Perceptions of, and willingness to consider, total joint arthroplasty in a population-based cohort of individuals with disabling hip and knee arthritis


  • Gillian A. Hawker,

    Corresponding author
    1. University of Toronto, Sunnybrook and Women's College Health Sciences Centre, University Health Network Research Institute, and Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada
    • Women's College Campus, Sunnybrook and Women's, 76 Grenville Street, 10th floor, Room 1010, Toronto, Ontario, Canada, M5S 1B2===

    Search for more papers by this author
  • James G. Wright,

    1. Clinical Epidemiology and Health Care Research Program, Hospital for Sick Children, and University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • Elizabeth M. Badley,

    1. Clinical Epidemiology and Health Care Research Program, University of Toronto, and University Health Network Research Institute, Toronto, Ontario, Canada
    Search for more papers by this author
  • Peter C. Coyte,

    1. University of Toronto and Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada
    Search for more papers by this author
  • Toronto Arthroplasty Health Services Research Consortium



To examine perceptions of total joint arthroplasty (TJA) and how they relate to willingness to consider TJA.


A population-based survey in Oxford County, Ontario, Canada identified 1,735 subjects ≥55 years with disabling hip or knee arthritis; 435 English-speaking respondents with no prior TJA and not on a TJA waiting list were invited to participate, and 379 (37.1%) agreed. We assessed demographics; comorbidity; evaluated and perceived arthritis severity; perceived risks, benefits, indications for, and knowledge of TJA; preferred decision-making style; self efficacy; and willingness to consider TJA by questionnaire.


Participants' mean age was 67.6 years; 33.5% were willing (definitely or probably) to consider TJA as a treatment option. Willingness was independently associated with younger age (odds ratio [OR] <75 versus ≥75 years 2.42, P = 0.01); worse perceived arthritis severity (OR per unit increase 1.30, P < 0.001); perceiving TJA as appropriate for moderate, controlled joint pain (OR 3.29, P = 0.004); walking limited to <1 block (OR 1.99, P = 0.015); the risk of revision as acceptable (OR 3.73, P < 0.001); and friends as an important health information source (OR 2.49, P = 0.01).


Participants overestimated the pain and disability needed to warrant TJA. These misperceptions were strongly associated with unwillingness to consider TJA and should be addressed at a population level.


Arthritis of the hip and knee is a major cause of long-term disability (1, 2). When medical therapy fails to control the pain or when physical functioning reaches an unacceptable level, total joint arthroplasty (TJA) of the hip or knee is indicated (3–5). TJA results in substantial and sustained improvement in pain, functioning, and quality of life (3–7). However, we previously found that among appropriate candidates for TJA (defined by arthritis symptoms, disability, and radiographic joint appearance), only one-third were definitely or probably willing to consider this procedure as a treatment option (8).

Treatment choices are influenced by a range of factors (9), including individuals' tolerances and importance attached to their symptoms (10, 11), the manner in which risks and benefits of the treatment are presented (referred to as framing) (12), and the individuals' attitude toward risk (13). Prior knowledge of someone who has had the procedure in question has also been shown to influence individuals' health care decisions, as does the patient-provider relationship (14). Finally, geography (i.e., proximity to care) (15, 16) and the individuals' preferred role in clinical decision making (17–24) influence health care decisions. The major limitation of these studies, however, is that they have been based on clinical decisions made in the health care setting. In the case of individuals with debilitating hip or knee arthritis, many have never seen an orthopedic surgeon or discussed TJA as a treatment option (8). Individuals who receive care, or who seek an opinion from a surgeon, are likely different from those who do not.

The purpose of this study was to assess, in a population-based cohort with disabling hip and knee arthritis, perceptions about TJA and how these opinions and other factors relate to willingness to consider TJA as a treatment option.


Details regarding the design of this 3-phase study have been published elsewhere (8, 25, 26). The current article focuses on findings from Oxford County, Ontario, Canada. Ontarians have comprehensive, universal health insurance coverage; thus, barriers to health care based on insurance status are not an issue.

