Secondary Logo

Journal Logo

Pediatric Anesthesiology: Research Report

Anesthesiologist- and System-Related Risk Factors for Risk-Adjusted Pediatric Anesthesia-Related Cardiac Arrest

Zgleszewski, Steven E. MD*; Graham, Dionne A. PhD; Hickey, Paul R. MD*; Brustowicz, Robert M. MD*; Odegard, Kirsten C. MD*; Koka, Rahul MD*; Seefelder, Christian MD*; Navedo, Andres T. MD*; Randolph, Adrienne G. MD, MSc

Author Information
doi: 10.1213/ANE.0000000000001059

Anesthesia-related mortality is, fortunately, now an uncommon event in most industrialized nations. Quality-of-care benchmarks for anesthesia must therefore expand to include clinically important but nonfatal anesthesia-related serious adverse events.1 Cardiac arrest (CA) in patients undergoing anesthesia is a potential quality indicator because it can lead to death and severe sequelae, including cognitive and motor impairment in survivors.2–7 Although some CAs in the perioperative period are related to surgical complications or underlying patient disease, previous studies show that up to 76% of reported arrests are likely related to the anesthetic process.2–13 Therefore, anesthesia-related cardiac arrest (ARCA) in the pre-intraoperative or recovery period is a potentially preventable anesthetic-related adverse outcome reflective of the safety dimension of anesthetic care quality.14

The earliest published studies on ARCA show that children are at higher risk of ARCA than adults.9,10,12,13,15 More recent published pediatric studies on the incidence of ARCA and its risk factors have come from single-center cohorts5–7,16 and across institutions participating in a voluntary registry.2,11 Despite safety improvements in anesthesia from the 1950s until the present, 2 risk factors in ARCA remain constant: young age and ASA physical status (ASA-PS) classification.17 Using the ASA-PS classification, children with severe systemic disease or worse (ASA-PS levels ≥III) have a higher risk of ARCA than children with no or minimal systemic disease (levels I–II).2,6,7,11,16,18,19 Studies have also reported a higher rate of ARCA in infants <1 year old.2,5–7,9,11–13,16,18,19 Some, but not all, studies report that children undergoing cardiac procedures6,7,20 and children undergoing emergency procedures (compared with elective procedures) are at higher risk for ARCA.2,5,7,11,18

Increasingly, as hospitals are being held accountable for their patient health outcomes, institutions are asked to publicly report their rates of potentially preventable complications.21–23 Because case-mix varies markedly across institutions, to make unbiased comparisons, it is necessary to adjust complication rates for population and procedural risk.24 Previous pediatric ARCA studies evaluating system-related and anesthesiologist-related risk factors did not rigorously adjust for patient-related risk, making it difficult to generalize the results of these studies to all hospital populations, including those based in the community and at academic institutions. Given the low frequency of pediatric ARCA, evaluation of provider-level risk factors requires a very large data set that includes providers performing sufficient cases to assess risk. Using data from over a decade of anesthetics delivered at a large, tertiary, pediatric institution, we aimed to identify system- and anesthesiologist-related risk factors for pediatric ARCA after controlling for patient-related risk.

METHODS

This study was approved by the IRB at Boston Children’s Hospital. A waiver of informed consent was granted. Using a prospectively collected administrative database of anesthetic clinical records for all anesthetics performed at Boston Children’s Hospital from January 1, 2000, through December 31, 2011, we identified the cohort of patients at risk. All anesthetics performed in the operating rooms (ORs) in the main and satellite hospitals, radiology department, endoscopy suites, recovery rooms, and other procedure rooms were included. Throughout the duration of the study, the attending anesthesiologists were engaged in the full range of coverage patterns (alone, certified registered nurse anesthetist, fellow, resident) in an academic environment. We excluded anesthetics for patients >18 years. We also excluded anesthetics in the cardiac catheterization laboratories where data with regard to pediatric ARCA was incomplete for the time period from 2000 to 2003. In addition, catheter-based interventional procedures and surgery with cardiopulmonary bypass in the cardiac OR are 2 separate patient populations. Indications for these procedures are different between the catheterization laboratory and the OR.25

