home-related activities may be underestimated. The major strength of the study was the comparability in the methodology of the study between the two countries.

This study provides an in-depth look at similarities and differences in the predictors of arthritis and associated activity limitations between the US and Canada. Although the prevalence of arthritis and AAL was slightly higher in the US than in Canada, the differences between these countries were within the range of regional differences. There was considerable internal variation in the prevalence of arthritis, with a range of 35.8–17.8% (for persons age >18 years) in 2002 for US states (40) and 23.3–11.6% (for persons age >15 years) in 2000 for Canadian provinces (8). It might be valuable for the targeting of public health strategies to explore to what extent regional differences in obesity and physical inactivity could contribute to these differences.

The results of this study suggest that the higher prevalence of arthritis and AAL in the US may be accounted for by the higher prevalence of obesity and physical inactivity there, particularly in women. This is reflected by high PAR% for these outcomes related to obesity in both countries, and to physical inactivity in the US. This study adds further support for the inclusion of arthritis in public health initiatives that promote healthy weight and physical activity, because this might have the potential to reduce the prevalence of arthritis and AAL.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Badley had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Badley, Ansari.

Analysis and interpretation of data. Badley, Ansari.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The authors thank Anthony Perruccio and Mayilee Canizãres for their help and advice on the analysis.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES