The Effect of Institution Teaching Status on Perioperative Outcomes After Robotic Partial or Radical Nephrectomy

Abstract

Objectives:

To test rates over time of robotic partial and radical nephrectomy (RPN and RRN) at teaching vs nonteaching institutions and to examine associated complication rates and length of stay.

Materials and Methods:

Within the National Inpatient Sample (2008–2013), after stratification according to institutional teaching status, we examined the rates of robotic, open, and laparoscopic PN and RN. Subsequently, we tested complication rates and length of stay associated with RPN or RRN according to institutional teaching status. We relied on estimated annual percentage change (EAPC) with the least squares linear regression to test temporal trends and on multivariable logistic regression (MLR) and Poisson regression models to test complication rates and length of stay.

Results:

Overall, 4070 and 1683 RPN and RRN were identified. In MLR, RPN rates were lower at teaching vs nonteaching institutions (odds ratio [OR]: 0.79, p < 0.001). RPN increased at similar rates over time at teaching and nonteaching institutions (EAPC: +32.1% and +30.1%, all p < 0.05). In MLR, RRN rates were similar at teaching and nonteaching institutions (p: 0.4). RRN rate increase over time was of similar magnitude at teaching vs nonteaching institutions (EAPC: +35.5% and +43.0%, all p < 0.05). RPN at teaching institutions yielded higher genitourinary complication rates (OR: 1.46, p: 0.01). RRN at teaching institutions resulted in lower respiratory complications (OR: 0.66, p: 0.04) and shorter length of stay (rate ratio [RR]: 0.93, p: 0.01), but higher intraoperative complication rates (OR: 3.39, p: 0.04).

Conclusion:

Despite statistically significant differences in rates of RPN use, according to the institutional teaching status and despite statistically significant difference in selected complications, no meaningful differences distinguish teaching vs nonteaching institution when RPN and RRN are considered.

Introduction

The future standard of care for partial and radical nephrectomy (PN and RN) will likely consist of robotic surgery. Indeed, the rates of robotic PN (RPN) and RN (RRN) are increasing in both teaching and nonteaching institutions.^1,2

Historical data suggest that the institutional teaching status represents an important predictor of perioperative outcomes and that perioperative outcomes may be better when surgery is performed at teaching institutions.^3

–6

We challenged this historic hypothesis and tested for contemporary differences in rates of RPN and RRN, as well as associated complications and length of stay. We tested our hypothesis within the National Inpatient Sample (NIS) registry.

Materials and Methods

Data source and study population

Within the NIS registry (2008–2013), we focused on patients with a primary diagnosis of renal cell carcinoma (ICD-9-CM code 189.0) aged ≥18 years. Patients with a secondary diagnosis of metastatic disease were excluded (ICD-9-CM code 197.x and 198.x). Primary procedure codes were used to identify partial (ICD-9-CM code 55.4) and radical (ICD-9-CM code 55.5; 55.51; 55.52; 55.54) nephrectomy patients. Robotic procedure was identified according to the modifier codes 17.4 and 17.42.

Laparoscopic procedures were identified according to the modifier codes 54.21 and 54.51. All the other procedures were considered as open nephrectomies. Patients with more than one modifier were excluded from the analyses.

Variables definition

All analyses were stratified according to institutional teaching status (teaching vs nonteaching). Hospital status was coded as a teaching according to the American Hospital Association Annual Survey of Hospitals. Hospitals were considered teaching institutions if they had an American Medical Association approved residency program, were a member of the Council of Teaching Hospitals, or had a ratio of 0.25 or higher for full-time equivalent interns and residents relative to non-nursing home beds.⁷

Covariates consisted of age at diagnosis (18–49, 50–64, 65–74, ≥75 years), gender, race (African American, Caucasian, other/unknown), Charlson comorbidity index (CCI)⁶ (0, 1, 2, ≥3), residence region (Midwest, Northeast, South, West), insurance status (Medicaid, Medicare, Private, Other), bed size (Large, Medium, Small), and annual hospital volume in tertiles (High, Medium, Low).

Outcomes definition

Complication rates were defined using secondary ICD-9 diagnostic codes, as previously described.^6,8 Intraoperative complications consisted of accidental blood vessel and/or nerve and/or organ puncture or laceration during the procedure.⁶ Postoperative complications consisted of cardiac, respiratory, vascular, operative wound, genitourinary, transfusion, miscellaneous medical, and miscellaneous surgical (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/end). Length of stay, provided by the NIS, was calculated by subtracting the admission date from the discharge date.⁶ In-hospital mortality rate information was coded from the disposition of the patient.⁷

Statistical analyses

All statistical analyses were performed on two separate data samples. First data sample focused on RPN patients among all PN patients. The second data sample focused on RRN patients among all RN patients.

