Disparities in Access to Robotic Technology and Perioperative Outcomes Among Patients Treated with Radical Prostatectomy

Abstract

Background:

Most radical prostatectomies are completed with robotic assistance. While studies have previously evaluated perioperative outcomes of robot-assisted radical prostatectomy (RARP), this study investigates disparities in access and clinical outcomes of RARP.

Study Design:

The National Cancer Database (NCDB) was used to identify patients who received radical prostatectomy for cancer between 2010–2017 with outcomes through 2018. RARP was compared to open radical prostatectomy (ORP). Odds of receiving RARP were evaluated while adjusting for covariates. Overall survival was evaluated using a propensity-score matched cohort.

Results:

Overall, 354,752 patients were included with 297,676 (83.9%) receiving RARP. Patients who were non-Hispanic Black (82.8%) or Hispanic (81.3%) had lower rates of RARP than non-Hispanic White (84.0%) or Asian patients (87.7%, p<0.001). Medicaid or uninsured patients were less likely to receive RARP (75.5%) compared to patients with Medicare or private insurance (84.4%, p<0.001). Medicaid or uninsured status was associated with decreased odds of RARP in adjusted multivariable analysis (OR 0.61, 95% CI 0.49–0.76). RARP was associated with decreased perioperative mortality and improved overall survival compared to ORP.

Conclusion:

Patients who were underinsured were less likely to receive RARP. Improved access to RARP may lead to decreased disparities in perioperative outcomes for prostate cancer.

Graphical Abstract

Introduction

There has been rapid adoption of minimally invasive surgery (MIS) throughout many surgical specialties.(1, 2) However, most of those studies were carried out at pioneering centers early in the MIS era.(3–6) and have compared short-term perioperative outcomes with fewer studies examining longer-term oncologic outcomes.(7–11) Less is known about outcomes in broader practice when this technology extends beyond the confines of selected high-volume experienced centers.(12–15) The use of robot-assisted surgery has increased more than 3-fold in the past decade, and the United States is now the largest market for this technology in the world.(16) Several studies have found that this rapid growth was not evenly distributed.(17–19)

The diffusion of robotic-assisted surgical procedures is concentrated within the fields of urology, thoracic surgery, general surgery, and gynecology. The technology is often marketed as a tool to mitigate some of the technical or anatomic challenges associated with specific surgical procedures. However, there is often disparate access to the implementation of robotic surgery. In thoracic surgery, prior research has demonstrated that low-income status, lack of insurance, or treatment at a rural hospital were all independent contributors to reduced likelihood of undergoing robotic lobectomy.(20) Similarly, the rates of implementation of MIS for general surgery procedures including appendectomy, cholecystectomy, hemicolectomy, and hernia repair have been significantly lower for the same patient population.(21, 22) Among gynecologic procedures, robotic hysterectomy has grown tremendously in use over the last decade, however, rurality and lack of private insurance were again correlated with reduced likelihood of MIS.(23)

Currently, most radical prostatectomies in the United States are performed with robotic assistance. However, it is unknown if disparities in access to robotic technology impact outcomes. The purpose of this study is to investigate access and outcomes disparities in patients who undergo robot-assisted radical prostatectomy (RARP).

Methods

Data Source and Study Cohort

The National Cancer Database (NCDB) is a large cancer registry that includes approximately 50.8% of prostate cancer in the United States.(24) More than 1,500 hospitals are approved to contribute to the NCDB. The NCDB was used to identify surgically treated patients with prostate cancer diagnosed from 2010–2017 for which all-cause mortality outcome data were available through 2018. Patient information in the NCDB is deidentified and this study is exempt from review by our institution’s institutional review board. The guidelines for Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) were followed for this retrospective cohort study.(25)

Geographic Characteristics

The United States Department of Agriculture Economic Research Service publishes a nine level Rural-Urban continuum code (RUCC). RUCC levels 1–3 are defined as metropolitan counties, 4–7 are defined as nonmetropolitan urban counties, and 8–9 are defined as rural counties. Patients with missing rurality data were included as a separate variable in analyses.

