Background: In the United Kingdom, organizations involved in health-care commissioning have recently introduced legislation limiting access to total knee arthroplasty through the introduction of arbitrary thresholds unsupported by the literature and based on body mass index. This study aimed to establish the relationship between body mass index and patient-reported specific and general outcomes on total knee arthroplasty.
Methods: Using national patient-reported outcome measures (PROMs) linked to the National Joint Registry, we identified 13,673 primary total knee arthroplasties performed for the treatment of osteoarthritis. The PROMs project involves the collection of condition-specific and general health outcomes before and at six months following total knee arthroplasty. The relationships between body mass index and the Oxford Knee Score, EuroQol 5D index, and EuroQol 5D Visual Analogue Scale were assessed with use of scatterplots and linear regression. The improvement in these measures was compared for three distinct groups based on body mass index (Group I [15 to 24.9 kg/m2], Group II [25 to 39.9 kg/m2], and Group III [40 to 60 kg/m2]) with use of multiple regession analysis to adjust for differences in age, sex, American Society of Anesthesiologists grade, general health rating, and number of comorbidities.
Results: The preoperative and postoperative patient-reported outcome measures declined to a similar extent with increasing body mass index. The gradient of the linear regression equation relating to the change in scores was positive in all cases, indicating that there was a tendency for scores to improve to a greater extent as body mass index increased. After adjustment, the changes in patient-reported outcome measures in Group I and Group III were equivalent for the Oxford Knee Score (mean difference, 0.5 point [95% confidence interval, −0.5 to 1.5 points]; p = 0.78), the EuroQol 5D index (mean difference, 0.014 point [95% confidence interval, −0.021 to 0.048 point]; p = 1.00), and the EuroQol 5D Visual Analogue Scale (mean difference, 1.9 points [95% confidence interval, −0.4 to 4.1 points]; p = 0.13). Wound complications were significantly higher (p < 0.001) at a rate of 17% (168 of 1018 patients) in Group III compared with 9% (121 of 1292 patients) in Group I.
Conclusions: The improvements in patient-reported outcome measures experienced by patients were similar, irrespective of body mass index. Health policy should be based on the overall improvements in function and general health gained through surgery. Obese patients should not be excluded from the benefit of total knee arthroplasty, given that their overall improvements were equivalent to those of patients with a lower body mass index.
Level of Evidence: Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.
1Institute of Cellular Medicine, Newcastle University, Framlington Place, Newcastle upon Tyne NE1 7RU, England. E-mail address: email@example.com
2Wansbeck General Hospital, Woodhorn Lane, Ashington, Northumberland NE63 9JJ, England. E-mail address for T. Petheram: firstname.lastname@example.org. E-mail address for M. Reed: email@example.com
3James Cook University Hospital, Marton Road, Middlesbrough TS4 3BW, England. E-mail address for S. Jameson: firstname.lastname@example.org. E-mail address for P. Gregg: email@example.com
4Royal Victoria Infirmary, Queen Victoria Road, Newcastle upon Tyne NE1 4LP, England. E-mail address: firstname.lastname@example.org
Obesity is an ever-worsening problem in the developed world. According to the National Health Service (NHS) Information Centre for Health and Social Care, 22% of men and 24% of women in the United Kingdom were classified as obese (body mass index [BMI], ≥30 kg/m2) in 2009, an increase from 13% and 16%, respectively, in 19931. Obesity is an independent causative factor in the development of knee osteoarthritis2. The number of obese patients undergoing total knee arthroplasty in the United States increased between 1990 and 20052. The average BMI of patients undergoing total knee arthroplasty in England and Wales (30.7 kg/m2) is also increasing and is now in the obese range3. While the obesity trend continues, there are concurrent reports that organizations involved in health-care commissioning in the United Kingdom are introducing policies that have no supportive evidence and that withhold funding for joint arthroplasty on the basis of BMI thresholds4-6. These thresholds are being used to limit access to total knee arthroplasty to control budgets and contain NHS spending.
Studies investigating the outcome of total knee arthroplasty in the obese population have produced conflicting results. Smaller improvements in knee scores (International Knee Documentation Committee score and Knee Society Score)7-10, inferior implant survival rates8-10, greater rates of complications7,10, a greater prevalence of radiolucent lines on radiographs10, and a greater proportion of patellofemoral symptoms11 have been reported for patients who are obese (BMI, ≥30 kg/m2) or morbidly obese (BMI, ≥40 kg/m2) when compared with patients who have a “normal” BMI (≤25 kg/m2). Alternative studies have demonstrated equivalent knee function (Knee Society Score, Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC])12-15, general and mental health (WOMAC mental health component, Nottingham Health Profile)12,15, patient satisfaction15,16, rates of radiolucent lines on radiographs14, and rates of implant survival14 for these groups. These studies are all limited in size. Two recent reviews concluded that the complication rate is higher and that the implant survival rate is inferior in obese and morbidly obese patients17,18. However, insufficient evidence is currently available to definitively determine the effect on functional and quality-of-life outcomes in obese patients undergoing total knee arthroplasty17.
In view of the lack of information on the relationship between patient outcomes and BMI, we evaluated patient-reported outcome measures collected by the National Joint Registry of England and Wales. We aimed to establish the nature of the relationship between BMI and preoperative and postoperative patient-reported outcome measures and investigated if the improvement of these scores following surgery was dependent on BMI. We hoped to determine whether there is any evidence to support the idea of a BMI cutoff as a basis for withholding total knee arthroplasty surgery because of functional outcomes. Our hypothesis was that similar levels of improvement would be seen following total knee arthroplasty, irrespective of the patient preoperative BMI.
Materials and Methods
With use of a combined request strategy involving both the National Joint Registry and the NHS Information Centre, we accessed records of patient-reported outcome measures relating to knee arthroplasty and linked these data to the corresponding patient-level National Joint Registry records. Data collection of patient-reported outcome measures commenced in April 2008 and involved both preoperative and postoperative assessments of outcomes with use of validated measures of knee-specific health (Oxford Knee Score) and general health (EuroQol 5D [EQ-5D]). Data for this analysis were collected on knee arthroplasties performed from May 1, 2008, to September 1, 2010.
In total, 40,925 patients were registered with both the National Joint Registry and the Patient Reported Outcome Measures (PROMs) project as of September 2010. From this cohort, we excluded 8043 patients who were missing either the preoperative or the postoperative PROMs questionnaire; 2676 patients who had missing dates of completion for the PROMs questionnaires; 5195 patients who had completed the preoperative questionnaire more than ninety days prior to surgery or who had completed the postoperative questionnaire <180 days or >365 days after surgery; 1618 patients who had undergone a unicondylar, patellofemoral, or revision knee arthroplasty; and 595 patients who had a primary indication that was not osteoarthritis. From the remaining 22,798 patients, we excluded 9125 patients who had BMI data that were either missing or outside the range of 15 to 60 kg/m2. This range is in line with the policy of BMI analysis used by the National Joint Registry as part of their annual reporting. In total, 13,673 patients fulfilled these criteria and were included in the analysis. To assess whether there was an inherent difference between the patients who had their BMI recorded and those who did not, we compared patient demographic characteristics and preoperative patient-reported outcome measures for the 13,673 patients who had their BMI recorded and the 9125 patients who were missing this variable (see Appendix). These two groups were comparable, suggesting that the group who had their BMI recorded was representative of the wider PROMs project population.
The PROMs questionnaire utilizes the Oxford Knee Score as its knee-specific outcome measure19. It is a reliable, valid, and responsive outcome measure, recommended for the assessment of large knee arthroplasty databases in a cross-sectional population20. It consists of twelve individual questions that are combined to generate an overall score of 0 to 48 points, with lower scores indicating more severe problems. General health is assessed with use of the EQ-5D, which provides a simple, generic measure of health for clinical and economic appraisal by evaluating five different aspects of general health (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). Responses to questions for these five elements can be combined with use of population weightings to produce a single index value for health status. In addition to the index, the EQ-5D health scale records general health on a visual analogue scale (VAS).
As part of the postoperative assessment, patients were asked whether they had developed a complication (wound problems, bleeding problems, readmission, or reoperation). These questions form part of the patient experience of surgery questionnaire, which previously has been used to audit complications after day surgery21,22. The interpretation of what constitutes a complication is at the discretion of the patient and no attempt is made to verify the presence of these complications from the medical records. This information is particularly useful when comparing groups for which the interpretation of a complication is expected to be similar.
For the purposes of analysis, the variables of interest (BMI, Oxford Knee Score, and EQ-5D) were considered to be continuous. We were interested in establishing whether the relationship between BMI and patient-reported outcome measures is linear and, if not, whether functional outcome plateaus or declines above a specific BMI threshold. Declining functional outcome with increasing BMI would give credence to the idea of using BMI as a means of rationing care. To determine the relationship between BMI and patient-reported outcome measures, we produced scatterplots of BMI against the preoperative, postoperative, and change in score for each of the outcome variables stratified by age (sixty-five years or younger or older than sixty-five years) and sex and calculated their associated linear regression equations.
Additional statistical analysis was undertaken by categorizing the data into three groups (Group I [BMI, 15 to 24.9 kg/m2], Group II [BMI, 25 to 39.9 kg/m2], and Group III [40 to 60 kg/m2]) on the basis of the World Health Organization upper cutoff for normal weight (BMI, 25 kg/m2) and lower cutoff for morbid obesity (BMI, 40 kg/m2)23. To limit the confounding effect of patient variables known to influence patient-reported outcome measures (age, sex, American Society of Anesthesiologists [ASA] grade24, number of comorbidities, and general health rating), we used multiple linear regression to adjust the changes in the Oxford Knee Score and EQ-5D scores observed for the three BMI groups. This adjustment effectively accounts for any differences in the distribution of these covariates and their respective influences on the changes in the patient-reported outcome measures within each of the groups.
Continuous variables were analyzed with use of one-way analysis of variance (ANOVA). Post hoc comparisons between groups were performed with use of the Bonferroni method. Counts of binary variables were compared with use of chi-square and Fisher exact tests. A p value of <0.05 was considered significant. This project was performed as a service evaluation without need for formal ethical approval after discussion with the relevant local ethics and research committees.
Source of Funding
This work was funded by a fellowship from the National Joint Registry. These funds were used to pay the salary of two of the authors (P.B. and S.J.) during their one-year tenure as National Joint Registry fellows.
The mean BMI (and standard deviation) for the 13,673 patients identified was 31.0 ± 5.5 kg/m2; 7771 patients (57%) had a BMI of ≥30 kg/m2. The postoperative PROMs questionnaires were collected at a mean of seven months (range, six to twelve months) following surgery. Patient demographic characteristics are summarized in the Appendix.
Scatterplots, stratified by age and sex, of BMI plotted against the preoperative, postoperative, and change in score for the Oxford Knee Score and the EQ-5D index are given in Figures 1 through 4. Initial analysis of the EQ-5D VAS demonstrated that patients reported only minimal improvements, which were not thought to be clinically important (mean improvement, 3.0 points [95% confidence interval (CI), 2.7 to 3.3 points] on a scale of 0 to 100 points). We therefore chose not to present the graphical output stratified by age and sex for this variable. For all patient-reported outcome measures, a consistent trend for decreasing preoperative and postoperative scores as BMI increased was observed, irrespective of age and sex. Overall, the gradients of the linear regression line for the Oxford Knee Score (+0.028 [p = 0.08]), EQ-5D index (+0.001 [p = 0.005]), and EQ-5D VAS (+0.033 [p = 0.35]) were positive, indicating that although the preoperative and postoperative scores decreased with BMI, the trend was for the change in score to improve, albeit minimally, as BMI increased.
The patient-reported outcome measures for the categorized BMI groups prior to adjustment are given in Table I. The observed values followed a pattern similar to that observed in Figures 1 through 4, with Group III having the lowest preoperative and postoperative Oxford Knee Score and EQ-5D index score but the greatest change in those scores when Group III was compared with the other two groups. The changes for the EQ-5D VAS were similar for all three groups.
