Abstract
Background
Higher hospital volume is associated with lower rates of adverse outcomes after revision total joint arthroplasty (TJA). Centralizing revision TJA care to higher-volume hospitals might reduce early complication and readmission rates after revision TJA; however, the effect of centralizing revision TJA care on patient populations who are more likely to experience challenges with access to care is unknown.
Questions/purposes
(1) Does a hypothetical policy of transferring patients undergoing revision TJA from lower-to higher-volume hospitals increase patient travel distance and time? (2) Does a hypothetical policy of transferring patients undergoing revision TJA from lower- to higher-volume hospitals disproportionately affect travel distance or time in low income, rural, or racial/ethnic minority populations?
Methods
Using the Medicare Severity Diagnosis Related Groups 466-468, we identified 37,147 patients with inpatient stays undergoing revision TJA from 2008 to 2016 in the Statewide Planning and Research Cooperative System administrative database for New York State. Revisions with missing or out-of-state patient identifiers (3474 of 37,147) or those associated with closed or merged facilities (180 of 37,147) were excluded. We chose this database for our study because of relative advantages to other available databases: comprehensive catchment of all surgical procedures in New York State, regardless of payer; each patient can be followed across episodes of care and hospitals in New York State; and New York State has an excellent cross-section of hospital types for TJA, including rural and urban hospitals, critical access hospitals, and some of the highest-volume centers for TJA in the United States. We divided hospitals into quartiles based on the mean revision TJA volume. Overall, 80% (118 of 147) of hospitals were not for profit, 18% (26 of 147) were government owned, 78% (115 of 147) were located in urban areas, and 48% (70 of 147) had fewer than 200 beds. The mean patient age was 66 years old, 59% (19,888 of 33,493) of patients were females, 79% (26,376 of 33,493) were white, 82% (27,410 of 33,493) were elective admissions, and 56% (18,656 of 33,493) of admissions were from government insurance. Three policy scenarios were evaluated: transferring patients from the lowest 25% by volume hospitals, transferring patients in the lowest 50% by volume hospitals, and transferring patients in the lowest 75% by volume hospitals to the nearest higher-volume institution by distance. Patients who changed hospitals and travelled more than 60 miles or longer than 60 minutes with consideration for average traffic patterns after the policy was enacted were considered adversely affected. The secondary outcome of interest was the impact of the three centralization policies, as defined above, on lower-income, nonwhite, rural versus urban counties, and Hispanic ethnicity.
Results
Transferring patients from the lowest 25% by volume hospitals resulted in only one patient stay that was affected by an increase in travel distance and travel time. Transferring patients from the lowest 50% by volume hospitals resulted in 9% (3050 of 33,493) of patients being transferred, with only 1% (312 of 33,493) of patients affected by either an increased travel distance or travel time. Transferring patients from the lowest 75% by volume hospitals resulted in 28% (9323 of 33,493) of patients being transferred, with 2% (814 of 33,493) of patients affected by either an increased travel distance or travel time. Nonwhite patients were less likely to encounter an increased travel distance or time after being transferred from the lowest 50% by volume hospitals (odds ratio 0.31 [95% CI 0.15 to 0.65]; p = 0.002) or being transferred from the lowest 75% by volume hospitals (OR 0.10 [95% CI 0.07 to 0.15]; p < 0.001) than white patients were. Hispanic patients were more likely to experience increased travel distance or time after being transferred from the lowest 50% by volume hospitals (OR 12.3 [95% CI 5.04 to 30.2]; p < 0.001) and being transferred from the lowest 75% by volume hospitals (OR 3.24 [95% CI 2.24 to 4.68]; p < 0.001) than non-Hispanic patients were. Patients from a county with a lower median income were more likely to experience increased travel distances or time after being transferred from the lowest 50% by volume hospitals (OR 69.5 [95% CI 17.0 to 283]; p < 0.001) and being transferred from the lowest 75% by volume hospitals (OR 3.86 [95% CI 3.21 to 4.64]; p < 0.001) than patients from counties with a higher median income. Patients from rural counties were more likely to be affected after being transferred from the lowest 50% by volume hospitals (OR 98 [95% CI 49.6 to 192.2]; p < 0.001) and being transferred from the lowest 75% by volume hospitals (OR 11.7 [95% CI 9.89 to 14.0]; p < 0.001) than patients from urban counties.
Conclusion
Although centralizing revision TJA care to higher-volume institutions in New York State did not appear to increase the travel burden for most patients, policies that centralize revision TJA care will need to be carefully designed to minimize the disproportionate impact on patient populations that already face challenges with access to healthcare. Further studies should examine the feasibility of establishing centers of excellence designations for revision TJA, the effect of best practices adoption by lower volume institutions to improve revision TJA care, and the potential role of care-extending technology such as telemedicine to improve access to care to reduce the effects of travel distances on affected patient populations.
Level of Evidence
Level III, prognostic study.
Introduction
THAs and TKAs are considered highly effective and are performed routinely across the United States at hospitals both large and small. However, revision arthroplasty, performed for a variety of reasons, complicates a substantial proportion of these procedures with a revision burden relative to primary TKA or THA of approximately 7% to 12% [29]. Revisions are associated with a high risk of complications, increased length of stay, higher cost, and higher risk of long-term patient morbidity compared to primary total joint arthroplasty (TJA) [7, 12, 33]. Given the harms and costs associated with revision TJA, efforts to minimize its frequency are warranted. The association between higher hospital and surgeon volume and lower rates of adverse outcomes after primary TJA is well established [1, 18, 21, 24, 26, 27, 37, 47]. Similar to the primary TJA setting, this volume-outcomes association also appears to occur in the setting of revision TJA [22, 38, 49]. Given these findings, the centralization of surgical care is one possible strategy to take advantage of the association between improvements in outcomes and higher surgical volumes.
Although centralization of care has not been formally adopted for revision TJA, primary TJA and other fields such as surgical oncology and bariatric surgery have tried different strategies to centralize care to higher-volume centers, with varying degrees of success [2, 4, 20, 23, 30, 36, 43, 45]. Most strategies have used a centers of excellence model, consisting of a combination of volume thresholds at either the hospital and/or surgeon level, standardization of care processes, and collection and reporting of patient outcomes [2, 4, 30, 35, 36, 45]. It is important for any policy that centralizes surgical care to minimize unintended negative effects on the patient population it is addressing. For example, centralizing TJA care may increase patient travel distances, especially in rural populations, or disproportionately reduce surgical volume in hospitals serving patient populations with less access to healthcare, such as low-income patients or racial or ethnic minorities [13, 15, 17, 25, 45, 46]. Centralizing revision TJA care to higher volume hospitals might reduce early complication and readmission rates after revision TJA [22, 38, 48]; however, the impact of centralizing revision TJA care on patient populations with existing inferior access to care is unknown.
We therefore sought to determine: (1) Does a hypothetical policy of transferring patients undergoing revision TJA from lower- to higher-volume hospitals increase patient travel distance and time? (2) Does a hypothetical policy of transferring patients undergoing revision TJA from lower- to higher-volume hospitals disproportionately affect travel distance or time in low income, rural, or racial/ethnic minority populations?
Patients and Methods
Data Source
We used the 2008 to 2016 Statewide Planning and Research Cooperative System (SPARCS) database to study the effect of centralizing revision TJA services on patients undergoing revision TJA [32]. This dataset is an all-payer administrative dataset derived from the billing records of inpatient, outpatient, ambulatory, and emergency surgery episodes from healthcare facilities in New York State [38]. These datasets contain encounter-level information about patient demographics, diagnostic and surgical codes, and services provided, and include patient, hospital, and provider identifiers [38]. The SPARCS database includes observations from facilities certified through Article 28 of the Public Health Law and is managed by the New York State Department of Health, which provides elaborate guidelines for the submission of data (https://www.health.ny.gov/statistics/sparcs/training/). These files were linked to the Annual American Hospital Association Survey database to obtain hospital-level characteristics. We used the SPARCS database for our study because of several advantages relative to other available databases: comprehensive catchment of all surgical procedures in New York State, regardless of payer; each patient has a unique identifier, allowing them to be followed across episodes of care and hospitals in New York State and allowing the tracking of readmissions across hospitals; and New York State has an excellent cross-section of hospital types for TJA, including rural and urban hospitals, critical access hospitals, and some of the highest-volume centers for TJA in the United States [38].
Patient and Hospital Cohort
We queried New York State’s SPARCS using Medicare Severity Diagnosis-related Groups and the ICD-9 and ICD-10 Clinical Modification diagnosis and procedure codes to identify inpatient episodes of revision hip and knee arthroplasty in New York State (n = 37,147) from 2008 to 2016. Revisions with missing or out-of-state patient identifiers representing 10% (3474 of 37,147) of the cohort or those associated with closed or merged facilities representing 0.5% (180 of 37,147) of the cohort were excluded. The final analytic cohort included 33,493 inpatient episodes for 28,756 patients who received care in 147 hospitals (Table 1). The mean number of revision procedures per hospital during the study period was 228.
Table 1.
Characteristics of hospitals performing revision TJA in New York State by volume
| Variable | Quartile 1 (n = 37) | Quartile 2 (n = 37) | Quartile 3 (n = 36) | Quartile 4 (n = 37) | Total group (n = 147) | p valuea |
| Revisions per hospital | 16 (5-29) | 65 (47-74) | 170 (126-229) | 480 (351-645) | 105 (35-258) | |
| Ownership | < 0.001 | |||||
| Government | 49 (18) | 11 (4) | 3 (1) | 8 (3) | 18 (26) | |
| Not-for-profit | 49 (18) | 86 (32) | 94 (34) | 92 (34) | 80 (118) | |
| For-profit | 3 (1) | 3 (1) | 3 (1) | 0 (0) | 2 (3) | |
| Medical school affiliation | 38 (14) | 38 (14) | 61 (22) | 84 (31) | 55 (81) | < 0.001 |
| Geographic location | < 0.001 | |||||
| Rural | 46 (17) | 27 (10) | 11 (4) | 3 (1) | 22 (32) | |
| Urban | 54 (20) | 73 (27) | 89 (32) | 97 (36) | 78 (115) | |
| Bed size | < 0.001 | |||||
| Small (< 200 beds) | 70 (26) | 54 (20) | 36 (13) | 30 (11) | 48 (70) | |
| Medium (200-400 beds) | 27 (10) | 35 (13) | 47 (17) | 22 (8) | 33 (48) | |
| Large (> 400 beds) | 3 (1) | 11 (4) | 17 (6) | 49 (18) | 20 (29) | |
| Total of index revision stays | 633 | 2417 | 6273 | 24,170 | 33,493 |
Data presented as median (range) or % (n); all pairwise tests used a Bonferroni correction.
p values for chi-square tests comparing the overall distribution of the variable across the four hospital groups.
We divided all hospitals performing revision arthroplasty into quartiles based on the mean hospital-level annual revision volume, which was considered to be a reliable indicator of a hospital’s relative expertise in revision THA and TKA [38]. This resulted in four groups of 36 to 37 hospitals each, based on the annual volume of procedures. The hospitals performing the lowest 25th percentile by volume (median [range] of 16 surgeries per hospital [5 to 29]) were classified as quartile 1; those performing in the 25th to 50th percentile by volume (median of 65 per hospital [47 to 74]) were classified as quartile 2; those performing in the 50th to 75th percentile (median of 170 per hospital [126 to 229]) were deemed quartile 3; and the hospitals performing 75th to 100th percentile (median of 480 per hospital [351 to 645]) were categorized as quartile 4 ( Table 1). Quartile 1 hospitals were more likely to be government-owned than higher-volume hospitals. Quartile 4 hospitals were more likely to be medical school–affiliated than hospitals in lower-volume quartiles. Hospitals in quartiles 1 and 2 were more likely to be in rural counties than higher-volume hospitals, and they were more likely to have fewer than 200 beds. Patients receiving care in quartile 1 and quartile 2 hospitals were more likely to be older, nonwhite, Hispanic, and have government insurance (Table 2). Mechanical loosening was more likely to be the revision indication in quartile 4 hospitals than in lower-volume institutions. The median county income was lower in quartile 1 and quartile 2 hospitals than in quartile 3 and quartile 4 hospitals.
Table 2.
Demographic characteristics of patients undergoing revision TJA in New York State
| Variable | Hospital volume | |||||
| Quartile 1 (n = 37) | Quartile 2 (n = 37) | Quartile 3 (n = 36) | Quartile 4 (n = 37) | Total (n = 147) | p value | |
| Index revision stays | 633 | 2417 | 6273 | 24,170 | 33,493 | |
| Age in years | 69 ± 13 | 67 ± 13 | 66 ± 13 | 66 ± 13 | 66 ± 13 | 0.002 |
| Women | 64 (405) | 60 (1451) | 61 (3797) | 59 (14,235) | 59 (19,888) | 0.008 |
| Race | < 0.001 | |||||
| White | 65 (410) | 75 (1821) | 80 (5047) | 79 (19,098) | 79 (26,376) | |
| Black | 22 (141) | 13 (306) | 10 (596) | 11 (3759) | 11 (3759) | |
| Other | 13 (82) | 12 (290) | 10 (630) | 10 (2356) | 10 (3358) | |
| Hispanic ethnicity | 9 (58) | 7 (163) | 5 (334) | 6 (1386) | 6 (1944) | < 0.001 |
| Admission type | < 0.001 | |||||
| Emergency | 31 (196) | 18 (428) | 18 (1136) | 12 (2820) | 14 (4580) | |
| Urgent | 8 (48) | 6 (154) | 4 (243) | 4 (1035) | 4 (1480) | |
| Elective | 61 (389) | 76 (1833) | 78 (4890) | 84 (20,298) | 82 (27,410) | |
| Other | 0 (0) | 0 (2) | 0 (4) | 0 (17) | 0 (23) | |
| Primary payer | < 0.001 | |||||
| Government | 66 (416) | 59 (1418) | 57 (3591) | 55 (13,231) | 56 (18,656) | |
| Private | 29 (181) | 31 (750) | 35 (2180) | 39 (9464) | 37 (12,575) | |
| Other | 6 (36) | 10 (249) | 8 (502) | 6 (1475) | 7 (2262) | |
| Diagnosis category | < 0.001 | |||||
| Dislocation | 16 (104) | 13 (318) | 14 (873) | 13 (3242) | 14 (4596) | |
| Infection | 19 (120) | 15 (361) | 12 (732) | 9 (2105) | 10 (3343) | |
| Mechanical loosening | 17 (107) | 24 (585) | 27 (1682) | 31 (7598) | 29 (9869) | |
| Other mechanical complication | 13 (83) | 21 (512) | 21 (1317) | 22 (5378) | 22 (7293) | |
| Periprosthetic fracture | 11 (68) | 5 (130) | 6 (371) | 6 (1351) | 6 (1910) | |
| Stiffness | 0 (2) | 0 (6) | 0 (29) | 1 (86) | 1 (123) | |
| Unknown diagnosis | 24 (149) | 21 (505) | 20 (1269) | 18 (4410) | 19 (6359) | |
| Number of comorbidities | 3 ± 2 | 3 ± 2 | 3 ± 2 | 3 ± 2 | 3 ± 2 | 0.986 |
| Surgeon annual revision volume | 4.7 ± 6 | 10 ± 12 | 12 ± 9 | 28 ± 19 | 23 ± 18 | < 0.001 |
| County median income (per USD 1000) | 43 ± 9 | 43 ± 7 | 47 ± 9 | 46 ± 9 | 46 ± 9 | < 0.001 |
Data presented as mean ± SD or % (n); all pairwise tests used a Bonferroni correction.
Determining Travel Distance and Time
The initial travel distance was calculated in miles using the patient’s home ZIP code and the ZIP code of the facility where they received their index revision TJA care. When assessing hypothetical policy scenarios that centralize revision TJA care, we chose the nearest available facility by distance, and the distance in miles was recalculated using the hypothetical receiving facilities’ ZIP codes. To account for the possibility of traffic affecting travel times, expected changes to the ideal travel time without traffic while traveling at the speed limit were assessed at the county level. This was performed using the difference between the expected and ideal travel times on Google Maps during rush hour traffic using historical travel patterns from 8 AM at random samples of days (n = 70 different travel routes within a given county and across counties) during the Monday-to-Friday work week versus the ideal travel times [3]. We found that average driving times in city counties and Long Island (Bronx, Kings, Nassau, New York, Queens, Richmond, and Suffolk counties) were approximately 50% longer than ideal times. Driving times through neighboring counties (Rockland and Westchester counties) were approximately 25% longer than ideal times. In upstate counties, ideal travel times were roughly similar to driving times incorporating typical traffic patterns. Travel time was then calculated by assigning a traffic factor based on a patient’s origin and destination counties (either 0.5 for city counties, 0.25 for neighboring counties, or 0 for upstate counties), and by using the following formula: ideal travel time x (origin county traffic factor + destination county traffic factor + 1).
Policy Scenarios
To assess the theoretical impact of centralization of revision TJA care, we divided hospitals into quartiles based on the mean revision TJA volume during the study period [29]. Three hypothetical policy scenarios were examined: (1) transferring patients receiving care in the lowest 25% by volume or quartile 1 hospitals to the nearest higher-volume or quartile 2 to 4 institution by distance; (2) transferring patients in the lowest 50% by volume or quartile 1 and quartile 2 hospitals to the nearest higher-volume or quartile 3 and 4 institution by distance; and (3) transferring patients in the lowest 75% by volume or quartile 1 to 3 hospitals to the nearest higher-volume or quartile 4 institution by distance.
Primary and Secondary Outcomes
The primary outcome of interest was a binary indicator of the increase in travel distance or the increase in travel time, including traffic, for patients undergoing revision TJA under the three different hypothetical policy scenarios. Patients who were transferred to a different hospital than their original treating institution according to a given policy scenario and who experienced an increased travel distance of more than 60 miles or an expected travel time by driving of more than 60 minutes were considered adversely affected with an increased travel burden. Patients who were expected to travel fewer than 60 miles or for less than 60 minutes to the nearest centralized facility under the various policy scenarios were classified as not being affected by increased travel distance. This time was chosen because 95% of patients travel less than 50 miles for their joint replacement care, and prior studies in different areas of surgery suggest that median travel times are approximately 20 to 30 minutes for routine care [13, 25, 26, 46]. Patients who did not have to change hospitals upon adoption of the new policy were considered to have no policy effect, regardless of travel time or distance, and were not included in the analysis of adversely affected patients.
Explanatory Variables
We evaluated patient characteristics that might be associated with an adverse patient travel burden because of facility centralization. These included continuous measures of the patient’s age and number of comorbidities on admission and categorical indicators of race, ethnicity, gender, primary payer, rural versus urban area of residence, and patient income. Race was specified as a categorical variable (white, Black, or other), ethnicity was specified as a categorical variable (Hispanic, non-Hispanic, or unknown), gender was specified as a categorical variable (women or men), primary payer was specified as a categorical variable (government, private, or other), and patient income was specified as a categorical variable (low: median income of a patient ZIP code of residence below the median income for New York State and high: median income of a patient ZIP code of residence above the median county income for New York State).
Ethical Approval
This study used publicly available databases. All patient data in this study were anonymous, and no protected health information was used; therefore, this study was exempt from institutional review board approval.
Statistical Analysis
The unit of analysis was the inpatient stay. Unadjusted descriptive statistics compared hospital and patient characteristics across hospital quartiles using chi-square tests for categorical variables and Kruskal-Wallis tests for continuous variables. Chi-square tests were also used for pairwise comparisons of categorical variables, and Mann-Whitney U tests were used for pairwise comparisons of continuous variables. We estimated multivariable logistic regression models to determine the association of patient travel burden with increased travel distance or travel time more than 60 minutes and patient demographics for each policy. The outcome was given as an increasing categorical value based on an increasing adverse travel burden, and explanatory parameters were added to the model as a simple linear combination. All statistical analyses were performed using Stata 16 (StataCorp LLC).
