National Multi-Institutional Comparison of 30-Day Postoperative Complication and Readmission Rates Between Open Retropubic Radical Prostatectomy and Robot-Assisted Laparoscopic Prostatectomy Using NSQIP

Abstract

Background:

Many American hospitals will soon face readmission penalties deducted from Medicare reimbursements, which will place further scrutiny on techniques that may offer reduced postoperative morbidity. We aimed to perform the first multi-institutional study using the National Surgical Quality Improvement Program (NSQIP) database, to compare predictors of readmission within cohorts of open radical retropubic prostatectomy (RRP) and robot-assisted laparoscopic radical prostatectomy (RALRP) in a contemporary nationwide series of radical prostatectomy.

Methods:

All patients who underwent radical prostatectomy in 2011 were identified in the NSQIP database using procedural codes. As no patients in the analysis underwent LRP, patients were grouped as RRP or RALRP for analysis. Perioperative variables were analyzed using chi-squared and Student's t-tests as appropriate. Multiple logistic regression was used to identify readmission risk factors.

Results:

Of 5471 patient cases analyzed, 4374 (79.9%) and 1097 (20.1%) underwent RALRP and RRP, respectively. RRP and RALRP cohorts experienced different readmission rates (5.47% vs 3.48%, respectively; p=0.002). In addition, RRP experienced a higher rate of overall complications than RALRP (23.25% vs 5.62%, respectively; p<0.001), but not higher rates of reoperation (1.09% vs 0.96%, respectively; p=0.689). Overall predictors of readmission included operative time, dyspnea, and RRP or RALRP procedure type. Current smoking and patient age were predictive of readmission for RRP only, while dyspnea was predictive of readmission following RALRP only.

Conclusion:

This is the first multi-institutional retrospective study that examines readmission rates and procedural intracohort predictors of readmission for RRP in the contemporary United States. We report a significant difference in postoperative complication and readmission rates in RRP compared with RALRP. Further prospective analysis is warranted.

Introduction

In the last decade robot-assisted laparoscopic radical prostatectomy (RALRP) has largely replaced radical retropubic prostatectomy (RRP) in the United States, such that >75% of all radical prostatectomies are currently performed in this manner.^1
–3 Further, a recent report from the National Inpatient Sample (NIS) found that minimally invasive prostatectomy (MIRP; RALRP and LRP) increased from 1.4% in 2002, to 29.5% in 2008. Despite widespread acceptance, oncological outcomes and improved complication rates have not been proven over the open approach.

Beginning in 2013, 2217 American hospitals will face “readmission penalties” deducted from Medicare reimbursements.⁴ The Patient Protection and Affordable Care Act, passed in 2009, was designed to slow the increase in costs of health care delivery, and facilitate affordable, quality care to all Americans.⁵ At an additional cost of $17.4 billion per year, readmissions for Medicare patients occur at a reported rate of nearly 20% annually.⁶ The Hospital Readmission Reduction Program will define a readmission in excess of a ratio at a given hospital compared to the 30-day readmission national average as “excess” and will begin levying penalties on offending institutions.⁷

The National Surgical Quality Improvement Program (NSQIP) database was instituted by the American College of Surgeons in 2004 and provides a comprehensive database of pre-, intra-, and postoperative information for major surgical procedures at over 400 institutions across the United States. Participating hospitals must meet rigorous reporting standards in order to participate in the program and data have been shown to be accurate and reliable.^8,9

In this study, we aimed to perform the first multi-institutional study using the NSQIP database, comparing postoperative morbidities, readmission rates, and predictors of readmission within RRP and RALRP cohorts in a contemporary nationwide series of radical prostatectomy. We also examined predictors of readmission overall.

