Surgical Referrals for Total Hip Arthroplasty in Primary Care: The Usefulness of the AAOS' Appropriate Use Criteria

Abstract

Background It has been proven that primary care physicians (PCPs) have a lack of clarity regarding the indications for total hip arthroplasty (THA), making the process of surgical referral prone to variability and inconsistency. The American Academy of Orthopaedic Surgeons' (AAOS) Appropriate Use Criteria (AAOS-AUC) is an evidence-based decision support tool that assists clinicians to select for whom treatment should be indicated. The present study aims to compare the rate of THA surgical referrals made by PCPs using the AAOS-AUC tool and the rate of referrals resultant after formal physician's education based on the current osteoarthritis treatment standards.

Materials and Methods Using a crossover design, 22 PCPs evaluated 2 rounds of 10 clinical cases each, generating 440 simulated clinical encounters of patients with hip osteoarthritis. In 220 simulated encounters, the PCPs decided if a surgical referral was appropriate by using the AAOS-AUC tool. On the other 220 simulated encounters, that decision was made using the knowledge acquired after formal education. The rate of surgical referrals generated by both strategies was compared.

Results There was no difference in the rate of surgical referrals comparing simulated encounters using the AAOS-AUC tool (57.3%) versus those using clinical judgment after formal education (62.7%; p = 0.2). Neither were there differences when comparing PCPs who used the AAOS-AUC tool during their initial or second round of cases (60.7% versus 58.8% respectively; p = 0.68)

Conclusion In the hands of PCPs, the AAOS-AUC web tool performs as good as formal physician's education in the process of surgical referrals for THA. It is plausible to consider the AAOS-AUC as a practical decision support tool for patients with hip osteoarthritis seen in primary care.

Level of Evidence Level V.

Keywords

total hip arthroplasty - surgical referral - primary care - Appropriate Use Criteria

Introduction

For patients with hip osteoarthritis (OA) and functional limitations, total hip arthroplasty (THA) is not only a cost-effective[1] but a successful treatment,[2] especially when surgery is performed timely.[3] [4] Pondering that consultations related to hip OA have an annual incidence of 1.4 to 2.1/1,000 person-years in the primary care setting,[5] [6] the role of primary care physicians (PCPs) should be considered fundamental to guarantee appropriate and opportune surgical referrals. Unfortunately, it has been proven that PCPs present variability and “lack of clarity about surgical indications” for THA.[7] Despite remarkable efforts,[8] there is no gold standard for THA referrals amid patients managed in primary care.

As part of its quality programs and guidelines, the American Academy of Orthopaedic Surgeons (AAOS) published the Appropriate Use Criteria (AAOS-AUC) for the management of hip OA.[9] This tool is intended to assist clinicians in deciding for whom a procedure (conservative or surgical) should be indicated by implementing evidence-based guidelines. It can be used as an online tool or through its homologous smartphone application called OrthoGuidelines (AAOS, Rosemont, IL, United States).

Recently, a multidisciplinary research group led by Waugh et al.[7] emphasized “the need for decision support tools to inform PCP-patient decision making regarding referral for TJA [total joint arthroplasty].” On the other hand, in a randomized trial, Østerås et al.[10] demonstrated that a structured care model for the management of hip and knee OA in primary care not only improved the quality of care but also reduced the rate of surgical referrals to orthopedic surgery. In another study by Østerås et al.,[11] PCPs who applied the structured care model were formally educated by receiving oral and written information on recommended OA care, imaging modalities, and information about the appropriate time to refer to an orthopedic surgeon. This successful experience emphasizes that physician's education has a role in primary care to improve OA quality of care and surgical referrals for joint replacement.

We hypothesize that, in the hands of PCPs, the AAOS-AUC tool will not perform inferiorly to formal physicians' education for THA surgical referral. Thus, the present study aims to compare the rate of THA referrals made by PCPs using the AAOS-AUC tool and the rate of referrals resultant after formal physician's education.