Sensitivity analyses were conducted to examine the impact of using 30 miles and 30 minutes as cutoffs for increased travel distance and time, respectively, on the explanatory variables listed above and to examine travel distance and time as a continuous variable.
Results
Effect of Centralization on Patient Travel Distance and Travel Time
A hypothetical policy of transferring patients undergoing revision TJA from lower- to higher-volume hospitals had an increasing but overall small effect on patient travel distance and time as the proportion of patients referred increased. Transferring patients from the lowest 25% by volume hospitals relocated 2% (633 of 33,493) of patients, with only 0.1% (1 of 633) of patients affected by increased travel distance or travel time (Table 3). Transferring patients away from the lowest 50% by volume hospitals moved 9% (3050 of 33,493) of patients, and 10% (312 of 3050) of patients were affected by either increased travel distance or travel time (Fig. 1). The median travel distance was longer in patients with increased travel burden both before (6 miles [interquartile range 3 to 16] in not-affected patient visits versus 16 miles [IQR 4 to 34] in affected patient visits before transferring patients away from the lowest 50% by volume hospitals; p < 0.001) and after implementation of policy transferring patients away from the lowest 50% by volume hospitals (7 miles [IQR 4 to 17] versus 95 miles [IQR 75 to 112] in not-affected versus affected patient visits; p < 0.001). Travel times followed a similar pattern.
Table 3.
Patient visit travel distances and travel times before and after each policy scenario (n = 33,493)
| Transferring patients from lowest 25% by volume hospitals | Transferring patients from lowest 50% by volume hospitals | Transferring patients from lowest 75% by volume hospitals | ||||||||||
| Demographic variable | Did not change hospitals under given policy | Travel distance < 60 minutes or 60 miles | Travel distance > 60 minutes or 60 miles | p value | Did not change hospitals under given policy | Travel distance < 60 minutes or 60 miles | Travel distance > 60 minutes or 60 miles | p value | Did not change hospitals under given policy | Travel distance < 60 minutes or 60 miles | Travel distance > 60 minutes or 60 miles | p value |
| Number of affected revision stays | 98 (32,860) | 2 (632) | 0 (1) | NA | 91 (30,443) | 8 (2738) | 1 (312) | NA | 72 (24,170) | 25 (8509) | 2 (814) | NA |
| Median travel distance in miles before policy | 13 (6-31) | 4 (2-10) | 10 (NA) | < 0.001 | 14 (6-32) | 6 (3-16) | 16 (4-34) | < 0.001 | 15 (7-35) | 8 (4-18) | 21 (6-36) | < 0.001 |
| Median travel distance in miles after policy | 13 (6-31) | 5 (3-17) | 60 (NA) | < 0.001 | 14 (6-32) | 7 (4-17) | 95 (75-112) | < 0.001 | 15 (7-35) | 12 (5-20) | 78 (72-127) | < 0.001 |
| Median travel time in minutes before policy | 23 (12-46) | 9 (5-19) | 20 (NA) | < 0.001 | 24 (12-47) | 12 (7-27) | 26 (14-56) | < 0.001 | 25 (13-51) | 15 (9-29) | 31 (14-57) | < 0.001 |
| Median travel time in minutes after policy | 23 (12-46) | 11 (7-32) | 82 (NA) | < 0.001 | 24 (12-47) | 14 (10-29) | 134 (110-141) | < 0.001 | 25 (13-51) | 21 (11-33) | 110 (95-154) | < 0.001 |
Data presented as % (n) or median (IQR); NA = not enough affected patients to perform a comparison; IQR = interquartile range.
Fig. 1.
A-B (A) This heat map illustrates the distribution of patients at the county level in New York State who had an adverse travel burden of more than 60 miles or 60 minutes if revision TJA was centralized to the highest 50% by volume hospitals. (B) The New York City region is shown with magnification.
Transferring patients away from the lowest 75% by volume hospitals moved 28% (9323 of 33,493) of patients, with 9% (814 of 9323) of patients affected by either increased travel distance or travel time ( (Fig. 2). The median (IQR) travel distance was longer in patient visits affected by increased travel distance both before (8 miles [IQR 4 to 18] in not-affected patient visits versus 21 miles [IQR 6 to 36] in affected patient visits; p < 0.001) and after this policy implementation (12 miles [IQR 5 to 20] versus 78 miles [IQR 72 to 127] in not- versus affected patient visits; p < 0.001) (Table 3). Travel times followed a similar pattern.
Fig. 2.
A-B (A) This heat map illustrates the distribution of patients at the county level in New York State who had an adverse travel burden of more than 60 miles or 60 minutes if revision TJA was centralized to the highest 75% by volume hospitals. (B) The New York City region is shown with magnification.
Effect of Centralization on Patient Demographic Characteristics
A hypothetical policy of transferring patients undergoing revision TJA disproportionately affected white patients, Hispanic patients, patients without governmental or private insurance, and patients who reside in rural or low-income counties (Table 4). Only one patient visit was affected with an increased travel distance or travel time with transferring patients from the lowest 25% by volume hospitals, so this patient was excluded from further analyses. Nonwhite patients were less likely to encounter an increased travel distance or time when transferred from the lowest 50% by volume hospitals to a higher volume hospital than white patients (odds ratio 0.3 [95% CI 0.2 to 0.7]; p = 0.002) (Table 5). When transferred from the lowest 50% by volume hospitals to a higher-volume hospital, Hispanic patients were more likely to experience an increased travel distance or time than non-Hispanic patients (OR 12 [95% CI 5 to 30]; p < 0.001) (Table 5). When transferred from the lowest 50% by volume hospitals to a higher-volume hospital, patients from a county with a lower median income were more likely to experience increased travel distances or time than those from a county with a higher median income (OR 70 [95% CI 17 to 283]; p < 0.001) (Table 5). When transferred from the lowest 50% by volume hospitals to a higher-volume hospital, patients from rural counties were more likely to experience increased travel distance or time than those from urban counties (OR 98 [95% CI 50 to 192]; p < 0.001) (Table 5). When transferred from the lowest 50% by volume hospitals to a higher volume hospital, patients without government or private insurance were more likely to experience an increased travel distance or time (Table 5).
Table 4.