The at-risk period for CA studied was the perioperative period, which comprises the time the patient entered the OR to the time the patient exited the postanesthesia care unit or was transferred to the intensive care unit team. The Perioperative Cardiac Arrest Registry (POCA) definition of CA and the role anesthesia had in the genesis of the CA were used in this study2,11 (Supplemental Digital Content 1, Supplemental Table 1, https://links.lww.com/AA/B278). The practitioner assigned responsibility for a CA categorized as an ARCA was the anesthesiologist caring for the patient at the time of the event. ASA-PS was defined using the classification of physical status by Dripps.17 Annual days delivering anesthetics was defined as the number of days at least one anesthetic was performed by the practitioner starting January 1 in that calendar year.

Table 1 lists the multiple methods used to identify CAs. Potential patient-related ARCA risk factors included patient demographic information (patient age and ASA-PS). Potential system-related risk factors for ARCA included anesthetic location, day of the week, time of anesthetic, and responsible anesthesiologist. These data were extracted from the patient’s record for anesthetics where a CA occurred. The same information was extracted from a foundation billing database as a denominator for all anesthetics performed in the study period where a CA did not occur. Potential provider-related ARCA risk factors included anesthesiologist characteristics, such as years of anesthesia experience, annual number of days in a calendar year delivering anesthetics, case volume per calendar year, and academic rank at the time of the anesthetic. These data were generated by the Anesthesiology Department Chairman’s office.

T1-26
Table 1:
Methods Used to Identify Cardiac Arrests

Each CA was reviewed independently by 2 anesthesiologists for noncardiac surgical anesthetics. The CAs were then categorized as (1) anesthesia-related or (2) not anesthesia-related, according to the definitions used by the POCA registry.2,11 In cases of disagreement, a third reviewer adjudicated to achieve consensus. For noncardiac surgery arrests, only one CA required the third reviewer for adjudication. In this case, the cause of the CA was unknown and, therefore, not anesthesia-related. Subspecialists reviewed the cardiac surgical anesthetics, as previously described.20

Statistical Analysis

Rates of ARCA are expressed as number of events per 10,000 anesthetics. Continuous anesthesiologist characteristics were categorized based on quartiles of practitioners by year. Univariate associations between rate of ARCA and patient, practitioner, and systems factors are reported as odds ratios and were assessed with χ2 tests or Fisher exact tests, as appropriate. Multivariable logistic regression was used to derive predictive models for ARCA. Because of the small numbers of events, bootstrapping was used to estimate model coefficients and standard errors.26 The final multivariable logistic regression model was used to calculate the expected number of ARCA by year accounting for patient factors, and then ratios of the observed to the expected number of ARCA were calculated using the method of Breslow and Day27 to calculate the 95% confidence intervals (CIs). In all tests, statistical significance was achieved with a 2-sided P value <0.05. Statistical analyses were performed with SAS software, version 9.3, of the SAS System for Windows (SAS Institute, Cary, NC) and S-plus, version 8.0 (TIBCO Software, Palo Alto, CA).

Because we evaluated physician-level factors and patient-level factors for ARCA, it was important to rule out the potential effect of clustering of events within physicians (a.k.a., the “unit of analysis error”).28 We used multiple methods to control for potential clustering of events by provider(s). We tested for correlation of events within physicians and compared the results of 3 modeling approaches: (1) bootstrapping without generalized estimating equations (GEEs); (2) logistic regression (no bootstrapping); and (3) GEEs to account for clustering (no bootstrapping).