Materials and Methods

In March 2020, 30 PCPs were invited to participate in a 2-day course sponsored by our institution focused on the current management of musculoskeletal conditions. During the course, all participants attended a one-hour structured class that covered the contemporary management of hip OA. The class was prepared and dictated by a fellowship-trained adult reconstruction surgeon. The contents included: 1) radiographic diagnosis and classification of hip OA, as well as the morphologic features of femoroacetabular impingement and hip dysplasia; 2) alternatives for conservative management in primary care; 3) surgical indications and results of THA; and 4) the importance of modifiable and non-modifiable risk factors to obtain good surgical results.

AAOS-AUC Tool

The AAOS-AUC tool was developed as part of the quality programs of the AAOS, and it includes the management of hip OA, among other conditions. A multidisciplinary group of experts developed this tool by employing the RAND/University of California at Los Angeles appropriateness method. It is intended to make recommendations on a patient-specific level based on the AAOS clinical practice guidelines. Specifically, the hip OA AUC tool uses five observable/appreciable patient parameters that are classified to create 270 unique, different clinical scenarios.[12] The tool can be found at http://www.orthoguidelines.org/.

According to the scenario, the tool determines the appropriateness of nine different treatments, categorizing them as “appropriate,” “maybe appropriate,” or “rarely appropriate” ([Table 1]). For its purposes, the tool involves the use of clinician expertise and experience.

Table 1

Indication	Classification
Age in years	• Young (approximately < 40) • Middle-aged (approximately 40 to 65) • Elderly (approximately > 65)
Function-limiting pain	• Function-Limiting Pain at Moderate to Long Distances • Function-Limiting Pain at Short Distances • Pain at Rest or Night
Radiographic evaluation	• Minimal OA • Minimal OA with acetabular dysplasia • Minimal OA with FAI • Moderate OA • Severe OA
Limitation in range of otion	• Minimal • Moderate • Severe
Risk of negative outcome	• Modifiable risk factors present • No modifiable risk factors present
Treatments addressed by the tool: 1. Assessment and optimization of risk factors 2. Changes in activity 3. Assistive devices 4. Management of oral medication: non-Opioids (NSAIDs, acetaminophen) or tramadol 5. Intraarticular steroids 6. Physical therapy (as conservative treatment) 7. Arthroplasty 8. Hip preservation surgery 9. Arthrodesis

As stated by the AAOS,[12] “An appropriate treatment for osteoarthritis of the Hip is one for which the treatment is generally acceptable, is a reasonable approach for the indication, and is likely to improve the patient's health outcomes or survival.” One of the nine possible treatments in the tool is “arthroplasty”, which, according to the AAOS, “as a treatment option means total hip arthroplasty, occasionally resurfacing, rarely hemiarthroplasty or hip resection arthroplasty (girdlestone procedure).”[12] For the present study, only THA was considered.

One hour after the structured class, a consecutive series of 20 clinicoradiographic cases of patients older than 65 years with hip OA were presented. All the cases were prepared as a simulated clinical encounter (that is, a patient visiting the PCP's clinic complaining of hip pain, with X-rays available for evaluation). The clinical details included: age, gender, body mass index, comorbidities, level of activity, pain characteristics (location, severity, and functional compromise), as well as a description of the physical examination and hip range of motion. All cases included anteroposterior (AP) pelvis and/or AP and cross-table lateral hip radiographs.

The 20 cases were shown in 2 consecutive rounds of 10. Ninety minutes of lunchtime were assigned between rounds. To reduce the influence of confounders such as level of schooling, previous experience, and digital literacy,[13] we decided to employ a crossover design instead of a parallel design (in which the participants remain on the same branch throughout the study). A crossover design enables the participants to act as their own matched controls and to utilize the AAOS-AUC tool at some point during the study. Consequently, the PCPs were randomly divided into 2 groups: Group 1–15 PCPS who responded to the first ten cases assisted by the online version of the AAOS-AUC ([Figure 1]). This group responded to the second round only using their clinical judgment; Group 2–15 PCPs who responded to the first round employing their clinical judgment and the second round using the online version of the AAOS-AUC. We asked the PCPs, to the best of their efforts, to rely exclusively on the AAOS-AUC tool while using it.