Demographic characteristics of visits both affected and unaffected by increased travel time in a given policy scenario by either increased travel distance or travel time greater than 60 minutes
| Transferring patients from lowest 50% by volume hospitals | Transferring patients from lowest 75% by volume hospitals | |||||
| Demographic variable | Unaffected patients | Affected patients | p value | Unaffected patients | Affected patients | p value |
| % of affected revision stays | 90 (2738 of 3050) | 10 (312 of 3050) | 91 (8495 of 9323) | 9 (828 of 9323) | ||
| Age in years | 67 ± 13 | 68 ± 13 | 0.44 | 67 ± 13 | 69 ± 12 | 0.005 |
| Women | 62 (1694) | 52 (162) | < 0.001 | 61 (5176) | 58 (477) | 0.046 |
| Race | < 0.001 | < 0.001 | ||||
| White | 71 (1931) | 96 (300) | 76 (6479) | 96 (799) | ||
| Nonwhite | 29 (807) | 4 (12) | 24 (2016) | 4 (29) | ||
| Ethnicity | < 0.001 | < 0.001 | ||||
| Not Hispanic | 86 (2355) | 70 (217) | 90 (7656) | 79 (656) | ||
| Hispanic | 6 (175) | 15 (46) | 6 (507) | 6 (48) | ||
| Unknown | 8 (208) | 16 (49) | 4 (332) | 15 (124) | ||
| Primary payer | < 0.001 | < 0.001 | ||||
| Private | 31 (848) | 27 (83) | 34 (2870) | 29 (241) | ||
| Government | 61 (1666) | 54 (168) | 58 (4932) | 60 (493) | ||
| Other (such as self or workers compensation) | 8 (224) | 20 (61) | 8 (693) | 11 (94) | ||
| Number of comorbidities | 3 ± 2 | 3 ± 2 | 0.65 | 3 ± 2 | 3 ± 2 | 0.56 |
| Admission type | < 0.001 | < 0.001 | ||||
| Elective | 74 (2018) | 64 (204) | 76 (6466) | 78 (646) | ||
| Emergency or trauma | 22 (589) | 11 (35) | 20 (1685) | 9 (75) | ||
| Urgent | 5 (129) | 23 (73) | 4 (338) | 13 (107) | ||
| Other | 0 (2) | 0 (0) | 0 (6) | 0 (0) | ||
| Patient ZIP code median incomea | < 0.001 | < 0.001 | ||||
| High | 55 (1500) | 1 (2) | 61 (5167) | 25 (207) | ||
| Low | 45 (1238) | 99 (310) | 39 (3328) | 75 (621) | ||
| Patient geographic location | < 0.001 | < 0.001 | ||||
| Urban | 85 (2315) | 4 (11) | 91 (7732) | 40 (332) | ||
| Rural | 15 (423) | 96 (301) | 9 (763) | 60 (496) | ||
Data presented as % (n) or mean ± SD. aHigh: > median income for New York State (USD 43,100); low: < median income for New York State.
Table 5.
Odds ratios from multivariable logistic regression models examining the demographics of patients with increased travel distances or times under different closure policy scenarios
| Transferring patients from lowest 50% by volume hospitals | Transferring patients from lowest 75% by volume hospitals | |||
| Demographic variable | OR (95% CI) | p value | OR (95% CI) | p value |
| Age | 1.0 (0.9-1.0) | 0.31 | 1.0 (1.0-1.0) | < 0.03 |
| Gender | ||||
| Women | ref | ref | ||
| Men | 1.2 (0.9-1.7) | 0.26 | 1.0 (0.9-1.2) | 0.82 |
| Race | ||||
| White | ref | ref | ||
| Nonwhitea | 0.3 (0.2-0.7) | 0.002 | 0.1 (0.1-0.2) | < 0.001 |
| Ethnicity | ||||
| Not Hispanic | ref | ref | ||
| Hispanic | 12 (5-30) | < 0.001 | 3.2 (2.2-4.7) | < 0.001 |
| Unknown | 3.3 (1.9-5.8) | < 0.001 | 8.9 (6.6-12) | < 0.001 |
| Primary payer | ||||
| Private | ref | ref | ||
| Government | 1.0 (0.7-1.5) | 0.95 | 1.0 (0.8-1.2) | 0.99 |
| Other (such as self or workers compensation) | 3.5 (1.9-6.5) | < 0.001 | 1.7 (1.2-2.2) | < 0.001 |
| Comorbidities | 1.0 (0.9-1.1) | 0.90 | 1.0 (1.0-1.1) | 0.89 |
| Patient ZIP code median incomeb | ||||
| High | ref | ref | ||
| Low | 70 (17-283) | < 0.001 | 3.9 (3.2-4.6) | < 0.001 |
| Patient geographic location | ||||
| Urban | ref | ref | ||
| Rural | 98 (50-192) | < 0.001 | 12 (10-14) | < 0.001 |
Other races include Asian, Native American, Pacific Islander, and unknown.
High: > median income for New York State (USD 43,100); low: < median income for New York State.
When transferred from the lowest 75% by volume hospitals to a higher volume hospital, nonwhite patients were less likely to encounter an increased travel distance or time than white patients (OR 0.1 [95% CI 0.1 to 0.2]; p < 0.001) (Table 5). When transferred from the lowest 75% by volume hospitals to a higher volume hospital, Hispanic patients were more likely to experience an increased travel distance or time than non-Hispanic patients (OR 3.2 [95% CI 2.2 to 4.7]; p < 0.001) (Table 5). When transferred from the lowest 75% by volume hospitals to a higher volume hospital, patients from a county with a lower median income were more likely to experience increased travel distances or time than those from a county with a higher median income (OR 3.9 [95% CI 3.2 to 4.6]; p < 0.001) (Table 5). When transferred from the lowest 75% by volume hospitals to a higher volume hospital, patients from rural counties were more likely to experience increased travel distance or time than those from urban counties (OR 12 [95% CI 10 to 14]; p < 0.001) (Table 5). When transferred from the lowest 75% by volume hospitals to a higher volume hospital, patients without government or private insurance were more likely to experience an increased travel distance or time (Table 5). Using 30 miles and 30 minutes for defining increased travel distance and time cutoffs, respectively, did not substantially affect prior conclusions (Appendix Table 1; http://links.lww.com/CORR/A698). Additionally, using travel distance and travel time as continuous variables did not affect prior conclusions (Appendix Table 2; http://links.lww.com/CORR/A699).
Discussion
Higher-volume hospitals have lower complication and readmission rates after revision TJA; however, policies to centralize revision TJA to higher-volume hospitals may create an increased travel burden for patients and exacerbate existing disparities in care [22, 38, 48]. Our results suggest that hypothetical centralization policies in New York State that involve directing patients away from lower-volume institutions would not result in increased travel distances and times more than 60 miles or 60 minutes, respectively, for most patients, and even centralizing patients from the lowest 75% by volume hospitals to higher-volume institutions would result in less than 10% of patients having to transfer to a new hospital, increasing their travel times or distances. Despite the lack of an impact on a broad range of the population, we found that increasingly centralized policies may disproportionately increase travel distances for white, Hispanic, rural, and lower-income patients. As a result, policies that centralize revision TJA care will need to be carefully designed in partnership with receiving institutions, payers, and governmental organizations to minimize the effects on patient populations with existing challenges in access to care such as those in rural areas, low-income patients, and ethnic minorities.