Methods

Population

A retrospective analysis was performed on data collected from the 2011 NSQIP participant-use files. This was the first-year readmission information became available. The data-collection methods for NSQIP have been previously described in detail.^10
–12 Briefly, 240 variables, including patient demographics, comorbidities, laboratory values, perioperative details, and 30-day risk-adjusted outcomes, were prospectively collected for each patient. To ensure integrity and accuracy of the database, certified nurse reviewers are rigorously trained to collect patient information according to standardized definitions and data are audited regularly.

Data were collected on patients who underwent radical prostatectomy from 2011 only, as this is the first year with NSQIP readmission data available.¹³ Patients were identified using primary Current Procedural Technology codes 55840, 55842, and 55845 for RRP, and 55866 for MIRP. Inspection of the procedure characterization note in the database for MIRP revealed that all the procedures logged were RALRP.

Variables

The primary outcome of interest tracked by the 2011 NSQIP database is “Unplanned Readmission,” defined as hospital readmission within 30 days related to the principle procedure, but not planned at the time of the procedure.¹⁴ Other outcomes tracked include postoperative complications, reoperation, and mortality. Complications, as defined by NSQIP, include surgical site (e.g., superficial, deep, or organ-space surgical site infection [SSI] and wound disruption), hematologic (e.g., transfusion rates, deep venous thrombosis [DVT], or pulmonary embolism [PE]), respiratory (e.g., unplanned intubation, pneumonia, and mechanical ventilation for longer than 48 hours), cardiovascular (e.g., myocardial infarction and cardiac arrest), genitourinary (e.g., renal insufficiency, acute renal failure, and urinary tract infection [UTI]), neurologic (e.g., cerebrovascular accident, peripheral nerve deficit, and coma lasting >24 hours), and multisystem (e.g., sepsis and septic shock) complications. Reoperation is defined as a return to the operating room within 30 days of the original procedure, while mortality includes all deaths, regardless of etiology, occurring within 30 days of the original procedure.

Preoperative variables include demographics, lifestyle variables, and comorbidities. Medical comorbidities include diabetes, dyspnea, hypertension, chronic obstructive pulmonary disease (COPD), congestive heart failure, bleeding disorders, prior angioplasty or cardiac surgery, previous stroke or transient ischemic attack, radiotherapy within 90 days of operation, chemotherapy within 30 days of operation, and previous operations within 30 days of the procedure of interest. Demographic variables include age, body mass index, and race. The only lifestyle variables included in the analysis were smoking and alcohol use. Operative outcomes specific to radical prostatectomy, including continence, erectile dysfunction, and urine leak, are not reported in NSQIP and as such are not included in this analysis.

Statistical analysis

Perioperative variables were analyzed using chi-squared tests and Student's t-tests for categorical and continuous variables, respectively. Analysis was used within each cohort to identify differences between readmitted and nonreadmitted patients in order to discern predictors. All perioperative variables with n≥10 and p<0.2 were identified as possible predictors of readmission, and included in a multiple logistic regression to control for confounding variables. This regression yielded an independent odds ratio (OR), 95% confidence interval (CI), and p-value for each potential predictor. Hosmer-Lemeshow (H-L) and c-statistics were calculated to assess goodness of fit and discriminatory capability, respectively. In the context of logistic regressions and risk analysis, these statistics serve to assess the model's ability to predict the data set from which it was derived.^15
–17 Significance was defined as p<0.05. All analysis was performed using SPSS version 20 (IBM Corp., Armonk, NY).

Results

Population demographics

A total of 5471 patients met inclusion criteria. One thousand ninety-seven patients (20%) and 4374 (80%) underwent RRP and RALRP procedures, respectively. Table 1 shows the distribution of demographic data and comorbidities between RRP and RALRP cohorts. The RRP cohort had higher rates of alcohol use (3.83% vs 2.17%, p=0.002) and chemotherapy not limited to hormonal therapy administered 30 days prior to surgery (0.55% vs 0.14%, p=0.02). The RRP group was also more likely to be older (age 70–80: 17.0% vs 12.3%; age >80 years: 1.6% vs 0.4%, p<0.001) and non-Caucasian (Caucasian: 61.53% vs 81.02%, p<0.001). Conversely, the percentage of patients aged <50 and between 50 and 60 was higher in the RALRP cohort (4.8% vs 3.37%, p=0.042, and 32.85% vs 29.08%, p=0.017, respectively). The RALRP group also had higher rates of hypertension (50.76% vs 46.95%, p=0.023).