The changes in Oxford Knee Score, EQ-5D index, and EQ-5D VAS for each group, adjusted for age, sex, ASA grade, number of comorbidities, and general health rating, are given in Table II. After adjustment, the mean change in Oxford Knee Score was greatest for Group III. The mean difference in the change in Oxford Knee Score when Group III was compared with Group I was 0.5 point (95% CI, −0.5 to 1.5 points) (p = 0.78), and the mean difference in the change when Group III was compared with Group II was 0.9 point (95% CI, 0.1 to 1.6 points) (p = 0.03). The mean difference between Group I and Group II was 0.4 point (95% CI, −0.3 to 1.1 points) (p = 0.57). For the adjusted EQ-5D index, the mean change was also greatest for Group III, although these differences were not significant. The mean difference in the change in the EQ-5D index when Group III was compared with Group I was 0.014 point (95% CI, −0.021 to 0.048 point) (p = 1.00), and the mean difference in the change when Group III was compared with Group II was 0.019 point (95% CI, −0.008 to 0.045 point) (p = 0.29). The mean difference between Group I and Group II was 0.005 point (95% CI, −0.021 to 0.031 point) (p = 1.00). For the adjusted EQ-5D VAS, the mean change was greatest for Group I, although these differences were not significant. The mean difference in the change in EQ-5D VAS when Group I was compared with Group III was 1.9 points (95% CI, −0.4 to 4.1 points) (p = 0.13), and the mean difference in the change when Group I was compared with Group II was 0.5 point (95% CI, −1.0 to 2.1 points) (p = 1.00). The mean difference between Group II and Group III was 1.3 points (95% CI, −0.4 to 3.1 points) (p = 0.20). For all PROMs, the differences observed for the changes in scores between groups were small and below the levels, indicating clinical significance: 3 points for the Oxford Knee Score, 0.1 point for the EQ-5D index, and 5 points for the EQ-5D VAS.
Patient-reported complications are given in Table III. The rates of wound complications significantly increased (p < 0.001) as BMI increased. Specifically, the rate was 9% (121 of 1292 patients) for Group I, 12% (1351 of 11,363 patients) for Group II, and 17% (168 of 1018 patients) for Group III. The rates of other complications (bleeding problems, readmission, reoperation) were similar for the three groups.
We found that although patients with high BMIs had lower postoperative knee and general health scores, the improvements that they experienced were comparable in magnitude with those of patients with a lower BMI. Obese patients gained as much benefit from knee replacement as patients with a “normal” BMI, even if they did not end up at a similar postoperative level. Other than wound problems, the rates of complications were not significantly different for morbidly obese patients (Group III) when compared with those for “normal” patients (Group I).
The major limitation of this analysis is its short follow-up period. A postoperative collection of national patient-reported outcome measures is routinely undertaken at six months following surgery and is overseen by the UK Department of Health. Because of the financial and logistical constraints of national sampling, there are currently no plans to extend the PROMs collection to include further assessments at a longer follow-up period. We have limited the effect of the short follow-up period by ensuring a minimum follow-up period of six months. By this time point, the majority of improvement in functional scores has already occurred and scores are starting to plateau25. In addition, this study was intended not to show the final improvements gained from total knee arthroplasty but to compare outcomes between patients who had differing BMIs at an equivalent time point. A failure to demonstrate a difference in the improvement in scores at six months is therefore valid, especially as the scores are likely to improve only minimally thereafter. The cohort of patients who had their BMI recorded were comparable with those who did not, suggesting that the recording of BMI was not biased by specific patient characteristics. When adjusting the scores, we were only able to adjust for variables available within the National Joint Registry and PROMs datasets. As such, we could not include additional variables such as mental health rating and educational status, which are known to influence patient-reported outcomes.
The strengths of the study are that there were a large number of patients available for analysis and that the data were extracted from a well-established registry reflecting current national practice. In addition, whereas other outcome studies have used surgeon-derived outcome measures, this is the first analysis, to our knowledge, that has focused exclusively on patient-reported outcome measures.
