Abstract
Background
Demand for hip arthroscopy (HA) has increased, but shortfalls in HA training may create disparities in care access. This analysis aimed to (1) compare out-of-network (OON) surgeon utilization for HA with that of more common orthopedics sports procedures, including rotator cuff repair (RCR), partial meniscectomy (PM), and anterior cruciate ligament reconstruction (ACLR), (2) compare the HA OON surgeon rate with another less commonly performed procedure, meniscus allograft transplant (MAT), and (3) analyze trends and predictors of OON surgeon utilization.
Methods
The 2013–2017 IBM MarketScan database identified patients under 65 who underwent HA, RCR, PM, ACLR, or MAT. Demographic differences were determined using standardized differences. Cochran-Armitage tests analyzed trends in OON surgeon utilization. Multivariable logistic regression identified predictors of OON surgeon utilization. Statistical significance was set to p < 0.05 and significant standardized differences were >0.1.
Results
410,487 patients were identified, of which 12,636 patients underwent HA, 87,607 RCR, 233,241 PM, 76,700 ACLR, and 303 MAT. OON surgeon utilization increased for HA, rising from 7.98 % in 2013 to 9.37 % in 2017 (p = 0.026). Compared to RCR, PM, and ACLR, HA was associated with higher likelihood of OON surgeon utilization. Usage of ambulatory surgery centers (ASCs) was predictive of higher OON surgeon rates along with procedure year, insurance plan type, and geographic region. HA performed in an ASC was 13 % less likely to have an OON surgeon (p = 0.047).
Conclusion
OON surgeon utilization generally declined but increased for HA. HA was a predictor of OON surgeon status, possibly because HA is a technically complicated procedure with fewer trained in-network providers. Other predictors of OON surgeon status included ASC usage, PPO/EPO plan type, and Northeast geographic region. There is a need to improve access to experienced HA providers—perhaps with prioritization of HA training in residency and fellowship programs—in order to address rising OON surgeon utilization.
Keywords: Hip arthroscopy, Hip, Arthroscopy training, Insurance, Disparity
1. Introduction
The volume of hip arthroscopy (HA) procedures has increased by over 250 % from 2007 to 2011 in the United States (US),1 and these numbers are expected to grow even larger.2 Despite this rise in the volume of HA procedures, the learning curve for HA has been found to be “unexpectedly demanding”.3 HA is a more technically advanced procedure than other, more established sports medicine procedures, such as knee arthroscopy, and more experience is needed to better select patients for whom the procedure is indicated.4 For example, Kautzner et al. found that surgeons performed over 100 cases before seeing a significant decrease in more severe postoperative complications such as avascular necrosis of the femoral head and worsening of osteoarthritis necessitating conversion to total hip arthroplasty.4 However, an analysis of early-career orthopedic surgeons in the US from 2006 to 2015 found that 67.2 % of these surgeons had performed ≤5 HA procedures, and high-volume HA surgeons made up only 6.5 % of American Board of Orthopaedic Surgery (ABOS) candidates but were responsible for 34.6 % of all HA procedures.5 Furthermore, in a survey of 47 sports fellowship directors conducted in 2010, 83 % “reported less emphasis on HA training than on knee and shoulder surgery”.6 In comparison, from 2013 to 2019 the average number of shoulder arthroscopy procedures performed by graduating orthopedic surgery residents was 82.0 and the average number of knee arthroscopies was 124.8.7 This gap in training has not improved in more recent years. An analysis of graduating orthopedic surgery residents’ case log records revealed that the average number of HA procedures did not significantly increase from 2016 to 2020. In fact, a third of graduating residents carried out 2 HA cases or less.8
Thus, the concentration of HA training and experience in a fraction of surgeons could create disparities in access to care, especially as the demand for HA continues to increase.2,9 As a result, patients may need to utilize out-of-network (OON) providers. A recent study examining 549,868 cases of commonly performed elective, inpatient orthopedic procedures found that the rate of OON surgeon billing declined from 2010 to 2018.10 Over this period, 6.7 % of procedures had an OON orthopedic surgeon, and this was relatively lower than the percentage of inpatient services with OON providers, which reached 42 % in 2016.11 There is currently a notable knowledge gap regarding the proportion of patients seeking OON providers for elective sports medicine outpatient procedures, particularly for HA. Furthermore, there is a need for a more rigorous discourse exploring the underlying reasons for this trend if one exists.
With HA being relatively novel, the authors sought to evaluate the accessibility of HA by comparing the rate of OON surgeon billing for HA with the rate of OON surgeon billing for more common elective outpatient sports medicine orthopedic procedures: rotator cuff repair (RCR), partial meniscectomy (PM), and anterior cruciate ligament reconstruction (ACLR). Meniscus allograft transplantation (MAT) is also a procedure that is technically demanding, not commonly performed by the majority of orthopedic providers, and demonstrates a steep learning curve.12, 13, 14 The primary goals of this analysis were to (1) compare OON surgeon utilization for HA with that of more common orthopedics sports procedures, including RCR, PM, ACLR, (2) compare the HA OON surgeon rate with another less commonly performed procedure, meniscus allograft transplant (MAT), and (3) analyze trends and predictors of OON surgeon utilization. The authors hypothesized that the OON surgeon rate for HA would be higher than that of RCR, PM, and ACLR, would be comparable to that of MAT, and would decline over the study period from 2013 to 2017. The authors also hypothesized that facility type, specifically, procedures taking place in an ambulatory surgical center (ASC), would be the strongest predictor of OON surgeon rate.
