Differences in expectations of outcome mediate African American/white patient differences in “willingness” to consider joint replacement

Authors

  • Said A. Ibrahim,

    Corresponding author
    1. Center for Health Equity Research and Promotion, VA Pittsburgh HealthCare System, Pittsburgh, Pennsylvania
    • Center for Health Equity Research and Promotion, VA Pittsburgh HealthCare System, University Drive C, 11-East (130 A-U), Pittsburgh, PA 15240
    Search for more papers by this author
    • Dr. Ibrahim is the recipient of a VA Career Development Award in Health Services Research.

  • Laura A. Siminoff,

    1. Case Western Reserve University School of Medicine, Cleveland, Ohio
    Search for more papers by this author
  • Christopher J. Burant,

    1. Case Western Reserve University School of Medicine, Cleveland, Ohio
    Search for more papers by this author
  • C. Kent Kwoh

    1. Center for Health Equity Research and Promotion, VA Pittsburgh HealthCare System, and University of Pittsburgh, Pittsburgh, Pennsylvania
    Search for more papers by this author

Abstract

Objective

Joint replacement therapy is an effective treatment option for end-stage osteoarthritis (OA) of the knee and/or hip. There are marked racial/ethnic disparities in the utilization of this procedure. The reasons for these disparities are not known. We sought to determine whether African American patients differ from white patients in their “willingness” to consider joint replacement and to determine the factors that influence this relationship.

Methods

We performed a cross-sectional survey of 596 elderly, male, African American or white patients with moderate-to-severe symptomatic knee or hip OA who were receiving primary care at the Department of Veterans Affairs outpatient clinics.

Results

The groups were similar with respect to age, severity of arthritis measured by the Lequesne Scale and the Western Ontario and McMaster Universities Osteoarthritis Index, and scores on the Charlson Comorbidity Index and Geriatric Depression Scale. Compared with whites, African Americans were less likely to be employed or married or to have attained a high school education, but were more likely to report a median annual household income of <$10,000. They were also less likely than whites to be familiar with joint replacement and more likely to expect a longer duration of hospital course, pain, and functional disability following replacement surgery. African American patients were less “willing” than white patients to consider joint replacement (odds ratio 0.50, 95% confidence interval 0.30–0.84). However, this difference was explained by the between-group differences in expectations.

Conclusion

African American patients were less likely than white patients to express “willingness” to consider joint replacement if the procedure was needed and recommended. This difference was explained by differences between the groups in their expectations of hospital course, pain, and function following replacement surgery.

Joint replacement therapy is a cost-effective treatment option for end-stage osteoarthritis (OA) of the knee and/or hip (1). Ninety percent of those who undergo knee or hip joint replacement experience relief of pain and/or functional improvement—an important outcome(s) for patients (2). The procedure is associated with a mortality risk of ∼1%, making it one of the safest surgical procedures routinely performed (2). A single successful joint replacement treatment lasts for ∼10–20 years.

One in every 115 Americans over the age of 65 years has had a joint replacement, with ∼500,000 procedures being performed annually in the US (3). Considering the aging of the US population, the high prevalence of OA in all ethnic and demographic groups (4–6), and the fact that OA is a leading cause of disability among the elderly (7, 8), one would expect few disparities in utilization of a treatment that is as effective as joint replacement. Yet there are marked racial/ethnic disparities in the utilization of this procedure. White men are 3.0–5.1 times more likely than African American men to undergo knee joint replacement (9–11). Significant differences in favor of whites have also been reported for hip joint replacement (12). Moreover, the rate of total knee replacements performed on whites between 1980 and 1987 increased twice as rapidly as the rate among African Americans (12). Similarly, the rate of total hip replacements has increased 1.5 times more rapidly for whites than for African Americans (12). Differences in access to care cannot explain these disparities, since most of the candidates for these procedures are older persons with access to public health insurance, such as Medicare. Furthermore, there are no known financial disincentives for orthopedic surgeons in performing these procedures. It would seem, then, that patient-level factors may be responsible for part, if not most, of the utilization disparity.

As in other elective medical procedures (e.g., cardiovascular procedures), the decision to undergo joint replacement is influenced by clinical indications (usually conveyed to patients as recommendations for surgery) and patient “willingness” to comply. “Willingness” to consider joint replacement can best be viewed as a culmination of numerous clinical, social, and personal factors (e.g., familiarity with the procedure and outcome expectations). Indeed, previous studies of the utilization of elective cardiovascular procedures found African Americans to be more reluctant than whites to consent to these procedures; among other things, familiarity with the procedure was one important determinant of “willingness” (13, 14).