A recent review concluded that there was currently insufficient evidence to form a definitive view on functional and quality-of-life outcomes following total knee arthroplasty when comparing obese and non-obese patients17. This study found that BMI is related to both the preoperative and postoperative Oxford Knee Score and EQ-5D in a linear fashion, with morbidly obese patients having significantly lower scores than patients with a normal BMI. However, the improvements in both of these patient-reported outcome measures for morbidly obese patients were equivalent to patients with a normal BMI and did not decline above a certain BMI threshold. This observation suggests that there is as much benefit, in terms of improving knee function, general health, and quality of life, to be gained by operating on these patients. Considering solely the postoperative scores discriminates against patients with higher BMIs by suggesting that they cannot gain the success achieved by non-obese patients and by ignoring the fact that their functional gains are equivalent. We therefore question the validity of the arbitrarily determined BMI thresholds implemented by some health-care commissioning organizations as a barrier to surgery.
The observed trend of lower preoperative knee scores as BMI increases suggests that there may be a selection bias against obese patients, relating to the point at which surgery is offered. This finding may reflect current referral guidelines determined by health-care comissioners4-6 stating that patients over predefined BMI thresholds should lose weight before undergoing joint replacement or a reluctance on the part of the operating surgeon to undertake such cases. These “delays” to surgery allow a progression of the natural disease process and a worsening of symptoms. Additionally, this observation may relate to the direct influence that obesity has on general activities, such as difficulty with mobility or ascending and descending stairs due to associated comorbidities and body habitus.
Other than wound problems, the rates of complications among our three BMI groups did not significantly differ. These findings contrast with a number of previous studies showing higher rates of complications in the obese population7-10,26,27. These contrasting findings may reflect differences in patient and surgeon reporting of complications. Obese patients may be more likely to report wound problems if they are informed preoperatively that they are at greater risk of these problems, which could be a potential source of bias when using patient-reported wound problems to compare obese and non-obese patients. This study did not assess the impact of BMI on implant survival rate, which is inferior in morbidly obese patients who have five-year survival rates of only 74%10. Although it is important to consider complication rates and implant survival rate when making decisions regarding surgery, one must remember that it is the results in patient-reported outcome measures that are linked to improvements in quality of life. Thus, the value of total knee arthroplasty is primarily related to these scores. If patients are fully informed of the elevated risks for equivalent improvements in patient-reported outcome measures, then these factors should not, in themselves, be a reason to withhold surgery. Many factors (comorbidities, technical feasibility, anesthetic issues), each of which may preclude surgery, become more prevalent as BMI increases. However, BMI should not in itself be a barrier to surgery. The decision to proceed with total knee arthroplasty should be made by the treating orthopaedic surgeon and the patient after careful consideration of all aspects of the case and the likely impact of surgery at an individual level.
In conclusion, although increasing BMI is associated with poorer preoperative and postoperative patient-reported outcome measures, the improvement experienced by patients was similar, irrespective of BMI. It is important that those in control of health-care resources do not penalize obese patients on the basis of the limited functional outcome data available from previous studies and a selective view of postoperative scores alone, in which obese patients are disadvantaged by lower preoperative scores. Instead, the overall improvements in function and general health should be the barometer of success. Accordingly, we believe that obese patients should not be excluded from the benefit of undergoing total knee arthroplasty that is experienced by their fellow patients with lower BMIs.
A table showing patient demographic characteristics for the PROMs cohorts with and without BMI data is available with the online version of this article as a data supplement at jbjs.org.
NOTE: We thank the patients and staff of all of the hospitals in England and Wales who have contributed data to the National Joint Registry. We are grateful to the Healthcare Quality Improvement Partnership, the National Joint Registry Steering Committee, and the staff at the National Joint Registry Centre for facilitating this work. This work was funded by a fellowship from the National Joint Registry. The authors have conformed to the National Joint Registry’s standard protocol for data access and publication. The views expressed represent those of the authors and do not necessarily reflect those of the National Joint Register Steering Committee or the Health Quality Improvement Partnership (HQIP) who do not vouch for how the information is presented.
Investigation performed at the Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, England
2. Fehring TK Odum SM Griffin WL Mason JB McCoy TH. The obesity epidemic: its effect on total joint arthroplasty. J Arthroplasty. 2007 Sep;22(6 Suppl 2):71–6. Epub 2007 Jul 26.