Fig. 1 Screenshot of the online version of the AAOS-AUC tool.

At the end of each case, the participants were asked to respond: 1) Considering all the information given, would you recommend a THA for this patient?; and 2) Is there any other alternative that you would consider for the management of this patient? When using the AAOS-AUC tool, the PCPs only included the recommended treatment when it was deemed “appropriate.” Categories such as “may be appropriate” or “rarely appropriate” were not included. The participants used their smartphones to answer a customized questionnaire through Google Forms (Google LLC, Mountain View, CA, United States). For this project, only the question regarding surgical referral was considered. At the end of the questionnaire, all participants were able to use the AAOS-AUC tool and their clinical judgment in equal proportions when making decisions.

The present study was conceived to demonstrate the non-inferiority of the AAOS-AUC (experimental group) for THA referral when compared to formal education/clinical judgment (standard group). To calculate an adequate sample size, in a pilot study, the first author (a fellowship-trained, high-volume adult reconstruction surgeon with 5 years of experience) evaluated all 20 simulated clinical encounters on 2 different occasions, 3 weeks apart. In the first evaluation, only the clinical judgment was used, leading to a rate of 60% of surgical indications for THA. In the second evaluation, by using the AAOS-AUC online tool, a rate of 70% of surgical indications for THA was obtained. To demonstrate a 10% difference in favor of the experimental group (70% versus 60%), with 80% of power, alpha-error of 5%, and a non-inferiority limit of 10%, a total of 140 simulated clinical encounters (70 per group) were required.

Out of 30 original participants, responses from 26 PCPs were received: 15 from Group 1 and 11 from Group 2. Secondarily, 3 responses from group 2 and 1 response from Group 1 had to be excluded due to incomplete data. After this process, we had 22 PCPs with a complete set of answers (Group 1= 14; Group 2 = 8). Finally, a total of 440 simulated clinical encounters (22 PCPs completing 20 cases) were obtained, a number that was valid according to our sample size calculation. In 220 simulated clinical encounters, the decision for THA referral was made using the AAOS-AUC tool (experimental group). In 220 simulated clinical encounters, the decision for THA referral was made based on clinical judgment supported by formal education (standard group). The rates of THA referral and conservative treatment were compared between the groups using the Fisher exact test (p < 0.05). The Mann-Whitney test was used to compare group characteristics.

Results

The 22 PCPs included had a median of 2 (range: 1 to 10) years in practice. In total, 18 of them (82%) declared that the education received in medical school regarding hip OA was insufficient for their daily practice. No difference between the groups was found when comparing years in practice or satisfaction with education during medical school.

After analyzing the 440 simulated clinical encounters, there was no difference in terms of the rate of surgical referrals when comparing the experimental group using the AAOS-AUC tool (57.3%) and the group using only clinical judgment after education (62.7%; p = 0.2).

In the first round, the rate of surgical referrals made by Group 1 (experimental; 45.7%) and Group 2 (standard; 40%) demonstrated no difference (p = 0.4). During the second round, neither were there differences in the rate of surgical referrals made by Group 1 (standard; 75.7%) and Group 2 (experimental; 77.5%) (p = 0.8).

The overall rates of surgical referral made by Group 1 (60.7%) and Group 2 (58.8%) showed no statistical difference (p = 0.68). [Table 2] shows details regarding group comparisons.

Table 2

	All series
Tool	Referral (n)	%	No Referral (n)	%	Total (n)	Fisher exact test
AAOS-AUC	126	57.3	94	42.7	220	Two-tailed p-value = 0.2844
Education	138	62.7	82	37.3	220
Total (n)	264	60.0	176	40.0	440
	First Round
Tool	Referral (n)	%	No Referral (n)	%	Total (n)	Fisher exact test
AAOS-AUC	64	45.7	76	54.3	140	Two-tailed p-value = 0.4802
Education	32	40.0	48	60.0	80
Total (n)	96	43.6	124	56.4	220
	Second Round
Tool	Referral (n)	%	No Referral (n)	%	Total (n)	Fisher exact test
AAOS-AUC	62	77.5	18	22.5	80	Two-tailed p-value = 0.8693
Education	106	75.7	34	24.3	140
Total (n)	168	76.4	52	23.6	220
	Group 1 versus Group 2
	Referral (n)	%	No Referral (n)	%	Total (n)	Fisher exact test
Group 1	170	60.7	110	39.3	280	Two-tailed p-value = 0.6872
Group 2	94	58.8	66	41.3	160
Total	264	60.0	176	40.0	440

