Surgical Endoscopy

, Volume 27, Issue 5, pp 1674–1680

Patient safety and the diffusion of surgical innovations: a national analysis of laparoscopic partial nephrectomy

Authors

    • Division of UrologyUC San Diego Medical Center
    • VA San Diego Medical Center
    • Division of Urologic OncologyMoores UCSD Cancer Center
  • Kerrin Palazzi
    • Division of UrologyUC San Diego Medical Center
  • David Chang
    • Division of UrologyUC San Diego Medical Center
  • Sean P. Stroup
    • Division of UrologyUC San Diego Medical Center
    • Department of UrologyNaval Medical Center San Diego
Article

DOI: 10.1007/s00464-012-2655-z

Cite this article as:
Parsons, J.K., Palazzi, K., Chang, D. et al. Surg Endosc (2013) 27: 1674. doi:10.1007/s00464-012-2655-z

Abstract

Background

There are scant data on patient safety and the national diffusion of surgical innovations. Laparoscopic partial nephrectomy (LPN) provides an apt model for population-based analyses of patient safety and the propagation of complex surgical innovations.

Methods

In the nationwide inpatient sample, we identified patients undergoing partial nephrectomy for renal tumors from 1998 to 2009 and utilized patient safety indicators (PSI) to measure preventable adverse outcomes.

Results

Of the cases, 68,713 (87 %) were OPN and 9,842 (13 %) were LPN. The prevalence of LPN increased more than threefold from 2006 to 2009. Compared to open partial nephrectomy (OPN), LPN patients were more likely to be younger (p = 0.022), have lower Charlson comorbidity scores (p = 0.002), and undergo surgery at urban (p < 0.001) and teaching (p = 0.02) hospitals. On multivariate analysis, LPN was associated with a 28 % decreased probability of any PSI (adjusted odds ratio [ORadj] 0.72, 95 % confidence interval [CI] 0.55–0.96, p = 0.025), although this benefit did not attain significance when comparing robot-assisted LPN to OPN (ORadj 0.72, 95 % CI 0.44–1.16, p = 0.173). Overall mortality decreased from 0.9 % in 1998 to 0.1 % in 2009 (p < 0.001). There were no differences in adjusted mortality between LPN and OPN (p = 0.75).

Conclusions

During its initial national diffusion, LPN resulted in enhanced perioperative patient safety compared to OPN. Mortality for both LPN and OPN decreased over time. Further study is needed to elucidate and promote factors contributing to the safe diffusion of complex surgical innovations.

Keywords

Kidney cancerLaparoscopic partial nephrectomyOpen partial nephrectomyPartial nephrectomyPatient safetyPatient safety indicator

Abbreviations

AHRQ

Agency for healthcare research and quality

LPN

Laparoscopic partial nephrectomy

NIS

Nationwide inpatient sample

OPN

Open partial nephrectomy

PSI

Patient safety indicator

The diffusion of a surgical innovation into clinical practice often outpaces the availability of robust comparative evidence demonstrating its safety relative to the technique or technology it is intended to replace [1]. Although population-based analyses of patient safety have gained broad acceptance within the fields of medicine and health policy [1], they have not disseminated as widely among the surgical disciplines [2]. Because adverse safety events result primarily from problems inherent in processes and organizations [3], it is possible that flaws in the diffusion of surgical innovations may expose patients to heightened risk of medical error. Thus, population-based studies of patient safety and the diffusion of surgical innovations would inform understanding of these processes and potentially enhance the safe spread of innovative surgical technologies.

Laparoscopic partial nephrectomy (LPN) provides an apt model for studying surgical innovation, diffusion, and patient safety. LPN, which is most often performed for renal cancer, provides equivalent oncologic control but with diminished pain and decreased length of stay (LOS) compared to open partial nephrectomy (OPN) [46]. LPN, with or without robotic assistance, is a technically complex procedure demanding an acquired and rigorous surgical skills set. Perioperative safety concerns include excessive hemorrhage and ischemic renal damage [6]. As with most other innovative surgical procedures, there are no standardized training programs or uniform rules governing surgeon credentialing for either LPN or robot-assisted LPN [7].

Published data suggest that LPN and robot-assisted LPN have diffused relatively rapidly in recent years, primarily as a result of the increased use of the Da Vinci robotic platform (Intuitive Surgical Inc., Sunnyvale, CA) [8, 9]. Although single institution cohort studies have observed that perioperative complications are similar between LPN and OPN [46], comprehensive patient safety analyses utilizing validated measures have yet to be performed in large populations.