7. Dowsey MM Liew D Stoney JD Choong PF. The impact of pre-operative obesity on weight change and outcome in total knee replacement: a prospective study of 529 consecutive patients. J Bone Joint Surg Br. 2010 Apr;92(4):513–20.
8. Foran JR Mont MA Rajadhyaksha AD Jones LC Etienne G Hungerford DS. Total knee arthroplasty in obese patients: a comparison with a matched control group. J Arthroplasty. 2004 Oct;19(7):817–24.
9. Foran JR Mont MA Etienne G Jones LC Hungerford DS. The outcome of total knee arthroplasty in obese patients. J Bone Joint Surg Am. 2004 Aug;86-A(8):1609–15.
10. Amin AK Clayton RA Patton JT Gaston M Cook RE Brenkel IJ. Total knee replacement in morbidly obese patients. Results of a prospective, matched study. J Bone Joint Surg Br. 2006 Oct;88(10):1321–6.
11. Stern SH Insall JN. Total knee arthroplasty in obese patients. J Bone Joint Surg Am. 1990 Oct;72(9):1400–4.
12. Deshmukh RG Hayes JH Pinder IM. Does body weight influence outcome after total knee arthroplasty? A 1-year analysis. J Arthroplasty. 2002 Apr;17(3):315–9.
13. Griffin FM Scuderi GR Insall JN Colizza W. Total knee arthroplasty in patients who were obese with 10 years followup. Clin Orthop Relat Res. 1998 Nov;(356):28–33.
14. Spicer DD Pomeroy DL Badenhausen WE Schaper LA Jr Curry JI Suthers KE Smith MW. Body mass index as a predictor of outcome in total knee replacement. Int Orthop. 2001;25(4):246–9.
15. Stickles B Phillips L Brox WT Owens B Lanzer WL. Defining the relationship between obesity and total joint arthroplasty. Obes Res. 2001 Mar;9(3):219–23.
16. Krushell RJ Fingeroth RJ. Primary Total Knee Arthroplasty in Morbidly Obese Patients: a 5- to 14-year follow-up study. J Arthroplasty. 2007 Sep;22(6 Suppl 2):77–80. Epub 2007 Jul 26.
17. Dowsey MM Choong PF. Early outcomes and complications following joint arthroplasty in obese patients: a review of the published reports. ANZ J Surg. 2008 Jun;78(6):439–44.
18. Samson AJ Mercer GE Campbell DG. Total knee replacement in the morbidly obese: a literature review. ANZ J Surg. 2010 Sep;80(9):595–9. doi: 10.1111/j.1445-2197.2010.05396.x. Epub 2010 Aug 19.
19. Dawson J Fitzpatrick R Murray D Carr A. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br. 1998 Jan;80(1):63–9.
20. Dunbar MJ Robertsson O Ryd L Lidgren L. Appropriate questionnaires for knee arthroplasty. Results of a survey of 3600 patients from The Swedish Knee Arthroplasty Registry. J Bone Joint Surg Br. 2001 Apr;83(3):339–44.
21. The Audit Commission. Measuring quality: patient’s view of day surgery. NHS Occasional Papers. London: The Audit Commission, 1991.
22. Smith SC Cano S Lamping DL Staniszewska S Browne J Lewsey J van der Meulen J Cairns J Black N. Patient-Reported Outcome Measures (PROMs) for routine use in Treatment Centres: recommendations based on a review of the scientific evidence. Final report to the Department of Health. 2005 Dec.
25. KAT Trial Group, Johnston L, MacLennan G, McCormack K, Ramsay C, Walker A. The Knee Arthroplasty Trial (KAT) design features, baseline characteristics, and two-year functional outcomes after alternative approaches to knee replacement. J Bone Joint Surg Am. 2009 Jan;91(1):134-41.
26. Winiarsky R Barth P Lotke P. Total knee arthroplasty in morbidly obese patients. J Bone Joint Surg Am. 1998 Dec;80(12):1770–4.
27. Namba RS Paxton L Fithian DC Stone ML. Obesity and perioperative morbidity in total hip and total knee arthroplasty patients. J Arthroplasty. 2005 Oct;20(7 Suppl 3):46–50.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.Copyright 2012 by The Journal of Bone and Joint Surgery, Incorporated