SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Objective

Individuals with lower socioeconomic status (SES) receive less arthritis care, including joint arthroplasty. However, no studies have considered the expectations or needs of the patients. Our objective was to assess the effect of education and income on the potential need for, and the willingness to consider hip and knee arthroplasty.

Methods

Through a mail/telephone survey of 48,218 persons ages 55 years or older residing in 2 areas of Ontario, Canada, 3,307 individuals with moderate-to-severe hip/knee problems were identified. These individuals received a questionnaire to assess education, income, arthritis severity, and comorbidity. In a subset of these subjects, we conducted interviews to evaluate the willingness to consider arthroplasty, and we also performed clinical and radiographic examinations of the joints to validate self-reports of arthritis. The potential need for arthroplasty was defined as the presence of severe arthritis (as scored by the Western Ontario and McMaster Universities Osteoarthritis Index), with no absolute contraindications to surgery. Separate logistic regression models examined the independent effects of education and income on the potential need for, and definite willingness to consider arthroplasty, after controlling for age, sex, and region of residence. Potential unmet need was estimated as the proportion of subjects with the need for arthroplasty who were not already on a surgery waiting list, who were definitely willing to consider arthroplasty, and who had evidence of arthritis by examination and radiography.

Results

Response rates were at least 72% for all questionnaires and interviews. Less education (adjusted odds ratio [OR] 1.57 for less than high school versus postsecondary education, 95% confidence interval [95% CI] 1.17–2.11) and lower income (adjusted OR 1.83 for ≤$20,000 versus >$40,000, 95% CI 1.24–2.70) were independently associated with a greater likelihood of having the potential need for arthroplasty. Among the subjects with potential need, neither education nor income was independently associated with a definite willingness to consider arthroplasty. Thus, taking willingness into consideration, individuals with less education and/or lower income were more likely to have potential unmet need for arthroplasty.

Conclusion

Persons with lower SES had a greater need for, and were equally willing to consider arthroplasty, compared with those with higher SES. Thus, observed SES disparities in the rates of performed arthroplasties cannot be explained by a lower need or less willingness to undergo arthroplasty in those with lower SES.

Arthritis is a leading cause of permanent incapacity, resulting in extensive use of health care resources and significant economic burden (1–5). As for other chronic disabling conditions, studies have documented a relationship between worse self-reported symptoms of arthritis/disability and lower socioeconomic status (SES), using a variety of accepted measures, including level of education, household income, and wealth (6–10). Observed SES differences in disease severity persist after adjustment for the prevalence of disease risk factors (7, 8). Thus, differential access to and/or use of arthritis health care services has been proposed as the most likely explanation for these observations (11–13).

Total joint arthroplasty is a well-recognized, efficacious, and cost-effective treatment for advanced hip or knee arthritis (14–20). The majority (>80%) of hip and knee replacements are performed for osteoarthritis (21). In the US, lower rates of joint arthroplasty have been reported in persons with low income versus those with high income (22). The Canadian health care system is characterized by universal access to care according to need. Nonetheless, even in Canada, SES gradients in medical care have been observed (23, 24). Studies have suggested that individuals with lower SES are less able to negotiate the health care system, which thereby impacts their access to care (25, 26). The main limitation of these prior studies is their failure to consider individuals' willingness or desire to receive treatment. The objective of this study was to assess the effect of education and income on potential need for, and willingness to consider hip and knee arthroplasty as a treatment option.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Details regarding the study design have been published elsewhere (27, 28). This was a 3-phase study of 100% of the population ages 55 years and older in a high-utilization (high-rate) area for arthroplasty, Oxford County (n = 21,925), and a low-utilization (low-rate) area, East York (n = 26,293), in Ontario, Canada. The high-rate area is relatively rural and ethnically homogenous; the low-rate area is located in metropolitan Toronto, and is therefore urban and ethnically diverse. The province of Ontario, with a population of >11.4 million, was ideal for study, because Ontarians have comprehensive, universal health insurance coverage; thus, barriers to health care based on insurance status were not an issue.

Identifying individuals with moderate-to-severe hip/knee arthritis.

In phase I, a brief questionnaire was sent to the listed residents of the 2 areas. Participants were asked to report their age, sex, height, and weight (to calculate body mass index [BMI]), the presence of symptomatic joints on a diagram, and the presence (or absence) of specific functional disabilities (stair climbing, walking, standing, rising from a chair). Respondents were selected for phase II if they had at least moderately severe hip or knee problems, as defined by the following criteria: 1) difficulty in the last 3 months with each of stair climbing, arising from a chair, standing, and walking, 2) swelling, pain, or stiffness in any joint lasting at least 6 weeks, and 3) indication on a diagram that a hip and/or knee had been “troublesome.” Participants eligible for phase II were mailed the phase II questionnaire within 3 weeks of receipt of their completed phase I questionnaire.

Assessing potential need for arthroplasty.