Table 1.

Patient Demographics and Comorbidities

	RP		RALRP
Preoperative variable	n=1097	%	n=4374	%	p-Value
Age
<50	37	3.4	210	4.8	0.042^a
50–60	319	29.1	1437	32.9	0.017^a
60–70	536	48.9	2169	49.6	0.686
70–80	186	17.0	539	12.3	<0.001^a
>80	18	1.6	19	0.4	<0.001^a
White	675	61.5	3544	81.0	<0.001^a
Black	43	3.9	392	9.0	<0.001^a
Asian	10	0.9	64	1.5	0.157
Other	6	0.5	10	0.2	0.081
Active smoker	123	11.2	493	11.3	0.956
>2 Drinks/day within 2 weeks of admission	42	3.8	95	2.2	0.002^a
Steroid treatment	15	1.4	40	0.9	0.179
Radiotherapy <90 days	0	0.0	2	0.0	0.639
Chemotherapy <30 days (not hormonal therapy)	6	0.5	6	0.1	0.02^a
Previous operation <30 days	2	0.2	9	0.2	0.616
Dyspnea (difficult, painful, or labored breathing)	37	3.4	125	2.9	0.368
Diabetes	130	11.9	463	10.6	0.228
History of severe COPD (functional disability, hospitalization, chronic bronchodilator therapy, or FEVI of <75%)	19	1.7	81	1.9	0.791
Hypertension requiring medication	515	46.9	2221	50.8	0.023^a
Previous cardiac surgery or PCI	36	3.3	113	2.6	0.204
History of transient ischemic attacks	10	0.9	21	0.5	0.089
Bleeding disorders	5	0.5	46	1.1	0.041^a
Resident physician participation in surgery	423	66.4	1237	64.3	0.341
Percent admitted from inpatient setting	9	0.8	255	5.8	<0.001^a
History of MI within 30 days of surgery	0	0.0	1	0.0	0.799
Cancer has spread beyond the primary site and multiple metastases indicate the cancer is widespread, fulminant, or near terminal	9	0.8	17	0.4	0.06
Congestive heart failure within 30 days of surgery	5	0.5	8	0.2	0.1

	RP		RALRP
	Mean	SD	Mean	SD	p-Value
Age (years)	63.1	7.4	61.7	7.2	<0.001^a
Work-relative value units	28.3	2.6	32.5	0.0	<0.001^a
Operative time (minutes)	174.0	84.7	212.3	73.7	<0.001^a
BMI	28.5	5.2	28.6	4.8	0.042^a

Significant value, p<0.05.

BMI=body mass index; COPD=chronic obstructive pulmonary disease; FEVI=forced expiratory volume in one second; MI=myocardial infarction; PCI=percutaneous coronary intervention; RALRP=robot-assisted laparoscopic radical prostatectomy.

Comparison of outcomes between cohorts

Outcome distributions across the RRP and RALRP groups are displayed in Table 2. The overall complication rate for RRP patients was 23.2% and RALRP was 5.62%. The surgical complication rate for RRP was 3.37% and RALRP was 0.91%. RRP patients were more likely to be readmitted (5.47% vs 3.48%, p=0.002), suffer increased overall complication rates (23.25% vs 5.62%, p<0.001), and increased surgical site complications (3.37% vs 0.91%, p<0.001) (including superficial SSI [2.01% vs 0.48%, p<0.001], organ space infection [0.82% vs 0.3%, p=0.014], and wound disruption [0.55% vs 0.16%, p=0.019]).

Table 2.