To our knowledge, no studies have compared African American and white potential candidates for joint replacement with respect to their “willingness” to undergo such procedures or with respect to the factors that might determine such “willingness.” A recent study from Canada indicated that having had a discussion on joint replacement treatment with a physician was one of only two factors (the other being younger age) that were found to predict “willingness” to consider joint replacement when the procedure was clinically indicated (15). Although the purpose of that study was to compare regional variations in the utilization of joint replacement in Ontario, Canada, there were not enough African Americans in the sample for it to be considered a true comparison of African Americans and whites.

In this study of elderly male patients at the Department of Veterans Affairs (VA) with moderate-to-severe symptomatic knee or hip OA, we addressed two issues. First, we investigated whether African American patients differed from white patients in their “willingness” to consider joint replacement if the procedure was needed and recommended by a doctor. Second, we examined the patient-level demographic, psychosocial, and clinical factors that influenced this decision.

PATIENTS AND METHODS

Patient population.

From the list of patients scheduled for primary care visits at VA outpatient clinics, study staff approached potential participants who met the age eligibility criteria as they entered the waiting room to be seen. Patients were initially asked a series of questions regarding the presence, duration, and severity of hip or knee pain (from the Arthritis Supplement of the National Health and Nutrition Examination Survey I [16]): “Have you ever had pain in and around your knee/hip on most days for at least one month? Over the past month, have you had pain in the knee when walking or standing at least half of the day?” Patients who answered “yes” to both of these questions were screened as positive for chronic knee/hip pain consistent with the presence of symptomatic OA. To be eligible, patients had to be ≥50 years of age and to have had moderate-to-severe pain for >6 months (evaluated on the basis of the Lequesne Scale [17]). Patients who had already had knee/hip replacement were excluded.

Between May 1997 and March 2000, 1,351 patients were approached. Seven hundred seventy-six patients met the study criteria. Thirty-eight patients refused to participate. The first 600 patients who were eligible and willing to participate were included in the study. Four patients were excluded from this analysis because of partially missing data. We have no information on the racial distribution of the 38 patients (5% of those who were eligible) who refused to participate in the study.

Data collection.

Baseline demographic information. Using field-tested questionnaires, interviewers gathered demographic information, such as age, education level, employment status, annual household income, and marital status. Patients were asked to identify their ethnicity. Chart and VA clinical computer database abstractions provided information on medications, comorbidity, and health care utilization.

Radiographic evaluation of the more symptomatic joint (hip or knee) was obtained to aid in the confirmation of the diagnosis of hip or knee OA. All knee and/or hip radiographs were read by two blinded observers and graded using the Kellgren/Lawrence (K/L) scale (18). This scoring system has been extensively used in research and has good reliability and validity (19, 20). Reliability was improved further with the use of an updated guide for standardizing the interpretation of radiographs prior to and during clinical trials (20) as the basis of grading the severity of each of the radiographic features of OA.

Study measures. To assess familiarity with joint replacement therapy for OA, patients were asked the following set of questions: 1) Have you ever heard of surgery to replace a hip or knee that has been damaged by arthritis? 2) Do you have family or friends that have had a hip or knee replaced by surgery? 3) Do you have a good understanding of what happens to somebody when they undergo hip or knee replacement surgery? Response options for all questions were “yes” or “no.”

To assess outcome expectations, patients were asked the following series of questions: 1) How often do you think someone dies as a result of hip or knee replacement surgery? Response categories ranged from “never” to “often.” 2) How long do you think someone who has hip or knee replacement surgery would be in the hospital or other health care facility? Response categories ranged from “1 to 3 days” to “more than 2 weeks.” 3) How long do you think it would take someone to fully recover from hip or knee replacement surgery? Response categories ranged from “less than 2 weeks” to “more than 12 months.” 4) Sometimes people who have extreme pain in their hip or knee choose to have joint replacement surgery. How much pain do you think people will still have after they have recovered from their surgery? Response categories ranged from “none” to “an extreme amount.” 5) Sometimes people who have extreme difficulty walking choose to have joint replacement surgery. How much difficulty walking do you think people will still have after they have recovered from their surgery? Response categories ranged from “none” to “an extreme amount.”

To assess “willingness” to consider joint replacement, patients were asked, “If your pain were to get severe, would you consider surgery to replace your knee/hip if your doctor recommended it?” Response options were “yes” or “no.” These questions and their complete response options are shown in Table 1.

