Introduction
ACLR reconstruction (ACLR) is a relatively common procedure, with approximately 130,000 procedures performed annually [4]. ACLR is generally successful, with 83% of high-level athletes returning to competition [11]. These favorable outcomes and the frequency with which ACLRs are performed may cause patients to have high expectations for ACLR outcomes. Eighty-one percent of patients undergoing primary ACLR expect no instability at any activity level, and 16% of patients expect only occasional instability during high-impact sports [7]. Furthermore, 100% of patients undergoing primary or revision ACLR expect normal or near-normal knee function [7]. These expectations contrast with evidence that 15% to 24% of patients who undergo ACLR experience abnormal postoperative knee function [10, 12, 13].
Recent evidence supports an association between achieved expectations and patient satisfaction [3, 8, 19]. Although limited, there is evidence that improving expectations can improve long-term surgical outcomes. In a randomized controlled study, Rief et al. [17] demonstrated that the development of realistic expectations through psychological intervention can improve patient outcomes 6 months after heart surgery. Moreover, unrealistic expectations can predispose patients to decreased postoperative protocol compliance, worse outcomes, and ultimately poor satisfaction [16]. In light of this evidence, it is important to both understand patient expectations and effectively manage those expectations. Although patients may have inflated expectations of ACLR, surgeon expectations regarding pain and knee function 6 months after ACLR and meniscectomy may be more accurate than patient expectations [18]. However, further research is needed to determine the relationship between expectations of ACLR and outcomes after complete recovery (18 to 24 months, postoperatively). Furthermore, the relationship between specific patient parameters and expectations is unknown.
In this prospective investigation, we aimed to answer the following questions: (1) Are patient and surgeon expectations of knee function 18 months after ACLR similar? (2) Are patients’ and surgeons’ expectations of outcomes similar to patients’ actual 2-year outcomes? (3) Is there a relationship between preoperative or intraoperative parameters and expectations? (4) Is there a relationship between preoperative or intraoperative parameters and 2-year ACLR outcomes?
Patients and Methods
Study Design and Setting
Institutional review board approval was obtained before patient enrollment. This prospective cohort study was conducted at one private orthopaedic institution with multiple clinics. All surgeries were performed in the outpatient setting. Participants were recruited between October 2013 and January 2017.
Participants
All patients undergoing primary ACLR were eligible for inclusion in this study. Exclusion criteria were a previous surgery on the operative knee, concomitant ligamentous repair (posterior cruciate ligament, collateral ligaments), pregnancy, workers compensation claims, or pending litigation.
Cohort Demographics
The seven orthopedic surgeons included in this study were board-certified, sports medicine fellowship-trained surgeons (MC, JS, SC, BT, KF, FT, MP). The surgeons were all males with a median age of 50 years and a median of 18 years of experience after fellowship completion. On average, each surgeon performed approximately 60.8 ACLRs (range 34.2-101.4) per year.
Between October 2013 and January 2017, the aforementioned surgeons performed 1340 ACLRs. Of these, 1239 were eligible for inclusion; however, most patients could not be contacted due to limited research staff spread across multiple locations. A total of 129 patients (76 men and 53 women) agreed to participate. The median (range) age of patients on the day of surgery was 33.6 years (16.8-64.1) and median (range) BMI was 26.6 kg/m2 (19.6-43.0). Seventy-four percent (95 of 129) of patients injured their ACL during sports-related activities, 23% (30 of 129) were chronic injuries, and 28% (36 of 129) demonstrated cartilage damage during arthroscopic evaluation (Table 1). Forty-three patients demonstrated signs of concomitant ligamentous damage on magnetic resonance images (MRI); however, no patients required surgical management of concomitant ligamentous injury.
;)
Table 1.: Patient demographics
The final follow-up rate at 2-years postoperative was 76% (98 of 129). The 31 patients who did not complete follow-up could not be contacted. Forty-five percent (14 of 31) of patients lost to follow-up had chronic injuries, compared with 16% (16 of 98) of patients who completed followed-up (p = 0.001) (Table 1). Patients who were lost to follow-up also had worst preoperative knee function (IKDC 39.1 [range 9.2 to 87] versus 47.4 [range 16.5 to 86.2]; p = 0.016); however, there was no difference in knee function between the follow-up and non-follow-up groups at 1-year postoperative (IKDC 90.7 [range 27.6 to 98.0] versus 86.6 [range 34 to 100]; p = 0.791).
Description of Experiment, Treatment, or Surgery
All patients were verbally counseled preoperatively by an attending orthopaedic surgeon regarding when they would be able to participate in light activity and rigorous physical activity, as well as what future limitations they may have as a result of their injury. A standardized patient counseling script was not used. A consistent perioperative protocol was used for all patients. All patients underwent ACL reconstruction with a single-bundle technique, using anatomic tibial and femoral graft insertion sites. ACLR graft sources included autologous patellar tendon (n = 48), allograft patellar tendon (n = 60), and autograft hamstring tendon (n = 21).
Postoperative Rehabilitation
Multiple physical therapists were involved in the rehabilitation of patients; however, rehabilitation protocols and goals were similar for all patients. All patients were cleared for sports participation by 12 months postoperatively.
Variables, Outcome Measures, Data Sources, and Bias
The primary outcome was patient subjective knee function, measured with the International Knee Documentation Committee (IKDC) Subjective Knee Evaluation questionnaire. The IKDC has been validated for use in patients with acute knee injuries [9]. The minimal clinically important difference (MCID) for the IKDC has previously been identified as 11.5 points [6], and represents the minimal change in IKDC score required for a clinically noticeable difference in knee function. Patients completed the IKDC questionnaires preoperatively and at 3, 6, 12, and 24 months postoperative. Patients were contacted via telephone, written mail, or in-person interviews for follow-up. Patients completed an additional IKDC questionnaire preoperatively pertaining to their expected knee function at 18 months postoperative. This timepoint was selected because all patients were expected to have maximal improvement and be cleared for full activity by 18 months postoperative. The expectation surveys were completed after patients were counseled as described above. Immediately after surgery, surgeons completed IKDC questionnaires about how they expected the patient to feel or function in 18 months. Surgeons and patients were blinded to each other’s responses. Surgeons were asked to consider the following parameters when completing expectation questionnaires: patient age, lifestyle, severity of injury, chronicity of injury, associated injuries, graft type, comorbidities, and surgical difficulty. Expectations were compared with 2-year postoperative outcomes. We chose 2-year outcomes to ensure patients were completely rehabilitated. Demographics, preoperative parameters, and intraoperative parameters were collected for analysis with expectations and outcomes (Table 2). MRI reports were available for 98% (126 of 129) of patients.
Table 2.: Preoperative and intraoperative parameters analyzed
Statistical Analysis, Study Size
The statistical analysis was performed using SPSS version 25.0 (IBM Corp, New York, NY, USA). We analyzed the distributions of expected and actual patient outcomes to determine the normality of the data. IKDC scores were not normally distributed; therefore, we performed nonparametric analyses. IKDC scores are presented as median (range). We analyzed expected and actual patient-reported outcomes using the Mann-Whitney U test and paired Wilcoxon signed ranks tests, as appropriate. Spearman’s nonparametric correlations were also performed. Patients were stratified by age (< or ≥ median age), gender, ACLR graft type (autograph or allograph) and surgeon expectation (< or ≥ IKDC score of 88.5). A cutoff of 88.5 was used for the surgeon expectation because this represents one completed MCID from a perfect score. Preoperative and intraoperative parameters were analyzed for associations with expectations and outcomes. Due to the non-linear distribution of patient- and surgeon-expected IKDC scores, we converted expectations to ordinal ranks (< 85, ≥ 85 and < 95, and ≥ 95) to enable ordinal logistic regression analysis of factors associated with expectations. Generalized linear repeated measure models with Bonferroni adjusted post-hoc comparisons were constructed to assess changes in outcomes over time. An α value of 0.05 was used. Missing data for each separate analysis were excluded from the analysis.
The required sample size was determined based on the use of the MCID for the IKDC questionnaire. An anticipated effect size of 0.55 was calculated based on the MCID of 11.5 points and published normative data for our patient population [2, 6]. The power analysis revealed that enrollment of 28 patients was required to detect a difference in paired IKDC scores with 80% power through paired, two-tailed t-tests. Enrollment was increased to account for anticipated participant attrition and to power subgroup analysis.