Final Data Set and Univariate Analysis

There were 314,589 total anesthetics delivered over the 12-year period. Of these, we excluded 18,920 (6%) because of patient age >18 years and 19,460 (6%) that were delivered in the cardiac catheterization laboratory. The final data set included 276,209 anesthetics delivered by 115 practitioners. CA occurred in 142 anesthetics (incidence 5.1/10,000 anesthetics; 95% CI, 4.4–6.1) and 72 anesthetics (2.6/10,000 anesthetics; 95% CI, 2.1–3.3) were classified as anesthesia-related. As shown in Supplemental Digital Content 1, Supplemental Table 2 (https://links.lww.com/AA/B278), our ARCA rate was similar to other studies reported in the literature.2,5,6,11 All CAs identified and those judged as ARCA are shown in Supplemental Digital Content 1, Supplemental Table 1 (https://links.lww.com/AA/B278), along with the location and procedure type. Table 2 lists the characteristics of the 72 cases of ARCA, including ARCA mechanism, timing, and clinical outcome. There was no significant difference in ARCA rates by year (P = 0.71).

T2-26
Table 2:
Characteristics of Anesthesia-Related Cardiac Arrests (ARCA)

Table 3 shows the relationship between the rate of ARCA and the investigated patient, practitioner, and system factors. We verified that there was a stepwise association between ARCA and ASA-PS and that ARCA was higher in infants <1 year of age with the highest rates in those ≤6 months of age. A multivariable bootstrap logistic regression model containing only age ≤6 months old and ASA-PS ≥III (Supplemental Digital Content 1, Supplemental Table 3, https://links.lww.com/AA/B278) was created and used to adjust for the risk of ARCA using patient factors alone.

T3-26
Table 3:
Univariate Associations Between Anesthesia-Related Cardiac Arrest (ARCA) and Patient, Practitioner, and Systems Factors

Most of the ARCAs occurred in the main ORs (n = 66, 92%) during the daytime hours (n = 71, 99%) and during weekdays (n = 70, 97%), where and when the great majority of anesthetics were delivered. There were 6 ARCAs (8%) in the radiology department and none in the endoscopy suites.

All anesthesiologists were either board certified by the American Board of Anesthesiology or had foreign board certification and additional training and/or extensive experience performing pediatric anesthesia. Forty-eight of the 115 anesthesiologists (42%) had at least 1 ARCA. Of the staff who had ≥1 ARCA event, the median years of experience delivering anesthetics at the time of the ARCA event was 11.4 years (interquartile range, 4.4–17.0). Their academic rank at the time of the ARCA event was 51% instructor, 31% assistant professor, 14% associate professor, and 4% professor.

As shown in Table 3, the rate of ARCA increased significantly with more years of experience at the time of the anesthetic and with academic rank (P = 0.01). Anesthetics delivered by anesthesiologists that performed <150 anesthetics per year also had a higher rate of ARCA. Using the number of days in a calendar year that practitioners delivered at least one anesthetic as the proxy for time spent delivering anesthetics, the rate of ARCA decreased as the number of annual days delivering anesthetics increased.

RESULTS

As shown in Figure 1, age at surgery was associated with ASA-PS with approximately 50% of cases of infants ≤180 days being ASA-PS ≥III compared with <20% of infants and children 181 days or older. The great majority of cardiac patients were also categorized as ASA-PS ≥III. Case complexity also rose with academic rank. Practitioners with the lowest caseload had the highest percentage of complex patients. There was also a trend toward higher case complexity for pediatric anesthesiologists who delivered anesthetics fewer days during the year.

F1-26
Figure 1:
Association of ASA physical status by patient, systems, and practitioner factors. d = days, y = years, OR = operating room.

Table 4 shows the effect of risk adjustment for patient ASA-PS ≥III and age ≤180 days on the association between pediatric ARCA and practitioner factors and anesthetic setting. The odds ratios were no longer significant for most risk factors we identified. Practitioners with lower annual days of delivering anesthetics still had a higher incidence of ARCA after risk adjustment.

T4-26
Table 4:
Effect of Adjustment for Patient ASA-PS and Age (≤180 Days) on Association Between ARCA and Practitioner Factors and Anesthetic Setting

Because ARCA was a rare event, and because only a small number of anesthesiologists had >1 ARCA, the likelihood of clustering was low. However, we performed multiple additional tests for clustering of events. Within-physician correlation of events was negligible at −5 × 10−6. The multivariable model estimates for the 3 modeling approaches used to control for clustering achieved similar results (Table 5).