In phase I, we surveyed all Oxford County residents age ≥55 years (n = 21,925) to identify those with disabling hip or knee arthritis, defined as difficulty in the last 3 months with each of stair climbing, arising from a chair, standing, and walking; and swelling, pain, or stiffness in any joint lasting at least 6 weeks; and indication on a diagram that a hip or knee was currently troublesome. In phase II, we assessed demographics, arthritis severity using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (27), comorbidity (health problems for which they were receiving treatment or had seen a physician in the past year), general health status using the Short Form 36 (SF-36) (28), and whether they were on a waiting list for TJA.

In phase III, a subsample of English-speaking phase II respondents who had neither undergone prior TJA nor were on a TJA waiting list completed a standardized in-person interview. The WOMAC was readministered to this subsample. Phase III participants were asked several questions: if they knew anyone who had undergone TJA and if so, whether the surgery was successful; whether family or friends had recommended they speak with a physician about TJA and whether they had done so; their perceived hip or knee arthritis severity (on a 10-point Likert scale); whether they had medical problems that were more troublesome than their arthritis; which information sources they considered most important when making a nonurgent health care decision; their usual health decision-making style (usually do what the physician recommends, discuss the pros and cons with physician, seek other opinions before deciding, or ask for another medical opinion) (19); their opinion about what constitutes an appropriate candidate for TJA (indicated or not indicated for 5 levels of severity from least to greatest, for each of the following: difficulty climbing stairs, walking, joint pain, and general health status; and 6 categories of age and body weight from low to high); their perceptions of the usual outcomes following TJA (individuals walk unaided [yes/no], have none or very mild joint pain [yes/no], have significantly improved quality of life [yes/no]); and the acceptability of potential surgical risks (1% risk of death, 2% risk of major complications including myocardial infarction and stroke, and possible need for revision surgery). Finally, we assessed their self-efficacy after 10 years (29).

Participants were then read a standardized script, written at a grade 6 level and intended to match the surgeon-patient informed consent process, that included the expected consequences of not having TJA, alternative treatments, and the benefits and risks of TJA (based on the opinions of Ontario orthopedic surgeons and a published metaanalysis) (30, 31). Potential risks were outlined in greater detail than is usually provided by orthopedic surgeons. After reading the script, participants were asked to indicate their willingness to consider having TJA now: definitely not willing, probably not willing, unsure, probably willing, and definitely willing.

Statistical analyses.

Multivariate logistic regression was used to determine the independent correlates of willingness to consider TJA as a treatment option. Willingness was dichotomized as willing (definitely or probably) versus unwilling (unsure, definite or probable unwillingness). Development of the model was guided by O'Connor's Ottawa Decision Framework (32–34). This conceptual model knits together 3 aspects of individuals' decision making in elective health care decision situations: their perceptions of the decision itself (e.g., tolerances and importance attached to their symptoms and knowledge of TJA, including perceived indications for treatment and probabilities of specific treatment outcomes), perceptions of others (e.g., perceived support for decision, perceived social norms), and personal and external resources (e.g., income, health beliefs, proximity to care, sex, and different decisional styles). Using this framework as a guide, potential correlates of willingness were categorized as relating to perceptions around the decision itself, perceptions of others, and personal and external resources. Perceptions around the decision itself included such variables as age, perceived arthritis severity, presence of more important medical problems, perceptions of the indications for and risks and benefits associated with TJA, and health status. Health status was determined by WOMAC scores, body mass index, SF-36 mental component score, and number of self-reported comorbidities. Perceptions of others included such variables as perceived importance of family physician and friends as health information sources; having been recommended to discuss TJA with a physician by friends or family, and having done so; and knowledge of anyone who has undergone TJA, specifically someone with a successful outcome. Personal and external resources included such variables as sex, education (less than a high school education versus high school graduate), income (≤$20,000 versus >$20,000), living circumstances (alone versus with others), preferred decision-making style (prefers to discuss the pros and cons with physician or seek other opinions before deciding), and self efficacy.

