Patient Education and Counseling xxx (xxxx) xxx Contents lists available at ScienceDirect Patient Education and Counseling journal homepage: www.journals.elsevier.com/patient-education-and-counseling Gender differences in the decision-making process for undergoing total knee replacement Ramon Sebastian Torrente-Jimenez a, Maria Feijoo-Cid b, c, *, Amado Javier Rivero-Santana d, e, Lilisbeth Perestelo-Pérez e, f, Alezandra Torres-Castaño d, e, Vanesa Ramos-García d, e, Amaia Bilbao e, g, h, Pedro Serrano-Aguilar e, f a Department of Medicine, Faculty of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain b Department of Nursing, Faculty of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain c Grup de Recerca Multidisciplinar en Salut i Societat (GREMSAS), (2017 SGR 917), Barcelona, Spain d Fundación Canaria Instituto de Investigación Sanitaria de Canarias (FIISC), Canary Islands, Spain e Health Services Research on Chronic Patients Network (REDISSEC), Spain f Evaluation Unit of the Canary Islands Health Service (SESCS), Canary Islands, Spain g Osakidetza Basque Health Service, Basurto University Hospital, Research and Innovation Unit, Bilbao, Spain h Kronikgune Institute for Health Services Research, Barakaldo, Spain A R T I C L E I N F O A B S T R A C T Keywords: Objective: To assess gender differences in the decision-making process for treatment of knee osteoarthritis (OA). Knee osteoarthritis Methods: A secondary analysis of a randomized trial was conducted (n = 193). Knowledge of OA and total knee Total knee replacement replacement (TKR), decisional conflict, satisfaction with the decision-making process, treatment preference and Patient-centered care TKR uptake 6 months later were compared by gender. Multivariate regression models were developed to identify Shared decision-making Health disparities gender-specific predictors. Results: Women showed less knowledge (MD = − 7.68, 95% CI: − 13.9, − 1.46, p = 0.016), reported less satis­ faction (MD = − 6.95, 95% CI: − 11.7, − 2.23, p = 0.004) and gave more importance to avoiding surgery (U = 2.09, p = 0.019). In women, more importance attributed to the time needed to relieve symptoms significantly reduced the odds of surgery (OR = 0.76, p = 0.016). Conclusion: The provision of information and/or promotion of shared decision-making could be of lower quality in female patients, although other explanations such as differences in information needs or preference for involvement in decision-making cannot be ruled out with the current evidence. Given the study’s limitations, especially regarding the sample size, further confirmation is needed. Practice implications: A systematic, shared decision-making approach in consultation is needed to avoid potential gender-based biases. 1. Introduction platelet-rich plasma remains unclear [3–5]. Total knee replacement (TKR) is the main option when pain and functional limitations persist Knee osteoarthritis (OA) is the leading cause of disability in devel­ despite the use of conservative treatments [6]. The prevalence of TKR is oped countries [1] and its prevalence has doubled since the mid-20th estimated to be around 1.5%, exceeding 10% at around 80 years of age century [2]. Currently, recommended treatments to manage the symp­ [7]. Some projections expect these procedures to double by 2030 and toms of knee OA include biomechanical interventions, exercise, strength even quadruple by 2040 [8]. training, weight management, painkillers and anti-inflammatory drugs, Many studies have investigated gender differences in the prevalence, whereas the efficacy of other treatment options such as glucosamine, incidence and severity of OA [9–11]. Prevalence, clinical pain and chondroitin, intra-articular corticosteroids, hyaluronic acid and inflammation is greater in women[12–14] and they use more health care * Corresponding author at: Department of Nursing, Faculty of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain. E-mail addresses: ramonsebastian.torrente@autonoma.cat (R.S. Torrente-Jimenez), maria.feijoo@uab.cat (M. Feijoo-Cid), amado.riverosantana@sescs.es (A.J. Rivero-Santana), lilisbeth.peresteloperez@sescs.es (L. Perestelo-Pérez), atorrcas@sescs.es (A. Torres-Castaño), vanesa.ramosgarcia@sescs.es (V. Ramos- García), amaia.bilbaogonzalez@osakidetza.eus (A. Bilbao), pseragu@gobiernodecanarias.org (P. Serrano-Aguilar). https://doi.org/10.1016/j.pec.2022.08.014 Received 19 April 2022; Received in revised form 18 August 2022; Accepted 21 August 2022 Available online 28 August 2022 0738-3991/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Please cite this article as: Ramon Sebastian Torrente-Jimenez, Patient Education and Counseling, https://doi.org/10.1016/j.pec.2022.08.014 R.S. Torrente-Jimenez et al. Patient Education and Counseling xxx (xxxx) xxx before undergoing TKR [15]. Although most studies have not found Data were managed and analyzed using SPSS version 25. Descriptive significant gender differences in willingness to undergo TKR [16–18], it statistics (means, standard deviations, and percentages) were calculated has been proven to be used significantly less frequently in women, who and compared by gender with t-tests and chi-square tests for continuous show worse post-operative outcomes [19–22]. Differences in the use of and categorical variables, respectively. Knowledge, decisional conflict, TKR may be explained by personal factors such as different beliefs, ex­ and satisfaction with the decision-making process were analyzed both pectations, and treatment preferences, but also by contextual factors, continuously and categorically. For the latter analysis, the variables such as interactions with health professionals [20]. Nonetheless, studies were dichotomized based on a score of 60 for knowledge (previous assessing gender differences in the patient-physician interaction when studies have used this value [31] or 66 [32] to define an adequately making decisions about OA treatment are scarce. Results have shown