Discussion

By utilizing simulated clinical encounters, the present study demonstrated that, in the hands of PCPs, the AAOS-AUC tool performs as good as formal physician's education in the process of surgical referrals for THA. Therefore, it is plausible to consider the AAOS-AUC a practical decision support tool to be useful in patients with OA of the hip seen in primary care.

The literature on surgical referrals for THA in primary care is not abundant. In a study conducted in the United Kingdom, McHugh et al.[14] reported that only 50% of the patients referred by PCPs to orthopedic surgeons for consideration of hip replacement ended up in surgery within a year. In Canada, where PCPs make most of the referrals for orthopedic surgery, Waugh et al.[7] determined the perceptions of 212 PCPs regarding the clinical indications, contraindications, effectiveness, and risks of total joint replacement, as well as their confidence in referring appropriate patients for joint replacement. The authors[7] found that PCP perceptions were substantially variable in all domains. More relevant, the confidence that PCPs had in deciding for whom to refer TJA was only moderate. The authors[7] concluded that there is a need to define who is an appropriate candidate for joint arthroplasty, and PCPs need guidance regarding surgical referral. A recently published study[8] funded by the Health Technology Assessment program of The National Institute for Health Research in the United Kingdom developed a tool called the Arthroplasty Candidacy Help Engine (ACHE). Using economic modeling, the authors[8] calculated that using the ACHE tool to identify patients with a ≥ 70% chance of a good outcome from THA would result in 13 thousand additional referrals to surgical assessment and around 5 thousand additional arthroplasty procedures in England each year. This strategy would increase the total number of operations by 7%, the total cost of referrals and operations by around £25 million, but would gain 16 thousand quality adjusted life years. Future ACHE tool studies focused on its acceptability among general practitioners and its role in supporting referral decisions are expected.

To the best of our knowledge, the present is the first study evaluating the performance of the AAOS-AUC tool in the hands of clinicians regarding surgical referral for THA. Recently, Riddle and Perera[15] evaluated the structure of the tool using mathematical modeling. They[15] found that the appropriateness of hip arthroplasty in this particular system is driven almost exclusively by age and radiographic severity. It is important to disclose that the AAOS-AUC tool does not include the compromise of quality of life as a variable to be considered during decision making, which, in our opinion, is a disadvantage. The authors[15] declared that the tool is therefore limited, and further validation is needed. Our data support the usefulness of the AAOS-AUC tool in primary care as it demonstrated to be not inferior to formal education in the process of surgical referral. Additionally, in all of our hypothetical clinical cases, we used patients older than 65 years of age, which are categorized as “elderly” in the AAOS-AUC tool, in order to avoid the influence of age in decision making. It is essential to disclose that, in Chile, patients older than 65 years of age who are considered candidates for THA enter an unobstructed pathway to access surgery, with a delay no longer than 8 months.

The present study has some limitations. First and foremost, we used simulated clinical encounters instead of real patients. However, all cases were prepared as if they were patients attending a primary care clinic complaining of hip pain due to OA. We believe that the cases were sufficiently well organized, including relevant information to make clinical decisions as if it was in real life. The second limitation is that the study was conducted only one hour after the physicians took a class, which may, in theory, create a bias, making PCPs more prone to remember the contents provided. This phenomenon is called the carry-over effect in crossover studies.[16]

We believe that the main strength of the present study is the number of clinical encounters evaluated, which was sufficient to support our methodology and results. From our standpoint, the use of crossover design is also a strong point, for it enables the participants to act as their own controls, thus decreasing selection bias. We gave all the participants the opportunity to respond using the AAOS-AUC tool, which may be attractive and engaging. We believe the ninety minutes of lunchtime between rounds was useful to decrease the carry-over effect that may have influenced the responses of those PCPs that first used the AAOS-AUC tool.

