Abstract
Objective
The purpose of this study is to analyze how the largest insurance companies support their medical necessity policies regarding osteochondral allograft transplantation (OCA) and to determine whether the literature they cite in their policies is of a high level of evidence (LOE).
Design
The 10 largest national health insurance companies were identified. Each payer was contacted via phone or email to obtain their coverage policy regarding OCA. For each policy, the medical necessity criteria were recorded, and all cited references were screened. For all references applicable to OCA, the LOE was recorded, and each reference was screened to determine whether they mentioned the specific criteria reported in the policies.
Results
The medical policies for 6 of the 10 national health insurance companies were identified. These 6 policies cited a collective total of 102 applicable references. Most of these studies were an LOE of IV (n = 58, 56.9%) and an LOE of V (n = 18, 17.6%). There were similarities amongst the medical necessity criteria between different commercial payers; however, most criteria were poorly supported by the cited literature.
Conclusions
Our results demonstrate that commercial insurance companies utilize studies that are of a low LOE when justifying their medical necessity criteria. Moreover, these cited studies infrequently support or mention the commercial payers’ criteria. Future studies should continue to explore how well-supported insurance policies are with the goal of potentially increasing access and authorization for well-supported treatment modalities.
Keywords
Introduction
Injury to articular cartilage is often problematic due to a combination of pain, functional impairments, and long-term morbidity secondary to degenerative arthritis.1,2 Unfortunately, these chondral lesions are highly prevalent, having been reported in up to two thirds of patients undergoing knee arthroscopy.3-5 The particular challenge with chondral injuries is in large part due to a limited potential for replication, and therefore healing, of chondrocytes after injury. 6 It is estimated that up to 300,000 chondral-related surgeries occur per year, and although there lacks consensus on the ideal technique, most techniques seek to either stimulate the creation of fibrocartilage or reconstruct the defect with hyaline cartilage.5-7
Osteochondral allograft transplantation (OCA) is a treatment for full-thickness chondral or osteochondral defects and has demonstrated excellent survival rates with a 95% survival at 5 years and 80% survival at 10 years post transplantation. 8 However, survival rates over time have been shown to slowly decrease with 1 systematic review finding an 86.7% survival at 5 years and 67.5% at 20 years. 9 With regard to athletes, it has been estimated that 75% to 88% of athletes are able to return to some level of sports. 10 Alternatives to OCA include chondroplasty, microfracture, and osteochondral autograft transplantation for small lesions (<2 cm) and matrix-induced autologous chondrocyte implantation for larger lesions (>2 cm). Given the high costs of surgical interventions in the United States, most patients require reimbursement through their insurance payers, with many relying on private commercial payers. These commercial payers have made attempts to limit spending primarily through their policies, which determine the medical necessity of a particular intervention. Insurance companies rely on scientific literature to help develop these policies; however, it has been repeatedly demonstrated that the studies that substantiate these policies are often of low level of evidence (LOE).11,12 Austin et al. 12 recently reported that in the commercial payer policies determining coverage for total joint arthroplasty (TJA), only 14.2% of cited references were of level I or II and less than half of all references were applicable for patients who were candidates for TJA.
It has previously been reported that there lacks a consistency in the published criteria among private commercial payers for OCA coverage. 13 These policy criteria vary not only in chondral lesion size but also in other variables including patient body mass index and lesion location. 13 However, there was no determination of the LOE substantiating these policies and their criteria for coverage. Therefore, the purpose of this study is to determine the LOE of the references cited in the policies regarding OCA for the leading commercial payers in the United States and to determine whether these references investigated the criteria that have been established by the payers.
Methods
The 10 national largest commercial private health insurance payers were identified. These payers included Centene Corporation, United Healthcare, Humana, Health Care Service Corporation (HCSC), Cigna, Kaiser Permanente, WellCare Health Plans, The UPMC Health Plan, Anthem, and Aetna. Each payer was contacted either via phone or email to obtain their coverage policies regarding OCA. For each policy, all references related to OCA were collected for a standardized review, completed by 2 independent reviewers. The review collected information including the criteria set by each policy, LOE of each reference, and whether the cited reference investigated the commercial payers’ eligibility criteria for OCA.