In each data sample, we performed descriptive statistical analyses. These relied on tests for means and proportions and, respectively, t-tests and chi-square test were used. Subsequently, we compared RPN and RRN rates according to teaching vs nonteaching institutions in multivariable logistic regression (MLR) models. Time trend analyses focused on annual RPN and RRN rates. Then, we tested in MLR models complications and in-hospital mortality rates after RPN and RRN at teaching vs nonteaching institutions. Finally, Poisson regression models compared length of stay after RPN and RRN at teaching vs nonteaching institutions.

All statistical tests were two sided. The level of significance was set at p < 0.05. Analyses were performed using the R software environment for statistical computing and graphics (version 3.4.3).

Results

Descriptive characteristics and rates of RPN over time

Overall, 13,241 PN patients were identified. Of these, 30.7% (N = 4070) underwent RPN. Of all RPN patients, 73.2% were treated at teaching institutions. Most RPN patients were 50 to 64 years old (42.3%), male (62.9%), and Caucasian (68.4%). Most RPN patients harbored CCI of 0 (62.3%) and were privately insured (55.0%). RPN patients treated at teaching institutions were less frequently (7.8% vs 10.7%) among the oldest age category (≥75 years old). Moreover, RPN patients treated at teaching institutions less frequently harbored CCI 1 (26.8% vs 29.9%), CCI 2 (5.5% vs 7.4%), or CCI ≥3 (4.0% vs 4.2%). Furthermore, RPN patients were more frequently treated at teaching institutions in the Midwest (30.6% vs 21.3%) (Table 1).

Table 1.

Descriptive Characteristics of Patients with Renal Cell Carcinoma, Who Underwent Robotic Partial and Radical Nephrectomy, Stratified According to Institutional Teaching Status

	Robotic partial nephrectomy				Robotic radical nephrectomy
	Overall (4070)	Nonteaching (1092)	Teaching (2978)	p	Overall (1683)	Nonteaching (556)	Teaching (1127)	p
Year of diagnosis	2012 (2011–2013)	2011 (2011–2013)	2012 (2010–2013)	0.03	2012 (2010–2013)	2011.5 (2010–2013)	2012 (2010–2013)	0.7
Age of diagnosis
18–49	959 (23.6)	262 (24.0)	697 (23.4)	0.01	245 (14.6)	72 (12.9)	173 (15.4)	0.5
50–64	1720 (42.3)	431 (39.5)	1289 (43.3)		664 (39.5)	217 (39.0)	447 (39.7)
65–74	1043 (25.6)	282 (25.8)	761 (25.6)		471 (28)	160 (28.8)	311 (27.6)
≥75	348 (8.6)	117 (10.7)	231 (7.8)		303 (18)	107 (19.2)	196 (17.4)
Gender
Female	1509 (37.1)	396 (36.3)	1113 (37.4)	0.5	614 (36.5)	214 (38.5)	400 (35.5)	0.2
Male	2561 (62.9)	696 (63.7)	1865 (62.6)		1069 (63.5)	342 (61.5)	727 (64.5)
Race
Caucasian	2784 (68.4)	781 (71.5)	2003 (67.3)	0.02	1182 (70.2)	428 (77)	754 (66.9)	<0.001
African American	383 (9.4)	85 (7.8)	298 (10)		144 (8.6)	36 (6.5)	108 (9.6)
Other/unknown	903 (22.2)	226 (20.7)	677 (22.7)		357 (21.2)	92 (16.5)	265 (23.5)
CCI
0	2536 (62.3)	639 (58.5)	1897 (63.7)	0.01	963 (57.2)	296 (53.2)	667 (59.2)	0.1
1	1125 (27.6)	326 (29.9)	799 (26.8)		488 (29)	172 (30.9)	316 (28)
2	244 (6)	81 (7.4)	163 (5.5)		135 (8)	53 (9.5)	82 (7.3)
≥3	165 (4.1)	46 (4.2)	119 (4.0)		97 (5.8)	35 (6.3)	62 (5.5)
Residence region
Midwest	1143 (28.1)	233 (21.3)	910 (30.6)	<0.001	544 (32.3)	129 (23.2)	415 (36.8)	<0.001
Northeast	837 (20.6)	178 (16.3)	659 (22.1)		292 (17.3)	74 (13.3)	218 (19.3)
South	1485 (36.5)	437 (40.0)	1048 (35.2)		548 (32.6)	179 (32.2)	369 (32.7)
West	605 (14.9)	244 (22.3)	361 (12.1)		299 (17.8)	174 (31.3)	125 (11.1)
Insurance status
Medicaid	214 (5.3)	43 (3.9)	171 (5.7)	0.05	76 (4.5)	22 (4)	54 (4.8)	0.7
Medicare	1409 (34.6)	385 (35.3)	1024 (34.4)		796 (47.3)	272 (48.9)	524 (46.5)
Other	208 (5.1)	47 (4.3)	161 (5.4)		73 (4.3)	23 (4.1)	50 (4.4)
Private	2239 (55.0)	617 (56.5)	1622 (54.5)		738 (43.9)	239 (43)	499 (44.3)
Bed size
Large	2790 (68.6)	712 (65.2)	2078 (69.8)	0.02	1176 (69.9)	404 (72.7)	772 (68.5)	<0.001
Medium	745 (18.3)	225 (20.6)	520 (17.5)		326 (19.4)	115 (20.7)	211 (18.7)
Small	535 (13.1)	155 (14.2)	380 (12.8)		181 (10.8)	37 (6.7)	144 (12.8)
Annual hospital volume
Low	1069 (26.3)	445 (40.8)	624 (21.0)	<0.001	531 (31.6)	245 (44.1)	286 (25.4)	<0.001
Medium	1564 (38.4)	422 (38.6)	1142 (38.3)		679 (40.3)	250 (45)	429 (38.1)
High	1437 (35.3)	225 (20.6)	1212 (40.7)		473 (28.1)	61 (11)	412 (36.6)