Distance from treatment facility was calculated by the distance between the centroid of the patient’s zip code to the address of the reporting hospital. Travel distance was evaluated as a continuous variable and also categorized into travel distances of less than 40 miles versus 40 miles or greater.

Operative Approach

Data regarding operative approach (open versus robot-assisted versus laparoscopic) were available from 2010–2017. RARP was compared to open radical prostatectomy (ORP). Patients who underwent RARP and were converted to ORP were included in the RARP category for analysis. Laparoscopic approach was excluded.

Patient Characteristics

NCDB patient data included patient age, sex, race and ethnicity, income and educational characteristics, primary payor status, comorbidities, clinical characteristics, Gleason scores and patterns, and prostate-specific antigen (PSA) levels.

Patient age was grouped into patients 55 and under, 56–60, 61–65, 66–70, and greater than 70 years. Comorbidities in the NCDB are reported with the Charlson-Deyo score grouped by 0, 1, and 2 or greater. Income and education level in the NCDB was based on zip code tabulation area (ZCTA) estimates which are categorized into quartiles. Patients with missing data for income or education were categorized as missing and included in analyses. Primary insurance payor was categorized as uninsured or Medicaid versus Medicare, private, or other government insurance status versus unknown payor. Pathological stage was categorized as stage I, II, III, or missing. Gleason scores, Gleason patterns, and PSA categories were created in line with National Comprehensive Cancer Network (NCCN) guidelines for prostate cancer and were used to categorize NCCN Risk Stratification Groups.(26)

Hospital Characteristics

Facility type was determined by the Commission on Cancer (CoC) accreditation program and were categorized as academic cancer programs and nonacademic cancer programs. CoC accreditation is based on facility-level organization, capabilities, and volume of new cancer cases. Annual hospital surgical volume was categorized into quartiles (<8; 8 to <22; 22 to <58; ≥58).

Statistical Analysis

Bivariate analysis with χ² tests was performed to evaluate differences between patients who underwent RARP and ORP. Continuous travel distance was evaluated with Mann-Whitney-U tests. Changes over time in rates of RARP were evaluated with Cochran-Armitage tests. P-values were two sided and statistical significance was considered to be p<0.01.

Multivariable logistic regression with adjustment for patient, hospital, clinical, and geographic factors was used to determine odds of receiving RARP. Standard errors were adjusted for clustering of cases within hospitals. Odds ratios were reported with 95% confidence intervals. Considering previous research by Kim et al (2013) on access to hospitals with robotic technology, two identically adjusted models were created to evaluate the association between disparities in access to RARP and hospital annual surgical volume: model 1 omitted hospital annual surgical volume quartiles as an independent variable, while model 2 included hospital annual surgical volume quartiles as independent variables.(18) Akaike’s Information Criterion (AIC) values were used to compare fit and were reported for each model.

Differences in perioperative outcomes, including 30- and 90-day mortality, and 30-day unplanned readmission, were also evaluated with χ² tests. Post-operative length-of-stay was evaluated with Student’s T-tests. Differences in cost associated with post-operative length-of-stay differences between patients who underwent RARP versus ORP were determined (Supplemental Table 1).

Propensity Score-Matching

Propensity scores were calculated from the odds of receiving RARP and adjusted for year of diagnosis, age, race and ethnicity, rurality, income, education, Charlson-Deyo score, NCCN risk category, pathological stage, Gleason score, travel distance, and for clustering of observations at hospitals. Missing data were excluded. The caliper was calculated by multiplying the standard deviation of the propensity-score by 0.2.(27) The logit of the propensity score was used to perform nearest-neighbor matching without replacement. Matching was evaluated by comparing standardized mean differences of covariates with an allowable variation of less than 0.1.(28)

Survival Analysis

Cox proportional hazard models with adjustment for clustering were used to evaluate overall survival. Hazards ratios were reported with 95% confidence intervals, and both unadjusted and adjusted results were reported. Kaplan-Meier survival estimates with log-rank tests were performed to evaluate five-year survival. All analyses were done with Stata MP Version 17, College Station, TX.