Developed by the Agency for Healthcare Research and Quality (AHRQ), patient safety indicators (PSI) are evidence-based metrics designed for use in administrative databases. These robust, validated instruments provide safety performance measures with greater specificity and less bias than other methods [2, 1012]. Therefore, in a national cohort, we analyzed the diffusion of LPN over a 10-year period and compared the prevalence of PSIs in OPN, LPN, and robot-assisted LPN.

Materials and methods

Study population: nationwide inpatient sample

We performed a cross-sectional analysis of all inpatient hospital discharges after partial nephrectomy from 1998 to 2009. We obtained data on inpatient procedures and adverse events from the nationwide inpatient sample (NIS) data set of the Healthcare Cost and Utilization Project. The NIS is compiled from up to 12 million hospital stays each year and represents the largest hospital inpatient data set in the United States. It consists of a 20 % sample of inpatient discharges from over 40 states and nearly 1,100 hospitals. Weighted sampling allows estimates for national trends. The NIS provides information on age, gender, race, income, insurance type, the Charlson comorbidity index, hospital location (rural or urban), hospital type (teaching or nonteaching), and LOS. We utilized PSIs validated for surgery and cancer patients.

Partial nephrectomy case selection

We identified all discharges from January 1998 to December 2009 using principle or secondary procedure ICD-9CM codes for partial nephrectomy (55.4) divided among “laparoscopic” (concurrent 54.51, 54.21, or 17.42) and “open” (without concurrent laparoscopy codes). In October 2008, a separate code was introduced for robot-assisted LPN. We included robot-assisted partial nephrectomies in the overall laparoscopic cohort and also performed a separate sub-group analysis of robot-assisted LPN from October 2008 to December 2009. We excluded patients <18 years old; with an admission type other than elective; and with diagnoses of transplant (v59.4), renal pelvis tumor (189.1), and pyelonephritis (590.00, 590.01, 590.10, 590.11).

To reduce the potential for confounding by tumor complexity, we initially excluded from analysis patients undergoing the following concomitant procedures: splenectomy (41.4, 41.42, 41.43), liver resection (50.2, 50.21–26, 50.29, 50.3), pancreas resection (52.5, 52.51–53, 52.59, 52.6, 52.7), bowel or colon resection (45.5, 45.50–52, 45.61–62, 45.7, 45.71, 45.73–76, 45.79, 45.8, 45.81–83), or thrombectomy with vascular reconstruction (37.10, 38.05, 38.07, 38.45, 38.47, 38.65, 38.67, 38.75, 38.77, 38.87, 39.6, 39.61, 39.63, 39.66). This step resulted in exclusion of 1,088 patients, 55 LPN and 1,033 OPN.

Primary outcome: PSIs

We identified PSIs according to the list published by AHRQ and included anesthetic complications (designated by AHRQ as an “experimental variable”), pneumothorax, hemorrhage/hematoma, decubitus ulcer, venous thrombosis/pulmonary embolism, hip fracture, postoperative physiologic and metabolic derangements, postoperative respiratory failure, sepsis, and accidental puncture or laceration of surrounding organs during a procedure. PSIs represent acute events that occur during the same hospitalization as the surgery. We also evaluated in-hospital mortality.

Statistical analysis

We examined the prevalence of LPN and OPN by year to assess trends and determine relationships between PSI and type of renal surgery performed. We report national prevalence estimates generated by applying sample weighting to the 20 % NIS sample. We applied univariate tests including Chi square (Rao and Scott second-order correction) and Student’s t test to compare groups and assess factors associated with increased frequency of PSIs. We incorporated variables significant at the 5 % level or of clinical interest in binary logistic regression models to determine the associations of PSIs and surgical type. The SVY coding in Stata v 11.1 (StataCorp, College Station, TX) was used to account for NIS sampling methodology.