informed patient) and 37.5 for decisional conflict (higher values are that women discuss TKR with a physician to a lesser extent than men associated with delaying decisions [29]). For satisfaction, we also used a [16,23]. One study in Canada with standardized patients showed that score of 60 as a threshold. physicians applied fewer informed decision-making elements and Differences in the motives for choosing treatment were analyzed by showed poorer interpersonal skills when the patient was female [23,24]. means of Mann-Whitney’s U test, since most items showed highly Studies evaluating whether doctors are less likely to recommend surgery asymmetric distributions. A principal component analysis (PCA) was to women have yielded mixed results [18,23,25]. carried out on the 6 items assessing the importance of treatment out­ The aim of this study is to investigate whether there are gender comes/characteristics (i.e., serious and mild adverse effects, pain relief, differences in variables related to the decision-making process for knee being able to perform daily activities, recovery time, need for further OA treatment in patients facing the decision to undergo TKR: decisional treatment) in order to reduce the number of variables introduced in the conflict, satisfaction with the decision-making process, preference for subsequent regression models. The remaining three items (i.e., avoiding surgery and uptake of TKR. surgery, injections, and medication, respectively) were introduced individually. 2. Methods Multiple regression models were constructed for the whole sample and separately by gender to analyze the predictors of decisional conflict, We performed a secondary analysis of the data obtained from a satisfaction with the decision-making process (linear regression for randomized controlled trial conducted at one hospital and nine primary both), preference for TKR (after dichotomizing the variable into TKR vs. care centers in Tenerife (Spain) in 2018 [26]. The trial aimed to assess other) and undergoing TKR in the following 6 months (logistic regres­ the effectiveness of a decision aid for patients with knee OA as compared sion for both). The motive of “avoiding surgery” was only included in the to usual care. Patients were candidates for TKR facing the decision of models for decisional conflict and satisfaction, and these two latter joining the waiting list for surgery or delaying that decision and variables were included as predictors in the models of preference for continuing with conservative treatment. The results showed TKR and surgery uptake. To assess gender-specific predictors, the post-intervention improvements in knowledge of OA and TKR, deci­ interaction term of sex with each of the independent variables was sional conflict and satisfaction with the decision-making process. These introduced successively in the multiple regression models for the whole improvements were not significantly different by gender, although the sample; a significant value for the interaction indicates that beta co­ analyses were post-hoc. efficients for women and men differ. The trial did not include a baseline assessment (i.e., the intervention The study was approved by the Ethical Committee of the Hospital and control groups completed the questionnaires only once following Universitario Nuestra Señora de la Candelaria, Tenerife (Spain) (ref.: PI- application of the decision aid in the former). Therefore, these post- 24/16). It was carried out according to the principles of the Declaration intervention data were used in this study, along with the rate of TKR of Helsinki, and local and national regulations on data protection and observed 6 months after the decision aid intervention. anonymity. Study variables included: 3. Results • Sociodemographic variables: age, gender, education (dichotomized into no studies/primary vs. secondary/university). The study included 193 participants, whose characteristics are • Health-related quality of life, measured with the Spanish version of shown in Table 1. Mean age was 66.8 (SD = 8.42), most participants the EuroQol questionnaire (EQ-5D-5 L) [27]. Derived health utilities were women (72%) and 77.2% had received only primary education or range from 0 (worst possible health) to 1 (perfect health). no formal studies. No significant differences by gender were found for • Self-perceived general health, assessed with the visual analogue scale age or education. The sample showed a low to moderate level of deci­ (VAS) of the EQ-5D-5 L (range 0–100). sional conflict (36.8), values around 50 (on a 0–100 scale) in knowledge • Knowledge of OA and TKR, assessed with a 7-item questionnaire in of OA/TKR, and satisfaction with the decision-making process. Women which the number of correct answers represents the total score, scored significantly lower on these two variables (knowledge: MD = transformed to a 0–100 scale. − 7.68, 95% CI: − 13.9, − 1.46, p = 0.016; satisfaction: MD = − 6.95, 95% • Importance given to certain characteristics and potential outcomes CI: − 11.7, − 2.23, p = 0.004). Categorical analyses also yielded signif­ of OA treatments: avoiding severe adverse effects, avoiding mild icant results: the rate of adequately informed women (knowledge score adverse effects, reducing pain, being able to perform daily activities, ≥ 60) was lower than that of men (38.8% vs. 59.3%, p = 0.010), as well avoiding the need for additional treatment, time needed to experi­ as the rate of those scoring ≥ 60 in satisfaction (23.7% vs. 51.9%, p < ence symptom relief, avoiding surgery, avoiding injections, and 0.001). avoiding drugs (pills). Each item is graded by the patient from 0 (not Approximately half of the sample (48.7%) preferred TKR as a important at all) to 10 (extremely important). treatment option, whereas 18.8% were unsure about their preference. At • Satisfaction with the decision-making process, assessed with the 12- 6 months, 26.3% of participants had undergone TKR. There were no item scale developed by Barry et al. [28]. Scores are transformed to a significant gender differences in these variables. 