Limitations
We acknowledge several limitations to this study. First, the SPARCS database only contains care obtained in New York State, so patients traveling out of the state for care could not be assessed. The effect of centralization policies was the most prominent for rural areas in the periphery of New York State, and patients from these areas might be less affected if they could travel into neighboring states for revision TJA care, thus overestimating the impact of our centralization policies on these populations. Second, confirmation of our results in other states with different hospital populations (increased or decreased numbers of teaching hospitals that take care of revision THA and TKA), population densities (high-population-density coastal states versus more rural mountain or Midwestern states), and underlying patient demographics would be warranted. Third, we could not assess whether other access-to-care issues besides hospital distance (such as patient insurance limitations or transportation issues) exist, which may underestimate the impact of travel distance. Fourth, the SPARCS dataset is based on administrative coding, which has been shown to be accurate relative to the medical record [6] and is audited for accuracy by New York State; however, the impact of incorrect coding cannot be directly measured particularly with regard to the initial diagnosis at revision. Fifth, physicians may work at both high- and low-volume hospitals, and the effect of surgeon or process of care improvements at a lower-volume institution may decrease the benefit of patient transfer to a higher-volume institution. Further studies should continue to examine these issues.
Effect of Centralization on Patient Travel Distance and Travel Time
A hypothetical policy of transferring patients undergoing revision TJA from lower-volume to higher-volume hospitals had an increasing but relatively small effect on patient travel distance and time as the proportion of referred patients increased. These findings suggest that centralization policies for revision TJA may be achievable without placing an unreasonable travel burden on most patients seeking care. The goal of centralizing revision TJA services would be to provide value-based care by decreasing patient complication/reoperation rates and controlling episode of care costs, which have been shown in both primary and revision TJA to be lower as hospital and/or surgeon volume increases [1, 18, 19, 21- 24, 26, 27, 37, 38, 46, 49]. Different policy initiatives have been used through both governmental and private insurance programs with varying degrees of success to centralize surgical care. The most common example is the centers of excellence model (COE). These models typically include a combination of strict hospital and/or surgical volume cutoffs along with a standardization of care processes and outcomes collection/reporting with the goal of either formally or informally directing patients to higher performing centers [5, 14, 30, 31, 48]. The performance of COE designations for elective surgical care is controversial. A COE model from the Blue Cross Blue Shield Association used for primary THA reduced complication rates for patients undergoing THA compared with nondesignated hospitals [30]. In bariatric surgery, however, a 2006 policy that required a COE designation for Medicare reimbursement did not improve outcomes relative to nondesignated hospitals, and both groups of hospitals reduced their surgical complication rates over time resulting in policy changes away from these strict requirements [14]. Similarly, a COE model for spine surgery did not lower complication rates, 30-day readmission rates, or 90-day costs relative to nondesignated hospitals [31]. The use of a value-driven COE model that incorporates cost may provide a better definition of a COE as opposed to prior models; for example, a value-driven COE definition for spine surgery by a commercial insurer showed decreased complication rates relative to nondesignated facilities [48]. Other potential models to centralize elective surgical care at the payer level include the use of reference pricing [40]. The goal of reference pricing is to encourage patients to seek low-cost, high-value care by creating a predefined contribution limit from the employer or insurer with the patient being responsible for the difference in care cost [40]. For example, the use of reference pricing in the California Public Employees’ Retirement System (CalPERS) decreased the utilization of high-priced hospital facilities, increased the use of ambulatory surgery centers, and decreased overall costs for knee and shoulder arthroscopy [39]. Similarly, the use of reference pricing in the CalPERS system increased the selection of lower priced hospitals for primary TJA and reduced costs for health insurers without compromising quality of care [9]. Another possible policy intervention would be the creation of a revision TJA bundled payment. Bundled payment initiatives encourage facilities to control their costs and reduce patient complications/readmissions by providing a single payment to cover the episode-of-care. A Bundled Payment for Care Improvement initiative for Medicare patients undergoing revision TJA resulted in a shorter length of stay for patients involved in the program but did not ultimately lower episode-of-care costs relative to patients not enrolled in the program [11]. A more heterogeneous patient and surgical complexity relative to primary TJA, increased cost outliers, and increased implant costs make bundled payments for revision TJA a continued challenge [11, 16]. In summary, our findings suggest that it may be feasible to evaluate COE models or alternative payment models for revision TJA that incorporate volume thresholds in combination with other cost and quality initiatives without creating excessive travel burden for the population of patients seeking revision TJA care.
Effect of Centralization on Patient Demographic Characteristics
A hypothetical policy of transferring patients undergoing revision TJA disproportionately affected white patients, Hispanic patients, patients without governmental or private insurance, and patients who reside in rural or low-income counties. A similar result was found for primary TJA, where centralizing patients to higher-volume institutions would hypothetically increase travel time for patients in rural areas and result in the transfer of more racial/ethnic minorities and lower-income patients to higher-volume institutions in urban areas because of the disproportional numbers of lower-volume institutions serving these populations [17]. Similarly, prior studies examining centralization of surgical oncology procedures resulted in increased travel times for patients in rural areas, although most patients had less than a 10-mile increase [45]. In contrast, Medicare’s COE program for bariatric surgery did not affect access to care for nonwhite or rural patients, despite increased travel distances relative to non-Medicare patients, suggesting that each centralization program needs to be evaluated individually for its effects on different demographic populations [25]. One concern with increased travel distances might be an association with increasing complication/readmission rates; however, prior studies have not found a link between travel distance and complications in primary TJA [34, 44]. Additionally, increased travel times impose a cost on patients, and for primary TJA, patients were willing to pay an extra USD 11.45 out of pocket to avoid traveling per mile for arthroplasty care, which means patient preferences will need to be further evaluated [42]. Careful planning of centralization may mitigate the effects of travel distance on patients; for example, for pancreaticoduodenectomy, a centralized care model design predicted that all patients in California could be directed to a high-volume institution without creating additional travel burdens [13]. One possibility to mitigate the effects of centralization on access to care for rural populations and certain demographic groups is to improve the coordination of care across a region. For instance, standardized care processes may have a greater impact on patient outcomes than other factors such as strict volume cutoffs used by previous COE designations, suggesting that process improvements adopted by higher-volume hospitals may be transferrable to lower-volume hospitals and may result in improved patient outcomes while reducing the need for patient transfers [8]. Emerging technology such as telemedicine may allow higher-volume practitioners and hospitals to assist in the care of patients in lower-volume institutions both through direct care and staff training of best practices. For example, in both orthopaedic and general surgery care, initial clinic visits and follow-up visits have been successfully performed through telemedicine even in complex trauma or rural patient populations [28, 41]. Additionally, structured mentorship programs have been successful in training lower-volume, rural surgeons to safely perform complex procedures such as periacetabular osteotomy, suggesting that similar access to higher-volume revision TJA surgeons at lower-volume institutions may allow more patients to receive care close to their home [10].