Results

Study population

Between 1998 and 2009, we identified 78,554 cases of partial nephrectomy: 68,713 (87 %) OPN and 9,842 (13 %) LPN. The prevalence of both LPN and OPN increased substantially throughout the study period, although the magnitude of the increase in LPN was greater. In 2009, over 30 % of all partial nephrectomies were LPN (Fig. 1) and, of these, 72 % were performed with robotic assistance (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00464-012-2655-z/MediaObjects/464_2012_2655_Fig1_HTML.gif
Fig. 1

a Prevalence of partial nephrectomy by technique, 1998–2009. b Proportion of partial nephrectomy cases treated by open or laparoscopic approaches, 1998–2009

Patient demographics, LOS, and transfusions

Compared to OPN, patients undergoing LPN were younger (p = 0.022), had lower Charlson comorbidity scores (p = 0.002), were more likely to have private insurance (p = 0.001), and were more likely to undergo surgery at urban (p < 0.001) and teaching (p = 0.02) hospitals. There were no differences with respect to gender, race, or income level. Median LOS and risk of perioperative transfusion were substantially less for LPN (Table 1).
Table 1

Demographic and clinical characteristics of patients undergoing partial nephrectomy for renal cancer

Characteristic

Open

(n = 68,713)

Laparoscopic

(n = 9,842)

pa

Age (years), mean ± SE

59 ± 0.2

58 ± 0.4

0.022

Sex, n (%)

  

0.726

 Male

39,510 (57.5)

5,620 (57.1)

 

 Female

29,203 (42.5)

4,222 (42.9)

 

Race, n (%)

  

0.405

 White

41,140 (59.9)

5,896 (59.9)

 

 African American

4,940 (7.2)

482 (4.9)

 

 Other

22,632 (32.9)

3,464 (35.2)

 

Charlson index, n (%)

  

0.002

 0–2

48,288 (70.3)

7,307 (74.2)

 

 ≥3

20,424 (29.7)

2,535 (25.8)

 

Income by zip code, n (%)

  

0.234

 <$45,000

46,313 (67.4)

6,905 (70.2)

 

 ≥$45,000

22,399 (32.6)

2,937 (29.8)

 

Insurance, n (%)

  

0.001

 Private

37,644 (54.8)

5,980 (60.8)

 

 Medicare

24,608 (35.8)

3,092 (31.4)

 

 Medicaid

3,023 (4.4)

334 (3.4)

 

 Other

3,437 (5)

435 (4.4)

 

Teaching hospital, n (%)

  

0.02

 Nonteaching

20,066 (29.3)

2,013 (20.5)

 

 Teaching

48,526 (70.7)

7,801 (79.5)

 

Urban hospital, n (%)

  

<0.001

 Rural

3,295 (4.8)

155 (1.6)

 

 Urban

65,297 (95.2)

9,658 (98.4)

 

Length of stay (days), median (IQR)

4 (3–5)

3 (2–4)

<0.001

Transfusion, n (%)

28,895 (11.2)

2,081 (6)

<0.001

Mortality, n (%)

181 (0.3)

14 (0.1)

0.316

ap values were calculated by independent t test, Fisher’s exact test, and Chi-square test. Two-tailed significant values were <0.05

Patient safety indicators

The unadjusted prevalence of any PSI was higher, with marginal statistical significance, for OPN (Table 2). PSIs associated with OPN included iatrogenic pneumothorax, postoperative physiologic and metabolic derangements, and accidental puncture or laceration. There were no differences between OPN and LPN in the risk of postoperative hemorrhage, complications of anesthesia, respiratory failure, thrombosis or pulmonary embolism, sepsis, or wound dehiscence.
Table 2

Unadjusted prevalence of individual PSIs among patients undergoing partial nephrectomy for renal cancer

Patient safety indicator

Open

(n = 68,713)

Laparoscopic

(n = 9,842)

pa

Any

3,356 (4.9 %)

382 (3.9 %)

0.060

Complications of anesthesia

129/53,582 (0.2 %)

16/9,536 (0.2 %)

0.535

Death in low-mortality DRGs

 

Decubitus ulcer

57/20,246 (0.3 %)

10/1,521 (0.6 %)

0.279

Failure to rescue

81/1,907 (4.2 %)

9/229 (4.1 %)

0.963

Iatrogenic pneumothorax

522/51,929 (1 %)

19/9,441 (0.2 %)

0.001

Selected infections due to medical care

0/13,055 (0 %)

0/2,579 (0 %)

Postoperative hip fracture

0/52,235 (0 %)

0/5,537 (0 %)

Postoperative hemorrhage or hematoma

158/53,574 (0.3 %)

28/9,536 (0.3 %)

0.990

Postoperative physiologic and metabolic derangements

189/53,514 (0.4 %)

10/9,531 (0.1 %)

0.089

Postoperative respiratory failure

609/53,397 (1.1 %)

75/9,522 (0.8 %)