Data are shown as medians (interquartile range) for continuous variables or as counts and percentages (%) for categorical variables.

CCI = Charlson comorbidity index.

Between 2008 and 2013, at teaching institutions, RPN rates increased from 2.2% to 50.1% (estimated annual percentage change [EAPC]: +32.1%, confidence interval [CI]: +19.0% to +49.6%, p: 0.01). Similarly, at nonteaching institutions RPN rates increased from 3.1% to 48.6% (EAPC: +30.1%, CI: +16.2% to +48.8%; p: 0.02) (Fig. 1). Furthermore, RPN rates increased in all NIS regions, namely the Midwest (EAPC: +28.0%, CI: +15.8% to +44.1%, p: 0.01), the Northeast (EAPC: +42.4%, CI: +28.9% to +60.4%, p: 0.004), the South (EAPC: +31.1%, CI: +16.3% to +51.9%, p: 0.02), and the West (EAPC: +25.8, CI: +10.7% to +46.9%, p: 0.03) (Fig. 2).

FIG. 1.

Graphical presentation of temporal trends for open, laparoscopic, and robotic partial and radical nephrectomy according to institutional teaching status (solid line: open, dot line: laparoscopic, and dashed line: robotic).

FIG. 2.

Graphical presentation of temporal trends for open, laparoscopic, and robotic partial and radical nephrectomy according to National Inpatient Sample regions (solid line: open, dot line: laparoscopic, and dashed line: robotic).

MLR models showed lower rates of RPN rates at teaching vs nonteaching institutes (odds ratio [OR]: 0.79, CI: 0.72–0.87, p < 0.001). In addition, after multivariable adjustment RPN patients were less frequently aged ≥75 years (OR: 0.81, CI: 0.67–0.99, p: 0.04) and less frequently harbored CCI 2 (OR: 0.78, CI: 0.66–0.91, p: 0.003) or CCI ≥3 (OR: 0.75, CI: 0.62–0.91, p: 0.005). Moreover, lower rates of RPN were recorded in the Northeast (0.52, CI: 0.47–0.59, p < 0.001), the South (OR: 0.63, CI: 0.56–0.69, p < 0.001), and West (0.76, CI: 0.66–0.86, p < 0.001) compared to the Midwest.

In addition, open PN rates decreased both at teaching (EAPC: −10.9%, p < 0.001) and at nonteaching (EAPC: −11.0%, p: 0.001) institutions. Similarly, laparoscopic PN rates also decreased both at teaching (EAPC: −20.5%, p: 0.002) and at nonteaching (EAPC: −18.5%, p: 0.003) institutions (Fig. 1).