Results
Patient and Surgeon Expectations
There was no difference between patient and surgeon expectations; however, there was also no correlation between worse surgeon expectations and worse patient expectations. In the total cohort, there was no difference between patient and surgeon expectations (IKDC score 94.8 [range 47.4 to 100] versus 94.3 [range 46 to 100]; p = 0.283). Higher patient expectations did not correlate with higher surgeon expectations (r = 0.168; p = 0.078) (Fig. 1). Among patients with suboptimal surgeon expectations (expected IKDC < 88.5), patients had higher expectations than surgeons (IKDC score 92.0 [range 47.4 to 100] versus 75.6 [range 46.0 to 87.4]; p < 0.001). There was a difference in patient expectations between patients with optimal and suboptimal surgeon expectations; however, this difference was so small that it was unlikely to be perceived (IKDC score 93.7 [range74.7 to 100] versus 89.0 [range 47.4 to 100]; p = 0.016). Among older patients (> median age), higher surgeon expectations also correlated with higher patient expectations (r = 0.345; p = 0.01), otherwise no correlations were identified through subgroup analysis stratified by age, gender, or graft type (Table 3).
Fig. 1: Worse surgeon expectations did not correlate with worst patient expectations.
Table 3.: Correlation between IKDC expectations and outcomes
Patient and Surgeon Expectations Compared with 2-year Outcomes
There was no difference between actual 2-year outcomes and patient expectations or surgeon expectations; however, there was also no correlation between higher expectations and higher 2-year postoperative outcomes. The median (range) 2-year postoperative IKDC score was 89.7 (32.2 to 100), which was lower than the median (range) patient-expected IKDC score, 94.8 (47.4 to 100; p < 0.001) and surgeon-expected IKDC score, 94.3 (46 to 100; p < 0.001); however, this difference is so small that it is unlikely to be perceived (Fig. 2). There was no correlation between 2-year postoperative outcomes and patient expectations (r = 0.14; p = 0.186) or surgeon expectations (r = 0.019; p = 0.86) (Fig. 3A-B). Among women, higher patient expectations correlated with better 2-year postoperative outcomes (r = 0.405; p = 0.014). No other correlations were identified through subgroup analysis stratified by age, gender, or graft type (Table 3).
Fig. 2: Patient and surgeon expectations were higher than actual patient outcomes but the difference is too small to be perceived. Postoperative IKDC scores improved over time. Error bars represent median absolute deviation.
Fig. 3: A-B (A) Worse patient expectations did not correlate with worst 2-year patient outcomes. (B) Worse surgeon expectations did not correlate with worst 2-year patient outcomes.
Relationship Between Patient Parameters and Expectations
Arthroscopic evidence of cartilage damage was independently associated with worse patient and surgeon expectations. Surgeon expectations were also independently associated with the surgeons themselves. Univariate analysis demonstrated a relationship between higher patient expectations and lower patient BMI (r = -0.199; p = 0.03), non-Caucasian race (IKDC 97.1 [range 77.0 to 100] versus 93.1 [range 47.4 to 100]; p = 0.043), and no arthroscopic evidence of cartilage damage (IKDC 95.4 [range 77.0 to 100] versus 90.8 [range 47.4 to 100]; p = 0.002). Logistic regression demonstrated an independent relationship between arthroscopic evidence of cartilage damage and worse patient expectations (OR 0.446 [95% CI 0.202 to 0.985]; p = 0.046). Multiple factors demonstrated a univariate relationship with higher surgeon expectations including: younger patient age (r = -0.241; p = 0.009), sports-related injury (IKDC 95.4 [range 46-100] versus 88.7 [range 47-100]; p = 0.001), use of autologous patellar tendon graft (IKDC score 95.4 [range 56.3-100] versus 93.1 [range 46-100]; p = 0.044), and no arthroscopic evidence of cartilage damage (IKDC 95.2 [range 46-100] versus 92.7 [range 49.4-100]; p = 0.022). Postoperative expectations also varied between surgeons (p < 0.001). Logistic regression demonstrated that surgeon expectations were independently affected by multiple surgeons (p < 0.001) and arthroscopic evidence of cartilage damage (OR 2.919 [95% CI 1.071 to 7.956]; p = 0.0436).
Association Between Preoperative Factors and Patient Outcomes
No parameters were associated with differences in postoperative knee function.
Functional Outcomes Over Time
Patients reported improvement in knee function during the first postoperative year, but not thereafter. There was an improvement in IKDC scores over time (η2= 0.497; p < 0.001). There was an increase in IKDC scores at each time interval during the first 6 postoperative months (p < 0.001) and between the 6-month and 12-month period (p = 0.001); however, there was no difference in IKDC scores between 1-year and 2-years postoperatively (p = 1.00) (Fig. 1).