Results

Overall, 354,752 patients diagnosed between 2010–2017 who underwent radical prostatectomy met criteria for inclusion. The overall cohort was further sub-categorized by those who received RARP (n=297,676, 83.9%) and those who received ORP (n=57,076, 16.1%) (Figure 1). There were 37,474 included in the propensity score-matched cohort with 54.2% RARP and 45.8% ORP (Supplemental Table 2). Overall, RARP became more common over the study period and patients who were diagnosed in 2017 were more likely to undergo RARP (90.7%) than those treated in 2010 (75.6%, p<0.001; Table 1). The patient cohort was treated at 1,180 separate reporting hospitals over the entire study period (2010–2017). The number of hospitals reporting at least one annual RARP increased from 671 hospitals to 883 hospitals, while the number of hospitals reporting only ORP dropped from 320 hospitals to 96 hospitals. The proportion of Gleason 6 patients who received radical prostatectomy in our study cohort decreased from 33.4% in 2010 to 11.5% in 2017 (p<0.001). Patients who received ORP traveled a median of 13.2 (IQR 5.5–33.9) miles while those who received RARP traveled a median of 14.5 (IQR 6.6–36.8) miles (p<0.001). Patients from metropolitan areas traveled a shorter median distance (miles) for both RARP and ORP (11.4, IQR 5.6–23.4 RARP; 10.1, IQR 4.6–21.8 ORP), compared to those from nonmetropolitan (52.2, IQR 31.8–85.1 RARP; 39.6, IQR 21.7–71.2 ORP) and rural (66.3, IQR 43.8–103.6 RARP; 56.2, IQR 34.2–89.3 ORP) areas (all p<0.001). Residence in a metropolitan county was more likely for patients who were non-Hispanic Black (88.6%), Hispanic (93.4%), or Asian (92.9%) than non-Hispanic White (79.1%, p<0.001).

FIGURE 1.

Inclusion criteria flow diagram.

TABLE 1.

Characteristics of patients who underwent radical prostatectomy by operative approach

	Total	RARP	ORP
N	354,752	297,676 (83.9%)	57,076 (16.1%)
Parameter	%	%	%	P Value
Surgical Approach				<0.001
Robotic	295,369	100	0
Robotic Converted	2,307	100	0
Open	57,076	0	100
Rurality				<0.001
Metropolitan	288,300	84.3	15.7
Nonmetropolitan	44,693	81.3	18.7
Rural	6,028	78.7	21.3
Missing	15,731	85.7	14.3
Travel Distance (miles)				<0.001
<40	243,577	83.3	16.7
≥40	70,423	84.9	15.1
Missing	40,752	85.6	14.4
Year of Diagnosis				<0.001
2010	51,682	75.6	24.4
2011	54,180	80.3	19.7
2012	42,694	82.8	17.2
2013	41,116	83.6	16.4
2014	38,648	85.5	14.5
2015	40,823	87.2	12.8
2016	41,723	88.7	11.3
2017	43,886	90.7	9.3
Age at Diagnosis				<0.001
≤55	71,284	84.8	15.2
56 to 60	79,158	84.4	15.6
61 to 65	92,681	84.1	15.9
66 to 70	75,582	84.0	16.0
>70	36,047	80.3	19.7
Race and Ethnicity				<0.001
Non-Hispanic White	271,448	84.0	16.0
Non-Hispanic Black	44,994	82.8	17.2
Hispanic	14,037	81.3	18.7
Asian	6,596	87.7	12.3
Other or Unknown	17,677	85.9	14.1
Insurance Status				<0.001
Medicare, Private, or Other	337,678	84.4	15.6
Medicaid or Uninsured	13,604	75.5	24.5
Unknown	3,470	72.7	27.3
Income Quartiles				<0.001
1 Lowest	43,775	81.2	18.8
2	61,965	82.1	17.9
3	77,612	82.9	17.1
4 Highest	128,330	85.7	14.3
Missing	43,070	85.6	14.4
Education Quartiles				<0.001
1 Lowest	47,299	81.4	18.6
2	70,607	82.7	17.3
3	96,546	83.6	16.4
4 Highest	97,584	85.5	14.5
Missing	42,716	85.6	14.4
Charlson-Deyo Score				<0.001
0	288,546	84.1	15.9
1	54,458	83.3	16.7
2	8,613	81.8	18.2
3 or greater	3,135	83.8	16.2
NCCN Risk Stratification				<0.001
Very Low or Low	50,955	84.2	15.8
Intermediate	204,931	86.0	14.0
High or Very High	90,983	79.8	20.2
Missing	7,883	75.9	24.1
Pathological Stage				<0.001
I	12,097	72.1	27.9
II	200,219	83.4	16.6
III	83,155	84.8	15.2
Missing	59,281	86.8	13.2
Gleason Score				<0.001
6	81,460	80.1	19.9
7 (3+4)	167,808	85.5	14.5
7 (4+3)	63,864	85.8	14.2
8	17,128	83.5	16.5
9 or 10	21,134	82.2	17.8
Program Type				<0.001
Nonacademic	208,727	83.3	16.7
Academic	146,025	84.8	15.2
Hospital Annual Surgical Volume Quartiles				<0.001
<8	4,170	40.2	59.8
8 to <22	25,996	67.7	32.3
22 to <58	83,503	81.0	19.0
≥58	241,083	87.4	12.6