Postoperative Outcomes

	RP		RALRP
	n=1097	%	n=4374	%	p-Value
Overall complications	255	23.25	246	5.62	<0.001^a
Surgical complications	37	3.37	40	0.91	<0.001^a
Superficial infection	22	2.01	21	0.48	<0.001^a
Deep infection	1	0.09	0	0.00	0.201
Organ space infection	9	0.82	13	0.30	0.02
Wound disruption	6	0.55	7	0.16	0.03^a
Medical complications^b	238	21.70	221	5.05	<0.001^a
DVT	6	0.55	24	0.55	0.604
Pulmonary embolism	10	0.91	20	0.46	0.061
Unplanned re-intubation	4	0.36	10	0.23	0.303
Ventilator treatment >48 hours	3	0.27	5	0.11	0.204
Renal insufficiency	7	0.64	13	0.30	0.087^a
Acute renal failure	4	0.36	5	0.11	0.086
Stroke	2	0.18	3	0.07	0.264
Cardiac arrest	3	0.27	4	0.09	0.149
Myocardial infarction	7	0.64	5	0.11	0.004^a
Peripheral nerve injury	0	0.00	1	0.02	0.799
Pneumonia	7	0.64	9	0.21	0.027^a
UTI	35	3.19	75	1.71	0.002^a
Transfusion	194	17.68	82	1.87	<0.001^a
Sepsis/septic shock	19	1.73	28	0.64	<0.001^a
Death	3	0.27	6	0.14	0.263
Reoperation	12	1.09	42	0.96	0.689
Unplanned readmission	60	5.47	152	3.48	0.002^a

Significant value, p<0.05.

Patients may have more than one complication.

DVT=deep venous thrombosis; UTI=urinary tract infection.

RRP patients were more likely to suffer nonsurgical site complications (21.7% vs 5.05%, p<0.001), including myocardial infarction (0.64% vs 0.11%, p=0.004), pneumonia (0.64% vs 0.21%, p=0.018), UTI (3.19% vs 1.71%, p=0.002), blood transfusion requirements (17.68% vs 1.87%, p<0.001), and sepsis or septic shock (1.73% vs 0.64%, <0.001). Additional perioperative complications, such as venous thromboembolism and cardiac arrest, were comparable and less than one percent for both procedures.

Operating room time was significantly shorter for RRP (174 vs 212 minutes, p<0.001) and reoperation rates did not significantly differ between cohorts (1.09% vs 0.96%, p=0.689).

Regression analysis of readmissions within cohorts

Predictors of unplanned readmission following prostatectomy included operative time (OR=1.004, p<0.001), patient age (OR=1.135, p=0.035), and whether they underwent RRP or RALRP (OR=0.538, p=0.006) (see Table 3). Independent predictors of unplanned readmission following RRP included operative time (OR=1.005, p<0.001), current smoking (OR=2.03, p=0.049), and patient age (OR=1.039, p=0.035) (see Table 4). Independent predictors of unplanned readmission following RALRP included dyspnea (OR=3.308, p<0.001) and operative time (OR=1.004, p<0.001) (see Table 5).

Table 3.

Multivariate Regression for Unplanned Readmissions

		95% CI
Variable	OR	Lower	Upper	p-Value
Operation time	1.0	1.0	1.0	<0.001^a
Dyspnea	2.5	1.4	4.4	0.001^a
RRP vs RALRP	0.5	0.3	0.8	0.006^a
History of severe COPD	1.4	0.7	3.0	0.386
Hypertension	1.3	1.0	1.8	0.074
ASA classification 3&4	1.2	0.9	1.6	0.298
Work-relative value units	1.0	0.9	1.1	0.494
BMI >30	1.2	0.9	1.6	0.272
Diabetes	1.0	0.6	1.5	0.914
Current smoker	1.2	0.8	1.9	0.325
H-L value	0.468
c-Statistic	0.644

Significant value, p<0.05.