2. Methods
2.1. Data Source
The IBM MarketScan Commercial Claims and Encounters Database was utilized for our retrospective observational study. The database contains data from roughly 265 million privately insured patients under the age of 65.15 It excludes Medicare, Medicaid, and Tricare patients; however, MarketScan has been previously used in other studies analyzing trends in HA, RCR, PM, ACLR and MAT.15, 16, 17, 18, 19
2.2. Study Population
Patients under 65 undergoing outpatient HA, RCR, PM, ACLR, or MAT procedures were included. Patients were identified using Current Procedural Terminology (CPT) codes, generating an initial population of 658,831 patients in total (see Supplementary Table 1 for CPT codes). Duplicate instances of patient IDs were removed to exclude revision surgeries and account for potential miscoding from the dataset. Cases with missing surgeon network status and missing facility type were excluded from the study population. Patients with a “Region” variable that corresponded to “unknown” region were excluded from regional subanalysis. After these criteria were applied, 248,344 patients were excluded resulting in a final study population of 410,487 patients that consisted of the five final procedure cohorts. Procedure cohorts were then broken down into cohorts containing only claims with OON surgeons (referred to as “procedure-specific OON surgeon cohort”), which were used for subanalyses.
2.3. Study Variables
Categorical variables chosen for analysis from review of prior literature and clinical judgment included procedure type, patient sex, age group, region, insurance plan type, procedure year, and facility type.15,16,20 Patient age groups consisted of the following: 1) 0–17 years of age, 2) 18–34, 3) 35–44, 4) 45–54, and 5) 55+. Regions included for analysis were defined by the “region” variable in the MarketScan database, reflecting the four census regions established by the U.S. Census Bureau: Northeast, Midwest, South, and West. This study was primarily stratified by insurance plan type, which was consolidated into three categories based on similarities between insurance plans to improve clarity and ease of statistical analysis. These three categories were created according to methodology defined by Carbone et al.: 1) high deductible health plan (HDHP) and consumer-driven health plan (CDHP), 2) health maintenance organization (HMO), point of service (POS), and POS with capitation, and 3) preferred provider organization (PPO) and exclusive provider organization (EPO).16 Category 1 was chosen because these plans typically have higher premiums, but also offer patients greater flexibility in selecting their providers, and feature lower copays, deductibles, and coinsurance than category 3 plans. Category 2 plans are cheaper but offer limited OON coverage. Furthermore, patients with category 2 plans can only see specialists after receiving referrals but pay lower deductibles and premiums than those with category 1 or 3 plans. Compared with category 1, category 3 plans include lower premiums and higher patient out-of-pocket expenditures. “Basic/major medical” and “comprehensive” insurance plans were not included in this study due to the fact that the database did not clearly define these plans.16 Procedure years included in analysis were 2013, 2014, 2015, 2016, and 2017. Facility type included outpatient hospital (OH) or ASC because we chose to focus on outpatient surgical settings, which are used to perform the majority of orthopedic sports procedures.
2.4. Statistical Analyses
For each procedure-specific OON surgeon cohort, categorical variables (procedure type, patient sex, age group, region, insurance plan type, procedure year, facility type) were tabulated and described as a total count and percentage. For univariable analysis, standardized differences (Std. Diff.) compared categorical variables between cohorts; a value > 0.1 identified a statistically significant difference in the distribution of covariates.15 The overall OON surgeon rate was calculated by dividing the total number of claims with OON surgeon status by the total number of claims included in this study. The OON surgeon rate for each cohort was calculated by dividing the number of claims with OON surgeon status for each procedure type by the total number of claims within the procedure cohort. The OON surgeon rate was calculated for each procedure cohort over the study period and for each procedure year. Cochran-Armitage analysis was used to analyze whether there was a significant linear trend in the OON surgeon rate from 2013 to 2017 for each procedure type.
Multivariable logistic regression analysis analyzed overall predictors of OON surgeon status for all procedure types combined. Predictors included the aforementioned categorical variables. Multivariable logistic regression analysis also analyzed the HA procedure cohort to determine predictors of HA OON surgeon status. For each predictor category, the likelihood of the surgeon being OON was calculated in comparison to a reference and was reported as an odds ratio with the corresponding 95 % confidence interval and p-value. The reference was chosen by selecting the predictor with the largest number of OON surgeon cases across all procedure cohorts combined. The significance level for Cochran-Armitage analysis and multivariable logistic regression analysis was set to p < 0.05.
3. Results
3.1. Study Population Characteristics
410,487 patients were included: 12,636 undergoing HA; 87,607 undergoing RCR; 233,241 undergoing PM; 76,700 undergoing ACLR; and 303 undergoing MAT. There were 1044 HA patients with OON surgeons; 6629 RCR; 12,824 PM; 6046 ACLR; and 23 MAT (Table 1). Between 2013 and 2017, the number of patients undergoing RCR, PM, ACLR, and MAT generally decreased. Conversely, the number of HA procedures generally increased from 2013 to 2017 (Table 1). The OON surgeon rate over the study period for each procedure type was 8.26 % (1044/12,636) for HA, 7.57 % (6629/87,607) for RCR, 5.50 % (12,824/233,241) for PM, 7.88 % (6046/76,700) for ACLR, and 7.59 % (23/303) for MAT. The overall OON surgeon rate was 6.47 % (26,566/410,487).
Table 1.