RARP, Robot-Assisted Radical Prostatectomy

ORP, Open Radical Prostatectomy

Disparities in Receipt of RARP

Disparities in receipt of RARP were found when evaluating rurality, race and ethnicity, insurance payor status, income, educational attainment, and annual hospital surgical volume with bivariate analysis. Patients from rural or nonmetropolitan areas were less likely to receive RARP than those residing in metropolitan areas (78.7% rural vs 81.3% nonmetropolitan vs 84.3% metropolitan, p<0.001) (Table 1). When evaluating rates of RARP versus ORP, patients who were non-Hispanic Black (82.8% RARP) or Hispanic (81.3% RARP) were less likely to receive RARP than those who were Asian (87.7% RARP) or non-Hispanic White (84.0% RARP; p<0.001) (Table 1). Patients with Medicaid or no insurance were less likely to receive RARP than those with Medicare or private insurance (75.5% vs 84.4%, p<0.001) (Table 1; Figure 2). Patients residing in areas characterized by the lowest quartile of income or education were less likely to receive RARP than those from areas with the highest quartile of income (81.2% vs 85.7% RARP) or education (81.4% vs 85.5% RARP; p<0.001) (Table 1). Patients who underwent RARP were more likely to be treated at a hospital with the highest quartile of annual surgical volume (≥58 annual cases, 87.4% RARP) versus patients treated at a hospital with the lowest quartile of annual surgical volume (<8 annual cases, 40.2% RARP, p<0.001). Patients from nonmetropolitan areas (89.8%) or rural areas (89.3%) were slightly less likely to be treated at hospitals above the median annual surgical volume than those residing in metropolitan areas (91.5%, p<0.001). Similarly, patients who are non-Hispanic Black (90.3%), Hispanic (86.1%), or Asian (89.6%) are less likely to be treated at hospitals above the median annual surgical volume than non-Hispanic White patients (92.1%, p<0.001).

FIGURE 2.

Trends in rates of robot-assisted radical prostatectomy versus open radical prostatectomy by insurance payor status. P<0.01 for trend. RARP, Robot-Assisted Radical Prostatectomy.

