Presented at the Fourth Mediterranean Emergency Medicine Congress, Sorrento, Italy, September, 2007.
Assessment of a New Model for Femoral Ultrasound-guided Central Venous Access Procedural Training: A Pilot Study
Article first published online: 9 DEC 2009
© 2009 by the Society for Academic Emergency Medicine
Academic Emergency Medicine
Volume 17, Issue 1, pages 88–92, January 2010
How to Cite
Wadman, M. C., Lomneth, C. S., Hoffman, L. H., Zeger, W. G., Lander, L. and Walker, R. A. (2010), Assessment of a New Model for Femoral Ultrasound-guided Central Venous Access Procedural Training: A Pilot Study. Academic Emergency Medicine, 17: 88–92. doi: 10.1111/j.1553-2712.2009.00626.x
- Issue published online: 4 JAN 2010
- Article first published online: 9 DEC 2009
- Received August 17, 2009; revision received October 1, 2009; accepted October 2, 2009.
- procedural training;
- central venous access;
- residency education;
Objectives: Repetitive practice with feedback in residency training is essential in the development of procedural competency. Lightly embalmed cadaver laboratories provide excellent simulation models for a variety of procedures, but to the best of our knowledge, none describe a central venous access model that includes the key psychomotor feedback elements for the procedure, namely intravascular contents that allow for determination of correct needle position by either ultrasonographic imaging and/or aspiration or vascular contents.
Methods: A cadaver was lightly embalmed using a technique that preserves tissue texture and elasticity. We then performed popliteal fossa dissections exposing the popliteal artery and vein. Vessels were ligated distally, and 14-gauge catheters were introduced into the lumen of each artery and vein. The popliteal artery and vein were then infused with 200 mL of icterine/gel and 200 mL of methylene blue/gel, respectively. Physician evaluators then performed ultrasound (US)-guided femoral central venous line placements and rated the key psychomotor elements on a five-point Likert scale.
Results: The physician evaluators reported a median of 10.5 years of clinical emergency medicine (EM) experience with an interquartile range (IQR) of 16 and a median of 10 central lines placed annually (IQR = 10). Physician evaluators rated the key psychomotor elements of the simulated procedure as follows: ultrasonographic image of vascular elements, 4 (IQR = 0); needle penetration of skin, 4.5 (IQR = 1); needle penetration of vein, 5 (IQR = 1); US image of needle penetrating vein, 4 (IQR = 2); aspiration of vein contents, 3 (IQR = 2); passage of dilator into vein, 4 (IQR = 2); insertion of central venous catheter, 5 (IQR = 1); US image of catheter insertion into vein, 5 (IQR = 1); and overall psychomotor feedback of the simulated procedure compared to the evaluators’ actual patient experience, 4 (IQR = 1).
Conclusions: For the key psychomotor elements of central venous access, the lightly embalmed cadaver with intravascular water-soluble gel infusion provided a procedural model that closely simulated clinicians’ experience with patients.
ACADEMIC EMERGENCY MEDICINE 2010; 17:88–92 © 2009 by the Society for Academic Emergency Medicine
Deliberate practice, or repetition of a skill with expert observation, debriefing, and feedback, is essential to the development of competence in a variety of psychomotor activities and is well studied in instrumental music and sports.1 Postgraduate medical education, however, often neglects the importance of scheduled, repetitive practice sessions when teaching procedural skills. High-fidelity simulation has been suggested as a means to provide opportunities for deliberate practice, but the variable anatomy encountered in clinical practice is not reflected in commercially available high-fidelity simulation models, possibly resulting in suboptimal skills acquisition.2 Human cadavers, fresh or embalmed in a manner that preserves the tissue texture and elasticity encountered in clinical practice, provide a procedural training model that will allow for deliberate practice, with the tissue haptics and anatomic variations necessary to achieve an acceptable level of competence.
While fresh or lightly embalmed cadavers provide excellent models for a variety of procedures, to the best of our knowledge, none describe a central venous access model that includes the key psychomotor feedback elements for the procedure: intravascular contents that allow for determination of correct needle position by either ultrasound (US) imaging and/or aspiration of vascular contents.3,4 In addition to providing repetitive practice, the model must also be durable, allowing for multiple punctures of the target vein without loss of intravascular contents. We hypothesized that a lightly embalmed cadaver with an intravascular gel infusion would provide a US-guided central venous access training model that would approximate actual patient experience for key psychomotor elements.
