Secondary Logo

Journal Logo

National Surgical Quality Improvement Program: What Can Anesthesiologists Learn From Surgical Outcomes?

Moller, Daryn H. MD; Gan, Tong J. MD, MBA, MHS, FRCA, Li Ac

doi: 10.1213/ANE.0000000000003411
Editorials: Editorial
Free

From the Department of Anesthesia, State University of New York at Stony Brook, Stony Brook, New York.

Accepted for publication March 23, 2018.

Funding: None.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Daryn H. Moller, MD, Department of Anesthesiology, Stony Brook Medicine, HSC L4-060, Stony Brook, NY 11790. Address e-mail to Daryn.moller@stonybrookmedicine.edu.

The acronyms NSQIP (National Surgical Quality Improvement Program), TQIP (Trauma Quality Improvement Program), and VQI (Vascular Quality Initiative) sound like alphabet soup to most anesthesia providers. NSQIP represents one of the largest outcome databases ever created, with >4.5 million cases sampled from >680 hospitals in the past decade.1 Historically, NSQIP can trace its origins to the early 1990s and the development of the National Veterans Administration Surgical Risk Study. In an effort to address quality at Veterans Affairs Medical Centers, 30-day outcomes for morbidity and mortality were investigated among 44 centers. Twelve variables, ranging from preoperative activity level to laboratory values, were originally used to describe a patient’s perioperative risk for complications and to develop an observed-to-expected ratio for outcomes. Annual comparative performance reports were made available to each clinical site to allow for benchmarking. From 1991 to 1997, an almost 10% decline in 30-day mortality was observed. After the expansion of the program to include 133 Veterans Affairs Medical Centers in 1994, there was a 30% decline in perioperative complications.2 In 2002, there was a pilot program of 18 private sector hospitals, and additional private sector hospitals began enrolling in 2004. More recently, participation in NSQIP and feedback from their annual benchmarking reports have led to consistent improvements in morbidity and mortality outcomes among a variety of institutions, and the longer an institution has been part of NSQIP, the greater the improvement in outcomes.3 Furthermore, the magnitude of the improvement is frequently greater among poor-performing institutions.4

These same outcome data have been used to create a web-based tool that can be used to estimate a patient’s risk for various postoperative complications (https://riskcalculator.facs.org/RiskCalculator). Over 1.4 million patients at 393 different NSQIP-enrolled hospitals and 1557 different Current Procedural Terminology codes were compared with 20 individualized patient factors.5 Information such as patient age, American Society of Anesthesiologists score, congestive heart failure (CHF) within the past 30 days to smoking, and functional status is used to estimate 12 different outcomes such as risk for cardiac complications to length of stay. NSQIP-derived estimations of adverse cardiac events have been shown to be more accurate at predicting complications than other commonly used tools such as the revised cardiac risk index.6

But why should we as anesthesiologists care about a surgical database? After all, examination of any large-scale database is filled with potential inconsistencies, and with a focus on surgical outcomes, there is probably little applicability to anesthesia care. Anesthesiologists may often have a visceral response to NSQIP: “My patient has an aortic valve area of 0.8 cm2. There is no place in NSQIP to account for this. I am sure they are at much higher risk of perioperative complications for this laparotomy. How can any outcome data or risk prediction be accurate when there is no place to account for severe aortic stenosis?” Even among defined surgical outcomes, there are numerous published inconsistencies regarding NSQIP’s ability to predict surgical risk for specific patient populations. Two-fold differences in surgical site infections,7 overpredicting length of stay,8 and other complications such as urinary tract infection or pneumonia have all questioned the reliability of NSQIP data and the derived risk calculator to predict outcomes for specific procedures or at specific centers.

Herein lie some of the potential power and paradoxes of “Big Data.” In this issue, Freundlich et al9 have examined outcomes from data submitted from NSQIP-participating institutions for 2012–2013. Traditional large-scale studies that have examined perioperative anesthetic management, such as the value of preoperative testing for cataract surgery, have had >18,000 patients.10 Fruendlich et al9 have examined outcomes for approximately 1.2 million patients. Thankfully, perioperative mortality is a relatively rare event, 0.77% by their findings, but because of such a large sample size, they were able to examine 9255 mortalities and focus on 1887 that were “attributable” to specific postoperative complications. Only through such large multicenter data collection can potential meaningful trends be identified. As one looks at the results of this and other such large outcome studies, there is an innate desire to apply their findings to one’s own institution and practice. However, by their nature, these data are going to be heterogeneous; in Freundlich et al’s9 data, 374 hospitals were participating in NSQIP data collection. Outcome data examined are a compilation from the best-performing institutions as well as the poorest performing ones. Institutions may be high outliers for specific complication rates and not others. Practice patterns are likely to be different by individuals, geographic regions, teaching versus nonteaching, and physician-based versus anesthesia care team models. The difference in performance among institutions is likely the source of NSQIP risk calculators’ “inability” to predict outcomes in small size, single-institution studies as mentioned above.7,8