To identify the studies relevant to OCA in the references of the coverage documents, reference titles were individually screened for inclusion of “Osteochondral Allograft” OR “Articular Cartilage Allograft” OR “Articular Cartilage Lesions” OR “OCA.” Only studies that evaluated the procedure in the knee joint were included, while studies that investigated similar procedures in other joints including ankle, hip, and shoulder studies were excluded.
For each study, an individual reviewer recorded the LOE and type of study. The categories for type of reference included primary journal article, review, government guidelines/documents, society guideline, website, miscellaneous, and could not be found. The LOE categories included levels I to V, 14 variable, and miscellaneous. Each study was reviewed to see if they discuss or investigate the specific criteria described in individual payer policies. Review articles were considered to have investigated a specific criterion if they had discussed it in any capacity.
Results
Overall
The coverage policies for OCA were obtained from 6 of the 10 (60%) commercial payers that were contacted. The other 4 policies were excluded due to lack of response or no publicly available criteria. These 6 commercial payers included: Centene, United Healthcare, Humana, HCSC, Cigna, and Aetna. Initial screen of the commercial payers cited references for studies related to OCA yielded 102 applicable studies. Of these, 59 (57.8%) were primary journal articles, 27 were reviews or expert opinion (26.5%), 5 (4.9%) were society guidelines, 2 were government reports or guidelines (2.0%), and 9 (4.5%) were categorized as either a website, textbook, or miscellaneous category (Table 1). 12 When delineating by LOE, only 1 study was of LOE I (0.98%), 10 were of LOE II (9.8%), 4 were of LOE III (3.9%), 58 were of LOE IV (56.9%), 18 were of LOE V (17.6%), and the remaining 11 (10.8%) were categorized as variable depending on recommendation or could not be determined ( Table 2 ). The average date of publication for all accessible references was found to be 2012 ± 7 years.
Frequency of the Different Types of Studies That Were Cited by Commercial Payer’s Medical Necessity Policy.
Level of Evidence for the Relevant Literature Cited in Each Commercial Payer’s OCA Policy.
Centene
In their coverage policy, Centene utilized a total of 31 references. Seventeen of these references were applicable to OCA of the knee. Two studies (11.76%) were of an LOE II, 1 (5.88%) of an LOE III, 9 (52.94%) of an LOE IV, and 5 (29.41%) of an LOE V ( Table 2 ). Of these 17, 15 studies were available for full-text analysis. Centene listed 9 specific criteria that dictated whether or not OCA was a medical necessity: (1) age of 18 to 55 years or skeletal maturity if <18, (2) body mass index (BMI) <35 kg/m2, (3) grade III or IV articular cartilage defect, (4) unipolar lesion >2cm2, (5) disabling symptoms unresponsive to conservative treatment for at least 2 months, (6) no evidence of arthritis on tibial surface, (7) normal hyaline cartilage surrounding defect with minimally or absent articular cartilage, (8) normal knee alignment, and (9) not a candidate for knee replacement. These criteria varied with regard to how frequently they were either mentioned or discussed by cited references, ranging from criteria 5 having no references to criteria 8 being referenced by 8 (53.3%) of 15 total references ( Table 3 ). The average date of publication for references found in Centene’s coverage policy was 2016 ± 4 years.
All Medical Necessity Criteria Listed by Each Commercial Payer’s OCA Policy and the Number of Cited References That Mention or Support Each Specific Criterion.
Humana
In their coverage policy, Humana utilizes a total of 38 sources, with only 5 sources with both full texts available and specific to OCA of the knee. Of these, 1 (20%) study was of LOE V, while the remaining 4 (80%) were classified as variable depending on recommendations ( Table 2 ). Humana listed 9 criteria that determined the medical necessity of OCA: (1) BMI less than or equal to 35, (2) knee pain unresponsive to conservative treatment including steroid injection, modification of pain inducing activities, non–steroidal anti-inflammatory drugs (NSAIDS), orthotics, and physical therapy, (3) Outerbridge grade 3 or 4 articular cartilage defects, (4) documentation including radiological or arthroscopic evidence of chondral defect greater than 2 cm2, (5) if a smoker or nicotine user, planning on cessation, (6) no evidence of arthritis on the corresponding tibial surface, (7) normal-appearing hyaline cartilage surrounding defect, (8) normal knee alignment, and (9) skeletally mature to 55 years of age. Of these, the majority of these criteria (n = 5) were discussed by 1 (20%) of the available sources and criteria 8 was mentioned by 2 (40%) of the accessible sources ( Table 3 ). Within Humana’s coverage policy, the average date of publication was found to be 2018 ± 5 years.