0.155

Postoperative PE/DVT

346/53,540 (0.6 %)

64/9,536 (0.7 %)

0.906

Postoperative sepsis

33/7,994 (0.4 %)

0/755 (0 %)

0.444

Postoperative wound dehiscence

0/4,589 (0 %)

0/460 (0 %)

Accidental puncture or laceration

1,554/53,579 (2.9 %)

174/9,536 (1.8 %)

0.003

Transfusion reaction

PE/DVT pulmonary emobolish/deep venous thrombosis

ap values were calculated by independent Fisher’s exact test and Chi-square test. Two-tailed significant values were <0.05

Multivariate analysis

On multivariate analysis, LPN was associated with a 28 % decreased probability of any PSI compared to OPN. In the sub-group analysis that compared robot-assisted LPN to OPN, robot-assisted LPN was also associated with a 28 % decreased probability of any PSI, but this association did not attain significance (Table 3).
Table 3

Multivariable-adjusted ORs for the occurrence of any PSI for LPN and robot-assisted LPN

Model

OR

95 % CI

p

Laparoscopic compared to opena

   

 Surgery type

  

0.019

  Open

Ref.

 

  Laparoscopic

0.72

(0.548–0.946)

 

 Gender

  

0.208

  Male

Ref.

 

  Female

0.904

(0.774–1.058)

 

 Charlson index

  

<0.0001

  0–2

Ref.

  

  ≥3

1.605

(1.351–1.905)

 

 Hospital

  

0.264

  Nonteaching

Ref.

  

  Teaching

1.123

(0.916–1.376)

 

Robot-assisted laparoscopic compared to openb

   

 Surgery type

  

0.173

  Open

Ref.

 

  Robot-assisted laparoscopic

0.716

(0.443–1.158)

 

 Gender

  

0.885

  Male

Ref.

 

  Female

0.973

(0.673–1.408)

 

 Charlson index

  

0.0008

  0–2

Ref.

 

  ≥3

1.717

(1.148–2.567)

 

 Hospital

  

0.410

 Nonteaching

Ref.

 

 Teaching

1.19

(0.787–1.799)

 

OR odds ratio, PSI patient safety indicator, LPN laparoscopic partial nephrectomy, CI confidence interval

aModel is controlled for age and year in addition to all variables listed

bModel is controlled for age in addition to all variables listed

Univariate sensitivity analyses of PSIs stratified by each of the major demographic variables (sex, race, age, income, Charlson comorbidity, and hospital type) confirmed that, compared to OPN, LPN was associated with significantly decreased risk of PSI within all of these groups (data not shown). In a sensitivity analysis that included concomitant liver, spleen, intestinal, and vascular surgery, the association of LPN with likelihood of any PSI was similar to the primary analysis (adjusted odds ratio [ORadj] 0.75, 95 % confidence interval [CI] 0.576–0.986, p = 0.039).

Mortality

Overall mortality decreased from 0.9 % in 1998 to 0.1 % in 2009 (p < 0.001). There were no significant differences in unadjusted (Table 1) or adjusted mortality between OPN and LPN, but men were two and a half times more likely to die than women (Table 4).
Table 4

Multivariable-adjusted ORs of mortality for LPN and robot-assisted LPNa

Characteristic

OR

95 % CI

p

Surgery typeb

  

0.984

 Open

Ref.

 

 Laparoscopic

1.01

(0.27–3.78)

 

Gender

  

0.022

 Female

Ref.

 

 Male

2.51

(1.14–5.52)

 

Charlson index

  

0.001

 0–2

Ref.

 

 ≥3

3.10

(1.62–5.92)

 

OR odds ratio, LPN laparoscopic partial nephrectomy, CI confidence interval

aModel is controlled for age and year in addition to all variables listed

bBecause of the small number of events, analyses of robot-assisted laparoscopy could not be performed

Discussion

In this population-based analysis, we observed a 28 % (ORadj 0.72) decreased probability of adverse safety events occurring in LPN compared to OPN as measured by PSIs, independent of teaching hospital status. Our data suggest that LPN confers significant perioperative safety benefits to standard open surgery and demonstrate the safe adoption of a novel surgical technique into broader clinical practice. Although there were no differences in perioperative mortality between modalities, overall reductions in mortality during the study period intimate substantial improvements in the process of care for partial nephrectomy—both open and laparoscopic—over time.