Ironically, the contradictions noted in applying results from a heterogeneous Big Data study to a specific institution or group can actually be answered by the Big Data. In 2017, there were 688 participating institutions in the NSQIP database. Each of these institutions receives benchmarking data regarding its complication rate with an observed-to-expected ratio. Freundlich et al9 have demonstrated a link between specific complications and mortality. For institutions that are high outliers for those complications, significant mortality benefits may be observed by addressing processes and practice patterns revolving around those complications. For institutions that are low outliers for these complications, quality efforts might be better served addressing other complications.

However, this question remains: “Why should we as anesthesiologists care about a surgically driven and surgically focused outcomes database?” Freundlich et al9 found 3 postoperative complications that were most commonly attributable to mortality: bleeding, unplanned intubation, and septic shock. For NSQIP, bleeding necessitating transfusion is defined as transfusion of a packed red cell or whole blood product from the start of surgery up to 72 hours postoperatively. Typically, we think of blood loss associated with surgery as determined by the person holding the scalpel, but the anesthetic management, including blood pressure control, and medication selection, can alter this. In addition, how a patient is resuscitated from that blood loss has been shown to impact his or her mortality. Although the Pragmatic, Randomized Optimal Platelet and Plasma Ratios (PROPPR) trial failed to define an ideal ratio of packed red blood cells (PRBC) to fresh frozen plasma (FFP),11 a balanced resuscitation of products along with red cells has been shown to improve mortality in patients undergoing massive transfusion associated with trauma,12 cardiac,13 and vascular surgery. In a meta-analysis, the use of other agents such as tranexamic acid has been shown to improve mortality in trauma patients14 and reduce perioperative blood transfusion.

Freundlich et al9 found reintubation within 30 days as the second greatest perioperative complication associated with mortality. However, as the authors stressed, unanticipated reintubation is not a complication in and of itself; it is an indicator of and a rescue from respiratory failure. Physiological reasons leading to respiratory failure are not demonstrated in this study, but one would assume that they are diverse in nature. Perioperative myocardial infarction, CHF, and sepsis can all be elucidated from large-scale databases. However, something as simple and common place as poor postoperative pain control leading to inadequate pulmonary toileting is a much more difficult variable to categorize and incorporate into a database. Ironically, postoperative pain control is one of the areas where anesthesia providers, acting as perioperative physicians and pain specialists, can have the largest impact. Others have correlated unplanned reintubation with a host of seemingly obvious factors such as a diagnosis of chronic obstructive pulmonary disease, American Society of Anesthesiologists class IV or greater, and active CHF. While it has been demonstrated that almost 50% of reintubations are occurring with the first 72 hours after an operation, the correlation with risk factors is incredibly diverse. The odds ratio for reintubation was equal among sepsis, smoking, preoperative weight loss, and cancer. By far the strongest correlation was based on the surgical procedure, with very high-risk procedures noted to be esophageal, lung, neurosurgical, and gynecological malignancy, having almost double the rate of other seemingly extensive surgical procedures such as low anterior resection.15

For most anesthesiologists, outcomes and complications often have a real-time correlation; aspiration on induction or hypoxia from a failed intubation is seemingly in the forefront of every individual’s mind. While there are countless enhanced recovery studies demonstrating reductions in length of hospital stay to faster return of bowel function, many anesthesiologists are still lacking the connection between their day-to-day activities and long-term patient outcomes. Why remains unclear, perhaps because large-scale outcome data on major complications are yet to demonstrate that these or other anesthesia-driven quality improvements have reduced perioperative mortality. This void does not suggest an absence of effect; it merely suggests an opportunity for large-scale research.