Demographic characteristics of patients undergoing hip arthroscopy, rotator cuff repair, partial meniscectomy, ACL reconstruction, or meniscus allograft transplant with an out-of-network surgeon from 2013 to 2017.
| HA (n = 1044) |
RCR (n = 6629) |
PM (n = 12,824) |
ACLR (n = 6046) |
MAT (n = 23) |
||
|---|---|---|---|---|---|---|
| Count (%) | Count (%) | Count (%) | Count (%) | Count (%) | Std. Diff. | |
| Sex | 0.220 | |||||
| Male | 647 (61.97 %) | 2713 (40.93 %) | 5717 (44.58 %) | 2719 (44.97 %) | 13 (56.52 %) | |
| Female |
397 (38.03 %) |
3916 (59.07 %) |
7107 (55.42 %) |
3327 (55.03 %) |
10 (43.48 %) |
|
| Age Group | 1.321 | |||||
| 0-17 | 109 (10.44 %) | 12 (0.18 %) | 402 (3.13 %) | 1487 (24.59 %) | 1 (4.35 %) | |
| 18-34 | 390 (37.36 %) | 122 (1.84 %) | 1189 (9.27 %) | 2669 (44.14 %) | 15 (65.22 %) | |
| 35-44 | 245 (23.47 %) | 686 (10.35 %) | 1833 (14.29 %) | 1024 (16.94 %) | 3 (13.04 %) | |
| 45-54 | 226 (21.65 %) | 2277 (34.35 %) | 4306 (33.58 %) | 663 (10.97 %) | 3 (13.04 %) | |
| 55+ |
74 (7.09 %) |
3532 (53.28 %) |
5094 (39.72 %) |
203 (3.36 %) |
1 (4.35 %) |
|
| Region | 0.306 | |||||
| Northeast | 220 (21.07 %) | 1616 (24.38 %) | 3752 (29.26 %) | 1478 (24.45 %) | 5 (21.74 %) | |
| Midwest | 182 (17.43 %) | 1024 (15.45 %) | 2252 (17.56 %) | 851 (14.08 %) | 5 (21.74 %) | |
| South | 379 (36.3 %) | 3054 (46.07 %) | 5205 (40.59 %) | 2541 (42.03 %) | 6 (26.09 %) | |
| West |
263 (25.19 %) |
935 (14.1 %) |
1615 (12.59 %) |
1176 (19.45 %) |
7 (30.43 %) |
|
| Insurance Plan Type | 0.349 | |||||
| EPO/PPO | 676 (64.75 %) | 4639 (69.98 %) | 9664 (75.36 %) | 4069 (67.3 %) | 15 (65.22 %) | |
| HD/CDHP | 141 (13.51 %) | 660 (9.96 %) | 812 (6.33 %) | 784 (12.97 %) | 3 (13.04 %) | |
| HMO/POS | 186 (17.82 %) | 969 (14.62 %) | 1459 (11.38 %) | 913 (15.1 %) | 2 (8.7 %) | |
| Other |
41 (3.93 %) |
361 (5.45 %) |
889 (6.93 %) |
280 (4.63 %) |
3 (13.04 %) |
|
| Procedure Year | 0.348 | |||||
| 2013 | 181 (17.34 %) | 1464 (22.08 %) | 3410 (26.59 %) | 1507 (24.93 %) | 4 (17.39 %) | |
| 2014 | 194 (18.58 %) | 1359 (20.50 %) | 2841 (22.15 %) | 1332 (22.03 %) | 4 (17.39 %) | |
| 2015 | 180 (17.24 %) | 1449 (21.86 %) | 2782 (21.69 %) | 1148 (18.99 %) | 8 (34.78 %) | |
| 2016 | 238 (22.80 %) | 1232 (18.59 %) | 2034 (15.86 %) | 1123 (18.57 %) | 6 (26.09 %) | |
| 2017 |
251 (24.04 %) |
1125 (16.97 %) |
1757 (13.70 %) |
936 (15.48 %) |
1 (4.35 %) |
|
| Facility Type | 0.157 | |||||
| OH | 719 (68.87 %) | 3624 (54.67 %) | 6783 (52.89 %) | 3596 (59.48 %) | 14 (60.87 %) | |
| ASC | 325 (31.13 %) | 3005 (45.33 %) | 6041 (47.11 %) | 2450 (40.52 %) | 9 (39.13 %) | |
Bold text indicates statistical significance, standardized differences >0.1. Abbreviations: Out-of-Network (OON), Hip Arthroscopy (HA), Rotator Cuff Repair (RCR), Partial Meniscectomy (PM), ACL Reconstruction (ACLR), and Meniscus Allograft Transplant (MAT), Exclusive Provider Organization (EPO), Preferred Provider Organization (PPO), Health Maintenance Organization (HMO), Non-capitated/Capitated/Partially-Capitated Point-of-Service (POS), Consumer-Driven Health Plan (CDHP), High Deductible Health Plan (HDHP), Outpatient Hospital (OH), Ambulatory Surgery Center (ASC), Standardized Differences (Std. Diff.).