Descriptive characteristics and rates of RRN over time

Overall, 24,872 RN patients were identified. Of these, 6.8% (N = 1683) underwent RRN. Of all RRN patients, 66.9% were treated at teaching institutions. Most of RRN patients were 50 to 64 years old (39.5%), male (63.5%), and Caucasian (70.2%). Most of RRN patients harbored CCI 0 (57.2%) and had Medicare insurance (47.3%). No differences in age and comorbidities were identified between RRN at teaching vs nonteaching institutions. Furthermore, RRN patients were more frequently treated at teaching institutions in the Midwest (36.8% vs 23.2%) (Table 1).

Between the years 2008 and 2013, we noticed several important trends. At teaching institutions, RRN rates increased from 0.6% to 14.8% (EAPC: +35.5%, CI: +20.7% to 56.3%, p: 0.01). Similarly, at nonteaching institutions, RRN rates increased from 0.4% to 12.1% (EAPC: +43.0%, CI: +31.3% to 58.1%, p: 0.002) (Fig. 1). Similar observations were made in all four NIS regions. Specifically, RRN rates increased in the Midwest (EAPC: +31.0%, CI: +14.7% to +54.8%, p: 0.02), in the Northeast (EAPC: +55.6%, CI: +33.2% to +93.3%, p: 0.007), in the South (EAPC: +47.4%, CI: +33.6% to +65.8%, p: 0.002), and in the West (EAPC: +27.2%, CI: +8.5% to +56.2%, p: 0.06) (Fig. 2).

Moreover, MLR models failed to find differences in rates of RRN at teaching vs nonteaching institutions (OR: 0.95, CI: 0.84–1.07, p: 0.4). In addition after multivariable adjustment, no differences in RRN rates were identified according to age or comorbidities (all p > 0.05). However, lower rates of RRN were recorded in the Northeast (0.63, CI: 0.54–0.73, p < 0.001), the South (OR: 0.48, CI: 0.42–0.54, p < 0.001), and West (0.81, CI: 0.69–0.94, p: 0.007) compared to the Midwest region.

In addition, open RN rates decreased at both teaching (EAPC: −2.1%, p: 0.003) and nonteaching (EAPC: −1.6%, p: 0.02) institutions. Similarly, laparoscopic PN rates also did not increase at both teaching (−5.4%, p: 0.002) and nonteaching institutions (−5.9%, p: 0.07) (Fig. 1).

RPN outcomes according to teaching vs nonteaching institution status

Overall, 43 patients had intraoperative complications (1.1%), and 1075 had postoperative complications (26.4%) after RPN. Most common complication categories were miscellaneous medical (9.4%), genitourinary (7.9%), and respiratory (7.2%). The median length of stay was 2 (interquartile range [IQR] 2–3) days. In-hospital mortality rate was 0.1% (Table 2).

Table 2.

Intraoperative and Postoperative Outcomes During Hospitalization in Patients Who Underwent Robotic Partial (N = 4070) and Radical (N =1683) Nephrectomy Stratified According to Institutional Teaching Status

	Robotic partial nephrectomy				Robotic radical nephrectomy
Outcome of interest	Overall (N = 4070)	Nonteaching (N = 1092)	Teaching (N = 2978)	p	Overall (N = 1683)	Nonteaching (N = 556)	Teaching (N = 1127)	p
Intraoperative complication	43 (1.1)	18 (1.6)	25 (0.8)	0.04	20 (1.2)	4 (0.7)	16 (1.4)	0.3
Postoperative complication
Overall	1075 (26.4)	308 (28.2)	767 (25.8)	0.1	569 (33.8)	218 (39.2)	351 (31.1)	0.001
Cardiac	134 (3.3)	40 (3.7)	94 (3.2)	0.5	93 (5.5)	33 (5.9)	60 (5.3)	0.7
Respiratory	295 (7.2)	92 (8.4)	203 (6.8)	0.09	141 (8.4)	62 (11.2)	79 (7.0)	0.005
Vascular	43 (1.1)	10 (0.9)	33 (1.1)	0.7	29 (1.7)	9 (1.6)	20 (1.8)	0.9
Operative wound	6 (0.1)	2 (0.2)	4 (0.1)	0.9	13 (0.8)	6 (1.1)	7 (0.6)	0.5
Genitourinary	323 (7.9)	71 (6.5)	252 (8.5)	0.04	123 (7.3)	44 (7.9)	79 (7.0)	0.5
Transfusion	224 (5.5)	75 (6.9)	149 (5.0)	0.02	141 (8.4)	58 (10.4)	83 (7.4)	0.04
Miscellaneous medical	382 (9.4)	123 (11.3)	259 (8.7)	0.01	293 (17.4)	117 (21.0)	176 (15.6)	0.007
Miscellaneous surgical	120 (2.9)	43 (3.9)	77 (2.6)	0.03	67 (4.0)	21 (3.8)	46 (4.1)	0.9
Length of stay, days	2 (2–3)	3 (2–3)	2 (2–3)	0.1	3 (2–4)	3 (2–4)	3 (2–4)	0.003
In-hospital mortality rate	5 (0.1)	3 (0.3)	2 (0.1)	0.2	10 (0.6)	5 (0.9)	5 (0.4)	0.4