Table 1. Study questions and response options
QuestionResponse option
Familiarity with joint replacement 
 Have you ever heard of surgery to replace a hip or knee that has been damaged by arthritis?Yes/no
 Do you have family or friends that have had a hip or knee replaced by surgery?Yes/no
 Do you have a good understanding of what happens to somebody when they undergo hip or knee replacement surgery?Yes/no
Outcome expectations 
 How often do you think someone dies as a result of hip or knee replacement surgery?Never, extremely rare, sometimes, often
 How long do you think someone who has hip or knee replacement surgery would be in the hospital or other health care facility?1–3 days, 4–7 days, 1–2 weeks, >2 weeks
 How long do you think it would take someone to fully recover from hip or knee replacement surgery?<2 weeks, 2 weeks to 1 month, 1–2 months, 2–6 months, 6–12 months, >12 months
 Sometimes people who have extreme pain in their hip or knee choose to have joint replacement surgery. How much pain do you think people will still have after they have recovered from their surgery?None, a little, a moderate amount, an extreme amount
 Sometimes people who have extreme difficulty walking choose to have joint replacement surgery. How much difficulty walking do you think people will still have after they have recovered from their surgery?None, a little, a moderate amount, an extreme amount
“Willingness” to consider joint replacement 
 If your pain were to get severe, would you consider surgery to replace your knee/hip if your doctor recommended it?Yes/no

Other study measures. The Lequesne Scale was used during the screening process to assess patients' rating of the severity of disease. The Lequesne Scale is an arthritis severity measure with high internal consistency (Cronbach's α = 0.84) that has been validated and used in other studies (17). Severity of hip pain is rated on a scale of 0–24; knee pain is rated on a scale of 0–22. Scores ≥5 are considered to reflect moderate-to-severe arthritis.

The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) has been validated and utilized to assess severity of OA (21, 22). It is a reliable (Cronbach's α ≥ 0.80) and valid scale that was designed specifically to assess lower extremity pain and function in OA. Summary scores range from 0 (no pain or disability) to 100 (extreme pain or disability). The WOMAC normally has a ceiling of 96, but for this analysis, the scores were transformed to yield a maximum possible score of 100.

The Geriatric Depression Scale (GDS) is a validated 15-item scale used to screen for depression in the elderly (23). Scores of 0–4 are considered normal, scores of 5–9 indicate mild depression, and scores of 10–15 indicate moderate-to-severe depression (24, 25). The GDS was used to assess the presence of comorbid depression in our sample.

The Charlson Comorbidity Index was used to assess overall disease burden. The index is based on the mean number of comorbid diseases per patient, based on medical record abstraction; scores range from 0 to 13 (26).

The study was approved by the Institutional Review Board of the Louis Stokes VA Medical Center, Cleveland, OH. All participants consented to the study.

Statistical analysis.

Baseline comparisons between African Americans and whites in the sample were performed using the chi-square test for categorical variables and the t-test for continuous, normally distributed variables. To dichotomize responses to questions with Likert scale response options, we picked the cut points based on the a priori distribution of Likert scale responses (that is, regardless of racial differences in responses). Differences that achieved a P value of ≤0.05 were considered statistically significant for this analysis. Bivariate associations between individual study outcome measures and ethnicity were examined. Multivariate logistic regression models were used to control for potential confounders. Simultaneous equation models were created using all covariates, along with the independent variable of interest (i.e., ethnicity). Variable inclusion in the model was based either on significance level in bivariate comparisons or on the importance of the variable based on the a priori hypothesis. Covariates entered in the final models were age, education level, median annual household income, GDS score, WOMAC score, and K/L score.

RESULTS

Baseline characteristics.

African American and white patients in this cohort were similar with respect to age (mean ± SD 65 ± 10 years versus 66 ± 19 years, respectively; P = 0.50), severity of arthritis on the Lequesne Scale (mean ± SD score 11 ± 4 versus 11 ± 4; P = 0.22) and the WOMAC (mean ± SD score 46 ± 17 versus 45 ± 17, respectively; P = 0.32), and scores on the Charlson Comorbidity Index (mean ± SD 2.3 ± 2 versus 2.5 ± 2, respectively; P = 0.24) and the GDS (mean ± SD 4.5 ± 3.4 versus 5 ± 3.8, respectively; P = 0.07). African Americans were less likely to be employed (8% versus 15%; P = 0.01) or married (39% versus 56%; P ≤ 0.00) or to have attained a high school education (43% of African Americans versus 29% of whites had not graduated high school; P ≤ 0.001), but were more likely to report an annual household income of <$10,000 (41% versus 20%; P ≤ 0.001). K/L scores for African Americans and whites were comparable (P = 0.08) (Table 2).