When comparing the OON surgeon cohorts via univariable analysis, the following categorical variable distributions significantly differed: patient sex, age group, region, insurance plan type, procedure year, and facility type (Std. Diff. = 0.220, 1.321, 0.306, 0.349, 0.348, and 0.157, respectively) (Table 1). HA had the highest proportion of OON surgeon cases taking place at an OH (68.87 % vs 54.67 % for RCR, 52.89 % for PM, 59.48 % for ACLR, and 60.87 % for MAT). With the exception of MAT, most procedures performed were located in the Southern region (36.3 % for HA, 46.07 % for RCR, 40.59 % for PM, 42.03 % for ACLR, 26.08 % for MAT). For insurance plan type, 64.75 % patients (676/1044) undergoing HA were covered by EPO/PPO plans while the second most popular plan was HMO/POS insurance at 17.82 % (186/1044). Similar to the trends observed for HA, the results showed that EPO/PPO and HMO/POS were the two most popular insurance plans for RCR, PM, and ACLR procedures. Meanwhile, for patients undergoing MAT, the majority had EPO/PPO and HD/CDHP insurance plans.
3.2. Trends in Out-of-Network Surgeon Rates over Time
No significant trend was observed in the OON surgeon rate from 2013 to 2017 for RCR and MAT (p = 0.080 and p = 0.892, respectively). On the other hand, a significant linear trend over time was observed in the OON surgeon rate for HA, PM, and ACLR (p = 0.026, p = 0.0001, p = 0.006). More specifically, the OON surgeon rate increased from 7.98 % in 2013 to 9.37 % in 2017 for HA. The OON surgeon rate decreased from 5.75 % in 2013 to 4.86 % in 2017 for PM, and the OON surgeon rate decreased from 8.47 % in 2013 to 7.11 % in 2017 for ACLR (Fig. 1).
Fig. 1.
Trend in Out-of-Network Surgeon Rate Over Time for Hip Arthroscopy, Rotator Cuff Repair, Partial Meniscectomy, ACL Reconstruction, and Meniscus Allograft Transplant Procedure Cohorts, 2013 to 2017. Cochran-Armitage analysis was used to analyze whether there was a significant linear trend in the OON surgeon rate from 2013 to 2017 for each procedure type: HA, RCR, PM, ACLR, and MAT. * indicates statistical significance (p < 0.05). Abbreviations: Out-of-Network (OON), Hip Arthroscopy (HA), Rotator Cuff Repair (RCR), Partial Meniscectomy (PM), ACL Reconstruction (ACLR), and Meniscus Allograft Transplant (MAT).
3.3. Predictors of Out-of-Network Surgeon Status
Relative to the HA procedure cohort, the ACLR (OR 0.88, 95 % CI 0.82–0.94, p = 0.0003), PM (OR 0.59, 95 % CI 0.55–0.63, p = 0.0001), and RCR (OR 0.85, 95 % CI 0.80–0.92, p = 0.0001) procedure cohorts had decreased likelihood of the primary surgeon being OON on multivariable logistic regression analysis. Overall, procedures billed in the Northeast had higher likelihood of the surgeon being OON than those in the South (OR 1.64, 95 % CI 1.59–1.69, p = 0.0001). On the other hand, the Midwest (OR 0.71, 95 % CI 0.68–0.73, p = 0.0001) and West (OR 0.77, 95 % CI 0.74–0.80, p = 0.0001) regions had lower likelihood of having an OON surgeon than the South region. Other insurance plan types were more associated with having an OON surgeon than EPO/PPO insurance plans (OR 2.40, 95 % CI 2.27–2.54, p = 0.0001). HD/CDHP insurance plans (OR 0.52, 95 % CI 0.49–0.54, p = 0.0001) and HMO/POS insurance plans (OR 0.69, 95 % CI 0.66–0.72, p = 0.0001) were less associated with seeking an OON surgeon compared with EPO/PPO plans. Lastly, having a procedure in the years 2014 (OR 0.85, 95 % CI 0.82–0.88, p = 0.0001), 2016 (OR 0.92, 95 % CI 0.88–0.96, p = 0.0001) and 2017 (OR 0.89, 95 % CI 0.85–0.93, p = 0.0001) resulted in decreased likelihood of being treated by an OON surgeon than those that took place in the year 2013. Procedures in 2015 had higher likelihood of having an OON surgeon than procedures in 2013 (OR 1.24, 95 % CI 1.19–1.29, p = 0.0001). Procedures taking place in an ASC had higher likelihood of having an OON surgeon when compared with those happening in an OH (OR 1.13, 95 % CI 1.10–1.16, p = 0.0001) (Table 2).
Table 2.
Multivariable regression analyses of predictors of out-of-network surgeon status for combined orthopedic sports procedures.