ASA=American Society of Anesthesiology; CI=confidence interval; H-L=Hosmer-Lemeshow; OR=odds ratio; RRP=radical retropubic prostatectomy.

Table 4.

Predictors of Readmission in RRP Cohort

		95% CI
Predictor	OR	Lower	Upper	p-Value
Current smoker	2.0	1.0	4.1	0.049^a
Operative time	1.0	1.0	1.0	<0.001^a
Age of patient	1.0	1.0	1.0	0.035^a
Diabetes	1.1	0.5	2.4	0.756
Hypertension requiring medication	1.4	0.8	2.6	0.215
ASA classification 3&4	1.7	0.9	3.1	0.061
H-L value	0.498
c-Statistic	0.711

Significant value, p<0.05.

Table 5.

Predictors of Readmission in RALRP Cohort

		95% CI
Predictor	OR	Lower	Upper	p-Value
Dyspnea	3.3	1.8	6.0	<0.001^a
Total operative time	1.0	1.002	1.0	<0.001^a
Hypertension requiring medication	1.3	.91	1.8	0.151
ASA classification 3&4	1.0	.73	1.5	0.813
Obesity	1.2	.84	1.7	0.322
Age of patient	1.0	.98	1.0	0.518
H-L value	0.782
c-Statistic	0.615

Significant value, p<0.05.

Discussion

Our study represents a national multi-institutional analysis from the NSQIP database to compare contemporary complications and readmission rates within RRP and RALRP cohorts in the first 30 days postoperatively. By examining outcomes that occurred in a single year, we avoided interyear variance in the definitions of preoperative variables and postoperative outcome measures.¹⁸ We report significantly increased rates of readmission and postoperative complications in RRP, specifically wound infection, myocardial infarct, pneumonia, UTI, blood transfusion, and sepsis. Broadly, age, operative time, and undergoing RRP were significant predictors of readmission on multivariate analysis. This is consistent with the findings of Liu et al. who examined overall predictors of readmission in prostatectomy but declined to elucidate intracohort predictors.¹⁹

Retrospective analysis would suggest marked differences in men who chose to undergo RRP vs RALRP, and is a contemporary examination of nationwide complications and readmission rates. Rather than a single-institution series of excellence, a broad analysis from such a large cohort of institutions may reflect the more “real life” management in patients undergoing the different surgical modalities. Our study is similar to recent results reported from a cohort of 2741 Taiwanese men identified from a nationwide data registry similar to NSQIP.²⁰ In support of our findings, they too report a significantly lower readmission rate for patients who underwent RALRP compared with RRP at 90 days postoperatively, even after adjusting for confounding variables (patient characteristics, Charlson comorbidity index, surgeon age, and case volume).

Interestingly, a study by Nelson et al. that compared the nonsurgical outcomes between RRP (n=374) and RALRP (n=629) was reported from an early robotic series between 2003 and 2006.²¹ In this study, 94.3% of RRP patients and 97.5% of RALRP patients were discharged home on or before postoperative day 1. The mean length of stay, readmission rates, and unscheduled clinic or emergency room visits were the same in both cohorts. There was a nonsignificant trend for increased symptomatic ileus in the RALRP group. Of note, the RALRP group tended to have less comorbidity and less obesity in this study, indicating a probable surgical selection bias present in the early days of RALRP.

In contrast, a recently published single-institution series from Johns Hopkins examined the trends in perioperative morbidity and discharge after open and MIRP over the last 20 years.²² Of 18,049 men who underwent RRP between 1991 and 2011, 83.8% were RRP, 7.8% were RALRP, and 6.9% were LRP. Over the study period the length of stay for all surgical approaches dramatically reduced, and the authors acknowledge the influence MIRP had on this. However, complications including postoperative anemia requiring transfusion and operative and nonoperative urine leak were more common in RALRP than RRP of LRP. RRP had a greater proportion of men with hematuria requiring irrigation; however, rates of reoperation for bleeding were similar among RRP and RALRP. RALRP patients were more likely to experience postoperative ileus (nearly a sevenfold increase compared with RRP or LRP), and this is postulated to occur due to the transperitoneal approach of RALRP, exposure of the abdominal viscera to chemical irritant, exposure to CO₂, or patient positioning.