T5-26
Table 5:
Multivariable Model Estimates for the 3 Modeling Approaches Used to Control for Clustering

To evaluate whether misclassification of ASA-PS might influence the results, we had 2 raters independently categorize ASA-PS for the 72 cases with ARCA. We did not have sufficient resources to recheck ASA-PS for the 276,137 cases without ARCA. After review of the ARCAs, 17% (12/72) of the original ASA-PS classifications were raised from I–II to III, 1% (1/72) from I–II to IV–V, and 14% (10/72) were raised from III to IV–V. In 1 case (1%), the ASA-PS was lowered from III to I–II. We performed a sensitivity analysis to determine the effect of ASA-PS misclassification in the anesthetics without ARCA on the patient-related risk factor multivariable logistic regression model assuming the same misclassification rate. The results showed a slightly attenuated effect of ASA-PS, and age at surgery ≤180 days remained significantly associated with ARCA.

In a post hoc analysis, we used the risk adjustment model that included patient factors only (Supplemental Digital Content 1, Supplemental Table 3, https://links.lww.com/AA/B278) to calculate an annual risk-adjusted rate of ARCA. This showed no significant differences in the frequency of ARCA, despite an increase in caseload from 18,685 anesthetics in 2000 to 31,416 anesthetics in 2011 and an increase in the number of complex patients (ASA-PS ≥III 16% in 2000 vs 23% in 2011; P < 0.001).

In patients who experienced an ARCA, 92% (66/72) survived to hospital discharge. Five patients died during the hospitalization as a result of the ARCA event (ARCA-related mortality rate of 0.18/10,000). Two deaths occurred within 24 hours (ARCA-related 24-hour mortality rate of 0.07/10,000) and 2 additional deaths occurred between 24 hours and 30 days after the ARCA event (ARCA-related 30-day mortality rate of 0.14/10,000). There was no ARCA-related mortality in ASA-PS I–II patients.

DISCUSSION

Using a large data set of over a decade of pediatric anesthetics allowed us to assess the association between pediatric ARCA and system-related and pediatric anesthesiologist-related potential risk factors. The strong associations we identified between the higher odds of pediatric ARCA and the cardiac anesthesia and annual caseload were explained by patient case-mix. The strong association that we identified between increased risk of ARCA for pediatric anesthesiologists delivering anesthetics on fewer days annually was markedly attenuated after adjusting for patient age ≤6 months and ASA-PS ≥III. However, this association remained significant in multiple models used to adjust for clustering of events within physicians (unit of analysis error),28 including bootstrap methodology, logistic regression, and GEEs.26 Our findings highlight the need for rigorous adjustment for patient risk factors in anesthesia patient safety studies.

Age and ASA-PS are commonly collected at most institutions and could be used to adjust for patient case-mix. Although our unadjusted incidence of ARCA is higher than the unadjusted incidence reported in the POCA registry (1.4/10,000)2 and in a study from Flick et al.,6 which used the POCA registry definition of ARCA (1.6/10,000), neither of these comparison studies reported were risk-adjusted ARCA rates. The POCA registry is based on voluntary reporting, whereas we used multiple methods to ensure that no ARCA cases were missed.

We identified annual days devoted to delivering pediatric anesthesia as a potentially remediable systems-related pediatric ARCA risk factor: the odds of ARCA after adjustment for patient risk factors may be over twice as high for patients anesthetized by anesthesiologists who spend approximately <30% of annual days delivering anesthetics. Two European studies also identified clinician experience and anesthetic volume as factors associated with pediatric anesthetic complications, leading a group in France to recommend a minimum of 200 pediatric cases annually and a group in the United Kingdom to recommend a graduated scale requiring a minimum of 12 anesthetics per year for the youngest age group of <6 months for anesthesiologists treating infants and children.29,30 The association we identified between ARCA and anesthesiologists caseload was confounded by case length and complexity, which may be true in these previous studies. We are unable to rule out the possibility that unidentified confounders may explain the association between pediatric ARCA risk and annual days delivering anesthetics. Because this is a single-center study, this finding is not generalizable to other pediatric institutions and practices. We believe that a multicenter study should replicate this finding before influencing changes in staffing broadly across other institutions.