Eleven complications in 11 of 129 patients were reported during the 2-year follow-up period; however, complications of patients lost to follow-up may not be represented. Arthrofibrosis requiring débridement and lysis of adhesions occurred in six patients. Two patients had superficial surgical site infections that were treated with oral antibiotics, while another patient experienced an intra-articular infection, resulting in incision and drainage with graft retention. Two patients had deep vein thromboses requiring anticoagulation. During this 2-year analysis, there was one graft failure, which was treated nonoperatively. When comparing patients with and without complications, there was no difference in patient expectations (IKDC 92.7 [range 70.1 to 100] versus 94.3 [range 47.4 to 100]; p = 0.493) or surgeon expectations (IKDC 96.6 [range 77 to 100] versus 94.5 [range 46 to 100]; p = 0.687).
Discussion
Given the association between patient expectations and patient satisfaction, appropriate management of patient expectations is an important aspect of patient care [16]. There is limited knowledge about the relationship between patient and surgeon expectations, and the relationship between these expectations and knee function after complete recovery (18 to 24 months postoperatively). The results of this prospective study demonstrated that surgeons and patients generally have similar expectations, and that there is no clinical difference between these expectations and actual outcomes. However, there is no correlation between worse surgeon expectations and worse patient expectations or between worse expectations and worse outcomes. Although arthroscopic evidence of cartilage damage was associated with both patient and surgeon expectations, no factors were found to be associated with postoperative outcomes.
Even though it is informative, this study has limitations. Although the average age, BMI, and percentage of autograft use in the study cohort are similar to other populations of patients undergoing primary ACLR, a larger proportion of our population injured their ACL while playing a sport compared with previously reported patient populations [15]. Other parameters such as education level and economic status may have influenced our results but were not assessed. Similarly, the cohort of surgeons included in this study may be different than the general population. The surgeons themselves were associated with differences in surgeon expectations, and therefore it is possible that the reader may have a different expectation tendencies than the included surgeons. Although the inclusion of multiple surgeons and multiple graft types decreased the internal validity of our study, it also increased the generalizability of our results. Because seven surgeons were included, our results are not representative of only one surgeon’s personality or tendencies. Another limitation was the difference in the timing of expectation surveys. Patients were surveyed preoperatively, while surgeons were surveyed postoperatively, potentially biasing surgeon responses. This may have altered differences between surgeon and patient expectations; however, it allowed us to capture surgeon expectations after all possible information was considered. This study may have also been biased by the lack of complete follow-up of all enrolled patients, even though more than 75% of patients followed-up at 2 years. This study was powered to detect a difference between surgeon and patient expectations, as well as between expectations and actual outcomes. There were preoperative differences between patients who completed follow-up and those who did not, which may indicate bias in our results; however, IKDC scores at 1-year postoperative were not different between those who completed follow-up and those who did not. Finally, validated expectation surveys were not used in this study. We elected to use the IKDC to assess expectations because it is widely known to surgeons and asks specific questions about pain, function, ROM, and other parameters that provide meaningful insight into the function of the knee joint. In addition, by using functional scales, we can easily correlate expectations to actual, observed outcomes. The MCID for the IKDC is also available, which allow for clinical interpretation of the statistical results. These benefits are not available for with other expectation questionnaires.
In this study, patients and surgeons demonstrated high expectations of outcomes after ACLR; however, there was no difference between patient and surgeon expectations. Our findings contrast with the results of Khair et al. [9], who observed worse ACLR expectations among patients compared with surgeons. The difference in methodology may account for the difference in our results. Khair et al. [9] assessed expectations using the Hospital for Special Surgery Expectations Survey, whereas we used the IKDC questionnaire in this study. Our results support the results of Feucht et al. [7], which demonstrated that patients have high expectations for outcomes after ACLR. Our results demonstrate high patient expectations in both patients with high surgeon expectations and patients with suboptimal surgeon expectations. This indicates that although patient and surgeon expectations are similar, surgeons are not effectively communicating poor expectations to patients. This is further supported by the difference in surgeon and patient expectations observed in patients with suboptimal surgeon expectations, and the overall lack of correlation between surgeon and patient expectations. The similarity in patient and surgeon expectations is likely not due to individual counseling, but rather a similar perception of outcomes among patients and surgeons. Instead of the surgeon’s counseling having the most influence on patient expectations, counseling from friends and family with prior ACLR, online medical forums, and extrapolation from anecdotal media evidence may be more influential.