Conclusion
Although centralizing revision TJA care to higher-volume institutions in New York State did not appear to result in an increased travel burden in most patients, policies that centralize revision TJA care will need to be carefully designed to minimize the disproportionate impact on patient populations that already face challenges with access to healthcare. Further studies should examine: the feasibility of establishing COE designations for revision TJA, the effect of best practices adoption by lower-volume institutions on revision TJA care, and the potential role of care-extending technology such as telemedicine to improve access to care to reduce the effects of travel distances on affected patient populations.
Footnotes
Each author certifies that there are no funding or commercial associations (consultancies, stock ownership, equity interest, patent/licensing arrangements, etc.) that might pose a conflict of interest in connection with the submitted article related to the author or any immediate family members.
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research® editors and board members are on file with the publication and can be viewed on request.
Ethical approval for this study was waived by the University of Rochester Institutional Review Board.
Contributor Information
Gabriel Ramirez, Email: gabriel_ramirez@urmc.rocheste.edu.
Thomas G. Myers, Email: thomas_myers@urmc.rochester.edu.
Caroline P. Thirukumaran, Email: caroline_thirukumaran@urmc.rochester.edu.
References
- 1.Badawy M, Espehaug B, Indrekvam K, Engesæter LB, Havelin LI, Furnes O. Influence of hospital volume on revision rate after total knee arthroplasty with cement. J Bone Joint Surg Am. 2013;95:e131. [DOI] [PubMed] [Google Scholar]
- 2.Bae J, Shade J, Abraham A, et al. Effect of mandatory centers of excellence designation on demographic characteristics of patients who undergo bariatric surgery. JAMA Surg. 2015;150:644-648. [DOI] [PubMed] [Google Scholar]
- 3.Baker DW, Tschurtz BA, Aliaga AE, Williams SC, Jauch EC, Schwamm LH. Determining the need for thrombectomy-capable stroke centers based on travel time to the nearest comprehensive stroke center. Jt Comm J Qual Pat Saf. 2020;46:501-505. [DOI] [PubMed] [Google Scholar]
- 4.Barksfield R, Murray J, Robinson J, Porteous A. Implications of the getting it right first time initiative for regional knee arthroplasty services. Knee. 2017;24:1191-1197. [DOI] [PubMed] [Google Scholar]
- 5.Birkmeyer NJ, Dimick JB, Share D, et al. Michigan Bariatric Surgery Collaborative. Hospital complication rates with bariatric surgery in Michigan. JAMA. 2010;304:435-442. [DOI] [PubMed] [Google Scholar]
- 6.Bozic KJ, Bashyal RK, Anthony SG, Chiu V, Shulman B, Rubash HE. Is administratively coded comorbidity and complication data in total joint arthroplasty valid? Clin Orthop Relat Res. 2013;471:201-205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bozic KJ, Kamath AF, Ong K, et al. Comparative epidemiology of revision arthroplasty: failed THA poses greater clinical and economic burdens than failed TKA. Clin Orthop Relat Res. 2015;473:2131-2138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bozic KJ, Maselli J, Pekow PS, Lindenauer PK, Vail TP, Auerbach AD. The influence of procedure volumes and standardization of care on quality and efficiency in total joint replacement surgery. J Bone Joint Surg Am. 2010;92:2643-2652. [DOI] [PubMed] [Google Scholar]
- 9.Brodke DJ, Guo C, Aouad M, Brown TT, Bozic KJ. Impact of reference pricing on cost and quality in total joint arthroplasty. J Bone Joint Surg Am. 2019;101:2212-2218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Chou DTS, Solomon LB, Costi K, Pannach S, Holubowycz OT, Howie DW. Structured-mentorship program for periacetabular osteotomy resulted in few complications for a low-volume pelvic surgeon. Clin Orthop Relat Res. 2019;477:1126-1134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Courtney PM, Ashley BS, Hume EL, Kamath AF. Are bundled payments a viable reimbursement model for revision total joint arthroplasty? Clin Orthop Relat Res. 2016;474:2714-2721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Delanois RE, Mistry JB, Gwam CU, Mohamed NS, Choksi US, Mont MA. Current epidemiology of revision total knee arthroplasty in the United States. J Arthroplasty. 2017;32:2663-2668. [DOI] [PubMed] [Google Scholar]
- 13.Diaz A, Pawlik TM. Optimal location for centralization of hospitals performing pancreas resection in California. JAMA Surg . 2020;155:261-263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dimick JB, Nicholas LH, Ryan AM, Thumma JR, Birkmeyer JD. Bariatric surgery complications before vs after implementation of a national policy restricting coverage to centers of excellence. JAMA. 2013;309:792-799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dy CJ, Baty J, Saeed MJ, Olsen MA, Osei DA. A population-based analysis of time to surgery and travel distances for brachial plexus surgery. J Hand Surg Am. 2016;41:903-909.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fang CJ, Shaker JM, Ward DM, Jawa A, Mattingly DA, Smith EL. Financial burden of revision hip and knee arthroplasty at an orthopedic specialty hospital: higher costs and unequal reimbursements. J Arthroplasty. 2021;36:2680-2684. [DOI] [PubMed] [Google Scholar]
- 17.FitzGerald JD, Soohoo NF, Losina E, Katz JN. Potential impact on patient residence to hospital travel distance and access to care under a policy of preferential referral to high-volume knee replacement hospitals. Arthritis Care Res (Hoboken). 2012;64:890-897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Glassou EN, Hansen TB, Mäkelä K, et al. Association between hospital procedure volume and risk of revision after total hip arthroplasty: a population-based study within the Nordic Arthroplasty Register Association database. Osteoarthritis Cartilage. 2016;24:419-426. [DOI] [PubMed] [Google Scholar]
- 19.Halder AM, Gehrke T, Günster C, et al. Low hospital volume increases re-revision rate following aseptic revision total knee arthroplasty: an analysis of 23,644 cases. J Arthroplasty. 2020;35:1054-1059. [DOI] [PubMed] [Google Scholar]