| Variable | Reference Variable | OR (95 % CI) for OON Surgeon | p-value |
|---|---|---|---|
| Procedure Type | |||
| ACL | HA | 0.88 (0.82–0.94) | 0.0003 |
| MAT | HA | 0.86 (0.56–1.33) | 0.506 |
| PM | HA | 0.59 (0.55–0.63) | 0.0001 |
| RCR |
HA |
0.85 (0.8–0.92) |
0.0001 |
| Sex | |||
| Male |
Female |
1.03 (1.00–1.06) |
0.025 |
| Age Group | |||
| 18-34 | 0–17 | 1.03 (0.97–1.09) | 0.364 |
| 35-44 | 0–17 | 0.99 (0.93–1.05) | 0.784 |
| 45-54 | 0–17 | 0.97 (0.91–1.02) | 0.239 |
| 55+ |
0–17 |
0.97 (0.92–1.03) |
0.346 |
| Region | |||
| Northeast | South | 1.64 (1.59–1.69) | 0.0001 |
| Midwest | South | 0.71 (0.68–0.73) | 0.0001 |
| West |
South |
0.77 (0.74–0.8) |
0.0001 |
| Insurance Plan Type | |||
| HD/CDHP | EPO/PPO | 0.52 (0.49–0.54) | 0.0001 |
| HMO/POS | EPO/PPO | 0.69 (0.66–0.72) | 0.0001 |
| Other |
EPO/PPO |
2.40 (2.27–2.54) |
0.0001 |
| Procedure Year | |||
| 2014 | 2013 | 0.85 (0.82–0.88) | 0.0001 |
| 2015 | 2013 | 1.24 (1.19–1.29) | 0.0001 |
| 2016 | 2013 | 0.92 (0.88–0.96) | 0.0001 |
| 2017 |
2013 |
0.89 (0.85–0.93) |
0.0001 |
| Facility Type | |||
| ASC | OH | 1.13 (1.10–1.16) | 0.0001 |
Bold text indicates statistical significance, p < 0.05. Abbreviations: Out-of-Network (OON), Odds Ratio (OR), Confidence Interval (CI), Exclusive Provider Organization (EPO), Preferred Provider Organization (PPO), Health Maintenance Organization (HMO), Non-capitated/Capitated/Partially-Capitated Point-of-Service (POS), Consumer-Driven Health Plan (CDHP), High Deductible Health Plan (HDHP), Anterior Cruciate Ligament Reconstruction (ACLR), Meniscus Allograft Transplant (MAT), Partial Meniscectomy (PM), Rotator Cuff Repair (RCR), Hip Arthroscopy (HA), Outpatient Hospital (OH), Ambulatory Surgery Center (ASC).
3.4. Predictors of Out-of-Network Surgeon Status for Hip Arthroscopy
HA cases billed in the Northeast had higher likelihood of having an OON surgeon when compared to cases billed in the South (OR 1.78, 95 % CI 1.49–2.13, p = 0.0001) on multivariable logistic regression analysis. Additionally, HA cases with the EPO/PPO, HDHP/CDHP, and HMO/POS insurance plan type were more likely to have an OON surgeon compared to HA cases with EPO or PPO insurance (OR 1.65, 95 % CI 1.18–2.33, p = 0.004). HA procedures taking place in 2017 had higher likelihood of OON surgeon status than HA procedures in 2013 (OR 1.24, 95 % CI 1.01–1.52, p = 0.041). Predictors that decreased likelihood of HA OON surgeon status included being located in the Midwest relative to the South (OR 0.61, 95 % CI 0.51–0.73, p = 0.0001), being covered under HD/CDHP insurance plans relative to EPO/PPO (OR 0.78, 95 % CI 0.64–0.94, p = 0.009), and being brought to a ASC facility relative to an OH (OR 0.87, 95 % CI 0.76–1.00, p = 0.047) (Table 3).
Table 3.
Multivariable regression analyses of predictors of out-of-network surgeon status for hip arthroscopy.
| Value | Reference | OR (95 % CI) | p-value |
|---|---|---|---|
| Sex | |||
| Male |
Female |
1.03 (0.9–1.17) |
0.704 |
| Age Group | |||
| 18-34 | 0–17 | 1.1 (0.88–1.37) | 0.426 |
| 35-44 | 0–17 | 1.02 (0.81–1.29) | 0.873 |
| 45-54 | 0–17 | 1.13 (0.89–1.44) | 0.308 |
| 55+ |
0–17 |
0.94 (0.69–1.28) |
0.702 |
| Region | |||
| Northeast | South | 1.78 (1.49–2.13) | 0.0001 |
| Midwest | South | 0.61 (0.51–0.73) | 0.0001 |
| West |
South |
1.02 (0.86–1.21) |
0.795 |
| Insurance Plan Type | |||
| HD/CDHP | EPO/PPO | 0.78 (0.64–0.94) | 0.009 |
| HMO/POS | EPO/PPO | 1.11 (0.94–1.32) | 0.233 |
| Other |
EPO/PPO |
1.65 (1.18–2.33) |
0.004 |
| Procedure Year | |||
| 2014 | 2013 | 0.94 (0.76–1.16) | 0.543 |
| 2015 | 2013 | 1.1 (0.88–1.36) | 0.414 |
| 2016 | 2013 | 1.09 (0.89–1.34) | 0.400 |
| 2017 |
2013 |
1.24 (1.01–1.52) |
0.041 |
| Facility Type | |||
| ASC | OH | 0.87 (0.76–1.00) | 0.047 |
Bold text indicates statistical significance, p < 0.05. Abbreviations: Out-of-Network (OON), Odds Ratio (OR), Confidence Interval (CI), High Deductible Health Plan (HD), Consumer Driven Health Plan (CDHP), Health Maintenance Organization (HMO), Point-of-Service (POS), Hip Arthroscopy (HA), Outpatient Hospital (OH), Ambulatory Surgery Center (ASC).
4. Discussion
The principal findings of this study are as follows. Primarily, consistent with our hypothesis, the OON surgeon utilization rate for HA rose significantly (p = 0.026) from 7.98 % in 2013 to 9.37 % in 2017 and HA was independently associated with a higher likelihood of OON surgeon utilization compared to RCR, PM and ACLR, but not MAT. This is in contrast with ACLR and PM, whose rate of OON surgeon utilization significantly decreased in the same time span. By 2017, the final year included in the analysis, HA had the highest OON surgeon utilization rate compared to RCR, PM, ACLR, and MAT. Second, contrary to our hypothesis, while procedures performed in an ASC were a strong predictor of OON surgeon rate, it was not the only strong predictor and procedure year, insurance plan type, and region were also significant predictors. For HA, having surgery in the central North or Northeast regions of the United States was a significant predictor of OON likelihood as compared to having surgery in the South. Furthermore, having surgery at an ASC was a positive predictor of OON usage compared to an OH surgical setting across all procedure types, but HA procedures taking place in an ASC are 13 % less likely to occur with an OON surgeon (p = 0.047).