Health Care Service Corporation
In their coverage policy, HCSC cites a total of 62 references, of which 19 were applicable to OCA of the knee analysis. Of these 19, 1 (5.26%) was of an LOE I, 3 (15.79%) were of an LOE II, 3 (15.79%) were of an LOE III, 10 (52.63%) were of an LOE IV, 1 (5.26%) of an LOE V, and 1 (5.26%) of a variable level ( Table 2 ). Of these 19, 17 studies were available for full-text analysis. According to their coverage policy, the only specific criteria for OCA of the knee is a full-thickness chondral defect caused by trauma that is not amenable to other repair techniques including microfracture, osteochondral autografting, or autologous chondrocyte implantation. Of the 17 full-text articles obtainable, only 7 (42.18%) studies specifically mentioned this criterion ( Table 3 ). References found in HCSC’s coverage policy were also found to have an average date of publication of 2014 ± 4 years.
Aetna
In their coverage policy, Aetna cites a total of 121 references, of which 24 were accessible and relevant to OCA. Of these, none were of an LOE I, 3 (12.50%) were of an LOE II, none of an LOE III, 13 (54.17%) of an LOE IV, 7 (29.17%) of an LOE V, and 1 (4.17%) unable to be determined. Of these 24, 22 studies were available for full-text analysis. According to their policy, the specific criteria’s for OCA include (1) avascular necrosis lesions of the femoral condyle, (2) osteochondral dissecans lesions of the distal femur, (3) otherwise healthy and non-elderly individuals with osteochondral dissecans lesions who have either failed an arthroscopic procedures or is a non-candidate for arthroscopic procedures, (4) full-thickness depth lesions of 2 cm or more in diameter by imaging or arthroscopy (grade 3 or 4 lesions), (5) surrounded by normal, non-arthritic cartilage, (6) causing disabling localized knee pain that is unresponsive to conservative treatment, (7) either normal knee alignment initially or during allograft surgery, (8) stable and aligned knee with intact meniscus and normal joint space on x-ray, and (9) the opposing articular surface should be generally free of disease or injury. These criteria varied with regard to how frequently they were either mentioned or discussed by cited references from criteria 9 being referenced by 3 (13.64%) of the full-text studies to criterion 4 being mentioned in 13 (59.09%) of sources ( Table 3 ). Aetna references were also found to have an average date of publication of 2008 ± 9 years.
United Healthcare
In their coverage policy, United Healthcare cites a total of 5 relevant studies, of which 1 (20%) was of an LOE IV, 2 (40%) were of an LOE V, and 2 (40%) were of a variable or could not be determined LOE. With regard to allograft transplant of the knee, United Healthcare’s policy discussed the use of osteochondral allograft disks and plugs but reported that these techniques are not proven or medically necessary for treatment of articular cartilage damage. The average publication date of the 5 relevant studies was found to be 2014 ± 11 years.
Cigna
In their coverage policy, Cigna lists a total of 236 references of which 32 are accessible and relevant studies to OCA. Of these, none were of an LOE 1, 2 (6.25%) were of an LOE II, none of an LOE III, 25 (78.13%) of an LOE IV, 2 (6.25%) of an LOE V, 3 (5.26%) of either a variable LOE that was unable to be determined ( Table 2 ). Of these 32, 29 studies were available for full-text analysis. With regard to policy criteria, Cigna lists a total of 12 criteria to establish the medical necessity of OCA including (1) function-limiting pain, (2) a stable knee with intact or reconstructed ligaments and menisci, (3) normal tibial-femoral and/or patellofemoral alignment, (4) grade III/IV articular cartilage defects, (5) a lesion less than or equal to 10 cm2 in size, (6) absence of inflammatory arthritis or other systemic disease affecting the joints, (7) minimal to absent osteoarthritic changes in the surrounding articular cartilage, (8) normal articular cartilage at the lesion border, (9) absence of osteoarthritis or generalized tibial chondromalacia, steroid-induced cartilage, or bone disease, (10) for femoral and patellar chondral lesions, absence of a corresponding “kissing lesion” with a Modified Outerbridge Classification, (11) not a candidate for total knee arthroplasty, (12) BMI of less than 35, and (13) age of 49 years or younger. These criteria varied with regard to how frequently they were either mentioned or discussed by cited references from criterion 12 being referenced by only 1 (3.45%) of studies, while criterion 1 was referenced by 20 (68.97%) of studies. The average date of publication for references found within Cigna’s medical policy was 2009 ± 6 years.