Odds of undergoing RARP was evaluated among patients who underwent radical prostatectomy using multivariable logistic regression (Model 1) adjusted for year of diagnosis, age, race and ethnicity, rurality, income, education, Charlson-Deyo score, NCCN risk category, pathological stage, Gleason score, and travel distance (Table 2). Compared to patients residing in metropolitan areas, patients from nonmetropolitan areas (OR 0.77, 95% CI 0.62–0.96) or rural areas (OR 0.64, 95% CI 0.45–0.91) had decreased odds of receiving RARP. Patients with Medicaid or no insurance (OR 0.56, 95% CI 0.45–0.70 versus Medicare or private insurance) had decreased odds of receiving RARP versus ORP. There was no significant difference in odds of receiving RARP between patients who were non-Hispanic Black or Hispanic and non-Hispanic White (reference) in adjusted multivariable analysis (Table 2). Asian patients, and patients with other or unknown race were more likely to receive RARP compared to non-Hispanic White patients (Table 2).

TABLE 2.

Multivariable logistic regression evaluating association between rurality, insurance status, race and ethnicity, and hospital surgical volume with receipt of RARP versus ORP among patients who underwent radical prostatectomy

	Model 1*	Model 2
	Odds of RARP	Odds of RARP
Parameter	OR (95% CI)	OR (95% CI)
Akaike’s Information Criterion	300,131	290,183
Rurality
Metropolitan	Reference	Reference
Nonmetropolitan	0.77 (0.62–0.96)	0.86 (0.70–1.06)
Rural	0.64 (0.45–0.91)	0.75 (0.54–1.04)
Missing	1.00 (0.71–1.41)	0.98 (0.69–1.39)
Insurance Status
Medicare, Private, or Other	Reference	Reference
Medicaid or Uninsured	0.56 (0.45–0.70)	0.61 (0.49–0.76)
Unknown	0.51 (0.39–0.68)	0.55 (0.41–0.75)
Race and Ethnicity
Non-Hispanic White	Reference	Reference
Non-Hispanic Black	0.93 (0.84–1.07)	0.96 (0.85–1.09)
Hispanic	0.93 (0.77–1.13)	1.02 (0.85–1.22)
Asian	1.38 (1.04–1.83)	1.40 (1.10–1.78)
Other or Unknown	1.30 (1.06–1.60)	1.33 (1.10–1.62)
Hospital Annual Surgical Volume Quartiles
<8	Omitted	Reference
8 to <22		3.10 (2.19–4.39)
22 to <58		6.49 (4.64–9.08)
≥58		10.44 (7.45–14.64)

^*Model 1 without Hospital Annual Surgical Volume, Model 2 with Hospital Annual Surgical Volume. Models similarly adjusted for year of diagnosis, age, income, education, Charlson-Deyo score, NCCN risk category, pathological stage, Gleason score, and travel distance.

With the addition of hospital annual surgical volume quartiles (Model 2), there was no significant difference between odds of receiving RARP and nonmetropolitan or rural residence versus metropolitan residence (Table 2). There remained no significant difference in odds of receiving RARP between patients who were non-Hispanic Black or Hispanic and non-Hispanic White (reference). However, patients with Medicaid or no insurance had persistently decreased odds of RARP (OR 0.61, 95% CI 0.49–0.76; Table 2). Patients treated at hospitals with the highest quartile of annual surgical volume were at much greater odds of receiving RARP (OR 10.44, 95% CI 7.45–14.64 versus ORP).

Compared to ORP, RARP was associated with decreased rates of 30-day unplanned readmissions as well as decreased 30-day and 90-day mortality (Table 3). Length-of-stay after RARP was lower than ORP by a mean of 1.2 (SE 0.018, p<0.001) days, which was associated with an estimated cost-savings of $481.28 USD per case (Supplemental Table 3).

TABLE 3.

Perioperative outcomes of patients undergoing radical prostatectomy by operative approach