Ethics approval for the study was obtained from the University of Toronto review boards.


Study participants.

Of 588 Oxford County residents with disabling hip or knee arthritis approached to participate, 153 (26%) were ineligible. Of the remaining 435, 379 (87.1%) participated (Table 1).

Table 1. Characteristics of the phase 3 participants
Characteristic*Overall n = 379Willing n = 127Unsure/unwilling n = 252
  • *

    Denominators are given when the response rate was <100%.

  • P < 0.05.

  • High scores on the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) indicate worse arthritis severity.

  • §

    P < 0.0001.

  • Higher scores on the Short Form 36 (SF-36) indicate better health status.

Age, mean ± SD years67.6 ± 8.266.2 ± 7.468.3 ± 8.6
Female, no. (%)273 (72.0)91 (71.7)181 (71.8)
Living alone, no. (%)75/376 (20.0)26/126 (20.6)49/250 (19.6)
Body mass index, mean ± SD kg/m228.6 ± 5.028.6 ± 4.828.6 ± 5.2
White, no. (%)364 (96.0)123 (96.9)242 (96.0)
Education <high school, no. (%)117 (30.9)38 (29.9)79 (31.3)
Annual income ≤$20,000, no. (%)200/330 (60.6)68/107 (63.6)132/223 (59.2)
WOMAC summary score, mean ± SD ×/9643.5 ± 17.848.1 ± 19.843.9 ± 17.8
Perceived hip/knee arthritis severity, mean ± SD (range) ×/105.7 ± 2.1 (1–10)6.4 ± 1.9 (0–10)5.4 ± 2.0 (1–10)§
SF-36 mental component score, mean ± SD ×/10050.5 ± 11.150.7 ± 11.150.3 ± 11.2
Number of self-reported comorbidities, mean (range)1.7 (0–8)1.5 (0–8)1.8 (0–7)
Has medical problems more troublesome than hip/knee arthritis, no. (%)79 (20.9)22 (17.3)57 (22.6)

Perceptions and experiences around TJA.

Participants' perceptions and experiences around TJA are described in Table 2. Participants' felt that TJA candidates should be extremely disabled by their arthritis. For example, most participants felt TJA was indicated in the setting of inability to climb stairs (81.3%), severe pain uncontrolled by pain killers (92.9%), and walking limited to in-house only (87.6%). However, significantly fewer participants felt it would be indicated in the setting of great difficulty climbing stairs (68.3%), severe pain controlled by pain killers (48.8%), or walking limited to <1 block (62.3%). Although they felt the arthritis disability needed to warrant TJA should be extreme, they felt that appropriate candidates should be in good general health otherwise: 71.2% felt TJA was appropriate for someone with excellent, very good, or good health compared with only 30.9% who felt it was appropriate for someone in fair health. Most perceived that, following recovery from TJA, individuals would walk unaided (76.1% versus 21.6% who felt individuals would usually walk with a cane), have none or very mild pain in the replaced joint (87.2% versus 12.0% who felt there would be moderate pain), and experience a significant improvement in their quality of life (72.7% versus 26.1% who felt that quality of life would be only somewhat improved). More than 80% felt the potential risk of death or of revision TJA were acceptable, and 71.2% were accepting of the potential for major complications.

Table 2. Perceptions and experiences around total joint arthroplasty*
 Total no. (%)Willing no. (%)Unsure/unwilling no. (%)
  • *

    Denominators are given when the response rate was <100%.

  • Respondents were asked to indicate acceptability for each level of indication presented, thus numbers (%) add to more than 100%.

  • P ≤ 0.001.

  • §

    P < 0.05.

  • Self-efficacy scores range from 10 to 100; minimum and maximum values for all 3 scale scores in this cohort were 10 and 100; higher scores indicate greater self efficacy.