Discussion
Based on the results of this study, private commercial payers rely on low LOE studies for the development of coverage policies related to OCA surgery. The results showed a poor correlation between the cited publications and the coverage criteria used for OCA surgery by the payers. An additional finding was that the definition of conservative treatment of osteochondral defects varies among the commercial payers. Discrepancies were also identified between the results of cited publications and the specific criterion or definition utilized by the insurance policy to determine coverage.
A total of 110 citations were identified among the cited articles included in the coverage policy of 6 commercial payers (Centene, United Healthcare, Humana, Health Care Service Corporation, Cigna, and Aetna) and the majority (~75%) of cited articles were of LOE IV (61%) or V (15.4%) indicating an overall low quality of evidence utilized to shape the coverage policies for OCA. In addition, it is noteworthy that the average age of publication was 2,011.64 ± 7.31 years across the 110 references. These findings may be partly due to the fact that OCA became a more commonly used treatment in the last decade; however, insurance policies should ultimately rely on higher quality and more recent studies as they become more readily available and become more transparent in their determination of medical necessity. This would result in greater applicability of their criteria, more suitable policies, and reinforcement of evidence-based medicine practices.
An important observation in our study was that certain coverage criteria related to the characteristics of cartilage defects were not supported by the cited literature. For example, in the policy of HCSC, which has set the threshold of insurance coverage for OCA if the chondral defect is 2.5 cm2 or greater, erroneously cited the paper by Chiu et al. 15 which states that OCA is recommended for cartilage lesions >2 cm2. Without further clarification of why they designate 2.5 cm2 as their cutoff for coverage, it may be reasonable to conclude that these insurance companies attempt to minimize costs by creating more stringent criteria that result in fewer patients receiving authorization. Furthermore, a study by Tírico et al. 16 showed that lesion size had no effect on post-operative outcomes after OCA transplantation, demonstrating further evidence that was not cited by the insurance payers. Although it may be reasonable for commercial payers to cover OCA only for larger cartilage lesions, their policies should agree with their cited literature.
Perhaps the best example of the incongruence between commercial payers and current literature regards whether the presence of bipolar chondral lesions is a contraindication to OCA. Aetna’s criterion requires that the opposing articular surface shoulder be “free of disease or injury, including no arthritis on the corresponding tibial surface.” Interestingly, only 3 of the 24 studies (12.5%) cited in the insurance policy discuss this topic, and the findings of these studies did not necessarily provide sufficient evidence to be used in the development of insurance coverage criteria.17-19 Specifically, 1 study revealed that the Outerbridge classification of the surrounding or opposing surface cartilage did not correlate with functional outcome, while another study showed that while unipolar grafting was better, bipolar was viable.17,19 The third study that referenced these criteria collected their data in the 1980s and early 1991. 18 More recently, several studies have demonstrated the viability of OCA in the setting of bipolar lesions. Meric et al. 20 demonstrated that while bipolar lesions treated with OCA had higher failure rates than unipolar, those with surviving grafts demonstrated significant clinical improvement as measured by several metrics. Moreover, it has been demonstrated that bipolar OCA in certain settings can act to prolong the time to total knee arthroplasty, which may be particularly beneficial for younger patients who may not otherwise be the best candidates for joint replacement. 21
Based on the above, private insurance providers should perform rigorous and inclusive review of OCA literature when developing policies and frequent policy reviews to reflect a rapidly advancing field. Furthermore, we recommend that insurance providers develop policies where specific coverage criteria or cutoff values related to the characteristics of cartilage defects are based on current scientific evidence in order to develop evidence-based coverage policies that are also clinically applicable. Furthermore, insurance, for example, all 5 commercial payers with publicly obtainable policies write that OCA is only a medical necessity when there is normal knee alignment. Of these 5, only 1 (Aetna) expressed that knee alignment may be corrected during allograft surgery. Although it may be possible that the other commercial payers may approve of both a knee correction and allograft procedure concomitantly, these policies should be updated with language that suggests this possibility, given that allograft transplant is often successful if performed with other procedures including high tibial osteotomy. 22