Both patient and surgeon expectations were greater than actual patient postoperative outcomes; however, this difference is likely too small to be perceived by the patient or surgeon. However, with the exception of women’s patient expectations, neither patient nor surgeon expectations correlated with actual outcomes. Our results contrast the conclusions of Feucht et al. [7], which showed that patients’ expectations for knee function after ACLR are greater than actual outcomes. However, Feucht et al. [7] compared their expectation results with prior studies instead of postoperative outcomes from their cohort. In this study, postoperative outcomes were collected from the same patients who completed expectation surveys, allowing for the comparison of expectations with paired postoperative outcomes. Our results indicate that overall, patients and surgeons have appropriate expectations for ACLR outcomes. Patients and surgeons expect high outcomes and generally patients achieve outcomes clinically similar to expectations. Nevertheless, it is alarming that there is no correlation between surgeon expectations and actual outcomes. Previous studies have identified deficiencies in physician counseling [5, 14]; however, these results indicate that the source of the discordance between expectations and outcomes is beyond physician-patient counseling. If surgeons cannot accurately predict surgical outcomes, then preoperative counseling, no matter how effective, is of limited value. The inclusion of several surgeons in this study demonstrates that this is not simply the deficiency of one surgeon, but rather an area of weakness for orthopaedic surgeons in general. Further research is needed to determine if there is a lack of correlation between surgeon expectations and postoperative outcomes for other orthopaedic procedures and to identify reasons for discordance between surgeon expectations and actual outcomes.
Both patient and surgeon expectations were independently influenced by arthroscopic evidence of cartilage damage, which partially explains the similarity in patient and surgeon expectations. Interestingly, patients were surveyed preoperatively, and preoperative MRI evidence of cartilage damage did not have a relationship with patient or surgeon expectations. Therefore, arthroscopic evidence of cartilage damage may be a proxy for another, unmeasured influencing variable. Conversely, surgeons may have sensed an increased likelihood of cartilage damage preoperatively and discussed this assumption and its postoperative impact with the patient at that time. This may indicate that even though surgeons have difficulty conveying the extent to which they expect decreased postoperative function, they are able to communicate the relative effect of certain patient parameters. Although this is not ideal, it represents a step in the right direction when compared with prior evidence. For example, Cailliez et al. [5] demonstrated that only 27% of patients reported receiving information about expected outcomes during preoperative counseling, and Matava et al. [14] identified a lack of understanding of the ACL and ACLR procedures amoung patients undergoing ACLR. Surgeon expectations were also influenced by the surgeon themselves. This is to be expected as different surgeons have different levels of optimism or pessimism.
There was no relationship between patient parameters and 2-year postoperative outcomes. Our results contrast with those of Ahldén et al. [1], who identified worse functional outcomes among patients with concomitant intra-articular injuries (meniscal or chondral) at 1-year and 5 years postoperatively. It is important to note that patient follow-up only included 2 postoperative years.
Finally, patient functional outcomes improved during the study; however, there was no functional improvement between 1 and 2 years after surgery. This could indicate the readiness of athletes and patients to return to activities at 1 year because they have maximized improvement, or it could indicate an opportunity to further rehabilitate patients beyond 1 year to obtain better outcomes.
Patient and surgeon expectations are similar to each other and 2-year actual outcomes; however, there is no correlation between the patient and surgeon expectations and no correlation between expectations and actual outcomes. Although expectations are similar to actual outcomes for most patients, surgeons do not have accurate, patient-specific expectations of short-term outcomes after ACLR, and they cannot predict which patients will have worse outcomes; thus, preventing adequate patient counseling. Surgeons should be cautious when evaluating and counseling patients preoperatively and avoid assuming high expectations. Further research is needed to identify reasons for the discordance between surgeon expectations and actual outcomes and to determine better short-term prognostic factors, so that these factors can be incorporated into preoperative counseling.
References
1. Ahldén M, Samuelsson K, Sernert N, Forssblad M, Karlsson J, Kartus J. The Swedish National Anterior Cruciate Ligament Register. Am J Sports Med
. 2012;40:2230-2235.
2. Anderson AF, Irrgang JJ, Kocher MS, Mann BJ, Harrast JJ, International Knee Documentation Committee. The International Knee Documentation Committee Subjective Knee Evaluation Form. Am J Sports Med
. 2006;34:128-135.
3. Bourne RB, Chesworth BM, Davis AM, Mahomed NN, Charron KDJ. Patient satisfaction after total knee arthroplasty: who is satisfied and who is not? Clin Orthop Relat Res
. 2010;468:57-63.