Perceptions around the decision itself   
 Perceptions of the indications for TJA   
  Joint replacement is appropriate for someone who,   
   Is unable to climb stairs308 (81.3)104 (81.9)204 (81.0)
   Has great difficulty climbing stairs259 (68.3)101 (79.5)158 (62.7)
   Has some difficulty climbing stairs54 (14.2)19 (15.0)35 (13.8)
   Has severe pain not controlled by pain killers352 (92.9)114 (89.8)238 (94.4)
   Has severe pain controlled by pain killers185 (48.8)77 (60.6)108 (42.9)
   Has moderate pain controlled by pain killers44 (11.6)24 (18.9)20 (7.9)§
   Is limited to walking in-house only332 (87.6)108 (85.0)224 (88.9)
   Is limited to walking <1 city block236 (62.3)93 (73.2)143 (56.7)§
   Is limited to walking 1–5 city blocks106 (28.0)42 (33.1)64 (25.4)
   Is in excellent/very good/good health270 (71.2)94 (74.0)176 (69.8)
   Is in fair health117 (30.9)50 (39.4)67 (26.6)§
   Is in poor health29 (7.7)11 (8.7)18 (7.1)
   Is normal weight/mildly overweight239 (63.1)76 (59.8)163 (64.7)
   Is moderately overweight153 (40.4)53 (41.7)100 (39.7)
   Is severely overweight34 (9.0)18 (14.2)16 (6.3)§
   Is <70 years of age176 (46.4)54 (42.5)122 (48.4)
   Is 70–79 years of age141 (37.2)50 (39.4)91 (36.1)
   Is 80–89 years of age52 (13.7)25 (19.7)27 (10.7)§
   Is ≥90 years of age19 (5.0)9 (7.1)10 (4.0)
   Age is not important182 (48.0)69 (54.3)113 (44.8)
 Perceptions of the outcomes following TJA   
  After full recovery from TJA, most individuals …   
   Can walk unaided287 (76.0)105 (82.7)182 (72.2)§
   Can walk with a cane82 (21.6)20 (15.7)62 (24.6)
   Have no pain/very mild pain in the replaced joint327/375 (87.2)111/126 (88.1)216/249 (86.7)
   Have moderate pain in the replaced joint45/375 (12.0)14/126 (11.1)31/249 (12.4)
   Feel the surgery significantly improved their quality of life274/377 (72.7)97/126 (77.0)177/251 (70.5)
   Feel the surgery somewhat improved their quality of life99 (26.1)28/126 (22.2)71/251 (28.3)
 Perceptions of potential risks with TJA:   
  The potential risk of … is acceptable/very acceptable   
   1% risk of death317 (83.6)109 (85.8)208 (82.5)
   2% risk of myocardial infarction or stroke270 (71.2)95 (74.8)175 (69.4)
   Possibility of revision surgery in ≥10 years306 (80.7)116 (91.3)190 (75.4)
Perceptions of others   
 Has spoken with a physician about joint arthroplasty118 (31.1)52 (40.9)66 (26.2)§
 Knows someone with TJA352/380 (92.9)121 (95.3)231 (91.7)
 Knows someone with a successful TJA243/352 (69.0)90 (70.9)153 (60.7)
 Family/friends have suggested they discuss TJA with a physician94 (25.0)43 (33.9)51 (20.2)§
 Important sources of health information   
  Friends51 (13.5)25 (19.7)26 (10.3)§
  Physician (including primary care doctor)358/370 (96.8)120 (94.5)238 (94.4)
Personal and external resources   
 Preferred decision-making style   
  Do what physician recommends238 (62.8)74 (58.3)164 (65.1)
  Discuss pros and cons with physician and/or seek other opinions from family, friends, other sources of information before deciding114 (30.1)47 (37.0)67 (26.6)
  Ask for another medical opinion27 (7.1)6 (4.7)21 (8.3)
 Self-efficacy scale scores, mean ± SD ×/100   
  Pain score56.8 (19.0)54.3 (20.1)58.3 (18.3)
  Function score64.3 (21.5)63.6 (22.0)64.8 (21.4)
  Other score64.3 (20.4)63.0 (22.4)65.2 (19.4)