- 20.Hsia RY, Krumholz H, Shen YC. Evaluation of STEMI regionalization on access, treatment, and outcomes among adults living in nonminority and minority communities. JAMA Netw Open. 2020;3:e2025874. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jeschke E, Citak M, Günster C, et al. Are TKAs performed in high-volume hospitals less likely to undergo revision than TKAs performed in low-volume hospitals? Clin Orthop Relat Res. 2017;475:2669-2674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Jeschke E, Gehrke T, Günster C, et al. Low hospital volume increases revision rate and mortality following revision total hip arthroplasty: an analysis of 17,773 cases. J Arthroplasty . 2019;34:2045-2050. [DOI] [PubMed] [Google Scholar]
- 23.Kalson NS, Mathews JA, Miles J, et al. Provision of revision knee surgery and calculation of the effect of a network service reconfiguration: an analysis from the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. Knee. 2020;27:1593-1600. [DOI] [PubMed] [Google Scholar]
- 24.Katz JN, Barrett J, Mahomed NN, Baron JA, Wright RJ, Losina E. Association between hospital and surgeon procedure volume and the outcomes of total knee replacement. J Bone Joint Surg Am. 2004;86:1909-1916. [DOI] [PubMed] [Google Scholar]
- 25.Kuo LE, Simmons KD, Kelz RR. Bariatric centers of excellence: effect of centralization on access to care. J Am Coll Surg . 2015;221:914-922. [DOI] [PubMed] [Google Scholar]
- 26.Laucis NC, Chowdhury M, Dasgupta A, Bhattacharyya T. Trend toward high-volume hospitals and the influence on complications in knee and hip arthroplasty. J Bone Joint Surg Am. 2016;98:707-712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Manley M, Ong K, Lau E, Kurtz SM. Total knee arthroplasty survivorship in the United States Medicare population: effect of hospital and surgeon procedure volume. J Arthroplasty. 2009;24:1061-1067. [DOI] [PubMed] [Google Scholar]
- 28.Maurice AP, Punnasseril JEJ, King SE, Dodd BR. Improving access to bariatric surgery for rural and remote patients: experiences from a state-wide bariatric telehealth service in Australia. Obes Surg . 2020;30:4401-4410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.McGrory BJ, Etkin CD, Lewallen DG. Comparing contemporary revision burden among hip and knee joint replacement registries. Arthroplast Today. 2016;2:83-86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Mehrotra A, Sloss EM, Hussey PS, Adams JL, Lovejoy S, Soohoo NF. Evaluation of centers of excellence program for knee and hip replacement. Med Care. 2013;51:28-36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Mehrotra A, Sloss EM, Hussey PS, Adams JL, Lovejoy S, SooHoo NF. Evaluation of a center of excellence program for spine surgery. Med Care. 2013;51:748-757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.New York State Department of Health. Statewide Planning and Research Cooperative System (SPARCS). Available at: https://goo.gl/RWxHGA. Accessed October 1, 2020.
- 33.Nichols CI, Vose JG. Clinical outcomes and costs within 90 days of primary or revision total joint arthroplasty. J Arthroplasty. 2016;31:1400-1406.e3. [DOI] [PubMed] [Google Scholar]
- 34.Nwachukwu BU, Dy CJ, Burket JC, Padgett DE, Lyman S. Risk for complication after total joint arthroplasty at a center of excellence: the impact of patient travel distance. J Arthroplasty. 2015;30:1058-1061. [DOI] [PubMed] [Google Scholar]
- 35.O'Mahoney PRA, Yeo HL, Sedrakyan A, et al. Centralization of pancreatoduodenectomy a decade later: impact of the volume-outcome relationship. Surgery. 2016;159:1528-1538. [DOI] [PubMed] [Google Scholar]
- 36.Polonski A, Izbicki JR, Uzunoglu FG. Centralization of pancreatic surgery in Europe. J Gastrointest Surg. 2019;23:2081-2092. [DOI] [PubMed] [Google Scholar]
- 37.Ravi B, Jenkinson R, Austin PC, et al. Relation between surgeon volume and risk of complications after total hip arthroplasty: propensity score matched cohort study. BMJ. 2014;348:g3284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ricciardi BF, Liu AY, Qiu B, Myers TG, Thirukumaran CP. What is the association between hospital volume and complications after revision total joint arthroplasty: a large-database study. Clin Orthop Relat Res. 2019;477:1221-1231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Robinson JC, Brown TT, Whaley C, Bozic KJ. Consumer choice between hospital-based and freestanding facilities for arthroscopy: impact on prices, spending, and surgical complications. J Bone Joint Surg Am. 2015;97:1473-1481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Robinson JC, MacPherson K. Payers test reference pricing and centers of excellence to steer patients to low-price and high-quality providers. Health Aff (Millwood) . 2012;31:2028-2036. [DOI] [PubMed] [Google Scholar]
- 41.Sathiyakumar V, Apfeld JC, Obremskey WT, Thakore RV, Sethi MK. Prospective randomized controlled trial using telemedicine for follow-ups in an orthopedic trauma population: a pilot study. J Orthop Trauma . 2015;29:e139-145. [DOI] [PubMed] [Google Scholar]
- 42.Schwartz AJ, Yost KJ, Bozic KJ, Etzioni DA, Raghu TS, Kanat IE. What is the value of a star when choosing a provider for total joint replacement? A discrete choice experiment. Health Aff (Millwood). 2021;40:138-145. [DOI] [PubMed] [Google Scholar]
- 43.Sheetz KH, Dimick JB, Nathan H. Centralization of high-risk cancer surgery within existing hospital systems. J Clin Oncol. 2019;37:3234-3242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Shen TS, Bovonratwet P, Morgenstern R, Chen AZ, Su EP. The effect of travel distance on outcomes for hip resurfacing arthroplasty at a high-volume center. Arthroplast Today. 2020;6:1033-1037.e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Stitzenberg KB, Sigurdson ER, Egleston BL, Starkey RB, Meropol NJ. Centralization of cancer surgery: implications for patient access to optimal care. J Clin Oncol. 2009;27:4671-4678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Tanaka H, Ishikawa KB, Katanoda K. Geographic access to cancer treatment in Japan: results from a combined dataset of the patient survey and the survey of medical institutions in 2011. J Epidemiol. 2018;28:470-475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Wilson S, Marx RG, Pan TJ, Lyman S. Meaningful thresholds for the volume-outcome relationship in total knee arthroplasty. J Bone Joint Surg Am. 2016;98:1683-1690. [DOI] [PubMed] [Google Scholar]
- 48.Wu SJ, Ma Q, Martin P, Devries A. Finding the value in 'value' designation: evidence and opportunity in the United States. Manag Care. 2016;25:36-42. [PubMed] [Google Scholar]
- 49.Yapp LZ, Walmsley PJ, Moran M, Clarke JV, Simpson AHRW, Scott CEH. The effect of hospital case volume on re-revision following revision total knee arthroplasty. Bone Joint J. 2021;103:602-609. [DOI] [PubMed] [Google Scholar]