Data are shown as medians (interquartile range) for continuous variables or as counts and percentages (%) for categorical variables.

Bold values are statistically significant.

After adjustment for all the covariates, MLR models showed higher rates of genitourinary complications at teaching vs nonteaching institutions (OR: 1.46, CI: 1.09–1.95, p: 0.01) (Table 3).

Table 3.

Univariable and Multivariable Analyses Predicting Main Outcomes of Robotic Partial Nephrectomy (N = 4070) in Teaching (N = 2987) vs Nonteaching (N =1092) Institutions

	Univariable analyses		Multivariable analyses
Outcome of interest	Odds ratio (95% confidence interval)	p	Odds ratio (95% confidence interval)	p
Intraoperative complication	0.89 (0.69–1.16)	0.4	0.55 (0.29–1.05)	0.07
Postoperative complication
Overall	0.87 (0.82–0.92)	<0.001	1.01 (0.86–1.20)	0.9
Cardiac	0.86 (0.59–1.26)	0.4	0.87 (0.58–1.31)	0.5
Respiratory	0.79 (0.62–1.03)	0.08	0.85 (0.65–1.12)	0.2
Vascular	1.21 (0.62–2.60)	0.6	1.46 (0.69–3.07)	0.3
Operative wound	0.73 (0.14–5.29)	0.7	0.95 (0.15–5.92)	0.9
Genitourinary	1.35 (1.17–1.56)	0.005	1.46 (1.09–1.95)	0.01
Transfusion	0.71 (0.54–0.96)	0.02	0.79 (0.58–1.07)	0.1
Miscellaneous medical	0.75 (0.60–0.94)	0.01	0.88 (0.69–1.13)	0.3
Miscellaneous surgical	0.65 (0.45–95)	0.02	0.74 (0.49–1.10)	0.1
In-hospital mortality rate	0.24 (0.03–1.47)	0.1	0.31 (0.03–2.99)	0.3

	Rate ratio (95% confidence interval)	p	Rate ratio (95% confidence interval)	p
Length of stay	0.98 (0.96–0.99)	0.02	1.03 (1.01–1.05)	0.001

Analyses adjusted for year of diagnosis, age at diagnosis, race, gender, Charlson comorbidity index, insurance status, region, bed size, and annual hospital volume (Reference: nonteaching institution).

Bold values are statistically significant.

RRN outcomes according to teaching vs nonteaching institution status

Overall, 20 patients had intraoperative complications (1.2%), and 569 had postoperative complications (33.8%) after RRN. Most common complication categories were miscellaneous medical (17.4%), transfusions (8.4%), and respiratory (8.4%). Median length of stay was 3 (IQR 2–4) days. In-hospital mortality rate was 0.6% (Table 2).

After adjustment for all covariates, MLR models showed higher rates of intraoperative complications (OR: 3.39, CI: 1.02–11.3; p: 0.04) after RRN at teaching vs nonteaching institution. Conversely, MLR models showed lower rates of respiratory (OR: 0.66, CI: 0.44–0.98; p: 0.04) complications after RRN at teaching vs nonteaching institution. Similarly, the length of stay was shorter after RRN at teaching vs nonteaching institution (rate ratio [RR]: 0.93, CI: 0.88–0.98; p: 0.01) (Table 4).

Table 4.