Reassuringly, our risk-adjusted annual rate of ARCA from 2000 to 2011 was not significantly different, despite a 44% increase in the percentage of patients with ASA-PS ≥III and a 68% increase in total annual anesthetics. However, the reported rate of pediatric ARCA at an institution must be paired with ARCA-associated mortality. In contrast to the most recently published reports for pediatric ARCA,2,5,6,11,31 the great majority of our patients that had an ARCA survived to hospital discharge. All our ASA-PS I–II patients who had an ARCA survived to hospital discharge. Our frequency of pediatric ARCA-related mortality is lower than the most recently published studies of pediatric ARCA, despite our higher pediatric ARCA rate (Supplemental Digital Content 1, Supplemental Table 2, https://links.lww.com/AA/B278). This was also true for patients undergoing cardiac surgical procedures, because there was no pediatric ARCA-related mortality.

A major limitation of our study is that we only identified 72 ARCA events from the >300,000 anesthetics delivered, limiting its potential utility as an institutional benchmark. We used multiple methods to identify pediatric ARCA events, and our raw and risk-adjusted rates of ARCA from 2000 to 2011 were consistent over time. In addition, the data come from a single academic center and may not be representative of the total population of pediatric patients at other institutions. Another limitation of our study is that underclassification of ASA-PS occurred in 32% (23/72) of the ARCA cases, and we were unable to determine the extent of under- or overclassification in the non-ARCA cases because of the large volume of cases. However, a sensitivity analysis to evaluate the effects of ASA-PS misclassification on patient-related risk adjusters showed that our model was robust. It would likely be beneficial to develop a reliable ASA-PS score specifically for neonates, infants, and children.

We were unable to accurately document whether trainees (residents and fellows), certified registered nurse anesthetists, and student registered nurse anesthetists were involved at the time of the ARCA, an important potential systems factor. Multiple trainees and nurse anesthetists can sign the anesthetic record; however, they can potentially move to another anesthetic to provide breaks and any other coverage needs in the OR at that time. Because trainee and nurse anesthetist involvement is common in most of the cases at our institution, the control group of anesthetics in which trainees or nurse anesthetists were not involved would be very small.

CONCLUSIONS

We have shown that adjustment for patient risk factors—including very young age and higher pre-procedure level of systemic disease—is essential when comparing anesthetic adverse event rates. Most of the practitioner-related pediatric ARCA risk factors we identified in univariate analyses were no longer associated after adjustment for these patient-related risk factors. At our hospital, lower annual days spent by anesthesiologists administering anesthesia remained associated with a higher risk of pediatric ARCA. Although this association was attenuated after controlling for patient case-mix, it has influenced our staffing patterns because it remained robust in sensitivity analyses. Because pediatric ARCA events are rare and strongly influenced by patient case-mix, such metrics must be used with caution as a quality-of-care indicator.

DISCLOSURES

Name: Steven E. Zgleszewski, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Steven E. Zgleszewski has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the co-first author responsible for archiving the study files.

Name: Dionne A. Graham, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, write the manuscript, and meets criteria for co-first author.

Attestation: Dionne A. Graham has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Paul R. Hickey, MD.

Contribution: This author helped design the study and write the manuscript.

Attestation: Paul R. Hickey reviewed the analysis of the data and approved the final manuscript.

Name: Robert M. Brustowicz, MD.

Contribution: This author helped design the study, analyze the data, and write the manuscript.

Attestation: Robert M. Brustowicz approved the final manuscript.

Name: Kirsten C. Odegard, MD.

Contribution: This author helped write the manuscript and collect the data.

Attestation: Kirsten C. Odegard approved the final manuscript.

Name: Rahul Koka, MD.

Contribution: This author helped analyze the data and write the manuscript.

Attestation: Rahul Koka approved the final manuscript.