This study is one of the first to describe the national diffusion of a minimally invasive surgical technique using validated perioperative safety outcomes. Understanding the processes by which new surgical technologies diffuse into broader clinical practice is highly relevant to promoting the public health. Although we recently observed that laparoscopic radical nephrectomy also diffused safely [13], other data indicate that the introduction of minimally invasive radical prostatectomy was associated with a substantially increased risk of adverse perioperative patient safety events [14]. These observations suggest that there are specific factors associated with the safe propagation of innovative surgical techniques: factors that have yet to be elucidated and are critically important for understanding—and thus fostering—the safe diffusion of surgical innovations. The present study represents an important initial step in identifying these factors.

Notably, the enhanced safety outcomes for LPN manifested during a period of rapid diffusion: from 2008 to 2009, the proportion of LPN cases more than doubled. The onset of rapid diffusion, often referred to as the “tipping point” [15, 16] occurred in 2007. This phase of diffusion is when early adopters, who are typically expert opinion leaders, transition into the early majority (those who adopt a new technique immediately before the average surgeon) [17]. Not every surgical innovation, of course, achieves a tipping point. Adoption of a surgical innovation depends upon a number of factors, including physician knowledge of the new technique, persuasion (including marketing), perceived advantage, patient demand, feasibility and safety [15, 17]. The recent rapid growth of partial nephrectomy in general and LPN in particular is mostly likely attributable to the development of robot-assisted LPN, which many argue overcomes the technical challenges and difficult learning curve of pure LPN [18]. However, in this cohort, although the magnitude of the safety benefit was similar for robot-assisted LPN, it did not attain significance, potentially as a result of the relatively small number of measureable robot-assisted LPN cases. In addition, the heterogeneity of the diffusion process, characterized by its lack of uniform training and credentialing [7], did not appear to be a barrier to safety.

Ours is the first national study of partial nephrectomy to utilize validated, process-focused measures of patient safety. PSIs are important metrics for quality improvement and are valuable in assessing safety because they correlate with preventable adverse events while minimizing potential biases associated with variations in individual patients and caregivers [2, 1012]. A limited number of prior studies have observed comparable, if not superior, perioperative outcomes for LPN with respect to complications, but have largely been limited to single institution case series at academic centers [6]. A population-based study, which matched patients on baseline differences, noted that, compared to OPN, LPN was associated with a decreased likelihood of blood transfusions, postoperative complications, and extended hospital stays [17]. We also observed a substantially diminished likelihood of mortality in women compared to men (ORadj 0.42). The reasons underpinning this finding are unclear. However, because at least one prior study has noted an independent survival advantage for women among patients undergoing major gastrointestinal and traumatic injury surgery, it is possible that there are gender-specific physiological variations that confer a protective effect to women undergoing major abdominal surgery [19].

Although we did not adjust for tumor size or anatomy in the analytic models, there are at least two reasons why it is unlikely that differential bias in tumor characteristics between groups accounted for the observed differences. First, we excluded patients that underwent concomitant surgical procedures (thrombectomy and/or vascular reconstruction; pancreatic, splenic, or hepatic procedures; bowel resection) indicative of large, complex renal tumors requiring extensive resection and potentially associated with poorer perioperative outcomes related to aggressive disease. Notably, in sensitivity analyses, incorporation of these patients did not substantially alter the results. Second, there are data to suggest that the anatomic complexity of a renal tumor is not necessarily associated with the likelihood of an adverse perioperative event [20, 21].

Still, lack of patient-specific oncological data, including tumor size or tumor stage, is a potential limitation of our study. A second limitation is that variations in coding practices over the 10-year study period may have influenced the analyses. These changes included the addition, in more recent years, of laparoscopic- and robotic method–specific procedure codes. Still, there is no evidence to suggest that such variations would have been differential with respect to the primary study outcomes. Finally, we could only assess adverse outcomes that occurred during the period of hospitalization after surgery. It is possible that differential outcomes of adverse events occurring after discharge, but still within the perioperative period (i.e., 30 days), were not captured.

During its initial national diffusion, LPN resulted in enhanced perioperative patient safety outcomes compared to OPN. These data demonstrate, on a national level, the safe adoption of a technically demanding surgical innovation into broader clinical practice. Further study is needed to identify factors that promote the safe diffusion of surgical innovations.

Disclosures

Drs. Parsons, Chang, and Stroup and Ms. Palazzi have no conflicts of interest or financial ties to disclose.

Copyright information

© Springer Science+Business Media New York 2012