Our findings illustrate that between 2013 and 2017, the OON surgeon rates are generally declining for common orthopedic procedures such as ACLR, PM, RCR, and MAT likely due to rising healthcare costs and more widespread Medicaid and Medicare coverage with the Affordable Care Act, a comprehensive healthcare reform law. This finding is consistent with what is currently documented in the literature. A recent 2021 study of over 500,000 patients undergoing total knee arthroplasty, total hip arthroplasty, anterior cervical discectomy and fusion and posterior lumbar fusion found that OON surgeon utilization decreased from 12.7 % to 3.6 % from 2010 to 2018.10 However, when independently examining a more complicated procedure, like HA, which has both a steeper learning curve and fewer providers performing it, the opposite is true. Our study found that HA had the highest rate of OON provider utilization in 2017 compared to each of RCR, PM, ACLR, and MAT, and was the only one of these procedures to exhibit an increase in OON surgeon utilization from 2013 to 2017. Patients may be seeking out high volume, expert hip arthroscopists regardless of whether they are in-network or OON, possibly in pursuit of better results than they believe they would achieve with a lower volume, in-network surgeon.5,6,8 This allows for a few high-volume hip arthroscopists to be disproportionately responsible for meeting the demand for HA. According to a 2017 survey of nearly 7000 American Board of Orthopaedic Surgery candidates, 6.5 % of surgeons accounted for 34.6 % of all the hip arthroscopies performed.5 Surgeons who understand their position as uniquely experienced providers of HA may also be marketing themselves as such to capture a higher percentage of OON patients and to garner higher compensation; for example, Wang et al. found that OON reimbursements were 73 % higher for TKA and THA and 191 % higher for ACDF and PLF compared to in network reimbursements.10 Given that surgeons need to perform at least 100 hip arthroscopies to reach proficiency,4 yet a third of graduating residents performed 2 HA cases or less8 during their 5 years of training, this study evidences that in order to reduce the disproportionate rate of OON utilization for HA, orthopedic residencies in the United States need to increase training in this procedure for their trainees.
One of the significant predictors of OON surgeon utilization was geographic location. The Northeastern region of the US was more associated with OON surgeon utilization compared to the Southern region. This was also true for HA specifically. This geographic finding is consistent with a study of more than 500,000 patients undergoing common orthopedic procedures that found that Northeastern states had double the rate of OON utilization compared to Southern states.10 Another study of HA found that the procedure was most frequently performed in the West21 while yet another identified the Northeast and Northwest as the regions where hip arthroscopy is most frequently performed.6 However, these studies used databases other than IBM Marksetscan, so it is difficult to make accurate comparisons between the respective regional findings as databases use varying geographic definitions for regions. Our findings could be explained by regional differences in median income. A study by Fanelli et al.22 found that in regions with higher poverty levels such as the South, there were fewer foot and ankle surgeons compared to regions with higher median income such as the Northeast. The same trend would likely be true for HA, which is typically performed by specialized sports surgeons rather than generalist orthopedic surgeons. Thus, Southern patients needing HA may only be able to seek care from generalists, while patients in the Northeast have more access to highly specialized HA surgeons who may be more likely to be OON. There may also be cultural differences between the two regions, in that patients in the South are more likely to seek surgical care from in-network providers while patients in the North have a greater proclivity for researching and pursuing a specific, expert surgeon who does a high volume of the relevant surgery. As a result, geographic region is a significant predictor of in network versus OON surgeon utilization for HA.
Other major drivers of OON surgeon utilization found in this study include surgical setting and insurance coverage. Surgical setting has proven to be a salient predictor of patient out-of-pocket expenditures15 and total healthcare expenditures,20,23 but our study introduces it as a significant predictor of OON provider utilization. Our results found that HA was associated with lower likelihood of OON surgeon utilization at an ASC when compared to an OH setting. This could be explained by recent literature which has pointed to significant savings when an ASC is utilized versus an OH for common sports orthopedic procedures like ACLR.20,23, 24, 25, 26 These findings may also suggest that surgeons at an academic OH may be more inclined to be OON versus a private practice surgeon operating at an ASC who may not be as inclined to be OON. With regards to insurance, across all procedures together, PPO and EPO plans were associated with a higher likelihood of OON provider utilization compared to HMO, POS, HDHP, and CDHP plans. Patients with PPO/EPO insurance plans may be more inclined than those with HDHP and CDHP plans to seek OON HA providers due to the lower deductible and overall cost associated with using an OON provider for PPO/EPO plans than HDHP/CDHP. Thus, this study's findings, in the context of the literature, suggest that by increasing in-network provider HA expertise there will be less of a burden on patients to actively seek out OON care.