Name: Christian Seefelder, MD.

Contribution: This author helped write the manuscript and collect the data.

Attestation: Christian Seefelder approved the final manuscript.

Name: Andres T. Navedo, MD.

Contribution: This author helped write the manuscript and collect the data.

Attestation: Andres T. Navedo approved the final manuscript.

Name: Adrienne G. Randolph, MD, MSc.

Contribution: This author helped design the study, analyze the data, and write the manuscript.

Attestation: Adrienne G. Randolph reviewed the analysis of the data and approved the final manuscript.

This manuscript was handled by: Steven L. Shafer, MD.

ACKNOWLEDGMENTS

The authors thank Karen Gould for providing academic information for the staff anesthesiologists involved in this study.

REFERENCES

1. Cohen NA, Hannenberg AA Anesthesiology. 2010;113:1004–6
2. Morray JP, Geiduschek JM, Ramamoorthy C, Haberkern CM, Hackel A, Caplan RA, Domino KB, Posner K, Cheney FW. Anesthesia-related cardiac arrest in children: initial findings of the Pediatric Perioperative Cardiac Arrest (POCA) Registry. Anesthesiology. 2000;93:6–14
3. Newland MC, Ellis SJ, Lydiatt CA, Peters KR, Tinker JH, Romberger DJ, Ullrich FA, Anderson JR. Anesthetic-related cardiac arrest and its mortality: a report covering 72,959 anesthetics over 10 years from a US teaching hospital. Anesthesiology. 2002;97:108–15
4. Sprung J, Warner ME, Contreras MG, Schroeder DR, Beighley CM, Wilson GA, Warner DO. Predictors of survival following cardiac arrest in patients undergoing noncardiac surgery: a study of 518,294 patients at a tertiary referral center. Anesthesiology. 2003;99:259–69
5. Braz LG, Módolo NS, do Nascimento P Jr, Bruschi BA, Castiglia YM, Ganem EM, de Carvalho LR, Braz JR. Perioperative cardiac arrest: a study of 53,718 anaesthetics over 9 yr from a Brazilian teaching hospital. Br J Anaesth. 2006;96:569–75
6. Flick RP, Sprung J, Harrison TE, Gleich SJ, Schroeder DR, Hanson AC, Buenvenida SL, Warner DO. Perioperative cardiac arrests in children between 1988 and 2005 at a tertiary referral center: a study of 92,881 patients. Anesthesiology. 2007;106:226–37
7. Gobbo Braz L, Braz JR, Módolo NS, do Nascimento P, Brushi BA, Raquel de Carvalho L. Perioperative cardiac arrest and its mortality in children. A 9-year survey in a Brazilian tertiary teaching hospital. Paediatr Anaesth. 2006;16:860–6
8. Ellis SJ, Newland MC, Simonson JA, Peters KR, Romberger DJ, Mercer DW, Tinker JH, Harter RL, Kindscher JD, Qiu F, Lisco SJ. Anesthesia-related cardiac arrest. Anesthesiology. 2014;120:829–38
9. Tiret L, Desmonts JM, Hatton F, Vourc’h G. Complications associated with anaesthesia—a prospective survey in France. Can Anaesth Soc J. 1986;33:336–44
10. Keenan RL, Boyan CP. Cardiac arrest due to anesthesia. A study of incidence and causes. JAMA. 1985;253:2373–7
11. Bhananker SM, Ramamoorthy C, Geiduschek JM, Posner KL, Domino KB, Haberkern CM, Campos JS, Morray JP. Anesthesia-related cardiac arrest in children: update from the Pediatric Perioperative Cardiac Arrest Registry. Anesth Analg. 2007;105:344–50
12. Olsson GL, Hallén B. Cardiac arrest during anaesthesia. A computer-aided study in 250,543 anaesthetics. Acta Anaesthesiol Scand. 1988;32:653–64
13. Rackow H, Salanitre E, Green LT. Frequency of cardiac arrest associated with anesthesia in infants and children. Pediatrics. 1961;28:697–704
14. Haller G, Stoelwinder J, Myles PS, McNeil J. Quality and safety indicators in anesthesia: a systematic review. Anesthesiology. 2009;110:1158–75
15. Beecher HK, Todd DP. A study of the deaths associated with anesthesia and surgery: based on a study of 599, 548 anesthesias in ten institutions 1948-1952, inclusive. Ann Surg. 1954;140:2–35
16. Murat I, Constant I, Maud’huy H. Perioperative anaesthetic morbidity in children: a database of 24,165 anaesthetics over a 30-month period. Paediatr Anaesth. 2004;14:158–66
17. Dripps RD. New classification of physical status. Anesthesiology. 1963;24:111
18. Tiret L, Nivoche Y, Hatton F, Desmonts JM, Vourc’h G. Complications related to anaesthesia in infants and children. A prospective survey of 40240 anaesthetics. Br J Anaesth. 1988;61:263–9
19. Paterson N, Waterhouse P. Risk in pediatric anesthesia. Paediatr Anaesth. 2011;21:848–57
20. Odegard KC, DiNardo JA, Kussman BD, Shukla A, Harrington J, Casta A, McGowan FX Jr, Hickey PR, Bacha EA, Thiagarajan RR, Laussen PC. The frequency of anesthesia-related cardiac arrests in patients with congenital heart disease undergoing cardiac surgery. Anesth Analg. 2007;105:335–43
21. Meier BM, Stone PW, Gebbie KM. Public health law for the collection and reporting of health care-associated infections. Am J Infect Control. 2008;36:537–51
22. Passaretti CL, Barclay P, Pronovost P, Perl TMMaryland Health Care Commission Health Care–Associated Infection Technical Advisory Committee. . Public reporting of health care-associated infections (HAIs): approach to choosing HAI measures. Infect Control Hosp Epidemiol. 2011;32:768–74
23. McKibben L, Horan T, Tokars JI, Fowler G, Cardo DM, Pearson ML, Brennan PJHeathcare Infection Control Practices Advisory Committee. . Guidance on public reporting of healthcare-associated infections: recommendations of the Healthcare Infection Control Practices Advisory Committee. Am J Infect Control. 2005;33:217–26
24. Sessler DI, Sigl JC, Manberg PJ, Kelley SD, Schubert A, Chamoun NG. Broadly applicable risk stratification system for predicting duration of hospitalization and mortality. Anesthesiology. 2010;113:1026–37
25. Odegard KC, Bergersen L, Thiagarajan R, Clark L, Shukla A, Wypij D, Laussen PC. The frequency of cardiac arrests in patients with congenital heart disease undergoing cardiac catheterization. Anesth Analg. 2014;118:175–82
26. Efron B, Tibshirani RJ An Introduction to the Bootstrap. 2003 New York, NY Chapman & Hall/CRC Monographs on Statistics & Applied Probability
27. Breslow NE, Day NE Statistical Methods In Cancer Research. Volume II—The Design and Analysis of Cohort Studies. 1987 Oxford, UK Oxford University Press; IARC Scientific Publications:1–406
28. Calhoun AW, Guyatt GH, Cabana MD, Lu D, Turner DA, Valentine S, Randolph AG. Addressing the unit of analysis in medical care studies: a systematic review. Med Care. 2008;46:635–43
29. Lunn JN. Implications of the National Confidential Enquiry into Perioperative Deaths for paediatric anaesthesia. Paediatr Anaesth. 1992;2:69–72
30. Auroy Y, Ecoffey C, Messiah A, Rouvier B. Relationship between complications of pediatric anesthesia and volume of pediatric anesthetics. Anesth Analg. 1997;84:234–5
31. Ramamoorthy C, Haberkern CM, Bhananker SM, Domino KB, Posner KL, Campos JS, Morray JP. Anesthesia-related cardiac arrest in children with heart disease: data from the Pediatric Perioperative Cardiac Arrest (POCA) registry. Anesth Analg. 2010;110:1376–82

Supplemental Digital Content

© 2016 International Anesthesia Research Society