Univariable and Multivariable Analyses Predicting Main Outcomes of Robotic Radical Nephrectomy (N = 1683) in Teaching (N = 1127) vs Nonteaching (N = 556) Institutions

	Univariable analyses		Multivariable analyses
Outcome of interest	Odds ratio (95% confidence interval)	p	Odds ratio (95% confidence interval)	p
Intraoperative complication	1.99 (0.72–6.96)	0.2	3.39 (1.02–11.3)	0.04
Postoperative complication
Overall	0.70 (0.56–0.87)	<0.001	0.80 (0.63–1.02)	0.08
Cardiac	0.89 (0.58–1.39)	0.6	0.98 (0.59–1.60)	0.9
Respiratory	0.60 (0.42–0.85)	<0.001	0.66 (0.44–0.98)	0.04
Vascular	1.09 (0.51–2.55)	0.8	0.97 (0.40–2.36)	0.9
Operative wound	0.57 (0.19–1.78)	0.3	0.79 (0.23–2.72)	0.7
Genitourinary	0.88 (0.60–1.29)	0.5	1.00 (0.89–1.12)	0.9
Transfusion	0.68 (0.48–0.97)	0.03	0.69 (0.47–1.05)	0.08
Miscellaneous medical	0.69 (0.53–0.90)	0.006	0.87 (0.65–1.17)	0.4
Miscellaneous surgical	1.08 (0.65–1.87)	0.76	1.39 (0.77–2.50)	0.3
In-hospital mortality rate	0.76 (0.16–3.45)	0.73	1.14 (0.79–1.65)	0.5

	Rate ratio (95% confidence interval)	p	Rate ratio (95% confidence interval)	p
Length of stay	0.87 (0.82–0.92)	<0.001	0.93 (0.88–0.98)	0.01

Bold values are statistically significant.

Discussion

We hypothesized that robotic nephrectomies performed at teaching institutions are used equally frequently and result in similar complications, length of stay, and mortality rate profiles, when compared to nonteaching institutions. We tested this hypothesis in two separate patient cohorts, namely in patients treated with RPN or RRN. Our results demonstrated several noteworthy observations.

First, we noted lower rates of RPN rates at teaching vs nonteaching institutions. Conversely, we recorded statistically significant differences in RPN rates according to region of residence. Specifically, the proportion of RPNs were 28.1%, 36.5%, 14.9%, and 20.6% in, respectively, the Midwest, the South, West, and the Northeast. These proportion differences remained statistically significant even after adjustment for all covariates. These differences in overall rate of RPN use, stratified according to the region, reveal an increase in RPN rates in all examined regions. Specifically, these increases were of +42.4%, +31.1%, +28.0%, and +25.8% in, respectively, the Northeast, the South, the Midwest, and West.

Second, we also identified differences in age and comorbidity distribution among patients treated with RPN. Specifically, a smaller proportion of patients in the oldest age category (≥75 years) were treated with RPN than with other PN approaches. Similarly, patients treated with RPN less frequently harbored CCI 2 or CCI ≥3. Taken together, these observations indicate a tendency to select younger and healthier patients for RPN, instead of open or laparoscopic PN. This selection, according to age or comorbidities, was not reported in previous analyses.^9,10

It is also of note to see that patients treated with RPN at teaching institutions were younger and harbored lower CCI score. It may be argued that surgeons at teaching hospitals are more motivated to proactively recommend surgery^11,12; this tendency could be even more pronounced in young and healthy patients with small renal masses due to protruded life expectancy of these individuals. It is also possible that resection of small renal masses in young and healthy individuals is referred out from nonteaching hospitals to teaching institutions as second opinion. It is not uncommon that nonteaching institution obtains second opinion to validate their treatment choice; these second opinion referrals are virtually universally made to center of excellence that falls in the umbrella of teaching institutions. Indeed, many patients may opt for a treatment in teaching institutions instead of returning to the institution of origin.

Third, we made an important observation regarding rate of RPN (OR: 0.79, p < 0.001) at teaching institutions, which is lower than in nonteaching institutions and appears counterintuitive at first. However, it may be explained by self-selection of smaller and more favorably positioned renal masses that are treated at nonteaching institutions. Conversely, larger and less favorably anatomically situated renal masses may be more frequently referred to teaching institutions.¹³ Some of these may be treated with PN, but many others may unfortunately require a RN. Such treatment distribution may in consequence result in higher rates of RRN at teaching institutions.

Fourth, we also observed significant regional differences in RRN use. Indeed, the proportion of RRNs were 32.3%, 17.8%, 17.3%, and 32.6% in, respectively, the Midwest, West, the Northeast, and the South. Moreover, as for RPN, we also observed an important increase in RRN overtime in all four examined regions. The rate of increase of RRN differed according to the four examined regions and occurred most rapidly in the Northeast (EAPC: +55.6%) and least rapidly in West (EAPC: +27.2%). Taken together, these observations demonstrate important regional variability that is not accounted for by patient case mix characteristics neither in the overall analyses nor in the analyses overtime. In consequence, both the rate of robotics uptake for RN, as well as for PN, and its speed differ according to region without obvious cause. This said, the granularity of data within the NIS does not allow us to test for differences such as the RENAL score or others, which could potentially shed further light on this unexplained variability.