4.1. Limitations
As a retrospective database study, this study has inherent limitations. First, there is the potential for coding misclassification that could have affected the accuracy of a claims-based database such as IBM MarketScan. Second, the findings from this study can only be generalized to the patients represented by this database, which included commercially insured patients under 65 years of age. Thus, our results may not be applicable to Medicare, Tricare, or Medicaid patients; however, the procedures of interest in this study have been shown to occur in a younger, majority commercial insurance population.15 For example, a study on HA utilizing the New York State Ambulatory Surgery and Services database found that 79.3 % of patients utilized private insurance while only 2.3 % and 6.8 % utilized Medicare and Medicaid, respectively.27 Furthermore, an analysis of HA trends found that the mean patient age was 43.9 years.28 Lastly, we were not able to delineate whether the OON surgeon rate is attributed to many surgeons performing procedures OON or a few high-volume surgeons performing all procedures OON. Future research should strive to elucidate this and provide further information on OON surgeon utilization.
5. Conclusion
OON surgeon utilization generally declined but increased for HA. HA was an independent predictor of OON surgeon status, possibly because HA is a technically complicated procedure with fewer trained in-network providers. Other predictors of OON surgeon status included ASC usage, PPO/EPO plan type, and Northeast geographic region. There is a need to improve access to experienced HA providers—perhaps with prioritization of HA training in residency and fellowship programs—in order to address rising OON surgeon utilization.
Informed Consent
Informed consent was not required as we analyzed a publicly available and anonymized dataset.
Institutional Ethical Committee Approval
Ethics committee approval was not required as for this study as the data is publicly available and anonymized.
Funding Statement
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
CRediT authorship contribution statement
Ashley M. Rosenberg: Conceptualization, Project administration, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. Justin Tiao: Investigation, Methodology, Writing – review & editing. David Kantrowitz: Writing – original draft, Writing – review & editing, Supervision. Timothy Hoang: Writing – original draft, Writing – review & editing, Visualization. Kevin C. Wang: Writing – original draft, Writing – review & editing, Supervision. Nicole Zubizarreta: Methodology, Data curation, Formal analysis. Shawn G. Anthony: Conceptualization, Project administration, Writing – review & editing, Supervision.
Declaration of Competing Interest
None.
Acknowledgements
None.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jor.2023.11.075.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
References
- 1.Sing D.C., Feeley B.T., Tay B., Vail T.P., Zhang A.L. Age-related trends in hip arthroscopy: a large cross-sectional analysis. Arthroscopy. 2015;31(12):2307–2313.e2. doi: 10.1016/j.arthro.2015.06.008. [DOI] [PubMed] [Google Scholar]
- 2.Palmer A.J.R., Malak T.T., Broomfield J., et al. Past and projected temporal trends in arthroscopic hip surgery in England between 2002 and 2013. BMJ Open Sport Exerc Med. 2016;2(1) doi: 10.1136/bmjsem-2015-000082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mehta N., Chamberlin P., Marx R.G., et al. Defining the learning curve for hip arthroscopy: a threshold analysis of the volume-outcomes relationship. Am J Sports Med. 2018;46(6):1284–1293. doi: 10.1177/0363546517749219. [DOI] [PubMed] [Google Scholar]
- 4.Kautzner J., Zeman P., Stančák A., Havlas V. Hip arthroscopy learning curve: a prospective single-surgeon study. Int Orthop. 2018;42(4):777–782. doi: 10.1007/s00264-017-3666-0. [DOI] [PubMed] [Google Scholar]
- 5.Duchman K.R., Westermann R.W., Glass N.A., Bedard N.A., Mather R.C., 3rd, Amendola A. Who is performing hip arthroscopy?: an analysis of the American board of orthopaedic surgery part-II database. J Bone Joint Surg Am. 2017;99(24):2103–2109. doi: 10.2106/JBJS.17.00342. [DOI] [PubMed] [Google Scholar]
- 6.Colvin A.C., Harrast J., Harner C. Trends in hip arthroscopy. J Bone Joint Surg Am. 2012;94(4):e23. doi: 10.2106/JBJS.J.01886. [DOI] [PubMed] [Google Scholar]
- 7.Clark S.C., Miskimin C., Mulcahey M.K. ACGME case volume minimums decrease the number of shoulder and knee arthroscopies performed by residents. Sports Med Arthrosc Rehabil Ther Technol. 2021;3(3):e689–e694. doi: 10.1016/j.asmr.2021.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sudah S.Y., Michel C.R., Nasra M.H., et al. Hip arthroscopy procedural volume is low among graduating orthopaedic surgery residents. Sports Med Arthrosc Rehabil Ther Technol. 2022;4(3):e1179–e1184. doi: 10.1016/j.asmr.2022.04.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sienko A., Ekhtiari S., Khanduja V. The growth of hip preservation as a speciality. Knee Surg Sports Traumatol Arthrosc. 2023;31(7):2540–2543. doi: 10.1007/s00167-023-07409-9. [DOI] [PubMed] [Google Scholar]