Population	Total	RARP	ORP
Overall	354,752	297,676	57,076
Metro	288,300	243,115	45,185
Nonmetro	44,693	36,333	8,360
Rural	6,028	4,743	1,285
Medicare or Private	337,678	284,887	52,791
Medicaid or Uninsured	13,604	10,265	3,339
Parameter	%	%	%	P value
30-Day Mortality
Overall	0.18	0.14	0.39	<0.001
Metro	0.18	0.14	0.39	<0.001
Nonmetro	0.20	0.17	0.31	0.011
Rural	0.20	0.08	0.62	<0.001
Medicare or Private	0.18	0.14	0.39	<0.001
Medicaid or Uninsured	0.21	0.16	0.36	<0.001
90-Day Mortality
Overall	0.30	0.21	0.72	<0.001
Metro	0.28	0.21	0.69	<0.001
Nonmetro	0.37	0.28	0.79	<0.001
Rural	0.43	0.19	1.32	<0.001
Medicare or Private	0.29	0.21	0.73	<0.001
Medicaid or Uninsured	0.41	0.32	0.69	<0.001
30-Day Unplanned Readmission
Overall	2.31	2.19	2.92	<0.001
Metro	2.36	2.25	2.98	<0.001
Nonmetro	2.15	2.01	2.75	<0.001
Rural	2.82	2.74	3.11	0.48
Medicare or Private	2.25	2.17	2.67	<0.001
Medicaid or Uninsured	3.50	2.95	5.18	<0.001

RARP, Robot-Assisted Radical Prostatectomy

ORP, Open Radical Prostatectomy

Survival Analysis

Overall survival for surgical patients who underwent radical prostatectomy was superior for patients selected for treatment with RARP versus ORP in unadjusted (HR 0.56, 95% CI 0.54–0.58) Cox proportional hazards models, as well as in an adjusted models in the propensity score-matched cohort (HR 0.64, 95% CI 0.59–0.70, Supplemental Table 4). In the propensity score-matched cohort, patients with Medicaid or no insurance had increased risk of death (HR 1.37, 95% CI 1.20–1.56) compared to patients with Medicare or private insurance (reference). However, patients from nonmetropolitan and rural areas had no statistically significant difference in survival compared to patients from metropolitan areas, and non-Hispanic Black patients had no statistically significant difference in survival compared to non-Hispanic White patients.

Discussion

We used the NCDB to investigate national trends in the rates of RARP versus ORP performed in the United States. Our analysis has indicated several noteworthy changes to patterns in disparities in access to RARP. For patients from rural areas, access to RARP was associated with travel for care at high-volume hospitals (HVH) where patients have significantly increased odds of receiving RARP. However, disparities in access to RARP persist regardless of access to HVH for underinsured patients. RARP was associated with decreased adverse short-term outcomes compared to ORP. Long-term survival appears to be superior with RARP versus ORP, while underinsured status remains an independent risk factor for worse survival.

There is prior evidence that RARP is associated with improved perioperative outcomes across multiple domains, including decreased blood loss and transfusion rates.(29–33) Also, a randomized trial has demonstrated lower biochemical recurrence at two years, though the study was powered for the assessment of quality-of-life and health-related outcomes and not for oncological outcomes.(34, 35) However, the underlying influence of hospital annual surgical volume may be confounding the association between RARP and overall survival in our data. HVHs use RARP much more often than low-volume hospitals (LVHs), making the use of RARP a possible surrogate for the volume-outcome relationship which has been described extensively in the literature.(11, 36, 37) Though it is possible that for some patients RARP may afford better perioperative outcomes which may then be associated with improved overall survival compared to ORP, evidence generated by high-quality Randomized Clinical Trials (RCTs) with long-term follow-up is not available.(38)

There is an abundance of evidence in the literature related to racial and ethnic disparities in prostate cancer diagnosis, treatment, and outcomes – especially for non-Hispanic Black patients.(39, 40) Non-Hispanic Black patients have higher incidence of prostate cancer and are more likely to present with metastases than non-Hispanic White patients.(41) Furthermore, non-Hispanic Black patients are less likely to undergo surveillance or to receive treatment in the forms of radiation or surgery for prostate cancer and have worse outcomes compared to other patients – even after treatment in equitable environments.(41) Though previous research by Kim et al (2013) using data from 2006–2008 found non-Hispanic Black patients had decreased odds of undergoing RARP, when evaluating patients who underwent radical prostatectomy in 2010–2017 we found no statistically significant difference in odds of undergoing RARP versus ORP compared to non-Hispanic White patients in multivariable analysis.(42, 43) Even though non-Hispanic Black patients are less likely to undergo any type of treatment, including surgical treatment, for those who access surgical treatment they are no less likely to receive minimally invasive techniques. These findings may reflect the overall trend of increased uptake and availability of minimally invasive surgical technology in hospitals in metropolitan areas where the majority of non-Hispanic Black patients receive their care.