Touijer et al. previously reported a higher medical and surgical complication rate for LRP (n=1134) compared with RP (n=3458) (RALRP was not analyzed in this study), in which they used a very thorough complication reporting template from a single high-volume institution.²³ Touijer et al. also reported that LRP was associated with less blood loss and lower transfusion rate but a higher hospital emergency room and readmission rate. The emergency room attendance rate was 15% after LRP and 11% after RRP, and of these 4.6% of LRP patients were readmitted and 1.2% in the RRP group. Interestingly, the mean hospital stay was on average 1.3 nights shorter for men who received LRP compared with RRP and this may account for the higher readmission rate.

Although our analysis did not include any LRP, this minimally invasive approach is becoming less attractive due to a long learning curve and the technical challenges it presents to the surgeon. It is interesting that from over 400 institutions across the United States in our study not a single one performed LRP, probably reflecting the abundant availability of RALRP and/or patient or surgeon preference. Our findings are likely a broad reflection of national trends toward the adoption of robotic and minimally invasive techniques in response in increased safety record.²⁴ Perioperative safety of RALRP is driven by the large difference in perioperative blood loss and transfusion requirements between the groups and the decreased risk of infectious complications in patients who do not require additional blood transfusions in the perioperative period.

Our data show that the RRP and RALRP cohorts were similar with respect to most demographic indicators and comorbidities (Table 1). However, a few significant differences were noted. The RRP cohort had a higher average age, and was less likely to be Caucasian. Chung et al. observed similar results in Taiwan and were able to analyze their results against geographical information and surgeon biographical data to which their database had access. To better adjust for socioeconomic variables and surgeon experience, they included a comparison of the population undergoing prostatectomy in the more affluent North versus other as well as surgeon age and throughput.²⁰ They found that RALRP patients were more likely to reside in the North (p<0.001) and also more likely to have surgeons aged ≤40 years (p<0.001). RALRP patients also had a greater likelihood of having a surgeon with a higher number of prostatectomy cases per year (p<0.001).²⁰ Although our study is limited by a lack of access to such information because our other results are similar, these may help explain part of the disparity between the cohort groups. The disparity may also be indicative that these populations may be initially more apprehensive about new technology and procedures or that their access to it may be restricted.²⁵

The RRP patients appear to have increased risk of superficial SSI, organ space infection, and wound disruption. Higher rates of SSIs are not surprising because longer incisions result in a greater fascial defect and larger exposed surface area for microbe entry. Reduced SSIs in minimally invasive procedures have been previously reported.²⁶ Of importance are the other significantly increased complications observed in this analysis associated with the open approach, such as myocardial infarction, pneumonia, UTI, sepsis or septic shock, and significant bleeding (requiring transfusion). Hu et al. previously reported that patients undergoing MIRPs experienced a shorter length of stay and fewer overall surgical complications but greater genitourinary complications, incontinence, and erectile dysfunction.²⁷ Although our study could not examine these outcomes specifically, their results may have been reflective of surgeon inexperience with novel procedural techniques.

Retrospective studies should be interpreted with some caution. There are many confounding variables that could potentially explain these findings and those of others. These include patient selection bias, comorbidities, and longer periods of immobilization in the RRP group. Interestingly, while the risk of readmission increased with patient age in the RRP cohort, this was not the case for patients undergoing RALRP. The higher risk of readmission for RRP may arise from younger patients' ability to better respond to surgical trauma and resulting inflammation.²⁸