In phase II, respondents completed, by mail or telephone, a reliable and valid measure of the severity of their hip and knee symptoms, that is, the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (29), and a general health status measure, the Short Form 36 (SF-36) health survey (30, 31). They also reported (from a list of 18 items) concurrent health problems for which they were receiving treatment or had consulted a physician in the past year, whether or not they were currently under the care of a physician for their arthritis, their highest level of education attained, annual household income, racial background (white, nonwhite), and whether they spoke English (yes, no).

Because there are currently no universally accepted criteria by which to determine appropriateness for arthroplasty, for this study, “severe” arthritis was defined as a WOMAC summary score of ≥39 (of 100). The WOMAC summary score ranges from 0 (no pain or disability) to 100 (most pain and disability); the cut-off value of 39 represents the 25th percentile of scores for patients undergoing arthroplasty in Ontario (i.e., 25% of individuals undergoing joint arthroplasty have lower [better] WOMAC scores) (32), thus providing a conservative estimate of potential need for arthroplasty. People with scores of 39 or higher were considered potential candidates for arthroplasty, provided that they had no self-reported absolute contraindications to having the surgery. Absolute contraindications for arthroplasty, for which there was consensus among Ontario orthopedic surgeons (>90% agreement), were major mental illness, stroke with paralysis, or another major neurologic disorder (33).

Validation of self-reported severe arthritis.

In phase III (conducted within 3 months of phase II completion), we visited and examined a random sample of 200 low-rate area respondents with WOMAC scores ≥39 (severe arthritis) to determine the positive predictive value of our WOMAC cut-off for identifying individuals with “arthritis.” After readministration of the WOMAC, participants underwent a standardized joint examination (34) in their homes by a physiotherapist. Radiographs of the hips and knees were obtained at the local hospital and graded by radiologists in a blinded manner for arthritis presence and severity, using the Kellgren and Lawrence (K/L) classification criteria (35). Each hip and knee was classified as “arthritic” if there was stress pain, reduced range of motion, or deformity on examination, and if there was any evidence of osteophytes, joint space narrowing, or subchondral sclerosis on radiographs. We assumed that the proportion of patients with radiographically and clinically confirmed hip/knee arthritis would be the same in both regions. Thus, the radiography and joint examinations were performed only on the individuals in the low-rate area.

Willingness to consider arthroplasty.

In both areas, all respondents with severe hip or knee arthritis who had not undergone prior arthroplasty and were not on an arthroplasty waiting list completed a standardized in-person interview to determine their willingness to consider arthroplasty. The interview, written at a sixth-grade level, described the consequences of not having surgery, treatment alternatives, and the risks and benefits of arthroplasty, including the projected life span of the replaced joint (33, 36). Participants were additionally asked whether they knew anyone who had successfully undergone joint arthroplasty. They were also asked whether they had ever discussed arthroplasty with a physician, and if so, whether surgery was recommended. Participants indicated their willingness to consider arthroplasty on a 5-point scale: “definitely not willing,” “probably not willing,” “unsure,” “probably willing,” and “definitely willing.”

Defining potential need and potential unmet need for arthroplasty.

Potential need for arthroplasty was defined as severe arthritis (WOMAC score ≥39) in the absence of absolute contraindications to surgery (as outlined above). Potential unmet need for arthroplasty was then calculated as the number of individuals with potential need (from phase II) who were not already on an arthroplasty waiting list multiplied by a factor “X,” where “X” was the overall percentage of phase III participants with WOMAC scores ≥39 who had evidence of arthritis on both examination and radiograph, and by a factor “Y,” where “Y” was the percentage of phase III participants with WOMAC scores ≥39 who indicated “definite willingness” to consider arthroplasty. Factor “Y” was initially assessed for each region, and then, to examine for differences in unmet need by education and income, each region/education and region/income stratum was assessed.

Statistical analysis.

The demographic and socioeconomic characteristics of phase I respondents were compared with 1991 census data to assess the representativeness of our sample (37). In the primary analyses, multiple logistic regression modeling was used to examine the independent effects of education and income on 1) potential arthroplasty need (WOMAC score ≥39 and no absolute contraindications to surgery), and 2) definite willingness to consider having arthroplasty. To examine a potential interaction between sex and SES, we constructed several interaction variables (male versus female with less than a high school education; male versus female with a high school education or higher; male versus female with ≤$20,000 annual income; male versus female with >$20,000 annual income). In all analyses, we controlled for age and sex and, where appropriate, region of residence (high-rate versus low-rate region) and arthritis severity (WOMAC score). Additional independent variables considered were race (white, nonwhite), BMI, being under the care of a physician for arthritis (yes/no; in Ontario, referral by a physician is required to see a specialist, such as an orthopedic surgeon), having discussed arthroplasty with a physician (yes/no), knowing someone who has successfully undergone arthroplasty, and self-reported health status (SF-36 mental health component score).