Fifth, unlike for RPN, within the RRN population we did not identify statistically significant differences in age and CCI according to teaching vs nonteaching institution status. This observation is in agreement with our hypothesis about broader use of RRN at teaching institutions, as discussed above.

In the second part of our analyses, we focused on RPN and RRN complications, length of stay, and mortality rate. Regarding RPN, we identified differences in one of nine examined complication categories. Specifically, RPN resulted in higher rates of genitourinary complications at teaching institutions. The absolute rate differences were, however, marginal between teaching and nonteaching institutions. Indeed, it is necessary to treat 50 patients with RPN at teaching institutions to have a genitourinary complication more compared to nonteaching institutions. For RRNs, we also identified differences in two of nine complication categories. Specifically, the pulmonary complications and length of stay favored RRN at teaching institutions. Conversely, intraoperative complications favored RRN at nonteaching institutions. However, as for RPN the absolute rate differences were marginal. Indeed, it is necessary to treat with RRN about 143 patients to develop an intraoperative complication more at teaching compared to nonteaching institutions; similarly, it is necessary to treat with RRN about 23 patients to develop a respiratory complication more at nonteaching compared to teaching institutions. Taken together, these observations suggest no clinically meaningful advantage when either RPN or RRN were performed at teaching or nonteaching institutions. Moreover, such benefits may only result in marginal meaningful differences, when they are calculated for large patient cohorts. Conversely, the observed differences are unlikely of importance when individuals or even single institutions are considered. In consequence, based on epidemiologic considerations, teaching institutions appear to hold a marginal although statistically significant advantage over nonteaching institutions. However, from an individual perspective, RRNs do not appear to warrant individual referrals to teaching institutions.

It is also of interest that higher complication rates were recorded with RRN than RPN (33.8% vs 26.4%). This observation may be explained by more challenging nature of RRN cases from a surgical or anatomical or patient perspective. Such a negative selection bias may alter results in worse outcomes as was observed in our analyses. Moreover, the rate of genitourinary complications in RPN was higher at teaching hospitals than at nonteaching hospitals (8.5% vs 6.5%). A negative selection toward more anatomical, surgical, and patient characteristics, as outlined before, may account for these differences.

Our findings regarding the effect of institutional teaching status on RPN and RRN outcome corroborate those of previous studies that relied on smaller cohorts. Khene and colleagues showed longer surgery duration and warm ischemia time when RPNs were performed by posturologic residency trainees (fellows) under staff surgeon supervision relative to RPNs performed by staff surgeons alone in a series of 216 cases from French institution.¹⁴ Similarly, Cerantola and colleagues compared 46 vs 23 RPN performed by staff surgeon vs fellows. Their results showed no statistically significant differences in complication rates and surgical margins within the two groups.¹⁵

The comparison between previous and our series is not straightforward since many differences distinguish the current population from those of previous analyses. The main and foremost of these is the definition of teaching status; the latter rest on substantially broader inclusion within our report that extends well beyond the simple act of surgery.

Although no other contribution focused on the effect of teaching vs nonteaching status in the context of RPN or RRN, several investigators examined the effect of teaching institution status on outcomes of radical prostatectomy or other common urologic interventions^6,16,17; of these, radical prostatectomy demonstrated superior outcomes, when performed at teaching institutions.^6,16 Conversely, other common procedures such as circumcision, hydrocelectomy, transurethral prostatectomy, and mid-urethral sling showed longer operation time at teaching institution.¹⁷ In consequence, it is important to continue monitoring outcome differences between teaching and nonteaching institutions for at least those surgeries for which a historical evidence of teaching institution benefit existed.

Despite his novelty, our work is not devoid of limitations. More important, the NIS dataset does not code for tumor stage and grade nor is it possible to examine the temporality of diagnostic codes. Moreover, use of diagnostic code represents a well-established method for analyses in population-based studies. However, even the ICD-9-CM codes may not capture all the events and may have led to underestimation of complication rates. Moreover, an inherent grade of variability may exist within the single institution in residents' exposure and involvement to the different surgical procedures and in postoperative management.¹⁸ Furthermore, surgeon experience represents a key variable that should ideally be used for adjustment of outcomes along with hospital volume that represents an aggregate indicator of cumulative surgical experience within one institution. Moreover, the teaching institution status may influence also the bed assistant expertise and, consequently, the examined surgical outcomes. The NIS dataset does not allow to control for this possible confounder. However, conflicting results are reported in literature about the effect of bed assistant expertise on surgical outcomes.^13,19 Finally, all the limitations typical of population-based database, such as NIS or others, apply also to our study.