Conclusion

The AAOS-AUC tool is a useful instrument to aid PCPs to decide who is an appropriate candidate for THA. According to our data, its performance, measured as the rate of surgical referrals, is not inferior compared to formal physician's education focused on the current standards for the management of hip OA.

Conflicto de Intereses

Los autores no tienen conflicto de intereses que declarar.

• Referencias

• 1 Chang RW, Pellisier JM, Hazen GB. A cost-effectiveness analysis of total hip arthroplasty for osteoarthritis of the hip. JAMA 1996; 275 (11) 858-865
• 2 Learmonth ID, Young C, Rorabeck C. The operation of the century: total hip replacement. Lancet 2007; 370 (9597): 1508-1519
• 3 Mota REM. Cost-effectiveness analysis of early versus late total hip replacement in Italy. Value Health 2013; 16 (02) 267-279
• 4 Mujica-Mota RE, Watson LK, Tarricone R, Jäger M. Cost-effectiveness of timely versus delayed primary total hip replacement in Germany: A social health insurance perspective. Orthop Rev (Pavia) 2017; 9 (03) 7161
• 5 Yu D, Peat G, Bedson J, Jordan KP. Annual consultation incidence of osteoarthritis estimated from population-based health care data in England. Rheumatology (Oxford) 2015; 54 (11) 2051-2060
• 6 Prieto-Alhambra D, Judge A, Javaid MK, Cooper C, Diez-Perez A, Arden NK. Incidence and risk factors for clinically diagnosed knee, hip and hand osteoarthritis: influences of age, gender and osteoarthritis affecting other joints. Ann Rheum Dis 2014; 73 (09) 1659-1664
• 7 Waugh EJ, Badley EM, Borkhoff CM. et al. Primary care physicians' perceptions about and confidence in deciding which patients to refer for total joint arthroplasty of the hip and knee. Osteoarthritis Cartilage 2016; 24 (03) 451-457
• 8 Price A, Smith J, Dakin H. et al. The Arthroplasty Candidacy Help Engine tool to select candidates for hip and knee replacement surgery: development and economic modelling. Health Technol Assess 2019; 23 (32) 1-216
• 9 Rees HW. Management of Osteoarthritis of the Hip. J Am Acad Orthop Surg 2020; 28 (07) e288-e291
• 10 Østerås N, Moseng T, van Bodegom-Vos L. et al. Implementing a structured model for osteoarthritis care in primary healthcare: A stepped-wedge cluster-randomised trial. PLoS Med 2019; 16 (10) e1002949
• 11 Østerås N, van Bodegom-Vos L, Dziedzic K. et al. Implementing international osteoarthritis treatment guidelines in primary health care: study protocol for the SAMBA stepped wedge cluster randomized controlled trial. Implement Sci 2015; 10: 165
• 12 American Academy of Orthopaedic Surgeons. Appropriate Use Criteria for the Management of Osteoarthritis of the Hip 2017
• 13 Mesko B, Győrffy Z. The Rise of the Empowered Physician in the Digital Health Era: Viewpoint. J Med Internet Res 2019; 21 (03) e12490
• 14 McHugh GA, Campbell M, Luker KA. GP referral of patients with osteoarthritis for consideration of total joint replacement: a longitudinal study. Br J Gen Pract 2011; 61 (589) e459-e468
• 15 Riddle DL, Perera RA. Appropriateness and Total Hip Arthroplasty: Determining the Structure of the American Academy of Orthopaedic Surgeons System of Classification. J Rheumatol 2019; 46 (09) 1127-1133
• 16 Sibbald B, Roberts C. Understanding controlled trials. Crossover trials. BMJ 1998; 316 (7146): 1719-1720

Publication History

Received: 12 February 2022

Accepted: 28 September 2022

Article published online:
18 May 2023