- 10.Wang K.Y., Suresh K.V., Harris A., et al. Out-of-Network billing in privately insured patients undergoing elective orthopaedic surgery. J Am Acad Orthop Surg. 2021;29(24):1072–1078. doi: 10.5435/JAAOS-D-21-00081. [DOI] [PubMed] [Google Scholar]
- 11.Sun E.C., Mello M.M., Moshfegh J., Baker L.C. Assessment of out-of-network billing for privately insured patients receiving care in in-network hospitals. JAMA Intern Med. 2019;179(11):1543–1550. doi: 10.1001/jamainternmed.2019.3451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cvetanovich G.L., Yanke A.B., McCormick F., Bach B.R., Jr., Cole B.J. Trends in meniscal allograft transplantation in the United States, 2007 to 2011. Arthroscopy. 2015;31(6):1123–1127. doi: 10.1016/j.arthro.2014.12.020. [DOI] [PubMed] [Google Scholar]
- 13.Smith H.E., Lyons M.M., Patel N.M. Epidemiology of meniscal allograft transplantation at children's hospitals in the United States. Orthop J Sports Med. 2021;9(9) doi: 10.1177/23259671211034877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Choe J.S., Bin S.I., Lee B.S., Kim J.M., Song J.H., Cho H.K. Learning curve for lateral meniscal allograft transplantation: preventing meniscal extrusion. Arthroscopy. 2021;37(11):3326–3334. doi: 10.1016/j.arthro.2021.04.042. [DOI] [PubMed] [Google Scholar]
- 15.Tiao J, Wang K, Herrera M, et al. Hip arthroscopy trends: increasing patient out-of-pocket costs, lower surgeon reimbursement, and cost reduction with utilization of ambulatory surgery centers. Arthroscopy. 2023;39(11):2313–2324.e2. doi: 10.1016/j.arthro.2023.03.027. [DOI] [PubMed] [Google Scholar]
- 16.Carbone A.D., Wang K., Tiao J., et al. Trends in health care expenditures and patient out-of-pocket expenses in primary anterior cruciate ligament reconstruction. Am J Sports Med. 2022;50(10):2680–2687. doi: 10.1177/03635465221107082. [DOI] [PubMed] [Google Scholar]
- 17.Parikh N., Martinez D.J., Winer I., Costa L., Dua D., Trueman P. Direct and indirect economic burden associated with rotator cuff tears and repairs in the US. Curr Med Res Opin. 2021;37(7):1199–1211. doi: 10.1080/03007995.2021.1918074. [DOI] [PubMed] [Google Scholar]
- 18.Block A.M., Eisenberg M.T., Inclan P.M., Nepple J.J. Treatment trends in meniscal pathology in the setting of concomitant ACL injuries in pediatric and young adult patients: an insurance database study. Am J Sports Med. 2022;50(9):2367–2373. doi: 10.1177/03635465221098141. [DOI] [PubMed] [Google Scholar]
- 19.Bonazza N.A., Smuin D.M., Joshi R., et al. Surgical trends in articular cartilage injuries of the knee, analysis of the truven health MarketScan commercial claims database from 2005-2014. Sports Med Arthrosc Rehabil Ther Technol. 2019;1(2):e101–e107. doi: 10.1016/j.asmr.2019.08.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tiao J., Wang K., Carbone A.D., et al. Ambulatory surgery centers significantly decrease total health care expenditures in primary anterior cruciate ligament reconstruction. Am J Sports Med. 2023;51(1):97–106. doi: 10.1177/03635465221136542. [DOI] [PubMed] [Google Scholar]
- 21.Montgomery S.R., Ngo S.S., Hobson T., et al. Trends and demographics in hip arthroscopy in the United States. Arthroscopy. 2013;29(4):661–665. doi: 10.1016/j.arthro.2012.11.005. [DOI] [PubMed] [Google Scholar]
- 22.Fanelli M., Cush C., Zhang H., et al. Geographic distribution of foot and ankle orthopedic surgeons throughout the United States. Foot & Ankle Orthopaedics. 2021;6(2) doi: 10.1177/24730114211003555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cody J.P., Pfefferle K.J., Ammeen D.J., Fricka K.B. Is outpatient unicompartmental knee arthroplasty safe to perform at an ambulatory surgery center? A comparative study of early post-operative complications. J Arthroplasty. 2018;33(3):673–676. doi: 10.1016/j.arth.2017.10.007. [DOI] [PubMed] [Google Scholar]
- 24.Ford M.C., Walters J.D., Mulligan R.P., et al. Safety and cost-effectiveness of outpatient unicompartmental knee arthroplasty in the ambulatory surgery center: a matched cohort study. Orthop Clin N Am. 2020;51(1):1–5. doi: 10.1016/j.ocl.2019.08.001. [DOI] [PubMed] [Google Scholar]
- 25.Toy P.C., Fournier M.N., Throckmorton T.W., Mihalko W.M. Low rates of adverse events following ambulatory outpatient total hip arthroplasty at a free-standing surgery center. J Arthroplasty. 2018;33(1):46–50. doi: 10.1016/j.arth.2017.08.026. [DOI] [PubMed] [Google Scholar]
- 26.Wodowski A.J., Throckmorton T.W., Mihalko W.M., Toy P.C. Anterior supine intermuscular total hip arthroplasty at an ambulatory surgery center versus hospitalization: cost and adverse events. Orthop Clin N Am. 2021;52(3):209–214. doi: 10.1016/j.ocl.2021.03.011. [DOI] [PubMed] [Google Scholar]
- 27.Lu Y., Lavoie-Gagne O., Forlenza E.M., et al. Duration of care and operative time are the primary drivers of total charges after ambulatory hip arthroscopy: a machine learning analysis. Arthroscopy. 2022;38(7):2204–2216.e3. doi: 10.1016/j.arthro.2021.12.012. [DOI] [PubMed] [Google Scholar]
- 28.Schairer W.W., Nwachukwu B.U., McCormick F., Lyman S., Mayman D. Use of hip arthroscopy and risk of conversion to total hip arthroplasty: a population-based analysis. Arthroscopy. 2016;32(4):587–593. doi: 10.1016/j.arthro.2015.10.002. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.