Highest level of education attained was defined as “no formal education,” “elementary school only,” “high school,” or “university/college/postgraduate (postsecondary) education.” Because relatively few individuals had received no formal education or only elementary school education, these categories were combined in the category “less than high school.” Dummy variables were created to examine the independent effect of the less than high school and high school education levels compared with postsecondary education as the reference category. Alternatively, where simpler for presentation purposes, we dichotomized the level of education as low (less than high school) versus high (high school or higher). Annual household income was defined as ≤$20,000, $20,001–$40,000, $40,001–$60,000, or >$60,000. Relatively few individuals reported annual household incomes >$60,000; therefore, this income category was combined with $40,001–$60,000 for all subsequent analyses (>$40,000). Dummy variables were created to examine the independent effect of ≤$20,000 and $20,001–$40,000, compared with an annual income >$40,000 as the reference category. As for education, we also considered income as a dichotomized variable (≤$20,000 versus >$20,000). A composite education/income variable was created to explore a possible additive effect of having both low education and low income; respondents were categorized as having “low income/education” if they reported having an annual income ≤$20,000 and an education level less than high school (yes/no).

The colinearity of independent variables was assessed using Spearman's rank correlation coefficients. Highly correlated pairs of variables were not included together in the models. We did not impute missing data; rather, we ran all analyses including only those subjects for whom the data were complete. Due to the high item-nonresponse rate for income (18% of phase II respondents had missing values for income), we separately ran our regressions including “nonresponse to income” as a covariate to determine whether there was any independent effect. The goodness-of-fit of our final models was assessed using the Hosmer-Lemeshow test statistic (38). For all analyses, statistical significance was considered at a 2-tailed level of 0.05. Adjustment for multiple comparisons was made using the Bonferroni technique.

Ethics approval for all phases of the study was received from the Human Subjects Review Committee of the University of Toronto. Participants gave verbal consent to participate.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Identifying individuals with moderate-to-severe hip/knee symptoms.

After excluding those individuals who did not live in either of the 2 areas at the time of the survey, were too young (age <55 years), were deceased, or were unable to complete the surveys, there were 19,675 study-eligible individuals in the low-rate area and 19,658 in the high-rate area; of these, 14,082 (71.6%) and 14,369 (73.1%), respectively, completed the phase I questionnaire. When compared with the census data (37), phase I respondents had a higher income and lower likelihood of having moved residence in the prior year. Of the phase I respondents, 1,572 (11.2%) in the low-rate area and 1,735 (12.1%) in the high-rate area met the criteria for phase II, indicating that they had at least moderately severe hip or knee problems. Phase II response rates, adjusted for deceased and ineligible subjects, were 82.9% in the low-rate area (n = 1,089) and 81.0% in the high-rate area (n = 1,322). The characteristics of phase II respondents in the 2 regions studied are shown in Table 1.

Table 1. Characteristics of phase II respondents in low- and high-utilization areas for arthroplasty*
CharacteristicLow-rate area (n = 1,089)High-rate area (n = 1,322)
  • *

    Denominators are given where question-response rates were <100%. 95% CI = 95% confidence interval.

  • Higher scores on the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) indicate worse arthritis symptoms and disability.

Mean age, years (95% CI)71.4 (70.8–71.9)69.7 (69.2–70.2)
Female, n (%) (95% CI)809 (74.3) (71.6–76.9)942 (71.3) (68.7–73.7)
Level of education, n (%) (95% CI)  
 No formal schooling45/1,044 (4.3) (3.2–5.7)24/1,278 (1.9) (1.2–2.8)
 Elementary school257/1,044 (24.6) (22.0–27.4)421/1,278 (32.9) (30.4–35.6)
 High school502/1,044 (48.1) (45.0–51.1)626/1,278 (49.0) (46.4–52.0)
 Postsecondary school240/1,044 (23.0) (20.5–25.7)204/1,278 (16.0) (14.0–18.1)
Nonwhite, n (%) (95% CI)84/1,053 (8.0) (6.4–9.8)11/1,265 (0.9) (0.4–1.6)
Annual household income, n (%) (95% CI)  
 ≤$20,000560/881 (63.6) (60.3–66.8)691/1,088 (63.5) (60.6–66.4)
 $20,001–$40,000234/881 (26.6) (23.7–29.6)311/1,088 (28.6) (25.9–31.4)
 $40,001–$60,00057/881 (6.5) (4.9–8.3)66/1,088 (6.1) (4.7–7.7)
 >$60,00030/881 (3.4) (2.3–4.8)20/1,088 (1.8) (1.1–2.8)
Mean summary WOMAC scores (scale 0–100) (95% CI)41.6 (40.3–42.8)40.6 (39.6–41.7)

Assessing potential need for arthroplasty.