Our results underscore the reality that defining the surgical indications for OCA is challenging. However, in cases where consensus is not reached, obtaining expert opinion can often be beneficial. Görtz et al. 23 recently utilized the Delphi method to panel 23 experts, all of whom perform more than 30 OCA transplantations per year. In their study, 4 indications for OCA received unanimous support and 1 indication received 94% agreement. 23 However, none of these indications mentioned BMI, smoking history, or age, which are commonly expressed by commercial payers as possible contraindications. This is particularly noteworthy because these factors were also found to be some of the most poorly mentioned criteria in the policies’ referenced literature. One particular indication that received unanimous support was “Systematic autoimmune/inflammatory joint disease is not an absolute contraindication to OCA implantation,” which was a direct contrast to Cigna listing systematic joint disease as a contraindication to OCA in their policy. 23 The inconsistencies across these different commercial policies are partially the result of variability in the OCA literature itself, given the lack of consistency in study design and specification of inclusion and exclusion criteria among the different articles.7,24-27 As such, there is some burden on the scientific community to clearly define their surgical indications for OCA, including the definition of failure of nonsurgical modalities. Conversely, we hope that as the literature continues to further our understanding of the efficacy of OCA, that commercial payers continue to revise and update their policies and provide greater clarity and transparency regarding both the creation and explanations of their medical necessity criteria, such that both patients and providers understand the basis of their medical coverage.
Our study also demonstrated the difficulties in obtaining the proper policies from commercial payers. First, after contacted all commercial payers, we received a response on where to obtain the applicable policies from only 6 (60%) of commercial payers. When searching for these policies, we also observed that multiple versions of policies may exist. For example, United Healthcare has a published document titled “Surgery of the Knee” effective as of January 10, 2023, but we were able to find similar documents regarding osteochondral allografts online published in the year 2013 and 2019. More imperative, however, is the fact that while their document “Surgery of the Knee” lists osteochondral allograft by its appropriate CPT code (27415), their document focuses more heavily on commercial, rather than fresh, osteochondral allografts such as Cartiform or Chondrofix. Moreover, their definition of commercial allograft plugs (Chondrofix) is more consistent with the definition of an osteochondral autograft. These findings may demonstrate that commercial payers may potentially utilize 1 version of a procedure (in this case, commercialized grafts) to decline coverage. It is possible that there is a more recent policy document regarding fresh osteochondral allografts, but our difficulty in verifying and obtaining these documents illustrates the need for commercial payers to make these documents more readily accessible for patients and providers. The inaccuracies we have also highlighted further demonstrate that commercial payers should take more care during the creation of these documents so that their policies reflect accurate information.
This study is limited by several factors, which mainly stem from the lack of access to insurance company information and the discrepancies among definitions used in the coverage policies and scientific literature. Insurance companies that did not provide their policies or we were unable to access them were excluded from this study. It is also possible that there are variations in these policies and that policies have since been updated. Moreover, references cited in the policies were screened manually, and although a thorough attempt was made, it is possible some references were missed. We also included only references that were accessible through our institutional access. Although we included some of the leading commercial payers, our results may not be representative of other insurance companies. The surgical indications for OCA vary among different scientific articles, which creates challenges in the development of insurance coverage criteria. We attempted to highlight the discrepancy between the proposed coverage criteria used by the insurance carrier and the citations used to support those criteria using certain example, but we likely missed mentioning a number of them and/or misjudged the appropriateness of the criteria due to multiple reasons including the evolving surgical practices for OCA, the period of time that the insurance companies were developed, the aging of the population, changes in federal, and/or state health policies, and others. In addition, the analysis of each reference to mention a payer’s criterion can often be subjective in nature. To mitigate this, references were screened by 2 authors.
Footnotes
Acknowledgment and Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: F.A.P. reports a relationship with Exactech Inc. that includes consulting or advisory; reports a relationship with Stryker Orthopaedics that includes consulting or advisory; and reports a relationship with OSSIO Inc. that includes consulting or advisory. J.N.L. reports a relationship with Stryker Orthopaedics that includes speaking and lecture fees and reports a relationship with Innocoll Biotherapeutics N.A. Inc. that includes travel reimbursement. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