The predominant predictor for readmission shared within the two groups was operative time although this was significantly higher for the RALRP group (174 vs 212 minutes) despite a lower readmission rate. Operative time may be a surrogate for surgical complexity or intraoperative complications, and is associated with more postoperative complications.²⁹ However, operative time may also be indicative of unusual anatomy, previous surgery, or obesity. However, specific predictors for readmission in RALRP patients only included dyspnea. Patients with dyspnea may have a greater likelihood of suffering respiratory problems related to insufflation for laparoscopy.⁸ Specific predictors for readmission in RRP included current smoking and patient age. Smoking has been shown in multiple studies to predict readmission and surgical complications.³⁰

Our study has a number of limitations. NSQIP does not allow for the specific identification whether patients were on medications that could predispose to bleeding, severity of comorbidities, and other pathology information on disease severity, including tumor characteristics. Selection bias may exist for the surgical techniques offered to each particular patient and may reflect low-volume surgeon preference for RRP. A randomized controlled examination of both techniques would be required to produce a more robust matching of preoperative variables. Neither the specific conditions for reoperation nor for which “unplanned” readmissions occurred is tracked by the database. Not every readmission is identical; it was not possible to perform an analysis after stratifying for readmission.

It may be possible that the results may be explained by a residual confounding unmeasured socioeconomic variable. Experience of the surgeon has not been measured or indeed tumor variables that may have influenced the decision toward an open approach. Hospital readmission rates have been used as a tool in the evaluation of clinical care quality. We do not have data on how many men visited the emergency department and what percentage were admitted or discharged, and readmission policies or practices would differ significantly between a large cohort of institutions such as that included in our analysis.

Conclusions

We present a multi-institutional retrospective study of readmission rates after radical prostatectomy in contemporary United States to examine risk factors for readmission within cohorts of patients undergoing RRP and RALRP. We report a significant difference in postoperative complication and readmission rates in RRP compared with RALRP. Further prospective analysis is warranted.

Footnotes

Disclaimer

The NSQIP and the hospitals that participated in the NSQIP are the source of the data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors of this study.

Disclosure Statement

The authors report no relevant financial disclosures related to this current work.

Abbreviations Used

References

Walsh

. Radical prostatectomy versus watchful waiting in early prostate cancer. J Urol, 2005; 174:1291.

Cooperberg

, Broering

, Carroll

. Time trends and local variation in primary treatment of localized prostate cancer. J Clin Oncol, 2010; 28:1117–1123.

Orvieto

, Patel

. Evolution of robot-assisted radical prostatectomy. Scand J Surg, 2009; 98:76.

Rau

. Hospitals face pressure to avert readmissions. New York Times. November 26, 2012. Available at www.nytimes.com/2012/11/27/health/hospitals-face-pressure-from-medicare-to-avert-readmissions.html?_r=0 (last accessed December 24, 2013 ).

Patient Protection and Affordable Care Act. Pub L No. 111–148.

Jencks

. Rehospitalizations among patients in the medicare fee-for-service program (vol 360, pg 1418, 2009). N Engl J Med, 2011; 364:1582.

Readmissions Reduction Program. Available at http://cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Readmissions-Reduction-Program.html/ (last accessed January 5, 2013 ).

Hofer

, Meeks

, Cashy

, et al. Impact of increasing prevalence of minimally invasive prostatectomy on open prostatectomy observed in the national inpatient sample and National Surgical Quality Improvement program. J Endourol, 2013; 27:102–107.

Shiloach

, Frencher

Jr. , Steeger

, et al. Toward robust information: Data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement program. J Am Coll Surg, 2010; 210:6–16.

10.

American College of Surgeons, National Surgical Quality Improvement Program. User Guide for the 2011 Participant Use Data File [Internet]. Available at http://site.acsnsqip.org/wp-content/uploads/2012/03/2011-User-Guide_Final.pdf (last accessed April 8, 2013 ).

11.

Rowell

, Turrentine

, Hutter

, et al. Use of National Surgical Quality Improvement program data as a catalyst for quality improvement. J Am Coll Surg, 2007; 204:1293–1300.

12.