Phase II nonresponse rates were 3.7% and 18.3% for education and income, respectively. Among the phase II respondents, those who did not report income were more likely to be female (high-rate area only; P = 0.04), to have a nonwhite racial background (P < 0.0001), and to be less well educated (low-rate area only; P = 0.03). The characteristics of phase II respondents by level of education and annual household income are shown in Table 2. Adjusting for multiple comparisons, individuals with less than a high school education were more likely than those better educated to report an income at or below $20,000 and to be nonwhite, and were marginally older. Individuals who reported an income at or below $20,000 had similar characteristics to those with little education, with the exception that, compared with individuals with greater income, they were significantly more likely to be female.

Table 2. Characteristics of phase II participants by level of education and annual household income*
CharacteristicLevel of educationAnnual household income
<High school (n = 747)≥High school (n = 1,572)≤$20,000 (n = 1,251)>$20,000 (n = 718)
  • *

    NA = not applicable; SF-36 = Short Form 36 (see Table 1 for other definitions).

  • P = 0.05, with correction for multiple comparisons using the Bonferroni technique.

Female sex, n (%) (95% CI)524 (70.1) (66.7–73.4)1,167 (74.2) (71.9–76.2)1,031 (82.4) (78.7–83.1)410 (57.1) (53.4–60.8)
Annual household income >$20,000, n (%) (95% CI)142/628 (22.6) (19.4–26.1)568/1,311 (43.3) (40.6–46.1)NANA
Level of education ≥high school, n (%) (95% CI)NANA743/1,229 (60.5) (57.7–63.2)568/710 (80.0) (76.9–82.9)
Nonwhite, n (%) (95% CI)57 (7.6) (5.8–9.8)85 (5.4) (4.3–6.6)88 (7.0) (5.7–8.6)33 (4.6) (3.2–6.4)
Non–English speaking, n (%) (95% CI)76 (10.2) (8.1–12.6)49 (3.1) (2.3–4.1)76 (6.1) (4.8–7.6)12 (1.7) (0.9–2.9)
Mean age, years (95% CI)71.7 (71.0–72.4)69.8 (69.4–70.3)71.1 (70.6–71.7)69.1 (68.4–69.7)
Mean body mass index, kg/m2 (95% CI)28.2 (27.8–28.6)27.6 (27.3–27.9)28.0 (27.7–28.3)27.7 (27.3–28.0)
Mean SF-36 mental health component score (95% CI)48.1 (47.2–49.0)49.3 (48.7–50.0)47.0 (46.3–47.7)51.0 (50.1–51.9)

Correlates of potential arthroplasty need.

Among the 2,411 phase II respondents, 1,325 (55.0%) had WOMAC summary scores ≥39. Of these, 1,105 had no self-reported absolute contraindications to surgery and were therefore considered to have potential arthroplasty need. The proportion excluded on the basis of absolute contraindications to surgery was similar for all education and income subgroups (data not shown).

Adjusting for age, sex, region, and BMI, potential arthroplasty need (WOMAC summary score ≥39 and no absolute contraindications to surgery) was independently associated with both less education (less than high school) and lower income (annual income ≤$20,000). As shown in Table 3, after adjusting for education and income, sex was not an independent correlate of potential arthroplasty need. In separate analyses, we examined for a potential interaction between sex and SES. After adjusting for age, region, and BMI, we found that women with more than a high school education, and to a lesser extent, an annual income >$20,000, were more likely than men with a similar level of education or income to have potential arthroplasty need (adjusted odds ratio [OR] 1.84, 95% confidence interval [95% CI] 1.21–2.81 and adjusted OR 1.30, 95% CI 0.94–1.79, respectively). However, among men and women with low SES (less than a high school education or income ≤$20,000), no sex differences in potential arthroplasty need were observed (for women versus men adjusted OR 1.11, 95% CI 0.81–1.53 and adjusted OR 0.95, 95% CI 0.70–1.28, respectively).

Table 3. Logistic regression model to examine the effects of level of education and annual household income on potential need for arthroplasty*
Independent variableAdjusted odds ratio95% confidence intervalP
  • *

    Potential arthroplasty need was defined as a Western Ontario and McMaster Universities Osteoarthritis Index summary score ≥39 (of 100) and no absolute contraindication to having surgery.

Age (per year increase in age)1.021.01–1.03<0.0001
Sex (reference is female)0.860.69–1.070.16
Region (reference is low-rate region)1.040.86–1.260.21
Body mass index (per kg/m2 increase)1.021.00–1.040.03
Level of education (reference is postsecondary education)   
 High school only1.170.89–1.520.26
 <High school1.571.17–2.110.003
Annual household income (reference is >$40,000)   
 $20,001–$40,0001.360.91–2.030.13
 ≤$20,0001.831.24–2.700.003

Determining potential unmet need for arthroplasty, after taking willingness into consideration.

Seventy-eight individuals with potential arthroplasty need were already on an arthroplasty waiting list and were therefore excluded from our estimates of potential unmet need, leaving 1,027 phase II participants with potential arthroplasty need (447 in low-rate region, 580 in high-rate region). In phase III, 89.6% of low-rate area participants with WOMAC summary scores ≥39 had evidence of arthritis in at least one hip or knee on both joint examination and radiographs (41% had K/L grade 3 or 4). We therefore reduced the estimated total with potential arthroplasty need by 10.4%, leaving 920 (400 in the low-rate region, 520 in the high-rate region).

Phase III “willingness” interviews were performed in 456 individuals with potential arthroplasty need (211 in low-rate region, 245 in high-rate region). The mean age of participants was 68.7 years (range 55–96 years) and 72.6% were female. The mean WOMAC summary score among participants was 55.9 (range 39.6–100).

Among these 456 participants, 17.5% in the low-rate region and 21.5% in the high-rate region (P = 0.021) were “probably willing” to have joint arthroplasty, while only 8.5% and 14.9%, respectively (P = 0.03), were “definitely willing.” In the low-rate region, 4.4% with less than a high school education, 10.4% with a high school education, and 9.1% with more than a high school education were “definitely willing” to consider arthroplasty. For the high-rate region, the frequencies of response for definite willingness by education level were 12.3%, 16.3%, and 16.7%, respectively. By income, 8.0% in the low-rate region and 15.3% in the high-rate region with ≤$20,000 annual income, 10.8% and 8.1%, respectively, with income between $20,001 and $40,000, and 21.4% and 18.2%, respectively, with >$40,000 annual income were “definitely willing” to consider arthroplasty. Adjusting for age, sex, and region, “definite willingness” to consider arthroplasty was associated independently with having spoken with a physician about arthroplasty (Table 4). After taking these variables into consideration, neither education (adjusted OR for less than a high school versus postsecondary education 1.19, 95% CI 0.50–2.86; P = 0.70) nor income (adjusted OR for ≤$20,000 versus >$40,000 annual income 0.36, 95% CI 0.11–1.15; P = 0.08) was statistically significantly associated with definite willingness to consider arthroplasty.

Table 4. Logistic regression model to examine the effects of education and income on definite willingness to consider arthroplasty
Independent variableAdjusted odds ratio95% confidence intervalP
Age (per year increase in age)0.940.90–0.980.005
Sex (reference is female)0.470.19–1.130.09
Region (reference is low-rate region)1.360.68–2.700.38
Spoken with physician about having arthroplasty (reference is “no”)3.191.63–6.230.0007
Level of education (reference is postsecondary education)   
 High school only0.820.28–2.380.71
 <High school1.190.50–2.860.70
Annual household income (reference is >$40,000)   
 $20,001–$40,0000.490.17–1.390.18
 ≤$20,0000.360.11–1.150.08

Taking into consideration the proportion who indicated “definite willingness” to consider arthroplasty in each region, 12.0% of phase II respondents with potential arthroplasty need were estimated as having potential unmet need for arthroplasty. By level of education, using education/region-specific willingness, the proportion of phase II respondents with potential unmet need was 5.6% among individuals with less than a high school education, 4.5% among individuals with a high school level of education, and 3.4% among individuals with postsecondary education. Similarly, using income/region-specific willingness, the proportion of individuals with potential unmet need for arthroplasty was 4.5% among those with annual incomes at or below $20,000, compared with 3.5% among individuals with annual incomes between $20,000 and $40,000 and incomes above $40,000. Finally, using the proportion of individuals “definitely willing” to consider arthroplasty for each region and within each combined education/income strata, estimated need, taking into consideration willingness, was 4.7% among those with both less than a high school education and annual income at or below $20,000, compared with 2.1% for those with both postsecondary education and income above $40,000.

Due to the high item-nonresponse rate for income (18%), we reran our models for “potential arthroplasty need” and “definite willingness to consider arthroplasty” with “nonresponse to income” included as a covariate. Inclusion of this variable did not affect the reported relationships between income, education, and arthroplasty need or willingness, nor was “nonresponse to income” a significant correlate of either of these outcomes.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Previous studies have documented lower rates of joint arthroplasty among individuals with lower SES, but to date, none have considered an important potential explanation for this finding, that is, differences in individuals' desire for care or treatment preferences. Herein we estimated both the potential need for and willingness to consider undergoing hip and knee arthroplasty in a population-based Canadian cohort, to determine whether differences existed on the basis of 2 accepted measures of SES, level of education and annual household income. As others have reported, we found that even in a setting of universal access to health care, less education and lower income were independent correlates of worse arthritis symptoms and disability, thus creating greater potential need for hip and knee arthroplasty. More importantly, for the first time, we have shown that among the individuals with potential need for this procedure, willingness to undergo joint arthroplasty did not differ by SES, and therefore did not explain the observed differences in rates of use.

Chaturvedi and Ben-Shlomo (13) documented lower rates of hip arthroplasty among individuals with lower SES, compared with individuals with higher SES, who had consulted a general practitioner in the UK for osteoarthritis. This was despite the fact that the overall rates of consultation for osteoarthritis were greater for those with lower SES. In our study, adjustment for whether or not the individual was under the care of a physician for their arthritis did not eliminate the observed SES differences in potential arthroplasty need. However, we did not quantify the frequency of visits to, or treatment prescribed by these health care practitioners, nor did we assess individuals' adherence to physicians' recommendations. Therefore, we cannot be certain that clinical practice patterns or patient compliance did not differ according to the level of education or income of the patient, as others have suggested. Neuberger et al found that individuals with fewer years of formal education had less perception of the benefits from exercise compared with those who had attained more education, potentially reducing their motivation to adhere to exercise recommendations (12). Similarly, Dexter and Brandt found that the instruction in, and monitoring of therapeutic joint exercises was conducted in a more comprehensive manner in individuals with greater education than in those without (11).

It has been suggested that arthritis severity, and thus the potential need for arthroplasty, may be greater among those with less education or lower income due to a higher prevalence of exposure to risk factors for osteoarthritis. Identified risk factors for hip and knee osteoarthritis include older age, female sex, and obesity (39, 40). In our cohort, individuals with low education and/or low income were significantly more likely to be female, and were older and heavier. However, adjustment for these potential confounders in our analyses did not eliminate the observed education/income gradients. Specific occupational exposures have also been associated with incidence of osteoarthritis, in particular, occupations requiring repetitive knee bending (41). Because most of the subjects in our cohort were retired, we did not collect information on prior occupations. Therefore, we could not examine the effect of occupational exposures on arthritis prevalence or severity.

Neither the level of education nor income was an independent correlate of individuals' willingness to undergo joint arthroplasty, after taking into consideration whether or not the individual had ever discussed this procedure with a physician. However, since relatively few individuals in our cohort had discussed arthroplasty with a physician, we could not explore, with confidence, whether or not education or income predicted physicians' recommendations regarding surgery.

Women, compared with men, have disproportionately greater potential unmet need for hip/knee arthroplasty after taking into consideration age, disease severity, and willingness to consider arthroplasty (28). In examining the interaction between sex and SES, we found that among men and women with low SES, no sex differences in potential need for arthroplasty existed. However, among those with higher SES, women were more likely than men to have need for this procedure, even after adjusting for important confounders such as age and BMI. We have previously proposed that one explanation for this difference is that barriers (actual and/or perceived) exist at the level of the interaction between the primary care provider and the hip/knee arthritis patient in referral to orthopedic surgery (28). The findings in the present study suggest that these barriers are similar for men and women with low SES, but systematically greater for women compared with men at higher levels of SES.

In addition to sex, racial background has also been associated with access to care. In particular, studies in the US have documented lower rates of joint arthroplasty among nonwhites (22, 42). In our study, there was too little racial disparity to examine, with confidence, the interactions among SES, sex, and racial background with respect to their impact on potential unmet need for joint arthroplasty. Further study is warranted to investigate the mechanisms whereby these factors interact to influence referral for, and performance of hip and knee arthroplasty.

There are several potential limitations to this study, in addition to those already noted. First, in an effort to factor in the important influence of patients' preferences on our estimates of need for arthroplasty, we adjusted our estimates of potential need by the patient's degree of willingness to have arthroplasty. The latter was evaluated by means of a standardized interview, rather than through conversations with an orthopedic surgeon. We used this approach to provide a standard and comprehensive list of the potential risks and benefits associated with arthroplasty, which we developed in light of our prior research that documented substantial variation among orthopedic surgeons in their opinions regarding the indications for, and outcomes of arthroplasty (33). We believe that because we listed all possible risks associated with arthroplasty, we obtained conservative estimates of the willingness to undergo arthroplasty. Furthermore, our approach is unlikely to have had systematically different effects depending on the participant's level of education or income.

Second, because there are no standardized guidelines regarding when and in whom arthroplasty should be performed, our criteria for potential need were based on self-reported symptoms and disability, obtained with use of a reliable and valid instrument (the WOMAC) that is widely used in North America to evaluate the outcomes of arthroplasty. Since studies indicate that the primary reasons reported by patients for undergoing arthroplasty are joint pain and functional limitation (16), these criteria seem reasonable. Furthermore, the estimates of need were adjusted for the likelihood that persons with high scores had both clinical and radiographic evidence of arthritis, and for contraindications to surgery.

Third, we estimated potential need for, and willingness to have, arthroplasty in only 2 areas of Ontario. The province of Ontario comprises approximately half the total population of Canada, and we compared the characteristics of our phase I respondents with both Ontario and Canadian census data. Nevertheless, although these respondents appeared to be representative of the underlying population, we cannot assume that similar results would be found in areas not studied.

Fourth, self-reported level of education and annual household income were used as our measures of “socioeconomic status.” Although the nonresponse rate for education was only 4% of participants, 18% did not specify their annual household income. Those who did not report their income were significantly more likely than those who did report income to have received less education, and to be nonwhite and non–English speaking. For this reason, we also ran our models including “nonresponse to income” as a covariate. As noted above, inclusion of this variable did not affect the reported relationships between income, education, and arthroplasty need or willingness. Of note, we did not collect information regarding the number of persons in the household. Thus, we could not adjust for this in examining the effect of income on arthroplasty need or willingness.