Only 31.1% of participants had ever discussed TJA with a physician. Those who had discussed it had worse perceived arthritis severity (mean score 6.5 of 10 versus 5.4 of 10; P < 0.001) and were more likely to have had family or friends suggest they do so (73.4% versus 16.7%; P < 0.001). Most participants (69.0%) knew someone who had successfully undergone TJA. Those who did were more optimistic than those who did not regarding usual TJA outcomes (e.g., usually individuals walk unaided, 80.3% versus 67.2%, P = 0.005; and have significantly improved quality of life, 77.0% versus 63.5%, P = 0.005) and were more likely to perceive the associated risks as acceptable (e.g., revision TJA 84.0% versus 74.5%; P = 0.025), but did not differ in their opinions regarding TJA indications. Although most participants knew someone who had undergone TJA, only 25.0% had been recommended by family or friends to discuss TJA with their physician.

Most participants (96.8%) felt their family physician was an important health information source; only 13.5% felt friends were also an important information source. Consistent with this, participants generally preferred to follow their physician's recommendations (62.8%) without consulting others when making a nonurgent health care decision.

Multivariate analyses: willingness to consider TJA.

Of the 379 participants, 127 (33.5%) were probably or definitely willing to consider TJA, 63 (16.6%) were unsure, and 189 (49.9%) were probably or definitely unwilling. Willingness to consider TJA was independently associated with participants' perceptions of the decision itself and with the perceptions of others. Specifically, younger age (adjusted odds ratio [OR] for <75 years versus ≥75 years 2.42, P = 0.01), greater perceived arthritis severity (adjusted OR per unit increase 1.30, P < 0.001), fewer comorbidities (adjusted OR per number of conditions 0.74, P = 0.001), perceiving TJA to be appropriate (versus inappropriate) for someone with at least moderate hip or knee pain controlled by pain killers (adjusted OR 3.29, P = 0.004) or walking limited to <1 city block (adjusted OR 1.99, P = 0.015), and that the potential risk of revision surgery is acceptable (versus unacceptable, adjusted OR 3.73, P < 0.001) were independent correlates of willingness. Participants who reported that friends were an important health information source (compared with those who did not, adjusted OR 2.49, P = 0.01) were also significantly more likely to be willing to consider TJA as a treatment option (Table 3). Adjusting for these factors, personal or external resources were not found to be significantly associated with willingness.

Table 3. Independent correlates of willingness to consider total joint arthroplasty*
Independent variableDependent variable = willing to consider TJA n = 350
Adjusted odds ratio95% confidence intervalP
  • *

    Hosmer-Lemeshow statistic = 3.04; P = 0.93; c statistic = 0.76.

Age <75 years (reference is ≥75 years)2.421.24–4.730.01
Perceived disease severity, per unit increase on 10-point scale1.301.15–1.48<0.001
Number of self-reported comorbidities, per unit increase0.740.61–0.890.001
Friends are the most important source of health information, yes2.491.23–5.040.01
Perceive the risk of revision TJA as acceptable, reference is unacceptable3.731.79–7.80<0.001
Perceive TJA is indicated for someone with at least moderate pain controlled by pain killers, reference is no3.291.48–7.310.004
Perceive that TJA is indicated for someone who is limited to walking <1 city block, reference is no1.991.15–3.460.015


In this population-based study, we examined perceptions of TJA, and how these and other factors effect willingness to consider TJA among individuals with disabling hip or knee arthritis who have not already made the decision to have TJA. Using the Ottawa framework for health care decision making, we specifically examined the effects of 3 groups of factors (individuals' perceptions around the decision itself, the perceptions of others, and individuals' personal and external resources). Our findings indicate that the first 2 groups of factors—specifically within those groups, perceived arthritis severity, comorbidity, indications for TJA, the acceptability of possible revision surgery, and perceived importance of friends' opinions—are driving individuals' willingness to consider TJA as a treatment option.

