Share this article on:

2007 American College of Cardiology/American Heart Association (ACC/AHA) Guidelines on Perioperative Cardiac Evaluation Are Usually Incorrectly Applied by Anesthesiology Residents Evaluating Simulated Patients

Vigoda, Michael M. MD, MBA*; Sweitzer, BobbieJean MD; Miljkovic, Nikola; Arheart, Kristopher L. EdD§; Messinger, Shari PhD§; Candiotti, Keith MD*; Lubarsky, David MD, MBA*

doi: 10.1213/ANE.0b013e31820a1457
Economics, Education, and Policy: Research Reports
Chinese Language Editions

BACKGROUND: The 2007 American College of Cardiology/American Heart Association (ACC/AHA) Guidelines on Perioperative Cardiac Evaluation and Care for Noncardiac Surgery is the accepted standard for perioperative cardiac evaluation. Anesthesiology training programs are required to teach these algorithms. We estimated the percentage of residents nationwide who correctly applied suggested testing algorithms from the ACC/AHA guidelines when they evaluated simulated patients in common clinical scenarios.

METHODS: Anesthesiology resident volunteers at 24 training programs were presented with 6 scenarios characterized by surgical procedure, patient's risk factors, and patient's functional capacity. Scenarios and 5 possible recommendations per scenario were both presented in randomized orders. Senior anesthesiologists at 24 different United States training programs along with the first author of the 2007 ACC/AHA guidelines validated the appropriate recommendation to this web-based survey before distribution.

RESULTS: The 548 resident participants, representing 12% of anesthesiology trainees in the United States, included 48 PGY-1s (preliminary year before anesthesia training), 166 Clinical Anesthesia Year 1 (CA-1) residents, 161 CA-2s, and 173 CA-3s. For patients with an active cardiac condition, the upper 95% confidence bound for the percent of residents who recommended evaluations consistent with the guidelines was 78%. However, for the remaining 5 scenarios, the upper 95% confidence bound for the percent of residents with an appropriate recommendation was 46%.

CONCLUSIONS: The results show that fewer than half of anesthesiology residents nationwide correctly demonstrate the approach considered the standard of care for preoperative cardiac evaluation. Further study is necessary to elucidate the correct intervention(s), such as use of decision support tools, increased clarity of guidelines for routine use, adjustment in educational programs, and/or greater familiarity of responsible faculty with the material.

Published ahead of print March 8, 2011 Supplemental Digital Content is available in the text.

From the *Department of Anesthesiology, and §Department of Epidemiology & Public Health, Division of Biostatistics, University of Miami Miller School of Medicine, Miami, Florida; Department of Anesthesiology, University of Chicago, Chicago, Illinois; and Department of Anesthesiology, University of Nebraska Medical Center, Omaha, Nebraska.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.anesthesia-analgesia.org).

Funded by the Department of Anesthesiology at the University of Miami Miller School of Medicine.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Michael M. Vigoda, MD, MBA, Department of Anesthesiology, UM/JMH Hospital, Central 300, Miami FL 33136. Address e-mail to mvigoda@med.miami.edu.

Accepted November 22, 2010

Published ahead of print March 8, 2011

Cardiac complications in the perioperative period increase mortality13 and cost of care.4 The American College of Cardiology (ACC) in conjunction with the American Heart Association (AHA) has published evidence-based guidelines for perioperative cardiac evaluation. These guidelines have been endorsed by the American Society of Anesthesiologistsa and adopted by the Agency for Healthcare Research and Quality. b However, implementation of these guidelines has been problematic.57

Although the American Board of Anesthesiology requires anesthesiology residency training to teach these algorithms,c studies show that residents in a variety of other specialties lack knowledge of their specialty-specific clinical guidelines.811 In this investigation, we estimated the percentage of residents who correctly apply suggested testing algorithms from the ACC/AHA guidelines. This was accomplished by having residents evaluate simulated patients with common clinical scenarios, and select a recommendation from a set of choices.

Back to Top | Article Outline

METHODS

Twenty-four anesthesiology training programs participated in the study. IRB approval was obtained at each institution.

We designed a web-based survey assessing application of the 2007 ACC/AHA Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery (Fig. 1).1214 Six clinical scenarios were constructed as described below.

Figure 1

Figure 1

The ACC/AHA algorithm is represented by a flowchart with 5 steps, and 7 distinct end points. Steps 1 to 4 have a single end point. Step 5 is the most complex and requires both consideration of the type of surgery and the presence and number of clinical risk factors. One proceeds through the algorithm in a stepwise manner. Recommendations are determined by the first step that applies to a particular patient and procedure.

We excluded step 1 (the end point pertaining to the evaluation of a patient requiring emergency surgery) and considered 6 scenarios that focused on nonemergency, noncardiac surgery. The survey assessed application of the ACC/AHA guidelines in steps 2 to 4, and 3 distinct decision points in step 5:

  • Scenario 1.
  • Active cardiac condition (corresponds to step 2 of ACC/AHA algorithm).
  • Scenario 2.
  • No active cardiac conditions, low-risk surgery (corresponds to step 3 of ACC/AHA algorithm).
  • Scenario 3.
  • No active cardiac conditions, intermediate-risk surgery, good functional capacity, has 1 clinical risk factor (corresponds to step 4 of ACC/AHA algorithm).
  • Scenario 4.
  • No active cardiac conditions, intermediate-risk surgery, poor/unknown functional capacity, has 2 clinical risk factors (corresponds to step 5 of ACC/AHA algorithm).
  • Scenario 5.
  • No active cardiac conditions, vascular surgery (1 or 2 risk factors) (corresponds to step 5 of ACC/AHA algorithm).
  • Scenario 6.
  • No active cardiac conditions, intermediate-risk surgery, and no clinical risk factors (corresponds to step 5 of ACC/AHA algorithm).

Interpretation of results assumes that correct application of the algorithm to each scenario would result in correct recommendation in practice.

Back to Top | Article Outline

Subject Recruitment and Characteristics of Training Programs

A standardized message describing the study was sent via e-mail to each program coordinator who then forwarded the message to their residents. The e-mail contained an institution-specific URL so that the number of participating residents from each institution could be tracked and feedback could be provided to the program coordinators regarding their program's level of participation. Program coordinators, at their own discretion, publicized the study using a variety of methods including follow-up e-mails, announcements at resident meetings, and/or personal communication with individual residents.

The first page of the program-specific website indicated that “we are seeking your participation in a study which evaluates clinicians' level of understanding of the 2007 ACC/AHA Guidelines for Perioperative Cardiac Evaluation. You will be presented with 6 clinical scenarios. Please make your recommendations based on your understanding/recall of the ACC/AHA guidelines.” Residents were informed that “any and all data would not be reported to the management of your department, not be used to assess your status in your residency program, and not be used to make any decision regarding your employment status.” To participate, residents had to acknowledge that they read the consent form by clicking on the “Agree to Participate” button.

The application was coded as a web application using PHP5, an open-source programming language (http://php.net). Data were stored in a MySQL database (Sun Microsystems, Santa Clara, CA).

Back to Top | Article Outline

Scenario Presentation

Although a web-based approach available over an extended period of time permitted flexibility for participation and anonymity of participants, it allowed for the possibility of contaminated results because one resident could inform another of the scenarios (Appendix 1). To minimize this possibility, 3 variants of each scenario were constructed. In addition, the order of scenario presentation and the variant of the scenario presented were randomized for each survey participant (Online Appendix 1, see Supplemental Digital Content 1, http://links.lww.com/AA/A211).

Back to Top | Article Outline

Validation of Correct Recommendations

There is no validated tool for assessing the application of preoperative testing standards. We therefore designed our scenarios based on real-world clinical presentations and validated the correct responses with a panel of experts in the following manner. The principal author in conjunction with 3 directors of preanesthesia evaluation clinics (BS, KC, AJ, Online Appendix 5, see Supplemental Digital Content 5, http://links.lww.com/AA/A215) initially created the 6 clinical scenarios. After reaching agreement that the 6 scenarios represented the 6 desired end points in the ACC/AHA flowchart, 3 variants (i.e., variant A, variant B, and variant C) were created for each scenario. This was accomplished by changing the patient's age, clinical risk factors (if any), and the planned surgical procedure of the same risk type. The 18 different scenarios were then reviewed and, if necessary, modified by 5 anesthesiology program directors (RG, PK, MW, SK, MP), 4 additional directors of preanesthesia evaluation clinics (RO, GVN, AE, DR), 2 chiefs of anesthesia (GL, SE), 2 cardiac anesthesiologists (MC, JG), a neuroanesthesiologist (SB), 2 intensivists (JR, PK), and a director of quality improvement (SD). Finally, the lead author of the 2007 ACC/AHA guidelines, Dr. Lee Fleisher, reviewed the consensus recommendations for each scenario and agreed that they were consistent with the guidelines.

Back to Top | Article Outline

Analytic Data

Analytic data consist of 3 major components: identification, outcome, and survey design variables. The identification variables include a self-chosen nonidentifiable username, the program number, and the training year of the respondent (Post Graduate Year [PGY]-1, Clinical Anesthesia Year 1 [CA-1], CA-2, and CA-3). The outcome variables are the binary scores (1 = correct, 0 = incorrect) for each of the 6 scenarios. Survey design variables include the presentation order of the scenario, variant of each scenario, time taken (in seconds) to answer each scenario, number of times they revised their recommendation, and number of days from the beginning of the study to the time the survey was completed.

Back to Top | Article Outline

Statistical Methods

The data structure for this analysis was complex because of the clustering of respondents within programs and additional clustering of scenarios within respondent. The generalized linear mixed model accommodates the complex covariance structure resulting from the clustering of the data as well as the binary distribution of the dependent variable (Appendix 2).

The data are reported as the percentage and 95% confidence interval (CI) for the estimate of the correct response rate corresponding to each scenario and training year. The percentages are generalized over the program and survey design variables as shown in Appendix 3 (Calculation of Percent Correct). We used contrasts to make pairwise comparisons among the training years for each scenario. The 0.05 significance level was used to determine statistical significance for all comparisons. For readers who prefer Bonferroni corrected comparisons, the critical significant level is 0.008 for the comparisons of year within each scenario. SAS 9.2 (SAS Institute, Inc., Cary, NC) was used for all analyses.

Back to Top | Article Outline

RESULTS

The 548 respondents from the 24 participating programs included 48 PGY-1s (9%), 166 CA-1s (30%), 161 CA-2s (29%), and 173 CA-3s (32%). All programs required residents to participate in preoperative assessment clinics. Most programs had a yearly presentation on the guidelines and most coordinators reported that their programs emphasized the guidelines (Online Appendix 2, see Supplemental Digital Content 2, http://links.lww.com/AA/A212).

Table 1 presents the overall percentage of residents arriving at the correct recommendation along with the 95% CI corresponding to each scenario. Overall percent correct was highest for scenario 1 (95% CI: 54.6–78.4). For all others, the upper 95% CI for percent correct was ≤46.4%: i.e., scenario 2 (95% CI: 21.4–43.1); scenario 3 (95% CI: 13.7–32.0); scenario 4 (95% CI: 27.2–46.4); scenario 5 (95% CI: 18.8–39.0); and scenario 6 (95% CI: 12.6–28.8).

Table 1

Table 1

There were few significant differences in the percentages of residents with correct recommendations among the training year categories (Table 1). For scenario 2, more CA-3s arrived at the correct recommendation than CA-1s (38% vs 24%, P = 0.009). PGY-1s were less likely to reach the correct recommendation than any other training year when presented with scenario 6 (8% vs CA-1 21%, P = 0.034; CA-2 24%, P = 0.015; CA-3 32%, P = 0.001). CA-1s were less likely than CA-3s to make the correct recommendation for scenario 6 (P = 0.04).

The pattern of incorrect recommendation selection was similar across all years of training for each scenario (Online Appendix 3, http://links.lww.com/AA/A213).

Response time was not associated with the chance of a correct response, except for scenario 1 whereby residents who took longer to respond were more likely to provide an incorrect recommendation (Online Appendix 4, see Supplemental Digital Content 4, http://links.lww.com/AA/A214).

Back to Top | Article Outline

DISCUSSION

Over a span of 3 decades, literature regarding preoperative cardiovascular screening and testing has been reviewed and synthesized by the combined Task Force on Practice Guidelines of the ACC/AHA.1214 The 2007 ACC/AHA guidelines stipulate that relatively few patients will benefit from advanced preoperative testing. It is unclear how well they are being followed in actual practice. Changing individual providers' practices and hospitals' protocols and standardized orders will take time. Residents often learn testing strategies based on how their faculty approach problems. We tested whether residents applied suggested testing algorithms from the ACC/AHA guidelines when they evaluated simulated patients in common clinical scenarios.

For patients with an active cardiac condition (scenario 1), the upper 95% confidence bound for the percent of residents who recommended evaluations consistent with the guidelines was 78%. However, for the remaining 5 scenarios, the upper 95% confidence bound for the percent of residents with an appropriate recommendation was 46%. We do not know which of several possible causes explain our results: guidelines are too confusing to be useful, anesthesia practices are not matched to current recommendations, faculty are deficient in their knowledge, and/or educational programs need to be modified.

Our results are consistent with findings from a pilot study involving perioperative clinicians.d With the exception of the active cardiac condition scenario, whereby 87% of clinicians' recommended options that were consistent with the guidelines, recommendations were consistent with guidelines <50% of the time for the remaining 5 scenarios: i.e., scenario 2, 45%; scenario 3, 45%; scenario 4, 38%; scenario 5, 36%; and scenario 6, 45%. Regardless, the study's findings are very much at odds with societal expectations of near-universal adherence to clinical guidelines.

There are limitations to this study. Because residents self-selected, participation may have been skewed toward those with greater confidence in their ability to apply the guidelines. Although we considered block randomization as a method to select residents, we chose to solicit volunteers for 2 reasons: (1) to increase the number of participants, and (2) to ensure the anonymity of participants. We were concerned that we would not attain a sufficient number of participants if we limited our potential recruitment sample. Using a random selection method would have meant either a priori identification of a group of residents or having residents login and then determine whether they would be selected to view the scenarios. The former method would have identified which residents were not participants. This would have more easily allowed identification of participating residents and, as a result, some IRBs did not favor the randomization method. Finally, we believed that the study was, in part, an educational exercise for the residents. Our perspective was that if a resident wanted to participate then we would offer them the opportunity.

In summary, anesthesiology residents (representing 12% of all United States trainees) inaccurately applied the guidelines to simulated patients in common clinical scenarios. We estimate that fewer than half of anesthesiology residents nationwide correctly demonstrate the approach considered to be the standard of care for preoperative cardiac evaluation. Further study analyzing the usefulness of decision support tools and attending practice patterns in relation to these guidelines is essential.

Back to Top | Article Outline

DISCLOSURES

Name: Michael M. Vigoda, MD, MBA.

Contribution: Study design, data analysis, conduct of study, and manuscript preparation.

Name: BobbieJean Sweitzer, MD.

Contribution: Study design, data analysis, conduct of study, and manuscript preparation.

Name: Nikola Miljkovic.

Contribution: Study design and conduct of study.

Name: Kristopher L. Arheart, EdD.

Contribution: Data analysis and manuscript preparation.

Name: Shari Messinger, PhD.

Contribution: Data analysis and manuscript preparation.

Name: Keith Candiotti, MD.

Contribution: Study design, data analysis, conduct of study, and manuscript preparation.

Name: David Lubarsky, MD, MBA.

Contribution: Study design, data analysis, and manuscript preparation.

Name: Program Coordinators (Online Appendix 5 see Supplemental Digital Content 5, http://links.lww.com/AA/A215).

Contribution: Conduct of study.

Back to Top | Article Outline

ACKNOWLEDGMENTS

The authors thank Dr. Lee Fleisher for his review of the scenario descriptions (and the set of possible recommendations) and for his participation in the video that residents could view after completing the study. The authors also thank Sophie and Joan Leonard for their support and Dr. Vicente Behrens for his assistance with the preparation of the tables.

a See http://www.asahq.org/Knowledge-Base/Anesthesia-Practice/ASA/ACC-AHA-2007-Guidelines-on-Perioperative-Cardiovascular-Evaluation-and-Care-for-Noncardiac-Surgery.aspx. Accessed October 23, 2010.
Cited Here...

b See http://www.guideline.gov/summary/summary.aspx?doc_id=11510&nbr=005963&string=ACC%2fAHA+AND+preoperative. Accessed May 16, 2010.
Cited Here...

c Clinical Sciences A.1.1.2 American College of Cardiology/American Heart Association Guidelines for Perioperative Cardiovascular Evaluation. Available at: http://www.theaba.org/pdf/ITEContentOutline.pdf, page 12. Accessed October 23, 2010.
Cited Here...

d Vigoda MM, Sweitzer BJ, Miljkovic N, Jaffer A. Perioperative clinicians' adherence to 2007 ACC/AHA guidelines is low. 2010 Annual Meeting of the American Society of Anesthesiologists, A356.
Cited Here...

Back to Top | Article Outline

REFERENCES

1. Goldman L. Multifactorial index of cardiac risk in noncardiac surgery: ten-year status report. J Cardiothorac Anesth 1987;1:237–44
2. Lee TH. Reducing cardiac risk in noncardiac surgery. N Engl J Med 1999;341:1838–40
3. Auerbach AD, Goldman L. Beta-blockers and reduction of cardiac events in noncardiac surgery: scientific review. JAMA 2002;287:1435–44
4. Lawrence V, Hilsenbeck SG, Mulrow C, Dhanda R, Sapp J, Page C. Incidence and hospital stay for cardiac and pulmonary complications after abdominal surgery. J Gen Intern Med 1995;10:671–8
5. Hoeks SE, Scholte op Reimer WJ, Lenzen MJ, van Urk H, Jörning PJ, Boersma E, Simoons ML, Bax JJ, Poldermans D. Guidelines for cardiac management in noncardiac surgery are poorly implemented in clinical practice: results from a peripheral vascular survey in the Netherlands. Anesthesiology 2007;107:537–44
6. Hoeks SE, Scholte Op Reimer WJ, Schouten O, Lenzen MJ, van Urk H, Poldermans D. Statin use in the elderly: results from a peripheral vascular survey in The Netherlands. J Vasc Surg 2008;48:891–5
7. Hoeks SE, Scholte op Reimer WJ, van Gestel YR, Schouten O, Lenzen MJ, Flu WJ, van Kuijk JP, Latour C, Bax JJ, van Urk H, Poldermans D. Medication underuse during long-term follow-up in patients with peripheral arterial disease. Circ Cardiovasc Qual Outcomes 2009;2:338–43
8. Logie CI, Smith SE, Nagy P. Evaluation of resident familiarity and utilization of the ACR musculoskeletal study appropriateness criteria in the context of medical decision support. Acad Radiol 2010;17:251–4
9. Wolf JM, Athwal GS, Hoang BH, Mehta S, Williams AE, Owens BD. Knowledge of levels of evidence criteria in orthopedic residents. Orthopedics 2009;32:494
10. Agrawal V, Ghosh AK, Barnes MA, McCullough PA. Awareness and knowledge of clinical practice guidelines for CKD among internal medicine residents: a national online survey. Am J Kidney Dis 2008;52:1061–9
11. Karakousis PC, Sifakis FG, de Oca RM, Amorosa VC, Page KR, Manabe YC, Campbell JD. U.S. medical resident familiarity with national tuberculosis guidelines. BMC Infect Dis 2007;7:89
12. Fleisher LA, Beckman JA, Brown KA, Calkins H, Chaikof E, Fleischmann KE, Freeman WK, Froehlich JB, Kasper EK, Kersten JR, Riegel B, Robb JF, Smith SC Jr, Jacobs AK, Adams CD, Anderson JL, Antman EM, Buller CE, Creager MA, Ettinger SM, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Lytle BW, Nishimura R, Ornato JP, Page RL, Tarkington LG, Yancy CW. ACC/AHA 2007 guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery): developed in collaboration with the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, and Society for Vascular Surgery. Circulation 2007;116:e418–99
13. Fleisher LA, Beckman JA, Brown KA, Calkins H, Chaikof EL, Fleischmann KE, Freeman WK, Froehlich JB, Kasper EK, Kersten JR, Riegel B, Robb JF, Smith SC Jr, Jacobs AK, Adams CD, Anderson JL, Antman EM, Buller CE, Creager MA, Ettinger SM, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Lytle BW, Nishimura R, Ornato JP, Page RL, Riegel B, Tarkington LG, Yancy CW. ACC/AHA 2007 Guidelines on Perioperative Cardiovascular Evaluation and Care for Noncardiac Surgery: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) developed in collaboration with the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, and Society for Vascular Surgery. J Am Coll Cardiol 2007;50:1707–32
14. Fleisher LA, Beckman JA, Brown KA, Calkins H, Chaikof E, Fleischmann KE, Freeman WK, Froehlich JB, Kasper EK, Kersten JR, Riegel B, Robb JF, Smith SC Jr, Jacobs AK, Adams CD, Anderson JL, Antman EM, Buller CE, Creager MA, Ettinger SM, Faxon DP, Fuster V, Halperin JL, Hiratzka LF, Hunt SA, Lytle BW, Nishimura R, Ornato JP, Page RL, Riegel B, Tarkington LG, Yancy CW. ACC/AHA 2007 guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery: executive summary—a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery). Anesth Analg 2008;106:685–712

Scenario descriptions follow. When presented to the residents, scenario numbers and description were not displayed.

Although the order of possible recommendations was presented in a random order, for display purposes, the correct recommendation for each scenario is #1.

Back to Top | Article Outline

APPENDIX 1: SCENARIO DESCRIPTION

Scenario 1: Active Cardiac Condition

  • 74-year-old man with history of buttock claudication who has Type 2 diabetes and receiving insulin and has chronic renal insufficiency (Cr = 2.1 mg/dL) is scheduled for ileofemoral repair. He tells you that recently he has experienced chest pain and has shortness of breath going upstairs to his bedroom.
  • 68-year-old woman with abdominal pain after eating, history of hypertension, chronic renal insufficiency with a Cr = 2.1 mg/dL is scheduled for cholecystectomy. On preoperative evaluation, she reports that she finds that she tires more easily when cleaning her house and last night when arguing with her husband she had some chest pressure.
  • 62-year-old woman with history of high blood pressure (BP) has spinal stenosis with increasing back pain. She is scheduled for a laminectomy. She reports “feeling tired and run down” in the past month. She gives a history of angina that has been getting worse in the past few months.

The most appropriate recommendation is:

  1. Patient should have a stress test (dobutamine echo, exercise, or nuclear imaging) before surgery
  2. Patient needs a cardiac catheterization before I can complete evaluation
  3. Proceed with surgery if today's electrocardiogram (ECG) is unchanged from ECG performed 6 months ago
  4. Start β-blocker and proceed with planned surgery
  5. Proceed with surgery
Back to Top | Article Outline

Scenario 2: No Active Cardiac Conditions, Low-Risk Surgery

  • 62-year-old African American woman with history of hypertension, poorly controlled diabetes, and sickle cell trait is scheduled for Achilles tendon repair. Blood glucose levels range between 120 and 240 mg/dL. She walks to work daily. Heart rate (HR) 84 bpm, BP 151/84 mm Hg.
  • 65-year-old Hispanic woman with history of hypertension, poorly controlled diabetes, and osteoarthritis is scheduled for diagnostic laparoscopy. Blood glucose levels range between 100 and 250 mg/dL. She is able to work in the garden without problem. HR 82 bpm, BP 143/92 mm Hg.
  • 60-year-old Chinese man with history of tobacco abuse, poorly controlled diabetes, and hypercholesterolemia is scheduled for incisional hernia. He reports being able to climb 2 flights of stairs. Blood glucose levels range between 130 and 290 mg/dL. HR 82 bpm, BP 138/90 mm Hg.

The most appropriate recommendation is:

  1. Proceed with the surgery
  2. Order a stress test (dobutamine echo, exercise, or nuclear imaging)
  3. Start prophylactic atenolol therapy and delay surgery until the HR is adequately blocked
  4. Order hemoglobin A1c to assess diabetes control
  5. Patient needs an ECG before I can make a recommendation
Back to Top | Article Outline

Scenario 3: No Active Cardiac Conditions, Intermediate-Risk Surgery, Good Functional Capacity, 1 Clinical Risk Factor

  • 65-year-old Latino man with history of smoking 1 pack per day. After his myocardial infarction (MI) 10 weeks ago, he started cardiac rehab and has now returned to his passion—samba dancing twice a week. He was hospitalized for pneumonia 7 months ago and required supplemental oxygen for 4 days. You are evaluating him for a preoperative evaluation in preparation for a total knee replacement due to severe osteoarthritis.
  • 74-year-old Caucasian man with 50-pack-year history of smoking and longstanding diverticulitis. After his MI 2 months ago, he started cardiac rehab and now walks at least 15 minutes each day on his treadmill, averaging 4 miles per hour. You are evaluating him for a preoperative evaluation in preparation for a colectomy.
  • 72-year-old African American woman with history of smoking 1 pack per day. After her MI 2.5 months ago, she started cardiac rehab and now swims 20 minutes each day without stopping. You are evaluating her for a preoperative evaluation in preparation for a laminectomy due to severe spinal stenosis.

The most appropriate recommendation is:

  1. Proceed with planned surgery
  2. Patient should have a stress test (dobutamine echo, exercise, or nuclear image) before surgery
  3. Surgery should be delayed at least 3 months after an MI
  4. Delay surgery until the patient is at least 6 months out from their MI
  5. Need ECG and pulmonary function tests before I can make a recommendation
Back to Top | Article Outline

Scenario 4: No Active Cardiac Conditions, Intermediate-Risk Surgery, Poor/Unknown Functional Capacity, 2 Clinical Risk Factors

  • 72-year-old man with longstanding history of Type 2 diabetes resulting in retinopathy and nephropathy. He can get dressed and brush his teeth but is otherwise not very active. Preoperative creatinine is 2.3 mg/dL. He was recently diagnosed with colon cancer. You are evaluating him for a preoperative evaluation in preparation for a colectomy. HR 82 bpm.
  • 77-year-old woman with remote history of MI (>5 years ago). Preoperative creatinine is 2.6 mg/dL. She can walk on a flat surface for 1 to 2 blocks. You are evaluating her for a preoperative evaluation in preparation for a nephrectomy due to renal cell carcinoma. HR 82 bpm.
  • 65-year-old man with history of Type 2 diabetes and his functional capacity is quite limited. Preoperative creatinine is 2.6 mg/dL. He can shower and dress himself. You are evaluating him for a preoperative evaluation in preparation for the removal of an abdominal tumor. HR 82 bpm.

The most appropriate recommendation is to:

  1. Proceed to surgery as long as a positive stress test would not change perioperative plans
  2. Start preoperative β-blockers and delay surgery until the HR is adequately blocked
  3. Patient should have a 2-dimensional echocardiogram
  4. Need ECG before I can make a recommendation
  5. Stress test is absolutely indicated before going to operating room
Back to Top | Article Outline

Scenario 5: No Active Cardiac Conditions, Vascular Surgery, 1 or 2 Clinical Risk Factors

  • 64-year-old man with history of tobacco abuse and stroke is scheduled for an ileofemoral repair because of increasing lower extremity pain. The patient's BP is 172/83 mm Hg.
  • 68-year-old woman with history of hypertension and several years ago had 2 transient ischemic attacks, the last of which was 18 months ago. She is scheduled for an axillofemoral bypass. The patient's BP is 168/85 mm Hg.
  • 78-year-old man with history of hypertension, Type 2 diabetes, and claudication from peripheral vascular disease presents with an abdominal mass. An abdominal computed tomography with contrast shows 5.6-cm abdominal aortic aneurysm. The patient's BP is 158/80 mm Hg.

The most appropriate recommendation is:

  1. Proceed to surgery as long as a positive stress test would not change perioperative plans
  2. Need ECG before I can make recommendation
  3. Delay surgery until BP <140/70 mm Hg
  4. Cardiac consultation
  5. Stress test is absolutely indicated before going to operating room
Back to Top | Article Outline

Scenario 6: No Active Cardiac Conditions, Intermediate-Risk Surgery, No Clinical Risk Factors

  • 64-year-old obese man with history of tobacco abuse, recently diagnosed with laryngeal cancer, is scheduled for a laryngectomy. Father died at 45 years of age of an MI. The patient's BP is 172/93 mm Hg.
  • 68-year-old woman with history of hypertension is scheduled for a laminectomy. Patient is taking a statin for her hypercholesterolemia. The patient's BP is 168/95 mm Hg.
  • 78-year-old man with history of hypertension and tobacco abuse is scheduled for cholecystectomy. ECG shows nonspecific ST changes from 1 year ago, but there is no history of MI. The patient's BP is 165/94 mm Hg.

The most appropriate recommendation is:

  1. Proceed with planned surgery
  2. Patient needs to have a stress test (dobutamine echo, exercise, or nuclear imaging) before I can make a recommendation
  3. Need ECG before I can make recommendation
  4. Delay surgery until BP is <140/70 mm Hg
  5. Cardiac consultation
Back to Top | Article Outline

APPENDIX 2: GENERALIZED LINEAR MIXED MODEL

A mixed linear model consists of 2 distinct sets of variables: fixed and random. Programs, training year, and days are fixed effects between subjects within program whereas scenarios and other survey design effects are considered to be fixed effects within subjects. The random portion of the model is the error terms for the between- and within-subjects effects. Normally, the error terms would be subject nested within the program × year interaction for the between-subjects factors and scenario × subject within program × year for the within-subjects factors; however, the program × year interaction had empty cells and could not be fit. Therefore, the error terms used in this analysis were subject within program for between-subjects factors and scenario × subject within program for within-subjects factors. A block-diagonal covariance matrix with twenty-four 6 × 6 blocks was used for this analysis: 1 block per program to represent the respondents clustered with each of the 24 programs and the 6 × 6 unstructured covariance matrix to accommodate the clustering of the 6 scenarios with a subject.

The dependent variable in the analysis was the SCORE for each scenario, which was coded as 0 for incorrect or 1 for correct. Thus, the dependent variable has a binary distribution that requires a logistic regression to be performed by the generalized linear mixed model. The independent variables of interest in the model were YEAR of training and SCENARIO. Because we wanted to interpret the data by year for each scenario, we also included a first-order interaction of YEAR × SCENARIO. Covariates including PROGRAM and survey design effects were included in the model to remove their effects so that we could generalize the findings over these artifacts. Covariates included PROGRAM and the design effects of scenario VARIANT, presentation ORDER, response TIME, number of TRIES, and DAYS from the beginning of the study. First-order interactions between survey design effects and scenario were included in the initial model and eliminated in subsequent runs if they were not significant at the 0.05 level to simplify the model. The final model evaluated was:

CV

CV

where a variable within parentheses indicates a nested variable: VARIANT(SCENARIO) is read VARIANT nested within SCENARIO. Note that error term for the between subject factors is ID(PROGRAM), and for the within subject factors, it is SCENARIO × ID(PROGRAM). The error terms are not included in the model but are generated as random effects in the random portion of the mixed linear model.

To generalize the percent correct and 95% confidence intervals over program and design effects, it is necessary to adjust the percent by averaging over the effects of program and survey design. The steps involved in calculating the percent correct and 95% confidence intervals are outlined below followed by an example to illustrate the procedure.

Back to Top | Article Outline

APPENDIX 3: CALCULATION OF PERCENT CORRECT

Step 1

Form a linear combination of the regression parameters, including the effects of interest and averaging over the effects to be removed. The equation in matrix terms is Lβ where L is a matrix of coefficients and β is the vector of estimated regression parameter estimates. The SE of this linear combination is computed in matrix terms by taking the square root of the diagonal values of Lcov(β)L′ where L is the matrix of coefficients used to calculate the linear combination and cov(β) is the covariance matrix of the estimated regression parameters. In the instance of a logistic regression, which is used in this study, the first step computes the log-odds and its SE: ln[p/(1 − p)] ± SE, where p is the proportion of successes (correct responses) and SE is the standard error of the log-odds.

Back to Top | Article Outline

Step 2

Exponentiate the log-odds to obtain the odds and compute the 95% confidence bounds of the odds: odds = exp{ln[p/(1 − p)]} = p/(1 − p) and the 95% confidence bounds are CI = exp{ln[p/(1 − p) ± tα/2,df(SE)]}.

Back to Top | Article Outline

Step 3

Algebraically manipulate the odds and 95% confidence intervals to obtain the percent correct and its 95% confidence interval: percent correct = 100[odds/(1 + odds)] and 95% CI = 100[oddsCI/(1 + odds CI)].

Back to Top | Article Outline

Example

The calculations for the percent correct and 95% confidence intervals presented in Table 2 involve a 24 × 138 L matrix, a 138 row β vector, and a 138 × 138 cov(β) matrix. This is too much to present as an example, so we illustrate the calculations with a smaller model of 2 programs, 2 training years, 2 scenarios, and 2 scenario variants. Note that the data in the tables are from SAS output, and illustrated calculations below the tables in steps 2 and 3 may result in slightly different answers because of rounding error.

Table

Table

Table

Table

Table

Table

Step 1 The Lβ vector is:

Step 1 The Lβ vector is:

The square root of the diagonal values of lcov(β)L′ matrix is:

The square root of the diagonal values of lcov(β)L′ matrix is:

Step 2 The odds and 95% confidence bounds are:

Step 2 The odds and 95% confidence bounds are:

Step 3 The percent and 95% confidence bounds are:

Step 3 The percent and 95% confidence bounds are:

Back to Top | Article Outline

Appendix 4: Collaborators

Site Coordinators (Conduct of Study)

Shawn Beaman MD, (UPMC), Sergio Bergese, MD (Ohio State), Ben Boedeker, MD (Nebraska), Matthew Caldwell, MD (Michigan), Stevin Dubin, MD (Medical College of Georgia), Sunil Eappen, MD (Mass Eye and Ear Infirmary), Angela Edwards, MD (Wake Forest University), Jesse Ehrenfeld, MD, MPH (Massachusetts General Hospital), David Feinstein, MD (Beth Israel-Boston), Robert Gaiser, MD (University of Pennsylvania), Jeffrey Green, MD (Virginia Commonwealth University), Amir Jaffer, MD (University of Miami), Praveen Kalra, MD (University of Oklahoma), Suzanne Karan, MD (University of Rochester School of Medicine and Dentistry), Paul Kranner, MD (University of Wisconsin), Gary Loyd, MD (University of Louisville), Ronald Olson, MD (Duke University), Michael Pilla, MD (Vanderbilt University), Deborah Richman, MBChB, FFA(SA) (Stony Brook), James Rowbottom, MD (University Hospitals—Cleveland), Kip Robinson, MD (University of Tennessee), Gail Van Norman, MD (University of Washington), Marsha Wakefield, MD (University of Alabama-Birmingham).

Supplemental Digital Content

Back to Top | Article Outline
© 2011 International Anesthesia Research Society