Birkmeyer

, Shahian

, Dimick

, et al. Blueprint for a new American College of Surgeons: National Surgical Quality Improvement program. J Am Coll Surg, 2008; 207:777–782.

13.

American College of Surgeons National Surgical Quality Improvement Program. Available at www.acsnsqip.org (last accessed January 2013 ).

14.

American College of Surgeons. American College of Surgeons National Quality Improvement Program User Guide for the 2011 Participant Use File [NSQIP Online User Guide]. Available at http://site.acsnsqip.org/wp-content/uploads/2012/03/2011-User-Guide_Final.pdf (last accessed March 10, 2013 ).

15.

Merkow

, Bilimoria

, Hall

. Interpretation of the C-statistic in the context of ACS-NSQIP models. Ann Surg Oncol, 2011; 18 Suppl 3:S295; author reply S296, 2011.

16.

Cohen

, Bilimoria

, Ko

, et al. Development of an American College of Surgeons National Surgery Quality Improvement Program: Morbidity and mortality risk calculator for colorectal surgery. J Am Coll Surg, 2009; 208:1009–1016.

17.

Paul

, Pennell

, Lemeshow

. Standardizing the power of the Hosmer-Lemeshow goodness of fit test in large data sets. Stat Med, 2013; 32:67–80.

18.

Froehner

, Novotny

, Koch

, et al. Perioperative complications after radical prostatectomy: Open versus robot-assisted laparoscopic approach. Urol Int, 2013; 90:312–315.

19.

Liu

, Maxwell

, Panousis

, et al. Perioperative outcomes for laparoscopic and robotic compared with open prostatectomy using the National Surgical Quality Improvement Program (NSQIP) database. Urology, 2013; 82:579–583.

20.

Chung

, Kelle

, Huang

, et al. Comparison of 90-day re-admission rates between open retropubic radical prostatectomy (RRP), laparoscopic RP (LRP) and robot-assisted laparoscopic prostatectomy (RALP). BJU Int, 2012; 110:E966–E971.

21.

Nelson

, Kaufman

, Broughton

, et al. Comparison of length of hospital stay between radical retropubic prostatectomy and robotic assisted laparoscopic prostatectomy. J Urol, 2007; 177:929–931.

22.

Pierorazio

, Mullins

, Ross

, et al. Trends in immediate perioperative morbidity and delay in discharge after open and minimally invasive radical prostatectomy (RP): A 20-year institutional experience. BJU Int, 2013; 112:45–53.

23.

Touijer

, Eastham

, Secin

, et al. Comprehensive prospective comparative analysis of outcomes between open and laparoscopic radical prostatectomy conducted in 2003 to 2005. J Urol, 2008; 179:1811–1817.

24.

Kowalczyk

, Levy

, Caplan

, et al. Temporal national trends of minimally invasive and retropubic radical prostatectomy outcomes from 2003 to 2007: Results from the 100% Medicare sample. Eur Urol, 2012; 61:803–809.

25.

, Elliott

, Wilt

, et al. Practice patterns in benign prostatic hyperplasia surgical therapy: The dramatic increase in minimally invasive technologies. J Urol, 2008; 180:241–245.

26.

Varela

, Wilson

, Nguyen

. Laparoscopic surgery significantly reduces surgical-site infections compared with open surgery. Surg Endosc, 2010; 24:270–276.

27.

, Gu

, Lipsitz

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

28.

Skarecky

. Robotic-assisted radical prostatectomy after the first decade: Surgical evolution or new paradigm. ISRN Urol, 2013; 2013:157379.

29.

Tan

, Kalish

, Hamburg

, et al. Shorter duration of femoral-popliteal bypass is associated with decreased surgical site infection and shorter hospital length of stay. J Am Coll Surg, 2012; 215:512–518.

30.

Evangelista

, Doering

, Dracup

. Usefulness of a history of tobacco and alcohol use in predicting multiple heart failure readmissions among veterans. Am J Cardiol, 2000; 86:1339–1342.