0–100 scale, with higher values indicating greater satisfaction. Table 2 shows the relevance of the different motives when deciding • Decisional conflict, measured with the Decisional Conflict Scale [29, about treatment options. The most important were being able to perform 30], which includes 16 items with a Likert scale ranging from daily activities (mean 9.64), reducing pain (9.52) and avoiding severe 0 (totally agree) to 5 (totally disagree), transformed to a 0–100 scale. adverse effects (8.99). Only one significant difference was observed: • Treatment preference (unsure, drugs, injections or TKR). women gave more importance to avoiding surgery than men (U = 2.09, • Undergoing TKR in the following 6 months (yes/no). p = 0.019). The PCA of the 6 items dealing with treatment 2 R.S. Torrente-Jimenez et al. Patient Education and Counseling xxx (xxxx) xxx Table 1 Table 3 Characteristics of the sample. Principal component analysis of items assessing the relevance attributed to Total Women Men p- treatment outcomes/characteristics. (n ¼ 193) (n ¼ 139) (n ¼ 54) value Components h2 Age (years), mean (SD) 66.8 (8.42) 67.3 65.4 0.069 1 2 3 (8.22) (8.87) Education, n (%) 0.973 Reducing pain 0.905 -0.034 0.098 0.83 No studies 49 (25.4) 35 (25.1) 14 (25.9) Being able to perform daily activities 0.873 0.248 -0.007 0.82 Primary studies 100 (51.8) 73 (52.5) 27 (50) Avoiding mild adverse effects -0.051 0.880 0.116 0.79 Secondary studies 32 (16.6) 23 (16.5) 9 (16.7) Avoiding serious adverse effect 0.274 0.835 0.044 0.78 University studies 12 (6.2) 8 (5.9) 4 (7.4) Recovery time -0.059 0.021 0.849 0.73 EQ-5D-5 L, mean (SD) 0.62 (0.23) 0.60 0.65 0.228 Needing additional treatment 0.150 0.129 0.782 0.65 (0.24) (0.20) Kaiser-Meyer-Olkin = 0.543; Bartlett sphericity test: χ 2 = 236.9, p < 0.001. EQ-5D-5 L VAS, mean (SD) 66.6 (18.1) 65.6 69.2 0.154 (19.3) (14.4) Knowledge, mean (SD) 56.1 (19.9) 54.0 61.6 0.016 These components were included in the regression models. (20.4) (17.6) Tables 4–7 show the multivariate models for each outcome. For the SAT, mean (SD) 53.24 51.3 58.2 0.004 whole sample, significant predictors of decisional conflict (Table 4) were (15.2) (14.7) (15.3) DCS, mean (SD) 36.8 (12.9) 37.1 35.9 0.406 the importance attributed to symptoms and function improvement (B = (13.0) (12.6) 1.26, p = 0.013) and lesser knowledge (B = − 0.13, p = 0.006). There Treatment of preference, n 0.603 were no gender-specific predictors. (%) Regarding satisfaction with the decision-making process, significant Unsure 36 (18.7) 28 (20.1) 8 (14.8) predictors were gender (confirming the result of the previous univariate Drugs 30 (15.5) 21 (15.1) 9 (16.7) Injections 32 (16.6) 25 (18.0) 7 (13.0) analysis), education (B = 6.98, p = 0.004) and a lower level of decisional TKR 95 (49.2) 65 (46.8) 30 (55.5) conflict (B = − 0.51, p < 0.001) (Table 5). The importance attributed to a Operated at 6 months (n = 50 (26.3%) 38 12 0.419 rapid and durable recovery showed a significant interaction with gender 190) (27.9%) (22.2%) (t = 2.61, p = 0.010): its association with satisfaction was negative and SD = Standard deviation. non-significant for women (B = − 0.34, p = 0.344) and positive for men (B = 1.46, p = 0.022). Preference for TKR was significantly predicted by poorer quality of Table 2 life (OR = 0.05, p = 0.001), greater importance attributed to symptoms Relevance of motives for decision. and function improvement (OR = 1.57, p < 0.001), desire to avoid in­ Total Women Men p- jections (OR = 1.22, p = 0.004), lower relevance of a rapid and durable (n ¼ (n ¼ 139) (n ¼ 54) value recovery (OR = 0.78, p < 0.001), greater knowledge (OR = 1.05, p < 193) 0.001) and lower decisional conflict (OR = 0.95, p = 0.006). The Avoiding severe adverse effects, 8.99 9.14 8.63 0.193 importance attributed to avoiding adverse effects showed a significant mean (SD) median (1.95) (1.75) (2.37) interaction with gender (Wald = 6.48, p = 0.011): it was non- 10.0 10.0 10.0 Avoiding mild adverse effects, 7.59 7.71 7.28 0.385 significantly related to a lower preference for TKR in women (OR = mean (SD) median (2.03) (1.87) (2.37) 0.93, p = 0.312), whereas the opposite occurred in men (OR = 2.65, p = 8.00 8.00 8.00 0.051). Being able to perform daily 9.64 9.66 9.57 0.224 Finally, uptake of TKR 6 months after assessment was significantly activities, mean (SD) median (0.99) (0.98) (0.98) predicted by poorer quality of life (OR = 0.11, p = 0.011), lower rele­ 10.0 10.0 10.0 Avoiding the need for recurring 8.03 8.16 7.70 0.192 vance of a rapid and durable recovery (OR = 0.90, p = 0.032), lower or additional treatment, mean (2.31) (2.20) (2.56) decisional conflict (OR = 0.95, p = 0.006) and lower satisfaction (OR = (SD) median 9.00 9.00 9.00 0.95, p = 0.002). When preference for TKR was included in the model, it Reducing pain, mean (SD) 9.52 9.55 9.43 0.106 had a strong significant effect (OR = 8.55, p < 0.001), and quality of life median (1.14) (1.16) (1.09) 10.0 10.0 10.0 and relevance of a rapid and durable recovery were no longer signifi­ Time needed to experience 7.51 7.47 7.63 0.592 cant, indicating a mediating effect of preference for surgery (data not symptom relief, mean (SD) (2.16) (2.17) (2.13) shown). median 8.00 8.00 8.00 None of the interaction terms with sex was significant, although two Avoiding surgery, mean (SD) 5.51 5.76 4.54 0.019 variables obtained p-values lower than 0.10: quality of life (p = 0.082), median (3.66) (3.65) (3.53) 5.00 7.00 4.00 with a stronger effect in men (OR = 0.00, p = 0.022) than in women (OR Avoiding injections, mean (SD) 6.34 6.42 6.13 0.491 = 0.22, p = 0.099), and the importance of a rapid and durable recovery median (3.15) (3.08) (3.31) (p = 0.075), which was significantly related to a lower uptake of TKR in 7.00 7.00 7.00 women (OR = 0.84, p = 0.009 vs. OR = 1.09, p = 0.592 for men). When Avoiding drugs (pills), mean 6.36 6.55 5.87 0.097 (SD) median (2.52) (2.45) (2.63) the two items that form this latter variable (i.e., time to relief and need 7.00 7.00 6.00 for further treatment) were introduced separately, the interaction with sex was significant only for the importance of the time needed to SD = Standard deviation. experience relief (p = 0.022); it was related to a lower likelihood of TKR in women (OR = 0.76, p = 0.016), whereas a non-significant opposite characteristics and outcomes yielded 3 components with eigenvalues result was observed in men (OR = 1.78, p = 0.128). greater than 1, which explained 76.6% of the variance (Table 3). The Fig. 1 shows a theoretical model based on the results obtained first component included the importance of reducing pain and being (excluding preference for surgery so as not to increase the complexity of able to perform daily activities (thus labeled "Symptoms and function the model). improvement"). The second (labeled "Avoiding adverse effects") included avoiding both severe and mild adverse effects. The third component included the time needed to experience relief and avoiding the need for further treatment (labeled "Rapid and durable recovery"). 3 R.S. Torrente-Jimenez et al. Patient Education and Counseling xxx (xxxx) xxx Table 4 Multiple linear regression analysis for decisional conflict. Total (n ¼ 193) Women (n ¼ 139) Men (n ¼ 54) Differencea b B p B p B p t p Gender 0.36 0.865 Age -0.02 0.875 0.02 0.909 -0.30 0.198 -0.62 0.539 Education 0.07 0.977 -0.27 0.923 -3.26 0.468 0.38 0.707 Self-perceived general health (EQ-5D-5 L VAS) -0.08 0.156 -0.05 0.429 -0.28 0.058 -0.98 0.328 Health-related quality of life (EQ-5D-5 L) 1.91 0.660 -1.24 0.800 14.33 0.152 0.94 0.351 Relevance: Avoiding adverse effects 0.15 0.606 0.06 0.867 0.71 0.183 0.12 0.903 Relevance: Symptoms and function improvement 1.26 0.013 1.40 0.017 -0.16 0.888 -0.87 0.383 Relevance: Rapid and durable recovery -0.53 0.052 -0.30 0.383 -1.58 0.003 -1.23 0.219 Relevance: Avoiding injections -0.07 0.814 -0.06 0.868 -0.26 0.634 0.38 0.703 Relevance: Avoiding drugs 0.23 0.553 0.34 0.470 -0.47 0.525 -0.47 0.641 Knowledge -0.13 0.006 -0.16 0.007 -0.02 0.819 1.12 0.263 EQ-5D-5 L: EuroQol questionnaire (5 dimensions, 5 levels); VAS: Visual analogue scale. a t and p-value of the interaction between the corresponding variable and sex, introduced in the multiple regression model for the whole sample. Significant values indicate that beta coefficients for women and men differ. b Unstandardized coefficients. Table 5 Multiple linear regression analysis for Satisfaction with the decisional process. Total (n ¼ 193) Women (n ¼ 139) Men (n ¼ 54) Differencea Bb p B p B p t p Gender 6.23 0.006 Age 0.02 0.869 0.18 0.252 -0.09 0.732 -1.64 0.103 Education 6.98 0.004 7.10 0.013 10.23 0.043 0.55 0.586 Self-perceived general health (EQ-5D-5 L VAS) 0.03 0.662 0.01 0.829 0.12 0.473 -0.45 0.650 Health-related quality of life (EQ-5D-5 L) 1.89 0.679 2.98 0.551 -1.62 0.884 0.41 0.680 Relevance: Avoiding adverse effects 0.11 0.710 0.34 0.355 -0.57 0.342 -0.66 0.509 Relevance: Symptoms and function improvement -0.36 0.508 0.26 0.668 -1.07 0.399 -1.48 0.140 Relevance: Rapid and durable recovery 0.13 0.663 -0.34 0.344 1.46 0.022 2.61 0.010 Relevance: Avoiding injections -0.13 0.686 -0.16 0.687 -0.01 0.986 -0.21 0.833 Relevance: Avoiding drugs 0.19 0.632 0.02 0.967 0.98 0.236 1.03 0.306 Knowledge -0.00 0.987 0.04 0.513 -0.14 0.213 -1.38 0.169 Decisional conflict -0.51 < 0.001 -0.53 < 0.001 -0.33 0.053 0.27 0.786 EQ-5D-5 L: EuroQol questionnaire (5 dimensions, 5 levels); VAS: Visual analogue scale. a t and p-value of the interaction between the corresponding variable and sex, introduced in the multiple regression model for the whole sample. Significant values indicate that beta coefficients for women and men differ. b Unstandardized coefficients. Table 6 Multiple logistic regression model for preference for surgery. Total (n ¼ 193) Women (n ¼ 139) Men (n ¼ 54) Differencea OR p OR p OR p Wald p Gender 0.67 0.386 Age 1.03 0.349 1.01 0.723 1.23 0.080 0.60 0.440 Education 0.76 0.580 0.71 0.577 0.12 0.229 0.26 0.614 Self-perceived general health (EQ-5D-5 L VAS) 1.01 0.415 1.01 0.623 1.12 0.094 1.41 0.235 Health-related quality of life (EQ-5D-5 L) 0.05 0.001 0.06 0.008 0.00 0.075 0.29 0.591 Relevance: Avoiding adverse effects 1.03 0.671 0.93 0.312 2.65 0.051 6.48 0.011 Relevance: Symptoms and function improvement 1.57 < 0.001 1.53 0.003 1.75 0.389 0.26 0.614 Relevance: Rapid and durable recovery 0.78 < 0.001 0.78 0.004 0.45 0.050 0.07 0.786 Relevance: Avoiding injections 1.22 0.004 1.21 0.026 1.31 0.159 0.05 0.829 Relevance: Avoiding drugs 1.11 0.225 1.07 0.542 1.27 0.431 0.74 0.390 Knowledge 1.05 < 0.001 1.06 < 0.001 1.08 0.105 1.05 0.305 Decisional conflict 0.95 0.006 0.96 0.079 0.88 0.065 1.06 0.303 Satisfaction with the decisional process 1.00 0.964 1.00 0.807 1.09 0.173 0.83 0.363 EQ-5D-5 L: EuroQol questionnaire (5 dimensions, 5 levels); OR: Odds ratio; VAS: Visual analogue scale. a Wald statistic and p-value of the interaction between the corresponding variable and sex, introduced in the multiple regression model for the whole sample. Significant values indicate that beta coefficients (log OR) for women and men differ. 