Table 2. Demographic, clinical, and psychosocial characteristics of the study patients, by ethnic group*
VariableAfrican Americans (n = 262)Whites (n = 334)P
  • *

    African American and white patients constituted 44% and 56%, respectively, of the sample. WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; GDS = Geriatric Depression Scale.

  • 0 = normal, 1 = minimal, 2 = moderate, 3 = severe.

Age, mean ± SD years65 ± 1066 ± 190.50
Education, %   
 Less than high school43.029.0≤0.001
 High school diploma46.151.7 
 More than high school11.319.3 
Annual household income, %   
 <$10,00041.020.0≤0.001
 $10,000–14,99928.530.6 
 $15,000–29,99920.136.3 
 ≥$30,00010.012.7 
Employed, %8.015.00.01
Married, %39.056.0≤0.00
WOMAC score, mean ± SD (0–100 scale)46 ± 1745 ± 170.32
Kellgren/Lawrence score, %   
 020.620.90.08
 128.332.6
 229.227.6
 321.918.9
GDS score, mean ± SD (0–15 scale)4.5 ± 3.45 ± 3.80.07
Charlson Comorbidity Index score, mean ± SD (0–13 scale)2.3 ± 22.5 ± 20.24

Familiarity with joint replacement and outcome expectations.

African American patients were less likely than white patients to have ever heard of joint replacement (81% versus 87%), but this difference was not statistically significant (P = 0.065). They were less likely than white patients to have had family or friends who had undergone joint replacement (52% versus 78%; P ≤ 0.001) or to report a good understanding of joint replacement as a form of treatment (44% versus 61%; P ≤ 0.001). African Americans were more likely than whites to say “sometimes” or “often” when asked, “How often do you think someone dies as a result of hip or knee replacement surgery?” This difference, however, did not reach statistical significance (25% versus 20%; P = 0.119). African Americans were more likely than whites to believe that the hospital course after surgery could last for >2 weeks (45% versus 18%; P ≤ 0.001) and that the recovery period would last for >6 months (47% versus 40%), although the latter difference was not statistically significant (P = 0.086). African Americans were more likely than whites to expect moderate or extreme pain (62% versus 42%; P ≤ 0.001) and moderate-to-extreme difficulty with walking (64% versus 39%; P ≤ 0.001) following joint replacement.

Table 3 summarizes the adjusted differences that were significant or nearly significant at the bivariate level, between African American and white patients' responses to questions. We adjusted for age, median annual household income, education level, GDS score, disease severity measured by the WOMAC, and K/L score.

Table 3. Summary of significant or nearly significant adjusted differences between African American and white patients' responses to questions regarding familiarity and expectations of outcome from multivariate models*
Question (variable)African American:white adjusted OR (95% CI)
  • *

    OR = odds ratio; 95% CI = 95% confidence interval; JR = joint replacement.

Ever hear of JR0.64 (0.37–1.09)
Family/friends had JR0.39 (0.26–0.61)
Good understanding of JR0.62 (0.42–0.92)
Hospital course >2 weeks4.09 (2.57–6.54)
Moderate-to-severe pain after JR surgery2.61 (1.74–3.89)
Moderate-to-severe difficulty walking after JR surgery2.76 (1.83–4.16)

“Willingness” to consider joint replacement.

When asked whether they would consider hip or knee replacement surgery if their pain were to become more severe and the doctor recommended it, African American patients were more likely than white patients to respond “no” (odds ratio 0.50, 95% confidence interval 0.30–0.84). When we adjusted for age, median annual household income, education level, GDS score, disease severity measured by the WOMAC, and K/L score, this difference remained significant (Table 4). Because we hypothesized that familiarity with joint replacement has an effect on “willingness,” we included responses to questions about familiarity with joint replacement in the model; the difference between African American and white patients with respect to their “willingness” to consider joint replacement remained significant (Table 4). When outcome expectations were included in the model that had all other covariates, the difference between African American and white patients in their “willingness” to consider joint replacement was no longer statistically significant (Table 4); this suggests that differential expectations of postsurgical hospital course, pain, and function mediated the observed difference in “willingness.”