Although participants had reasonably accurate views of the outcomes following TJA, and were accepting of the potential risks with this surgery, the majority felt that arthritis pain and disability should be extreme before TJA should be considered. Furthermore, they felt TJA was appropriate only in the setting of otherwise good health. This is contrary to published guidelines for the management of hip or knee arthritis (35–38) and a U.S. National Institutes of Health consensus statement that says that TJA candidates should have “… radiographic evidence of joint damage and moderate to severe persistent pain or disability, or both…” (39). Individuals who held these misperceptions about when, and in whom, TJA should be performed were significantly more likely than those who did not to report unwillingness to consider TJA.

These misperceptions mirror the opinions of Ontario family physicians, who wait until their patients' symptoms and disability are extreme to refer for an opinion about TJA, and discount as potential candidates those who were “too old” or severely obese (40). Together, these findings are worrisome, particularly in light of data by Fortin and colleagues, which show that individuals referred for TJA late in the course of their disease do less well than those referred earlier (41). These misperceptions likely pose a major barrier to receipt of TJA, causing delays in decision making about this procedure and potentially compromising the results of surgery. Shared decision-making strategies will likely not address the problem because many individuals with disabling hip or knee arthritis will not see themselves as candidates for TJA and therefore not present to their physicians for consideration for TJA. An entirely different approach, targeting the general population and primary care physicians regarding when and in whom to consider TJA, will be needed.

The opinions of friends and family members also have a significant impact on participants' willingness to consider TJA. Those who had been recommended to discuss TJA with their physician by family or friends were more likely to have done so. Those who knew someone who had successfully undergone TJA were more accepting of the potential risk for revision surgery, positively influencing their willingness to consider TJA. Finally, those who felt friends were an important information source when making a nonurgent health decision were more than twice as likely as those who reported otherwise to be willing to consider TJA. Thus, to be successful, educational interventions must consider these key social networks.

After taking into consideration both the perceptions of the individual and of others around TJA, personal and external resource variables (including sex and socioeconomic status) were not associated with willingness to consider TJA as a treatment option. However, we have previously shown in our cohort that women and individuals with low socioeconomic status have worse arthritis symptoms and disability than do men and those with higher socioeconomic status (25, 26). This suggests that resource factors may indirectly influence individuals' willingness to consider having TJA through their impact on individuals' perceptions of the decision itself, and possibly also on social network factors, such as their exposure to individuals who have successfully undergone TJA.

There are several potential limitations to our study. First, we evaluated willingness to consider TJA using 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 of TJA. Second, the current estimates of factors affecting willingness to consider TJA are based on one region of Ontario. However, comparing the characteristics of our Oxford County phase I respondents with both Ontario and Canadian census data, these respondents appear representative of the underlying population. Finally, we did not consider employment status, cultural traditions, medical mistrust, preferences for different treatment approaches, social support, and the characteristics of the clinicians providing care, each of which may additionally impact willingness to consider TJA as a treatment option (42–47).

In conclusion, in a population-based cohort of individuals with disabling hip or knee arthritis, willingness to consider TJA as a treatment option was associated with individuals beliefs regarding the indications for, and acceptability of the outcomes following, TJA. Importantly, the majority of participants overestimated the degree of pain and disability at which TJA is indicated, relative to current guidelines for the management of hip and knee arthritis. This raises concerns that TJA is not being offered, or even considered by patients or their physicians at a time when it might result in greatest benefit. Interventions targeted to the general population and at primary care physicians, rather than only at individuals actively considering this surgery, are needed to address these misperceptions, and thereby reduce the population impact of disabling hip or knee arthritis.


Members of the Toronto Arthroplasty Health Services Research Consortium include Richard H. Glazier, MD, MPH, Bart Harvey, PhD, Annette L. Wilkins, BA, and Jack Williams, PhD.