Starting in 2012, the Centers for Medicare and Medicaid Services began publicly reporting NSQIP outcome data on a voluntary basis. Eighteen percent of hospitals involved in NSQIP voluntarily participated in this reporting. NSQIP data are uploaded to the Medicare website (https://www.medicare.gov/hospitalcompare/search.html), allowing potential patients and health care consumers to do side-by-side comparisons of surgical outcomes. Institutional survival is going to be based on the ability to provide consistent high-quality care with minimal morbidity and mortality. Anesthesiologists have traditionally examined their impact on immediate perioperative outcomes and complications. However, we have an obligation to our surgical colleagues, institutions, and, most importantly, our patients to take an active role in examining long-term surgical outcomes. By demonstrating an association among bleeding, reintubations, and sepsis, Freundlich et al9 have given anesthesiologists a potential place to focus efforts, which may result in the greatest mortality reductions.

Back to Top | Article Outline

DISCLOSURES

Name: Daryn H. Moller, MD.

Contribution: This author helped prepare the manuscript.

Name: Tong J. Gan, MD, MBA, MHS, FRCA, Li Ac.

Contribution: This author helped prepare the manuscript.

This manuscript was handled by: Thomas R. Vetter, MD, MPH.

Back to Top | Article Outline

REFERENCES

1. ACS National Surgical Quality Improvement Program. Available at: https://www.facs.org/quality-programs/acs-nsqip/program-specifics/participant-use. Accessed May 10, 2018.
2. Khuri SF, Daley J, Henderson WThe Department of Veterans Affairs’ NSQIP: the first national, validated, outcome-based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. National VA Surgical Quality Improvement Program. Ann Surg. 1998;228:491–507.
3. Cohen ME, Liu Y, Ko CY, Hall BLImproved surgical outcomes for ACS NSQIP hospitals over time: evaluation of hospital cohorts with up to 8 years of participation. Ann Surg. 2016;263:267–273.
4. Hall BL, Hamilton BH, Richards K, Bilimoria KY, Cohen ME, Ko CYDoes surgical quality improve in the American College of Surgeons National Surgical Quality Improvement Program: an evaluation of all participating hospitals. Ann Surg. 2009;250:363–376.
5. Bilimoria KY, Liu Y, Paruch JLDevelopment and evaluation of the universal ACS NSQIP surgical risk calculator: a decision aid and informed consent tool for patients and surgeons. J Am Coll Surg. 2013;217:833.e1–842.e1.
6. Gupta PK, Gupta H, Sundaram ADevelopment and validation of a risk calculator for prediction of cardiac risk after surgery. Circulation. 2011;124:381–387.
7. Burgess JR, Smith B, Britt R, Weireter L, Polk TPredicting postoperative complications for acute care surgery patients using the ACS NSQIP surgical risk calculator. Am Surg. 2017;83:733–738.
8. Cologne KG, Keller DS, Liwanag L, Devaraj B, Senagore AJUse of the American College of Surgeons NSQIP Surgical Risk Calculator for Laparoscopic Colectomy: how good is it and how can we improve it? J Am Coll Surg. 2015;220:281–286.
9. Freundlich RE, Maile MD, Sferra JJ, Jewell ES, Kheterpal S, Engoren MComplications associated with mortality in the National Surgical Quality Improvement Program Database. Anesth Analg. 2018;127:55–62.
10. Schein OD, Katz J, Bass EBThe value of routine preoperative medical testing before cataract surgery: study of medical testing for cataract surgery. N Engl J Med. 2000;342:168–175.
11. Holcomb JB, Tilley BC, Baraniuk SPROPPR Study Group. Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: the PROPPR randomized clinical trial. JAMA. 2015;313:471–482.
12. Borgman MA, Spinella PC, Perkins JGThe ratio of blood products transfused affects mortality in patients receiving massive transfusions at a combat support hospital. J Trauma. 2007;63:805–813.
13. Delaney M, Stark PC, Suh MMassive transfusion in cardiac surgery: the impact of blood component ratios on clinical outcomes and survival. Anesth Analg. 2017;124:1777–1782.
14. Roberts I, Shakur H, Coats TThe CRASH-2 trial: a randomised controlled trial and economic evaluation of the effects of tranexamic acid on death, vascular occlusive events and transfusion requirement in bleeding trauma patients. Health Technol Assess. 2013;17:1–79.
15. Ramachandran SK, Nafiu OO, Ghaferi A, Tremper KK, Shanks A, Kheterpal SIndependent predictors and outcomes of unanticipated early postoperative tracheal intubation after nonemergent, noncardiac surgery. Anesthesiology. 2011;115:44–53.
Copyright © 2018 International Anesthesia Research Society