Table 4. “Willingness” to consider joint replacement, adjusted for familiarity and expectations of outcome*
  • *

    See Table 3 for definitions.

  • Adjusted for familiarity in addition to other covariates.

  • Adjusted for expectation of outcomes, familiarity, and other covariates.

Crude OR (95% CI)0.50 (0.30–0.84)
Adjusted OR (95% CI)0.54 (0.30–0.96)
Adjusted OR (95% CI)0.53 (0.30–0.96)
Adjusted OR (95% CI)0.86 (0.45–1.63)

DISCUSSION

In this sample of male veterans with moderate-to-severe symptomatic OA of the knee or hip, we found that African American patients were less likely than white patients to express “willingness” to consider joint replacement if it was needed and recommended by a physician. This finding was explained by differences in the groups' expectations of the hospital course, pain, and function following replacement surgery. African American patients were more likely than white patients to perceive a longer duration of hospitalization, pain, and difficulty with walking following replacement surgery; they were also less familiar with joint replacement therapy. These differences remained significant after adjusting for important potential demographic, psychosocial, and clinical confounders.

Our findings are important in helping to understand the reasons for the marked racial/ethnic disparities in the utilization of joint replacement for knee/hip OA. They provide evidence supporting a finding by Katz suggesting that African American patients perceive, perhaps accurately, a higher risk of complications from joint replacement therapy (27). Utilizing Medicare claims data, Katz found that African American patients were more likely to receive joint replacement in low-volume hospitals, where the risks of complications or poor outcomes were higher. When volume and case mix were taken into account, African Americans still had greater rates of complications, according to Katz (27). We posit that our results hint at a possible mechanism by which the findings of Katz would translate into a clinical reality—that is, African American patients may be more likely than white patients to be exposed to knowledge of complications of joint replacement surgery through relatives, friends, and, perhaps, neighbors who have undergone the procedure. As a result, African American patients would be less “willing” than white patients to consider joint replacement even if it was needed and recommended—which is precisely what our results suggest.

It is not surprising that patient expectations shape attitudes and “willingness” to utilize recommended medical procedures. After all, “willingness” to consider or undergo a procedure is part and parcel of the medical decision-making process. Studies on decision making have shown that outcome bias or framing influences decision making (28–31), and that people assess the likelihood of an outcome—be it pain or function after replacement surgery—on the basis of similar experiences known to themselves, friends, relatives, or neighbors (30, 32). Also, the notion that expectations shape “willingness” is consistent with the health belief model, in that perceptions of efficacy affect decision-making (33).

The fact that African American and white patients differ in familiarity with joint replacement and outcome expectations is consistent with what other investigators found when studying the reasons for racial/ethnic variations in the utilization of elective cardiovascular procedures. For instance, Sedlis et al found that African American patients were more reluctant than white patients to agree to cardiovascular procedures (14). Similarly, Whittle et al found African American VA patients to be less likely than white VA patients to consent to elective cardiovascular procedures. However, they found that familiarity with the procedure, as opposed to race/ethnicity, was the single most important predictor of whether patients were willing to consent to procedures (13).

If future studies confirm that African American patients who undergo joint replacement therapy are more likely than white patients to receive this treatment in low-volume hospitals, where they are more likely to experience greater rates of complications, then the observed racial/ethnic disparities in the utilization of joint replacement may be understandable. Such evidence would be critical in the national effort to eliminate racial/ethnic disparities in the utilization of joint replacement. In that climate, interventions designed to improve quality of care for African American patients undergoing joint replacement surgery would be just as important as interventions targeting patient behavior and decision-making.

Several limitations must be considered in interpreting our results. First, we studied only male veterans in one VA location; therefore, our findings may not be generalizable to women and to veterans in other locations. Second, our measure of “willingness” is a single-item question with dichotomous response options, and it may not capture the true range of potential patient responses. Also, the question is framed in a hypothetical manner (i.e., it expects patients to imagine a scenario in which they need to undergo joint replacement and in which their doctor recommended the procedure). We are unable to predict whether their feelings about joint replacement would be the same if they were actually faced with the decision. Any one of these factors or a combination of them could have biased our results in one direction or the other.

In summary, this study of elderly male patients with symptomatic knee/hip OA found that African American patients were less likely than white patients to express “willingness” to consider joint replacement, even if the procedure was needed and recommended. This suggests that differential expectations of postsurgical hospital course, pain, and function mediate the observed differences in “willingness.” More studies are needed to assess the source of the observed differences between African American and white patients in outcome expectations for replacement surgery.

Ancillary