4. Discussion and conclusion epidemiology of knee OA and the outcomes of TKR, showing worse re­ sults for women, the research on gender differences in the decision- 4.1. Discussion making process about OA treatment is scarce. Our study was not designed for that aim, and therefore it has an exploratory nature and low Although many studies have evaluated gender differences in the internal validity. However, we think that the results are relevant enough 4 R.S. Torrente-Jimenez et al. Patient Education and Counseling xxx (xxxx) xxx Table 7 Multiple logistic regression model for TKR uptake at 6 months. Total (n ¼ 190) Women (n ¼ 136) Men (n ¼ 54) Differencea OR p OR p OR p Wald p Gender 1.27 0.601 — — — — — — Age 0.96 0.170 0.96 0.239 0.98 0.784 0.57 0.451 Education 0.99 0.988 1.01 0.991 0.19 0.247 0.03 0.870 Self-perceived general health (EQ-5D-5 L VAS) 1.02 0.055 1.02 0.095 1.06 0.156 0.06 0.801 Health-related quality of life (EQ-5D-5 L) 0.11 0.011 0.22 0.099 0.00 0.022 3.03 0.082 Relevance: Avoiding adverse effects 1.20 0.131 0.89 0.083 0.89 0.445 2.08 0.150 Relevance: Symptoms and function improvement 0.92 0.154 1.20 0.162 1.33 0.527 0.33 0.568 Relevance: Rapid and durable recovery 0.90 0.032 0.84 0.009 1.09 0.592 3.18 0.075 Relevance: Avoiding injections 1.03 0.633 1.01 0.888 1.10 0.548 0.27 0.601 Relevance: Avoiding drugs 1.07 0.371 1.06 0.534 1.42 0.094 0.72 0.395 Knowledge 1.01 0.507 1.01 0.655 1.02 0.524 0.27 0.602 Decisional conflict 0.95 0.006 0.96 0.024 0.94 0.233 0.48 0.488 Satisfaction with the decisional process 0.95 0.002 0.95 0.007 0.93 0.083 0.28 0.595 EQ-5D-5 L: EuroQol questionnaire (5 dimensions, 5 levels); OR: Odds ratio; VAS: Visual analogue scale. a Wald statistic and p-value of the interaction between the corresponding variable and sex, introduced in the multiple regression model for the whole sample. Significant values indicate that beta coefficients (log OR) for women and men differ. rheumatoid arthritis, the results available have shown a greater desire for information and decision involvement in women [37]. Nor does the research support that women engage less than men in seeking out health information from sources other than their health provider, such as the Internet or friends; in fact, it is quite the contrary [38–40]. Therefore, the observed gap could be explained, at least to some extent, by un­ conscious gender biases in the provision of information and/or the promotion of shared decision-making by health professionals. In a study with standardized OA patients (thus controlling for individual prefer­ ences), Borkhoff et al. (2013) observed that physicians implemented shared decision-making elements to a lesser extent and displayed worse communication skills when the patient was a woman [24]. Experimental studies have shown that, compared to men, women’s pain is under­ estimated by health care providers [41] and lay people [42] and that this Fig. 1. Theoretical model based on the results obtained (preference for surgery underestimation is related to gender stereotypes about pain expression not included). (-): negative association; (+): positive association; (A): possible [42]. Participants also judged women as relatively more likely to benefit interaction with gender (positive for men, negative non-significant for women); from psychotherapy instead of medication [41,42]. In other health (B): possible interaction with gender (negative for women, positive non- conditions, it has been observed that women need to report higher pain significant for men): (C) possible interaction with gender (negative for both, stronger in men, non-significant in women). levels than male patients to make healthcare professionals document their symptom [43], and that physicians tend to inquire about them less frequently [44]. In our study, this interpretation is supported by the to merit confirmation in a representative sample of the appropriate size. significantly worse result for women in satisfaction with the decision Our results indicate that women have lower objective knowledge of process; however, this variable was not significantly related to objective TKR than men. The difference is small from a continuous perspective knowledge. Another explanation could be that the provision of infor­ (7.7 points in a 0–100 scale), but when the variable is dichotomized, the mation is adequate but there are gender differences in other aspects of rate of adequately informed women (score ≥ 60) is 35% lower than that physicians’ communication, such as affective ones [45,46], that could of men. A patient’s good, objective knowledge of the disease and negatively affect women’s information recall. Future studies should treatments is a fundamental requirement for any approach to medical include the appropriate variables to answer these questions (e.g., in­ decision-making that aims to be informed or shared. Although a formation needs and sources, verbal and non-verbal interaction between knowledge-based decision does not necessarily have to be “better” or patient and doctor, trust in the doctor). wiser than a decision just based on the trust in the professionals’ The analysis of gender-specific predictors of the dependent variables expertize, it can only be autonomous and concordant with the patient’s yielded few significant results, although these analyses have low sta­ preferences if they have adequate knowledge of treatment characteris­ tistical power and definitive conclusions cannot be drawn. The impor­ tics and potential outcomes. tance attributed to a rapid and durable recovery was related to more The difference observed in our data is not explained by differences in satisfaction with the decision-making process in men but not in women. education or disease severity. Assuming the non-spurious nature of this The same occurred with the association between the importance result, one possible explanation is that women are less interested in attributed to avoiding adverse effects and the preference for TKR; in being informed or actively participating in the decision-making process. men, the more importance given to avoiding adverse effects, the more In OA, research on information needs and participation preferences has likely they are to prefer surgery. This could be related to different beliefs been mostly qualitative and has not investigated gender differences or expectations about adverse effects of conservative treatment and [33]. Two recent studies, one with rehabilitation patients (two-thirds complications of TKR [23,47]. orthopedic, although the specific disease is not reported) and the other In women, greater importance attributed to the time needed to with non-OA orthopedic surgery patients, did not find significant dif­ experience relief significantly reduced the odds of