Finally, because the proportion of individuals with potential need for arthroplasty who reported definite willingness to consider having the procedure was low, we may have had insufficient power to detect significant differences in willingness by level of education or income. Among individuals with potential arthroplasty need who had low income, the likelihood of having indicated definite willingness to consider arthroplasty was much less, although not statistically so, than among those having a higher income. A larger study is needed to confirm our reported lack of an association between “willingness” and income.

In conclusion, we have documented greater potential need for hip/knee arthroplasty among individuals with less education and, in particular, low income. Among those with potential need, the proportion reporting definite willingness to consider joint arthroplasty was similar, regardless of level of education or income. As a result, there was an overall greater proportion of individuals who met our criteria for unmet arthroplasty need among those of low SES versus those of high SES. These results suggest that the previously reported lower rates of hip/knee arthroplasties performed in those with low SES (versus high SES) cannot be explained by either a lower prevalence of disabling hip/knee arthritis (i.e., demonstrable need for the procedure) or a greater unwillingness to consider arthroplasty as a treatment option in those with low SES. Instead, disparity in arthroplasty rates may be related to differences in physician management of hip/knee arthritis, and in particular, their opinions regarding when, and if, to refer for, or perform joint arthroplasty, according to the patients' level of education or income.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The authors are most grateful to Dr. Susan Jaglal for her methodologic and statistical advice.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES
  • 1
    Elders MJ. The increasing impact of arthritis on public health. J Rheumatol 2000; 27: 68.
  • 2
    Liang MH, Larson M, Thompson M, Eaton H, McNamara E, Katz R, et al. Costs and outcomes in rheumatoid arthritis and osteoarthritis. Arthritis Rheum 1984; 27: 5229.
  • 3
    Pincus T, Mitchell JM, Burkhauser RV. Substantial work disability and earnings losses in individuals less than age 65 with osteoarthritis: comparisons with rheumatoid arthritis. J Clin Epidemiol 1989; 42: 44957.
  • 4
    Badley EM, Rasooly I, Webster G. The impact of musculoskeletal disorders in the population: are they aches and pains? Findings from the 1990 Ontario Health Survey. J Rheumatol 1995; 22: 7339.
  • 5
    Badley EM, Rasooly I, Webster G. Relative importance of musculoskeletal disorders as a cause of chronic health problems, disability and health care utilization: findings from the 1990 Ontario Health Survey. J Rheumatol 1994; 21: 50514.
  • 6
    Leigh JP, Fries JF. Occupation, income and education as independent covariates of arthritis in four national probability samples. Arthritis Rheum 1991; 34: 98495.
  • 7
    Leigh JP, Fries JF. Correlations between education and arthritis in the 1971–1975 NHANES I. Soc Sci Med 1994; 38: 57583.
  • 8
    Hannan MT, Anderson JJ, Pincus T, Felson DT. Educational attainment and osteoarthritis: differential associations with radiographic changes and symptom reporting. J Clin Epidemiol 1992; 45: 13947.
  • 9
    Kaplan RM, Alcaraz JE, Anderson JP, Weisman M. Quality-adjusted life years lost to arthritis: effects of gender, race and socioeconomic status. Arthritis Care Res 1996; 9: 47382.
  • 10
    Badley EM, Ibanez D. Socioeconomic risk factors and musculoskeletal disability. J Rheumatol 1994; 21: 51522.
  • 11
    Dexter P, Brandt K. Relationships between social background and medical care in osteoarthritis. J Rheumatol 1993; 20: 698703.
  • 12
    Neuberger GB, Kasal S, Smith KV, Hassanein R, DeViney S. Determinants of exercise and aerobic fitness in outpatients with arthritis. Nurs Res 1994; 43: 117.
  • 13
    Chaturvedi N, Ben-Shlomo Y. From the surgery to the surgeon: does deprivation influence consultation and operation rates? Br J Gen Pract 1995; 45: 12731.
  • 14
    Chang RW, Pellissier JM, Hazen GB. A cost-effectiveness analysis of total hip arthroplasty for osteoarthritis of the hip. JAMA 1996; 275: 85865.
  • 15
    Bunker JP, Frazier HS, Mosteller F. Improving health: measuring effects of medical care. Milbank Q 1994; 72: 22558.
  • 16
    Hawker GA, Wright JG, Coyte PC, Paul JE, Dittus R, Croxford R, et al. Health-related quality of life after knee replacement surgery: results from the Knee PORT Study. J Bone Joint Surg (Am) 1998; 80-A: 16373.