4. Buller LT, Best MJ, Baraga MG, Kaplan LD. Trends in anterior cruciate ligament reconstruction in the United States. Orthop J Sport Med
. 2015;3:1-8.
5. Cailliez J, Reina N, Molinier F, Chaminade B, Chiron P, Laffosse J-M. Patient information ahead of anterior cruciate ligament reconstruction: Experience in a university hospital center. Orthop Traumatol Surg Res
. 2012;98:491-498.
6. Collins NJ, Misra D, Felson DT, Crossley KM, Roos EM. Measures of knee function: International Knee Documentation Committee (IKDC) Subjective Knee Evaluation Form, Knee Injury and Osteoarthritis Outcome Score (KOOS), Knee Injury and Osteoarthritis Outcome Score Physical Function Short Form (KOOS-PS), Knee Ou.
Arthritis Care Res. (Hoboken). 2011;63(Suppl 1):S208-28.
7. Feucht MJ, Cotic M, Saier T, Minzlaff P, Plath JE, Imhoff AB, Hinterwimmer S. Patient expectations of primary and revision anterior cruciate ligament reconstruction. Knee Surg Sport Traumatol Arthrosc
. 2016;24:201-207.
8. Hamilton DF, Lane J V., Gaston P, Patton JT, MacDonald D, Simpson AHRW, Howie CR. What determines patient satisfaction with surgery? A prospective cohort study of 4709 patients following total joint replacement. BMJ Open. 2013;3:1-7.
9. Khair MM, Ghomrawi H, Wilson S, Marx RG. Patient and surgeon expectations prior to anterior cruciate ligament reconstruction. HSS J
. 2018;14:282-285.
10. Kraeutler MJ, Bravman JT, McCarty EC. Bone–patellar tendon–bone autograft versus allograft in outcomes of anterior cruciate ligament reconstruction. Am J Sports Med
. 2013;41:2439-2448.
11. Lai CCH, Ardern CL, Feller JA, Webster KE. Eighty-three per cent of elite athletes return to preinjury sport after anterior cruciate ligament reconstruction: a systematic review with meta-analysis of return to sport rates, graft rupture rates and performance outcomes. Br J Sports Med
. 2018;52:128-138.
12. Lamblin CJ, Waterman BR, Lubowitz JH. Anterior cruciate ligament reconstruction with autografts compared with non-irradiated, non-chemically treated allografts. Arthrosc J Arthrosc Relat Surg
. 2013;29:1113-1122.
13. Li S, Chen Y, Lin Z, Cui W, Zhao J, Su W. A systematic review of randomized controlled clinical trials comparing hamstring autografts versus bone-patellar tendon-bone autografts for the reconstruction of the anterior cruciate ligament. Arch Orthop Trauma Surg
. 2012;132:1287-1297.
14. Matava MJ, Howard DR, Polakof L, Brophy RH. Public Perception Regarding Anterior Cruciate Ligament Reconstruction. J Bone Joint Surg Am
. 2014;96:e85.
15. Mitchell JJ, Cinque ME, Dornan GJ, Matheny LM, Dean CS, Kruckeberg B, Moatshe G, Chahla J, LaPrade RF. Primary versus revision anterior cruciate ligament reconstruction: Patient demographics, radiographic findings, and associated lesions. Arthroscopy. 2018;34:695-703.
16. Neuprez A, Delcour J-P, Fatemi F, Gillet P, Crielaard J-M, Bruyère O, Reginster J-Y. Patients’ expectations impact their satisfaction following total hip or knee arthroplasty. PLoS One. 2016;11:e0167911.
17. Rief W, Shedden-mora MC, Laferton JAC, Auer C, Petrie KJ, Salzmann S, Schedlowski M, Moosdorf R. Preoperative optimization of patient expectations improves long-term outcome in heart surgery patients : results of the randomized controlled PSY-HEART trial. BMC Med
. 2017:1-13.
18. Rosenberger PH, Jokl P, Cameron A, Ickovics JR. Shared decision making, preoperative expectations, and postoperative reality: Differences in physician and patient predictions and ratings of knee surgery outcomes. Arthroscopy
. 2005;21:562-569.
19. Younger ASE, Wing KJ, Glazebrook M, Daniels TR, Dryden PJ, Lalonde K-A, Wong H, Qian H, Penner M. Patient expectation and satisfaction as measures of operative outcome in end-stage ankle arthritis. Foot Ankle Int
. 2015;36:123-134.