surgery, whereas in ferences in participation preferences [34,35]. Nor was gender a signif­ men the odds were 78% higher (although non-significant, probably due icant predictor of elderly patients’ preference for involvement in the to the small sample size). This suggests that women may be concerned decision about taking analgesics for acute musculoskeletal pain [36]. In about interrupting their daily activities due to the recovery time needed 5 R.S. Torrente-Jimenez et al. Patient Education and Counseling xxx (xxxx) xxx after TKR, independently from their functional status. This is in line with Role of the funding source previous studies showing that women, compared with men, are more concerned about the interference that surgery and the recovery period This project was supported by Instituto de Salud Carlos III, Ministry could have on their caregiving roles [47–50]. In Spain, although gender of Health, Spain (grant number PI15/01264). disparities have been reduced over the past decades, women still spend more time than men performing unpaid work (i.e., housework, care of Data statement children, elderly and people with disabilities), and this difference is more pronounced in an older population such as knee OA patients [51]. The data that support the findings of this study are available from the This population still has a low education level on average, and most corresponding author upon reasonable request. women have always been exclusively responsible for housework, whereas many men are retired and have less time restrictions. Lastly, poorer quality of life was a stronger predictor of TKR uptake after 6 Conflict of interest months in men than in women. Although the interaction did not reach statistical significance (p = 0.087), the result is in line with the hy­ The authors disclose no conflicts of interest. pothesis that in women the decision to undergo surgery depends less exclusively on health status. Future studies should also include Acknowledgments OA-specific quality of life measures. We thank the clinical professionals who recruited patients and all the 4.1.1. Limitations patients who generously participated. We also thank the researchers of This study is a secondary and exploratory analysis and therefore it the GREMSAS group for their contributions to improving the discussion. has many limitations. The main one is the limited sample size, especially in the case of men. This results in low statistical power, principally for References the interaction analyses in the regression models. Along with the fact that most patients were treated by male professionals, the small sample [1] Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M, et al. Years lived size also precludes the analysis of the interaction between the gender of with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380: patients and professionals. The sample included mostly individuals with 2163–96. https://doi.org/10.1016/S0140-6736(12)61729-2. a low education level; although it seems representative of the OA pop­ [2] Wallace IJ, Worthington S, Felson DT, Jurmain RD, Wren KT, Maijanen H, ulation (i.e., middle aged and elderly) in a region with a low socioeco­ Woods RJ, Lieberman DE. Knee osteoarthritis has doubled in prevalence since the nomic level, the results should not be generalized to more educated mid-20th century. Proc Natl Acad Sci 2017;114:9332–6. https://doi.org/10.1073/ pnas.1703856114. populations. The knowledge questionnaire is short, and future studies [3] Hauk L. Treatment of knee osteoarthritis: a clinical practice guideline from the should assess more specific aspects of both conservative and surgical OA AAOS. Am Fam Physician 2014;89:918–20. https://doi.org/http://www.ncbi.nlm. treatment options. The follow up was not long enough to observe high nih.gov/pubmed/25077402. [4] Hochberg MC, Altman RD, April KT, Benkhalti M, Guyatt G, McGowan J, TKR rates since waiting lists are long in our region, and this could affect Towheed T, Welch V, Wells G, Tugwell P. American College of Rheumatology 2012 to some extent the analysis of TKR uptake predictors. We did not assess recommendations for the use of nonpharmacologic and pharmacologic therapies in relevant variables that could influence the decision-making process, osteoarthritis of the hand, hip, and knee. Arthritis Care Res (Hoboken) 2012;64: 465–74. https://doi.org/10.1002/acr.21596. such as information needs, trust in the physician or patients’ affective [5] McAlindon TE, Bannuru RR, Sullivan MC, Arden NK, Berenbaum F, Bierma- variables (e.g., anxiety, depression, stress). Zeinstra SM, Hawker GA, Henrotin Y, Hunter DJ, Kawaguchi H, Kwoh K, Lohmander S, Rannou F, Roos EM, Underwood M. OARSI guidelines for the non- surgical management of knee osteoarthritis. Osteoarthr Cartil 2014;22:363–88. 4.2. Conclusion https://doi.org/10.1016/j.joca.2014.01.003. [6] Frankel L, Sanmartin C, Conner-Spady B, Marshall DA, Freeman-Collins L, Wall A, Hawker GA. Osteoarthritis patients’ perceptions of “appropriateness” for total joint The conclusions are conditioned by the exploratory nature of the replacement surgery. Osteoarthr Cartil 2012;20:967–73. https://doi.org/10.1016/ study and its limited internal validity. Women show a lower level of j.joca.2012.05.008. [7] Maradit Kremers H, Larson DR, Crowson CS, Kremers WK, Washington RE, knowledge of OA/TKR and satisfaction with the decision-making pro­ Steiner CA, Jiranek WA, Berry DJ. Prevalence of total hip and knee replacement in cess, and their concerns about the recovery time needed after surgery the United States. J Bone Jt Surg-Am 2015;Vol. 97:1386–97. https://doi.org/ could be a barrier to undergoing TKR. Although explanations based on 10.2106/JBJS.N.01141. potential gender differences in information needs and preferences for [8] Singh JA, Yu S, Chen L, Cleveland JD. Rates of total joint replacement in the united states: future projections to 2020–2040 using the national inpatient sample. involvement cannot currently be ruled out, the results may indicate a J Rheumatol 2019;46:1134–40. https://doi.org/10.3899/jrheum.170990. gender bias in the provision of information and/or promotion of shared [9] Boyan BD, Tosi L, Coutts R, Enoka R, Hart DA, Nicolella DP, Berkley K, Sluka K, decision-making by health professionals. Studies with adequate statis­ Kwoh K, O^Connor MI, Kohrt W. Sex differences in osteoarthritis of the knee. J Am Acad Orthop Surg 2012;20:668–9. https://doi.org/10.5435/JAAOS-20-10-668. tical power and more refined measurement instruments are required to [10] Srikanth VK, Fryer JL, Zhai G, Winzenberg TM, Hosmer D, Jones G. A meta- confirm whether this bias actually exists in this population. analysis of sex differences prevalence, incidence and severity of osteoarthritis. Osteoarthr Cartil 2005;13:769–81. https://doi.org/10.1016/J.JOCA.2005.04.014. [11] Stevens-Lapsley JE, Kohrt WM. Osteoarthritis in women: effects of estrogen, 4.3. Practice implications obesity and physical activity. Women’s Health 2010;6:601–15. https://doi.org/ 10.2217/WHE.10.38. [12] Li D, Li S, Chen Q, Xie X. The Prevalence of Symptomatic Knee Osteoarthritis in It is important that professionals systematically apply a shared Relation to Age, Sex, Area, Region, and Body Mass Index in China: A Systematic decision-making approach adapted to the patients’ need for information Review and Meta-Analysis. Front Med 2020;7. https://doi.org/10.3389/ fmed.2020.00304. and their desired level of involvement. Patients’ objective knowledge [13] Migliore A, Picarelli G. Is osteoarthritis a gender-specific disease? Ital J Gend- must be adequately checked to avoid misunderstandings and false be­ Specif Med 2018;4:13–20. https://doi.org/10.1723/2968.29765. liefs about medical procedures. The standardized use of tools like patient [14] Elboim-Gabyzon M, Rozen N, Laufer Y. Gender differences in pain perception and functional ability in subjects with knee osteoarthritis. ISRN Orthop 2012;2012:1–4. decision aids could help to reduce the variability of clinical practice and https://doi.org/10.5402/2012/413105. improve the quality of patient-physician interaction and the provision of [15] Bawa HS, Weick JW, Dirschl DR. Gender disparities in osteoarthritis-related health information, thus avoiding potential biases based on gender or other care utilization before total knee arthroplasty. 2115-2118.e1 J Arthroplast 2016; patient characteristics. In the case of older women facing TKR, the 31. https://doi.org/10.1016/j.arth.2016.03.044. [16] Postler A, Goronzy J, Günther KP, Lange T, Redeker I, Schmitt J, Zink A, Callhoff J. involvement of their families in the decision-making process could help Which disease-related factors influence patients’ and physicians’ willingness to them to plan their daily activities after surgery. consider joint replacement in hip and knee OA? Results of a questionnaire survey 6 R.S. Torrente-Jimenez et al. Patient Education and Counseling xxx (xxxx) xxx linked to claims data. BMC Musculoskelet Disord 2020;21:1–11. https://doi.org/ [34] Lindsay SE, Alokozai A, Eppler SL, Fox P, Curtin C, Gardner M, Avedian R, 10.1186/S12891-020-03368-1/TABLES/6. Palanca A, Abrams GD, Cheng I, Kamal RN. VOICES health policy research [17] Vina ER, Cloonan YK, Ibrahim SA, Hannon MJ, Boudreau RM, Kwoh CK. Race, sex, investigators:, patient preferences for shared decision making: not all decisions and total knee replacement consideration: role of social support. Arthritis Care Res should be shared. J Am Acad Orthop Surg 2020;28:419–26. https://doi.org/ (Hoboken) 2013;65:1103–11. https://doi.org/10.1002/ACR.21925. 10.5435/JAAOS-D-19-00146. [18] Mota REM, Tarricone R, Ciani O, Bridges JF, Drummond M. Determinants of [35] Baumann LA, Brütt AL. Participation preferences of health service users in health demand for total hip and knee arthroplasty: a systematic literature review. BMC care decision-making regarding rehabilitative care in Germany—A cross-sectional Health Serv Res 2012;12:225. https://doi.org/10.1186/1472-6963-12-225. study. Heal Expect 2022;25:125–37. https://doi.org/10.1111/hex.13356. [19] Novicoff WM, Saleh KJ. Examining sex and gender disparities in total joint [36] Holland WC, Hunold KM, Mangipudi SA, Rittenberg AM, Yosipovitch N, Platts- arthroplasty. Clin Orthop Relat Res 2011;469:1824–8. https://doi.org/10.1007/ Mills TF. A prospective evaluation of shared decision-making regarding analgesics s11999-010-1765-y. selection for older emergency department patients with acute musculoskeletal [20] Katz JN, Wright EA, Guadagnoli E, Liang MH, Karlson EW, Cleary PD. Differences pain. Acad Emerg Med 2016;23:306–14. https://doi.org/10.1111/acem.12888. between men and women undergoing major orthopedic surgery for degenerative [37] Mattukat K, Boehm P, Raberger K, Schaefer C, Keyszer G, Mau W. How much arthritis. Arthritis Rheum 1994;37:687–94. https://doi.org/10.1002/ information and participation do patients with inflammatory rheumatic diseases art.1780370512. prefer in interaction with physicians? Results of a participatory research project. [21] Mehta SP, Perruccio AV, Palaganas M, Davis AM. Do women have poorer outcomes Patient Prefer Adherence 2019;Volume 13:2145–58. https://doi.org/10.2147/ following total knee replacement? Osteoarthr Cartil 2015;23:1476–82. https://doi. PPA.S209346. org/10.1016/j.joca.2015.05.007. [38] Chaudhuri S, Le T, White C, Thompson H, Demiris G. Examining health [22] Sveikata T, Porvaneckas N, Kanopa P, Molyte A, Klimas D, Uvarovas V, Venalis A. information–seeking behaviors of older adults. Comput Inform Nurs 2013;31:547. Age, sex, body mass index, education, and social support influence functional https://doi.org/10.1097/01.NCN.0000432131.92020.42. results after total knee arthroplasty. Geriatr Orthop Surg Rehabil 2017;8:71–7. [39] Bidmon S, Terlutter R. Gender Differences in Searching for Health Information on https://doi.org/10.1177/2151458516687809. the Internet and the Virtual Patient-Physician Relationship in Germany: [23] Borkhoff CM, Hawker GA, Wright JG. Patient gender affects the referral and Exploratory Results on How Men and Women Differ and Why. J Med Internet Res recommendation for total joint arthroplasty. Clin Orthop Relat Res 2011;469: 2015;17:e156. https://doi.org/10.2196/JMIR.4127. 1829–37. https://doi.org/10.1007/s11999-011-1879-x. [40] Manierre MJ. Gaps in knowledge: tracking and explaining gender differences in [24] Borkhoff CM, Hawker GA, Kreder HJ, Glazier RH, Mahomed NN, Wright JG. health information seeking. Soc Sci Med 2015;128:151–8. https://doi.org/ Influence of patients’ gender on informed decision making regarding total knee 10.1016/J.SOCSCIMED.2015.01.028. arthroplasty. Arthritis Care Res (Hoboken) 2013;65:1281–90. https://doi.org/ [41] Schäfer G, Prkachin KM, Kaseweter KA, de AC, Williams C. Health care providers’ 10.1002/acr.21970. judgments in chronic pain: the influence of gender and trustworthiness. Pain 2016; [25] Mandl LA. Determining who should be referred for total hip and knee 157:1618–25. https://doi.org/10.1097/j.pain.0000000000000536. replacements. Nat Rev Rheumatol 2013;9:351–7. https://doi.org/10.1038/ [42] Zhang L, Losin EAR, Ashar YK, Koban L, Wager TD. Gender biases in estimation of nrrheum.2013.27. others’ pain. J Pain 2021;22:1048–59. https://doi.org/10.1016/j. [26] Rivero-Santana A, Torrente-Jiménez RS, Perestelo-Pérez L, Torres-Castaño A, jpain.2021.03.001. Ramos-García V, Bilbao A, Escobar A, Serrano-Aguilar P, Feijoo-Cid M. [43] Falk H, Henoch I, Ozanne A, Öhlen J, Ung EJ, Fridh I, Sarenmalm EK, Falk K. Effectiveness of a decision aid for patients with knee osteoarthritis: a randomized Differences in symptom distress based on gender and palliative care designation controlled trial. Osteoarthr Cartil 2021;29:1265–74. https://doi.org/10.1016/j. among hospitalized patients. J Nurs Sch 2016;48:569–76. https://doi.org/ joca.2021.06.005. 10.1111/jnu.12254. [27] Herdman M, Gudex C, Lloyd A, Janssen M, Kind P, Parkin D, Bonsel G, Badia X. [44] Koens S, Marx G, Gras C, Scherer M, Lüdecke D, von dem Knesebeck O. Physicians’ Development and preliminary testing of the new five-level version of EQ-5D (EQ- information seeking behavior in patients presenting with heart failure symptoms – 5D–5L). Qual Life Res 2011;20:1727–36. https://doi.org/10.1007/s11136-011- Does gender of physician and patient matter? Patient Educ Couns 2020;103: 9903-x. 2437–42. https://doi.org/10.1016/j.pec.2020.05.022. [28] M.J. Barry, D.C. Cherkin, C. YuChiao, F.J. Fowler, S. Skates, A Randomized Trial of [45] Visser LNC, Tollenaar MS, de Haes HCJM, Smets EMA. The value of physicians’ a Multimedia Shared Decision-Making Program for Men Facing a Treatment affect-oriented communication for patients’ recall of information. Patient Educ Decision for Benign Prostatic Hyperplasia, Dis. Manag. Clin. Outcomes. 1 (1997) Couns 2017;100:2116–20. https://doi.org/10.1016/J.PEC.2017.06.005. 5–14. https://www.infona.pl//resource/bwmeta1.element.elsevier-1c72f9fd- [46] Westendorp J, Stouthard J, Meijers MC, Neyrinck BAM, de Jong P, van Dulmen S, d36e-31ff-90be-f168cfb83dda (accessed January 31, 2022). van Vliet LM. The power of clinician-expressed empathy to increase information [29] O’Connor AM. Validation of a decisional conflict scale. Med Decis Mak 1995;15: recall in advanced breast cancer care: an observational study in clinical care, 25–30. https://doi.org/10.1177/0272989X9501500105. exploring the mediating role of anxiety. Patient Educ Couns 2021;104:1109–15. [30] Urrutia M, Campos S, O’Connor A. Validación de una versión en español de la https://doi.org/10.1016/J.PEC.2020.10.025. Escala de Conflicto Decisional. Rev Med Chil 2008;136. https://doi.org/10.4067/ [47] Karlson EW, Daltroy LH, Liang MH, Eaton HE, Katz JN. Gender differences in S0034-98872008001100010. patient preferences may underlie differential utilization of elective surgery. Am J [31] Valentine KD, Cha T, Giardina JC, Marques F, Atlas SJ, Bedair H, Chen AF, Med 1997;102:524–30. https://doi.org/10.1016/s0002-9343(97)00050-8. Doorly T, Kang J, Leavitt L, Licurse A, O’Brien T, Sequist T, Sepucha K. Assessing [48] Chang HJ, Mehta PS, Rosenberg A, Scrimshaw SC. Concerns of patients actively the quality of shared decision making for elective orthopedic surgery across a large contemplating total knee replacement: differences by race and gender. Arthritis healthcare system: cross-sectional survey study. BMC Musculoskelet Disord 2021; Rheum 2004;51:117–23. https://doi.org/10.1002/ART.20073. 22:1–10. https://doi.org/10.1186/S12891-021-04853-X/TABLES/5. [49] Clark JP, Hudak PL, Hawker GA, Coyte PC, Mahomed NN, Kreder HJ, Wright JG. [32] Stacey D, Taljaard M, Dervin G, Tugwell P, O’Connor AM, Pomey MP, Boland L, The moving target: a qualitative study of elderly patients’ decision-making Beach S, Meltzer D, Hawker G. Impact of patient decision aids on appropriate and regarding total joint replacement surgery. J Bone Jt Surg Am 2004;86:1366–74. timely access to hip or knee arthroplasty for osteoarthritis: a randomized 〈http://www.ncbi.nlm.nih.gov/pubmed/15252082〉. controlled trial. Osteoarthr Cartil 2016;24:99–107. https://doi.org/10.1016/J. [50] Gandhi R, Razak F, Davey JR, Rampersaud YR, Mahomed NN. Effect of sex and JOCA.2015.07.024. living arrangement on the timing and outcome of joint replacement surgery. Can J [33] van der Sluis G, Jager J, Punt I, Goldbohm A, Meinders MJ, Bimmel R, van Surg 2010;53:37–41. http://www.ncbi.nlm.nih.gov/pubmed/20100411. Meeteren NLU, Nijhuis-van Der Sanden MWG, Hoogeboom TJ. Current Status and [51] Instituto Nacional de Estadística (Institution/Organization), Mujeres y hombres en Future Prospects for Shared Decision Making before and after Total Knee España, 2021. https://www.ine.es/ss/Satellite?L=es_ES&c=INEPublicacion_ Replacement Surgery—A Scoping Review. Int J Environ Res Public Health 2021; C&cid=1259924822888&p=%5C&pagename=ProductosYServicios% 18:668. https://doi.org/10.3390/ijerph18020668. 2FPYSLayout¶m1=PYSDetalleGratuitas (accessed July 18, 2022). 7