  • 17
    Williams JI, Llewellyn Thomas H, Arshinoff R, Young N, Naylor CD and the Ontario Hip and Knee Replacement Project Team. The burden of waiting for hip and knee replacements in Ontario. J Eval Clin Pract 1997; 3: 5968.
  • 18
    Lavernia CJ, Guzman JF, Gachupin-Garcia A. Cost effectiveness and quality of life in knee arthroplasty. Clin Orthop Rel Res 1997; 345: 1349.
  • 19
    Garellick G, Malachau H, Herberts P, Hansson E, Axelsson H, Hansson T. Life expectancy and cost utility after total hip replacement. Clin Orthop Rel Res 1998; 346: 14151.
  • 20
    Laupacis A, Bourne R, Rorabek C, Feeny D, Wong C, Tugwell P, et al. Costs of elective total hip arthroplasty during the first year. J Arthroplasty 1994; 9: 4817.
  • 21
    Katz BP, Freund DA, Heck DA, Dittus RS, Paul JE, Hawker GA, et al. Demographic variation in counts and the rate of knee replacement: a multi-year analysis. Health Services Res 1996; 31: 12540.
  • 22
    Gittelsohn AM, Halpern J, Sanchez RL. Income, race, and surgery in Maryland. Am J Public Health 1991; 81: 143541.
  • 23
    Humphries KH, van Doorslaer E. Income-related health inequality in Canada. Soc Sci Med 2000; 50: 66371.
  • 24
    Dunlop S, Coyte PC, McIssac W. Socio-economic status and the utilisation of physician' services: results from the Canadian National Population Health Survey. Soc Sci Med 2000; 51: 12334.
  • 25
    McIssac W, Goel V, Naylor D. Socio-economic status and visits to physicians by adults in Ontario, Canada. Journal of Health Services Research and Policy 1997; 2: 94102.
  • 26
    Roos NP, Mustard CA. Variation in health and health care use by socioeconomic status in Winnipeg, Canada: does the system work well? Yes and no. Milbank Q 1997; 75: 89111.
  • 27
    Hawker GA, Wright JG, Coyte PC, Williams JI, Harvey B, Glazier R, et al. Determining the need for hip and knee arthroplasty: the role of clinical severity and patients' preferences. Medical Care 2001; 39: 20616.
  • 28
    Hawker GA, Wright JG, Coyte PC, Williams JI, Harvey B, Glazier R, et al. Differences between men and women in the rate of use of hip and knee arthroplasty. N Engl J Med 2000; 342: 101622.
  • 29
    Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes following total hip or knee arthroplasty in osteoarthritis. J Ortho Rheumatol 1988; 1: 95108.
  • 30
    Stewart AL, Hays RD, Ware JE Jr. The MOS short-form general health survey: reliability and validity in a patient population. Medical Care 1988; 26: 72435.
  • 31
    Ware JE Jr, Snow KK, Kosinski M, Gandek B. SF-36 health survey: manual and interpretation guide. Boston: Health Institute, New England Medical Center; 1993.
  • 32
    Wright JG, Young NL. The patient specific index: asking patients what they want. J Bone Joint Surg (Am) 1997; 79: 97483.
  • 33
    Wright JG, Coyte PC, Hawker GA, Bombardier C, Cooke D, Heck DA, et al. Variation in orthopaedic surgeons' perceptions of the indications for and outcomes of knee replacement. CMAJ 1995; 152: 68797.
  • 34
    Bombardier C, Kinkhoff A, Bell M for the Canadian Clinical Epidemiology Research Group. Illustrated guide to a standard examination for joint tenderness and swelling. Toronto: Merrell Dow Research Institute; 1988.
  • 35
    Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthritis. Ann Rheum Dis 1957; 16: 494501.
  • 36
    Callahan CM, Drake BG, Heck DA, Dittus RS. Patient outcomes following tricompartmental total knee replacement: a meta-analysis. JAMA 1994; 271: 134957.
  • 37
    1996 census: national series tables. Ottawa, Ontario, Canada: Statistics Canada; 1997.
  • 38
    HosmerDW, LemeshowS, editors. Applied logistic regression. New York: John Wiley and Sons; 1989. p. 140.
  • 39
    Oliveria SA, Felson DT, Cirillo PA, Reed JI, Walker AM. Body weight, body mass index, and incident symptomatic osteoarthritis of the hand, hip and knee. Epidemiology 1999; 10: 1616.
  • 40
    Felson DT, Zhang Y, Hannan MT, Naimark A, Weissman B, Aliabadi P, et al. Risk factors for incident radiographic knee osteoarthritis in the elderly: the Framingham study. Arthritis Rheum 1997; 40: 72833.
  • 41
    Felson DT, Hannan MT, Naimark A, Berkeley J, Gordon G, Wilson PW, et al. Occupational physical demand, knee bending, and knee osteoarthritis: results from the Framingham study. J Rheumatol 1991; 18: 158792.
  • 42
    Wilson MG, May DS, Kelly JJ. Racial differences in the use of total knee arthroplasty for osteoarthritis among older Americans. Ethn Dis 1994; 4: 5767.