Though nonmetropolitan or rural residence was associated with decreased rates of RARP, there was no statistically significant difference compared to metropolitan residence in multivariable analysis adjusted with hospital surgical volume. However, patients from nonmetropolitan or rural areas were much more likely to travel long distances for care, and patients who traveled were more likely to receive RARP in previous research.(44) Though it is not possible in the NCDB to determine the rurality of the treating hospital, prior research suggests rurally-located hospitals are much less likely to perform robotic surgery.(20) It is probable that patients from nonmetropolitan or rural areas who have the resources and ability are traveling to regionalized high volume hospitals in metropolitan areas for surgical treatment, thus increasing their likelihood of receiving RARP.(44)

Of the sociodemographic variables in the NCDB, insurance payor status is the only one tied directly to the individual patient, while educational attainment and income are linked to geographic areas designated by Zip Code Tabulation Area (ZCTA). This makes insurance payor status unique in that it is a more direct predictor of individual patient access to care. For patients with Medicaid or who were uninsured, odds of receiving RARP and overall survival were both decreased. Also, perioperative short-term outcomes were improved for patients with Medicaid or no insurance who received RARP. This is consistent with previous research by Kim et al (2015) demonstrating lower odds of RARP for Medicaid patients treated in California between 2009–2011.(19) These findings can likely be attributed to known difficulties that community healthcare providers have in finding care referrals to appropriate specialists that are covered under Medicaid.(45) Furthermore, the disparity in access to urologic care has been specifically demonstrated for patients with Medicaid coverage seeking treatment at practices acquired by private equity firms.(46) Improvements in access to RARP may lead to improvements in morbidity and mortality in this population.(19, 47)

Limitations

There are some limitations to this study inherent to the NCDB. The NCDB is a large database that includes cases submitted by Commission on Cancer (CoC) approved hospitals. Though the contributing hospital network includes greater than 1,500 sites, only approximately 50% of prostate cancers are captured.(24, 48) Additionally, the geographic location of the reporting hospitals is not reported. Previous research revealed that non-CoC approved hospitals that do not contribute to the NCDB are often located in rural areas with low available resources dedicated to cancer care.(48) Because of this, patients from rural areas are less likely to be captured by the NCDB. Another limitation is that the NCDB is a hospital-based cancer registry and the incidence of uninsured or Medicaid patients represented in the database may be lower than that of the general population. As a result, relatively few patients in the study were underinsured or uninsured and it is possible that outcomes for these patients may be even worse than reflected in our results. Additionally, patients who are dual enrolled in Medicaid and Medicare are identified in the NCDB by their primary payor (Medicare in this instance). Finally, a growing body of experience and years of training for surgeons performing RARP may possibly account for improved outcomes over time, rather than this improvement being attributed solely to beneficial differences inherent to the robotic approach. Regardless, receipt of robotic surgery should still serve as a quality measure relevant for urologic oncology care disparities.

Conclusions

RARP has increased during the study period. Rural, non-Hispanic Black and Hispanic patients are less likely to be treated at high-volume robotic centers. When considering the procedural volume of a hospital, patients from rural areas or non-Hispanic Black and Hispanic patients were no less likely to receive RARP. However, underinsured patients were less likely to receive RARP at both high and low-volume hospitals. Because RARP was associated with lower odds of adverse perioperative outcomes and decreased length-of-stay, continued efforts to expand access to RARP to underinsured populations may further reduce disparities and improve outcomes.

Synopsis:

Robotic-assisted radical prostatectomy (RARP) was associated with decreased perioperative mortality and improved overall survival compared to open radical prostatectomy (ORP). Patients who were underinsured were less likely to receive RARP. Improved access to RARP may lead to decreased disparities in perioperative outcomes for prostate cancer.