Conclusion

Despite statistically significant differences in rates of RPN use according to the institutional teaching status and despite statistically significant difference in selected complications, no meaningful differences distinguish teaching vs nonteaching institution when RPN and RRN are considered.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Mohapatra

, Potretzke

, Weaver

, Anderson

, Vetter

, Figenshau

. Trends in the management of small renal masses: A survey of members of the Endourological Society. J Kidney Cancer VHL, 2017; 4:10–19.

Pak

, Lee

, Bilal

, Finkelstein

, Palese

. Utilization trends and outcomes up to 3 months of open, laparoscopic, and robotic partial nephrectomy. J Robot Surg, 2017; 11:223–229.

Ayanian

, Weissman

, Chasan-Taber

, Epstein

. Quality of care for two common illnesses in teaching and nonteaching hospitals. Health Aff (Millwood), 1998; 17:194–205.

Dimick

, Cowan

, Colletti

, Upchurch

. Hospital teaching status and outcomes of complex surgical procedures in the United States. Arch Surg, 2004; 139:137–141.

Allison

, Kiefe

, Weissman

, et al. Relationship of hospital teaching status with quality of care and mortality for Medicare patients with acute MI. JAMA, 2000; 284:1256–1262.

Trinh

, Sun

, Kim

, et al. The impact of hospital volume, residency, and fellowship training on perioperative outcomes after radical prostatectomy. Urol Oncol Semin Orig Invest, 2014; 32:29.e13–29.e20.

Trinh

, Karakiewicz

, Sammon

, et al. Venous thromboembolism after Major Cancer surgery: Temporal trends and patterns of care. JAMA Surg, 2014; 149:43–49.

, Gu

, Lipsitz

, et al. Comparative effectiveness of minimally invasive vs open radical prostatectomy. JAMA, 2009; 302:1557–1564.

Leow

, Heah

, Chang

, Chong

, Png

. Outcomes of robotic versus laparoscopic partial nephrectomy: An updated meta-analysis of 4,919 patients. J Urol, 2016; 196:1371–1377.

10.

Luciani

, Chiodini

, Mattevi

, et al. Robotic-assisted partial nephrectomy provides better operative outcomes as compared to the laparoscopic and open approaches: Results from a prospective cohort study. J Robot Surg, 2017; 11:333–339.

11.

Elliott

, Smith

, Raman

. Are urology residents adequately exposed to conservative therapies for managing small renal masses?. J Endourol, 2011; 25:129–133.

12.

Vigneswaran

, Lec

, Brito

, Turini

, Pareek

, Golijanin

. Partial nephrectomy for small renal masses: Do teaching and nonteaching institutions adhere to guidelines equally?. J Endourol, 2016; 30:714–721.

13.

Mitsinikos

, Abdelsayed

, Bider

, et al. Does the level of Assistant experience impact operative outcomes for robot-assisted partial nephrectomy?. J Endourol, 2017; 31:38–42.

14.

Khene

, Peyronnet

, Bosquet

, et al. Does training of fellows affect perioperative outcomes of robot-assisted partial nephrectomy?. BJU Int, 2017; 120:591–599.

15.

Cerantola

, Ploussard

, Kassouf

, Anidjar

, Bladou

. Does teaching of robotic partial nephrectomy affect renal function and perioperative outcomes?. Urol Oncol Semin Orig Invest, 2017; 35:227–233.

16.

Trinh

, Schmitges

, Sun

, et al. Radical prostatectomy at academic versus nonacademic institutions: A population based analysis. J Urol, 2011; 186:1849–1854.

17.

Welk

, Winick-Ng

, McClure

, Vinden

, Dave

, Pautler

. The impact of teaching on the duration of common urological operations. Can Urol Assoc J, 2016; 10:172–178.

18.

Cocci

, Patruno

, Gandaglia

, et al. Urology residency training in Italy: Results of the first national survey. Eur Urol Focus, 2016. Available at: https://www-sciencedirect-com.web.bisu.edu.cn/science/article/pii/S2405456916300669 Accessed August 17, 2016 .

19.

Potretzke

, Knight

, Brockman

, et al. The role of the assistant during robot-assisted partial nephrectomy: Does experience matter?. J Robot Surg, 2016; 10:129–134.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB