Evans, Leigh V. MD; Dodge, Kelly L. MD; Shah, Tanya D. MD; Kaplan, Lewis J. MD; Siegel, Mark D. MD; Moore, Christopher L. MD; Hamann, Cara J. MPH; Lin, Zhenqiu PhD; D'Onofrio, Gail MD, MS
Physicians often use central venous catheters (CVCs) in the care of critically ill patients in the emergency department (ED) and intensive care unit. In the United States, physicians insert more than five million CVCs every year,1 and investigators report that mechanical complications from these CVC insertions occur at a rate as high as 21%.2–6 Mechanical complications positively correlate with morbidity, mortality, and cost increases due to prolonged hospitalization.7
Traditional education in invasive procedures relies on the apprenticeship model of learning at the bedside without deliberate step-by-step standardization or skills assessment. Simulation recently emerged as a tool that may be used to reduce medical errors and is particularly attractive for teaching invasive procedures, such as CVC insertion, that require eye–hand coordination and ambidextrous maneuvers.8 Simulation allows physicians-in-training to repeatedly practice a procedure prior to performing it on an actual patient. Improved performance in CVC insertion after simulation training may lead to a decrease in insertion attempts in the clinical setting, which is important because higher numbers of venous cannulation attempts correlate with higher rates of mechanical complications.6,9–11 Whereas simulation training has potential benefits, well-designed studies to understand the effect of simulation on resident training will allow medical educators and physicians to use the technology more intelligently to improve provider performance, to reduce errors, and, ultimately, to improve patient safety.
This current prospective, randomized, single-blind control study assessed the effect of simulation training in ultrasound (US)-guided CVC insertion on resident performance on actual patients. We hypothesized that residents who completed competency-based simulation training on a partial task trainer would (1) perform CVC insertions in the clinical setting with improved first venous cannulation and CVC insertion success rates and (2) insert CVCs with fewer technical errors and mechanical complications.
We enrolled and observed postgraduate year 1 (PGY1) and postgraduate year 2 (PGY2) resident physicians over a 21-month period from January 2007 to September 2008 at a 944-bed tertiary care teaching hospital. Eligible resident physicians included those who rotated through the ED, medical intensive care unit (MICU), and/or surgical intensive care unit (SICU) during the study period and included residents from five different specialties (emergency medicine, internal medicine, general surgery, anesthesia, and obstetrics–gynecology). The combined Yale University / Yale University Hospital institutional review board granted full approval, and participating residents provided their informed, written consent. We randomly assigned residents to a control or intervention group. Both the control and intervention groups received CVC insertion training according to the traditional teaching hospital practices, that is, the bedside apprenticeship Halstedian model, often described as “see one, do one, teach one.”12 This apprenticeship model consisted of performing supervised CVC insertions on actual patients.
During the course of this study, the control group received no formal simulation training; nor did this group have access to a CVC partial task trainer in a simulation laboratory. Further, the control group did not receive formal didactic training on CVC insertion.
The intervention group received, in addition to traditional bedside teaching, competency-based simulation training in US-guided CVC insertion approximately two weeks prior to beginning clinical rotations in the ED, MICU, or SICU. Participants performed hands-on training on a partial task training simulator until they achieved competency. We defined competency as the successful completion of all criteria listed in Table 1 scored on a binary scale (yes/no).
At our institution, attending physicians, fellows, and senior residents primarily supervise junior residents and interns during CVC insertion; residents insert CVCs independently after inserting five CVCs under direct observation. Supervisors provide feedback at their own discretion. All residents record their procedures for graduation and board certification requirements.
On the clinical wards, from 8 am to midnight daily, blinded independent raters (IRs) observed CVCs inserted by residents in both the control and intervention groups in the ED, MICU, and SICU. IRs collected and recorded data on the resident performing the CVC insertion, the patient, and the procedure itself (see “Data collection” below for more details). Primary outcome measures were the number of venous cannulation attempts and the success rate at CVC insertion by the participating resident. Secondary outcome measures were the rate of technical errors and complication rates as determined by IR procedural checklist evaluation.
Enrollment/eligibility and randomization
The research coordinator (C.H.) approached eligible residents, explained the study protocol and study expectations, and consented and enrolled study participants. Residents received no incentive to participate in the study. We excluded more experienced residents (postgraduate year 3 or higher) because they often functioned as supervisors during CVC insertion at the hospital sites included in the study. The only other exclusion criterion was unwillingness to participate in the project. Residents' decisions not to participate—or their CVC insertion performance evaluations if they did participate—did not impact their advancement in the residency program.
The study biostatistician (Z.L.) generated the block randomization of the participant population. Two weeks prior to beginning a new clinical rotation, we randomly allocated residents rotating through the ED, SICU, and MICU into the control or intervention group stratified according to postgraduate year and residency training program, so that approximately equal numbers of residents from both PGY1 and PGY2 and from all five specialties were in the intervention and the control groups. Each participant completed the baseline assessment. The research coordinator (C.H.) scheduled participants in the intervention group for simulation training. Intervention group participants returned to the simulation laboratory as often as necessary until completing competency-based training as described below. Some participants rotated through the ED, MICU, and/or SICU more than once during the study period. Participants remained in their initially assigned group so that no crossover occurred between the control and intervention groups. We offered residents assigned to the control group the structured, competency-based simulation training at the conclusion of the study.
Baseline assessment of residents in both the control and intervention groups consisted of demographic data including age, gender, residency program affiliation, and postgraduate year. Both the control and intervention groups completed a questionnaire to assess their previous CVC insertion experience, prior US training, comfort with CVC insertion, and their perceived importance of gaining CVC insertion skills training.
The structured, hands-on, competency-based simulation training included the following elements:
1. a didactic lecture (slide presentation) focused on describing the anatomy, indications, contraindications, and potential complications of CVC insertion at three different sites (internal jugular vein, subclavian vein, and femoral vein), a review of the physics of US, US instrumentation and controls, imaging and use of US, and a review of the steps required for mastery of venous cannulation and CVC insertion including the fundamental landmarks and use of the Seldinger technique;
2. a videotape of US-guided venous cannulation (Advanced Medical Technologies, LLC, Kirkland, Washington);
3. a videotape of landmark CVC insertion using the Seldinger technique9;
4. a videotape of US-guided CVC insertion;
5. hands-on performance in US-guided needle cannulation of a simulated vein on a simulation vascular access US trainer (Blue Phantom anthropomorphic simulation US training model, Advanced Medical Technologies, LLC) observed by a faculty evaluator; and
6. hands-on performance of CVC insertion using the Seldinger technique on a central venous access simulation model (Blue Phantom central venous access head neck and upper torso simulation model, Advanced Medical Technologies, LLC) observed by a faculty evaluator.
The hands-on portion of the structured training program with a simulator used a kinesiology approach13 with deconstruction of skills into readily understood steps. Participants repeated simulated CVC insertion until performed correctly. Faculty served as instructors and evaluators during the training session. All faculty instructors were study coinvestigators with extensive US experience. All faculty instructors had inserted more than 50 CVCs on actual patients and had been observed to perform a CVC insertion on a partial task trainer with no technical errors. Competency criteria are described in Table 1. During US-guided simulation training, we used a Sonosite 180 Plus US machine with a 38-mm, high-frequency linear array vascular probe (Sonosite, Inc., Bothell, Washington) without the use of a needle guide. In the MICU and SICU, participants used a Sonosite 180 US machine. In the ED, they used a B-K Medical Hawk 2,102 XDI US Scanner (B-K Medical, Herlev, Denmark) or Philips Envisor US (Philips Medical, Andover, Massachusetts). In both the simulation and hospital, participants used Multi Lumen CVC Kits with Blue FlexTip, CA-15703 Arrow-Howes (Arrow International, Reading, Pennsylvania).
The research coordinator (C.H.) recruited and hired nonphysician IRs to collect data at all sites during two 8-hour shifts, seven days per week, excluding the hours from midnight to 8 am. Undergraduate students constituted 48% of the IRs, and the remaining IRs were nursing students, medical students, graduate students, and master's of public health students. The IRs had minimal exposure to invasive procedures prior to the study.14 We trained the IRs in data collection prior to the start date of the study. IR training included the same didactic lecture and videotapes of US-guided venous cannulation and CVC insertion techniques used for resident training.9 The IRs also viewed a videotape of a simulated CVC insertion under US guidance with an explanation of the correct technique for each step of the procedure. To increase interrater reliability, IRs completed a data collection sheet of 5 of 10 videotaped simulated scenarios of CVC insertion with choreographed technical errors. Each training simulation videotape incorporated specific technical errors to be detected by the rater. To be eligible for hire, IRs had to accurately assess the time of the procedure to within one minute, validate the procedural checkpoints to within 95% accuracy, and detect technical errors and complications within a 3% margin of error.14
IRs measured resident procedural skill performance in the hospital setting for both control and intervention groups. When a decision to insert a CVC occurred, residents contacted the IRs via beeper. The IRs were blinded to the randomization of the residents. Resident participation in the proposed study impacted neither the clinical decision making regarding CVC insertion nor the supervision level during the procedure in the hospital setting. Residents performed CVC insertions under the current clinical guidelines of the ED, MICU, or SICU (clinical guidelines at all three sites included full barrier sterile precautions and a “time out” prior to beginning the procedure).
The IRs collected demographic data (postgraduate year, specialty, etc.) on the resident performing the CVC insertion. In addition, because specific patient characteristics may lead to greater technical difficulty in CVC insertion and/or predispose the patient to the development of complications, the IR collected patient demographics at the time of the CVC insertion. The patient data the IRs collected included indications for CVC insertion, patient risk factors, and the calculation of the patient's Acute Physiology and Chronic Health Evaluation (APACHE) II scores. The APACHE II score reflects a commonly employed severity of illness scoring system.15 A higher score is associated with a higher mortality rate.16 If information was missing for any patient, a member of the research team subsequently reviewed the patient's medical records to obtain these data.
IRs completed a 50-point procedural checklist for all observed CVC insertions.14 Data points for performance included the site chosen for the CVC insertion (internal jugular, subclavian, or femoral vein), the type of catheter (triple lumen, antibiotic-coated triple lumen, or cordis), the use of US guidance (yes/no), the number of venous cannulation attempts, and CVC insertion success. IRs also recorded any technical errors performed and/or any mechanical complications occurring during the hospital procedure.
Technical errors included the following: break in sterile technique, improper positioning of the patient, removal of the hand from the wire during any portion of the procedure, inability to thread the wire, inability to flush all ports, improper suturing of the catheter, and improper sterile dressing. We considered some complications to be major (i.e., pneumothorax, hemothorax, hemomediastinum, vessel laceration, transient dysrhythmia, air embolus) and some to be minor (i.e., superficial hematoma, transient catheter malposition, and arterial puncture without significant hemorrhage). We reviewed all postprocedure chest X-rays to assess for pneumothorax, hemothorax, pneumomediastinum or hemomediastinum, and catheter position. We reviewed the catheter tip position for correct location at the superior vena cava–right atrial junction in each patient. We determined catheter-related bloodstream infections by reviewing all positive blood cultures in the study hospital and then cross-referencing the positive cultures with the presence of a CVC by chart review. The Department of Epidemiology uses the Center for Disease Control's National Health Safety Network definition of catheter-related bloodstream infection to define infectious complications.17 We also reviewed charts to search for the presence of air embolism.
If the resident could not successfully insert the CVC, the senior resident or attending clinician assumed responsibility. The supervising clinician determined the number of CVC insertion attempts permitted by the resident. The IR documented whether the procedure was successful or aborted. As the units of analysis, we used the number of cannulation attempts and whether a particular CVC insertion was successful (yes/no). Successful cannulation referred to the insertion of the cannulation needle into the vein with return of venous blood, whereas cannulation attempt referred to each needle pass. We defined successful CVC insertion by the operator as venous placement of the catheter with blood return in all three catheter ports; if a second resident participating in the study attempted the procedure after a failed insertion attempt, we recorded this as an independent procedure whereby the same data points were collected.
This was a cluster randomized trial with the resident as the unit of randomization. We applied the intervention at the resident level and measured outcomes at the patient level; thus, we nested patients within clusters of residents. We provided descriptive statistics of the characteristics of the residents in the two groups. We also described and compared the characteristics of the patients in the intervention and control groups. We reported categorical data as frequencies (percentages) and compared differences between the two groups using χ2 or the Fisher exact test. We reported continuous data as the mean and standard deviation or the median and interquartile range, and we tested differences between the intervention and control groups using the two-sample t test or the Wilcoxon rank sum test. We considered a P value of <.05 as significant.
As the primary outcome measure of the study, we used a patient-level dichotomous variable, indicating the success or failure of the CVC insertion. To assess the effect of the intervention on the failure rate, we used a two-level hierarchical logistic regression model: the patient (Level 1) nested within the residents (Level 2).18 We considered patient characteristics as Level 1 covariates and resident group assignments as Level 2 covariates. We analyzed secondary outcomes by multilevel logistic regression to assess the effects of the intervention on the rate of technical errors and the rate of complications, and we used multilevel linear regression to assess the effects of the intervention on the number of attempts.19 We conducted all analyses using SAS 9.1.3. (SAS Institute, Cary, North Carolina).
Between January 7, 2007, and September 30, 2008, we recruited 199 PGY1 and PGY2 residents to participate in the study. We excluded 11 residents (6%) because they declined to participate. Excluded residents cited time constraints affiliated with the intervention group. Of the 188 subjects randomized, we allocated 93 to the intervention group and 95 to the control group. We excluded three residents from analysis: Two withdrew from the study, citing training time constraints, and one participant withdrew from the residency training program. No differences existed in baseline characteristics between members of the intervention and control groups (Table 2).
The median number of prior CVC insertions among all participating residents was five (range 0–40). Of all the residents, 37 (20%) had no prior CVC experience, and 28 (15%) had inserted 11 or more CVCs. In the intervention group, 19 residents (21%) completed a CVC insertion on a partial task trainer with no technical errors on their first attempt, an additional 39 (43%) in two attempts, and an additional 15 (17%) in three attempts; the number of attempts to complete a CVC insertion with no technical errors on a partial task trainer ranged from 1 to 8. The median number of cannulation attempts of a simulated vein was 7 in residents having inserted no prior CVCs compared with 5.3 in residents who had previously inserted 25 or more prior CVCs. Prior CVC experience translated into fewer cannulation attempts of a simulated vein (P = .02). We found no association between increased experience and decreased attempts at CVC insertion on the partial task trainer (P = .15).
In-hospital CVC insertions
IRs observed 495 CVC insertion procedures performed by 115 residents (mean: 4.3 ± 4.0; median: 2.5 procedures per resident) participating in the study. Of the 185 residents participating in the study, 70 did not perform any CVC insertions during the study period because of rotating for only one month at one of the three study sites (ED, MICU, or SICU). We report no significant differences in the demographics of patients who received CVC insertions from residents in either the intervention or control group (Table 3). The mean number of cannulation attempts was 2.4 in the intervention group versus 2.9 in the control group (P = .05; 95% CI: −0.9 to 0.0). The success rate for first cannulation was 51% in the intervention group versus 37% in the control group (P = .03). The success rate for CVC insertion was 78% for the intervention versus 67% for the control group (P = .02) (Table 4). Hierarchical logistic regressions for the primary outcomes demonstrated that simulation training in CVC insertion was independently and significantly associated both with success at first cannulation with an odds ratio of 1.7 (95% CI: 1.1–2.8) and with successful CVC insertion with an odds ratio of 1.7 (95% CI: 1.1–2.8). These results were independent of the following three variables: (1) the use of US, (2) patient comorbidities, and (3) resident specialty.
Table 4 compares the incidence of technical errors between the intervention and control groups. We observed no difference in US prep, compliance with sterile precautions, ordering of insertion steps (threading guidewire to insertion of CVC), or break in sterile technique between the two groups. A similar total number of mechanical complications occurred between the intervention and control groups.
Discussion and Conclusions
This study is the first large prospective, randomized control study demonstrating that procedural skills for CVC insertion learned through a structured simulation training program transfer to improved clinical practice. We have demonstrated that rates of first cannulation and CVC insertion success are associated with simulation training, which is important as others have reported that fewer cannulation attempts and improved rates of successful insertion reduce mechanical complications.10 This innovative study design was novel in its (1) use of IRs blinded to participant randomization rather than potentially biased physician-supervisor raters, (2) training of residents in the intervention group to a predefined level of hands-on competency, (3) use of a 50-point procedural checklist to collect objective data rather than a global rating scale to assess improvement, and (4) inclusion of patient demographic data on all observed CVC insertions performed in the hospital.
CVC insertions are associated with significant risks to patients. Complications are associated with failed CVC insertion. In one study, 28% of patients experienced complications after a failed catheter insertion, as compared with only 7.2% experiencing complications after a successful catheterization.10 In this same study, the number of venous cannulation attempts was strongly associated with the rates of failure and complications, rising from 4.3% with one attempt to 24% with more than two attempts. In addition to complications, multiple cannulation attempts may create patient anxiety and discomfort.20 In our study, simulation training was significantly associated with reduced CVC insertion failures and a decreased number of cannulation attempts independent of the use of US, patient comorbidities, or resident specialty.
Multiple previous studies have demonstrated improved performance on a simulator after simulation training.21–23 Evidence showing improved performance in the clinical setting is now beginning to emerge.24–26 In the field of surgery, simulator training has been shown to improve performance in the operating room27–31 as well as to effectively increase proficiency of surgical skills in the simulation laboratory.27,31–33 Britt and colleagues34 evaluated CVC insertion performance on actual patients after simulation training. Simulation-trained residents self-reported a higher level of comfort and ability, and the investigators observed a significant decrease in mechanical complications, but Britt and colleagues considered failed CVC insertion to be a complication (we considered failed CVC insertion separately). Further, an unblinded trauma fellow or critical care surgeon, rather than an independent, blinded rater, measured performance in Britt and colleagues' study. In an observational cohort study, Barsuk and colleagues,25 like Britt and colleagues, found that a simulation-based mastery learning program increased residents' skills in simulated CVC insertion and decreased complications in actual patient care.
For our study, we report lower complication rates (i.e., 1.4%) than those reported elsewhere in the literature (i.e., 5%–26%),9 so we are unable to assess whether our approach—simulation training with US—resulted in fewer complications.
CVC simulation training
Previous studies demonstrated that it is possible to train residents in CVC insertion in a simulated setting by allowing repetition, instantaneous correction of errors, and the ability to err.35,36 Our training and the use of a checklist enabled residents to become proficient in the mechanics of cannulation of a simulated vein under US guidance before they encountered patients in the clinical setting. We found that, despite attempts at inserting a CVC prior to the study, only 20% of the residents in the intervention group could successfully complete the CVC insertion on the partial task trainer on their first attempt, demonstrating that the observed residents in this cohort, trained under the Halstedian model, did not receive adequate training and did not master the skills necessary to perform CVC insertions on patients. Simulation training was associated with an increased rate of successful CVC insertions. Previous studies demonstrated that 50 successful CVC cannulations are typically required to reduce the rate of complications by half.4 Simulated CVC insertions on partial task trainers can provide residents with that experience.
US-guided CVC insertion
The traditional approach for insertion of CVC relies on the use of external landmarks. Inherent drawbacks to this approach include the inability to predict aberrancy of the anatomy, the possibility of an unusually small vessel, and the occasional occlusion of the target vessel secondary to thrombosis. The multiple attempts often required may lead to inadvertent arterial puncture, injury to neighboring structures, and perivascular hematoma formation.37 We chose to train residents in US-guided CVC insertion based both on studies showing that residents trained in US guidance via simulators are more likely to use US in actual practice and on studies showing the superiority of the use of real-time US guidance during CVC placement, particularly for the internal jugular site, by demonstrating a lower technical failure rate, decreased rate of complications, faster access, and lower number of attempts.20,38,39 Utilization of the hierarchical logistic regression model controlled for US usage showed an independent significant, positive effect of the simulation training.38,39
The incidence of CVC infections in the intensive care unit is reported to be between 5% and 26%.9 Predictors for infection include inaccurate or poor site choice, an increased number of days with the catheter, catheter type, poor catheter maintenance, and multiple accessed lumens. Barsuk and colleagues40 reported fewer catheter-related bloodstream infections in the patients of residents trained through simulation to insert CVCs. We are unable to make any conclusions with regard to the impact of simulation on catheter-related infections because of several confounding factors: the inconsistent type of catheter used (antibiotic coated or not), the fact that only some (not all) catheters were changed over a wire, and the variable duration of time before the catheter was removed. In addition, we report a lower (1.4%) incidence of catheter-related infections, as defined stringently by the National Healthcare Safety Network,17 than that previously reported in the literature, although recent studies such as the Keystone study in Michigan41 showed the ability to virtually eliminate these infections with good technique. Possibly, the observation of residents—in both the intervention and control groups—by the IRs yielded more careful practice. However, our low overall infection rate in comparison with other institutions may also reflect our hospital's efforts both to enforce standard techniques with high adherence to maximal barrier precautions and to reduce the duration of empiric antibiotic use, another factor related to catheter-related infections.42
The difficulty of blinding the participants to the objectives of our study was a limitation. When they enrolled, residents received information that if they were randomized to the intervention group, they would have to commit to a minimum of three hours for simulation training. We did not expose residents to either the procedural checklist or the primary objectives of the study. Additionally, residents in the intervention group received didactic instruction on anatomy, and we are unable to assess whether their higher CVC insertion success rate could be partially attributed to this educational intervention. Another limitation of the study is the known difficulty in accurately determining the total number of CVC insertions performed during the study period,43 although to minimize this difficulty, we attached numbered packets with the data collection forms on all CVC kits in the ED, MICU, and SICU. There may have been a sampling bias because we included only CVCs inserted from 8 am to midnight in the study, and we could not determine the population characteristics of patients for whom CVC insertions occurred overnight. According to billing data and hospital bloodstream infection data, approximately 1,500 CVC insertions occurred over the 21-month period. We were not able to assess which CVCs represented missed procedures potentially eligible for inclusion in the study. We initially anticipated that the control group, lacking confidence and comfort in their ability to perform CVC insertions, would be less likely to call IRs; however, both groups inserted similar numbers of CVCs, making systemic bias less likely. We are aware that both our complication and infection rates are low. Nonetheless, we believe these results are reliable because we used trained IRs, reviewed all chest X-rays for mechanical complications, and checked for all possible infections. This low prevalence rate of complications in our population precluded our ability to detect statistically significant differences in the complication rates in our two study arms. To determine whether such differences actually exist, either investigating a larger population or performing the larger study in an environment with a higher background complication rate would be necessary.
In conclusion, a structured, competency-based simulation training protocol was associated with improved CVC insertion performance. Our study demonstrates that simulation training is an effective medical education tool, increasing physician skill and competency while protecting patient safety.
The authors wish to thank and acknowledge the following: Ingrid Tuckler for her administrative assistance, and all of the independent raters for participating in this study.
Financial support for this project was provided by a patient safety grant from the Agency for Healthcare Research and Quality (Grant # U18 HS16725).
This manuscript includes research data that meet Academic Medicine's considerations for ethical approval and institutional review board requirements (Yale University Human Investigation Committee approval #0605001388). This study is registered with ClinicalTrials.gov as “Simulation Training for Ultrasound Guided Central Venous Catheter Insertion,” Identifier NCT00919308.
This project was presented as an oral presentation at the plenary session of the Society for Academic Emergency Medicine in New Orleans, Louisiana, in May 2009.
1Raad I. Intravascular-catheter-related infections. Lancet. 1998;351:893–898.
2Defalque R. Percutaneous catheterization of the internal jugular vein. Anesth Analg. 1974;53:116–121.
3Bernard RW, Stahl W. Subclavian vein catheterizations: A prospective study I. Non-infectious complications. Ann Surg. 1971;173:184–190.
4Sznajder JI, Zveibil FR, Bitterman H, Weiner P, Bursztein S. Central venous catheterization. Failure and complication rates by three percutaneous approaches. Arch Intern Med. 1986;146:259–261.
5Steele R, Irvin C. Central line mechanical complication rate in emergency medicine patients. Acad Emerg Med. 2001;8:204–207.
6McGee WT, Ackerman BL, Rouben LR, Prasad VM, Bandi B, Mallory D. Accurate placement of central venous catheters: A prospective, randomized multicenter trial. Crit Care Med. 1993;21:1118–1123.
7Girard TD, Schectman JM. Ultrasound guidance during central venous catheterization: A survey of use by house staff physicians. J Crit Care. 2005;20:224–229.
8Patow CA. Medical simulation makes medical education better and safer. Health Manag Technol. 2005;26:40, 39.
9McGee DC, Gould M. Preventing complications of central venous catheterization. N Engl J Med. 2003;348:1123–1133.
10Mansfield PF, Hohn DC, Fornage BD, Gregurich MA, Ota DM. Complications and failures of subclavian vein catheterization. N Engl J Med. 1994;331:1735–1738.
11Eisen LA, Marasimhan M, Berger JS, Mayo PH, Rosen MJ, Schneider R. Mechanical complications of central venous catheters. J Intensive Care Med. 2006;21:40–46.
12Gorman PJ, Meier AH, Rawn C, Krummel T. The future of medical education is no longer blood and guts, it is bits and bytes. Am J Surg. 2000;180:353–356.
13Martin M, Scalabrini B, Berard D. Training for competence in critical surgical skills: The impact of kinesiology-based instruction on outcome. Paper presented at: The Society of University Surgeons Sixty-First Annual Meeting. Abstract no. 40.
14Evans LV, Morse JL, Hamann CJ, Osborne M, Lin Z, D'Onofrio G. The development of an independent rater system to assess residents' competence in invasive procedures. Acad Med. 2009;84:1135–1143.
15Knaus WA, Draper EA, Wagner DP, Zimmerman J. APACHE II: A severity of disease classification system. Crit Care Med. 1985;13:818–829.
16Knaus WA, Harrell FE Jr, LaBrecque JF, et al. Use of predicted risk of mortality to evaluate the efficacy of anticytokine therapy in sepsis. The rhIL-1ra Phase III Sepsis Syndrome Study Group. Crit Care Med. 1996;24:46–56.
17Horan TC, Anderus M, Dudeck M. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control. 2008;36:309–332.
18Omar RZ, Thompson SG. Analysis of a cluster randomized trial with binary outcome data using a multilevel model. Stat Med. 2000;19:2675–2688.
19Bryk AS, Raudenbush SW. Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, Calif: Sage Publications; 1992.
21Crofts JF, Attilakos G, Read M, Sibanda T, Draycott TJ. Shoulder dystocia training using a new birth training mannequin. Br J Obstet Gynaecol. 2005;112:997–999.
22Crofts JF, Bartlett C, Ellis D, Hunt LP, Fox R, Draycott TJ. Training for shoulder dystocia: A trial of simulation using low-fidelity and high-fidelity mannequins. Obstet Gynecol. 2006;108:1477–1485.
23Deering S, Poggi S, Macedonia C, Gherman R, Satin A. Improving resident competency in the management of shoulder dystocia with simulation training. Obstet Gynecol. 2004;103:1224–1228.
24Draycott TJ, Crofts JF, Ash JP, et al. Improving neonatal outcome through practical shoulder dystocia training. Obstet Gynecol. 2008;112:14–20.
25Barsuk JH, McGaghie WC, Cohen ER, O'Leary KJ, Wayne DB. Simulation-based mastery learning reduces complications during central venous catheter insertion in a medical intensive care unit. Crit Care Med. 2009;37:2697–2701.
26Wayne DB, Didwania A, Feinglass J, Fudala MJ, Barsuk JH, McGaghie W. Simulation-based education improves quality of care during cardiac arrest team responses at an academic teaching hospital: A case–control study. Chest. 2008;133:56–61.
27Seymour NE, Gallagher AG, Roman SA, et al. Virtual reality training improves operating room performance: Results of a randomized, double-blinded study. Ann Surg. 2002;236:458–464.
28Hyltander A, Liljegren E, Rhodin PH, Lönroth H. The transfer of basic skills learned in a laparoscopic simulator in the operating room. Surg Endosc. 2002;16:1324–1328.
29Seymour N. VR to OR: A review of evidence that virtual reality simulation improves operating room performance. World J Surg. 2008;32:182–188.
30Grantcharov TP, Kristiansen VB, Bendix J, Bardram L, Rosenberg J, Funch-Jensen P. Randomized clinical trial of virtual reality simulation for laparascopic skills training. Br J Surg. 2004;91:146–150.
31Andreatta PB, Woodrum DT, Birkmeyer JD, et al. Laparoscopic skills are improved with LapMentor training: Results of a randomized, double-blinded study. Ann Surg. 2006;243:854–863.
32Korndorffer JR Jr, Dunne JB, Sierra R, Touchard CL, Markert RJ, Scott DJ. Simulator training for laparoscopic suturing using performance goals translates to the operating room. J Am Coll Surg. 2005;201:23–29.
33Korndorffer JR Jr, Hayes DJ, Dunne JB, et al. Development and transferability of a cost-effective laparoscopic camera navigation simulator. Surg Endosc. 2005;19:161–167.
34Britt RC, Novosel TJ, Britt LD, Sullivan M. The impact of central line simulation before the ICU experience. Am J Surg. 2009;197:533–536.
35Huang GC, Newman LR, Schwartzstein RM, et al. Procedural competence in internal medicine residents: Validity of a central venous catheter insertion assessment instrument. Acad Med. 2009;84:1127–1134.
36Wayne DB, Barsuk JH, O'Leary KJ, Fudala MJ, McGaghie W. Mastery learning of thoracentesis skills by internal medicine residents using simulation technology and deliberate practice. J Hosp Med. 2008;3:48–54.
37Habib FA, McKenney M. Surgeon performed ultrasound in the ICU setting. Surg Clin N Am. 2004;84:1151–1179.
38Hind D, Calvert N, McWilliams R. Ultrasonic locating devices for central venous cannulation: Meta-analysis. BMJ. 2003;327:361.
39Randolph AG, Cook DJ, Gonzales CA, Pribble CG. Ultrasound guidance for placement of central venous catheters: A meta-analysis of the literature. Crit Care Med. 1996;24:2053–2058.
40Barsuk JH, Cohen ER, Feinglass J, McGaghie WC, Wayne D. Use of simulation-based education to reduce catheter-related bloodstream infections. Arch Intern Med. 2009;169:1420–1423.
41Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med. 2006;355:2725–2732.
42Namias N, Samiian L, Nino D, et al. Incidence and susceptibility of pathogenic bacteria vary between intensive care units within a single hospital: Implications for empiric antibiotic strategies. J Trauma. 2000;49:638–646.
43Wright SB, Huskins WC, Dokholyan RS, Goldmann DA, Platt R. Administrative databases provide inaccurate data for surveillance of long-term central venous catheter-associated infections. Infect Control Hosp Epidemiol. 2003;24:946–949.