This was a cross-sectional validation study using cadavers. The study protocol received approval from the institutional review board of the University of Nebraska.
Study Setting and Population
We conducted the study in the Advanced Anatomy Laboratory of the Department of Genetics, Cell Biology, and Anatomy of a state-supported college of medicine. Emergency medicine (EM) faculty members were recruited to serve as independent evaluators for the study. Physician evaluators recorded their past clinical experience in terms of years in practice postresidency and estimated the average number of central venous catheters placed per year.
A single cadaver was lightly embalmed using a modification of a previously described technique that preserves tissue texture and elasticity. The preparation of lightly embalmed cadavers for anatomical study was first described by Anderson.5 The donor used in our study was similarly prepared, using a glutaraldehyde-based rather than formaldehyde-based chemical. The carotid artery was first injected with approximately 8 L of Co-Inject Beta Factor (Champion Millenium, The Champion Co., Springfield, OH) diluted 1:16 with tap water. In contrast to the method described by Anderson, the internal jugular vein of the cadaver was opened prior to injection of Beta Factor to allow free drainage. Next, the carotid artery was injected with approximately 8 L of Arterial 24 Alpha Factor (Champion Millenium, The Champion Co.), diluted 1:10 to 1:16 with tap water. Both fluids were injected at a pressure of 500 mm Hg and a flow rate of 400–500 mL/min. The cadaver was then stored supine in a plastic bag at 4°C. Using the lightly embalmed cadaver, we performed popliteal fossa dissections to expose the popliteal artery and vein. Vessels were then ligated distally with 0 silk sutures, and 14-gauge catheters were introduced proximally into the lumen of each artery and vein and secured with 0 silk sutures. The popliteal artery and vein were then infused with 200 mL of icterine-colored water-based gel (Surgilube, E. Fougera and Co., Melville, NY) and 200 mL of methylene blue–colored gel, respectively. The intravascular gel mixture allows for multiple punctures of the vessel without significant loss of intravascular contents.6,7
Evaluators then performed US-assisted femoral central venous line placements on the gel-infused lightly embalmed cadaver by Seldinger technique. The evaluators were then asked to compare the psychomotor elements of the procedure to their actual patient experience with the procedure and to rate the key psychomotor elements on a five-point Likert scale. Specifically, evaluators were asked to rate their level of agreement with the following statement: “The model accurately simulated the following elements of central venous catheter insertion based on my past clinical experience with the procedure” (1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree).
Descriptive analyses were performed including medians and interquartile ranges (IQRs) for years of clinical experience and procedural experience of each evaluator, as well as for each psychomotor element. Decay of the model over the course of the study or evidence for linear trend was assessed using linear regression. Analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).
Ten physician evaluators participated in the study. The evaluators reported a median of 10.5 years of clinical experience, with an IQR of 16 years, and a median procedural experience of 10 central lines placed annually, with an IQR of 10.
The study model compared favorably to the evaluators’ procedural experience on actual patients for seven of the eight psychomotor elements measured. Only aspiration of intravenous contents failed to rate above “neutral.” For overall psychomotor feedback for the procedure, the evaluators agreed that the model closely simulated central venous access performed on actual patients, with a median rating of 4 (IQR = 1; see Table 1).
|Psychomotor Element||Median Likert Score||IQR|
|US image of vascular elements||4||0|
|Needle penetration of skin||4.5||1|
|Needle penetration of vein||5||1|
|US image of needle penetration of vein||4||2|
|Aspiration of vein contents||3||2|
|Passage of dilator into vein||4||2|
|Insertion of central venous catheter||5||1|
|US image of catheter insertion||5||1|
No significant model decay over repeated usage by 10 evaluators was found as demonstrated by linear regression. Only aspiration approached statistical significance at p-values of 0.052. p-values for other psychomotor elements ranged from 0.22 to 0.999. The linear trend in overall score was also not significant (p = 0.32; Table 2).
|Psychomotor Element||Linear Trend p-value|
|US image of vascular elements||0.84|
|Needle penetration of skin||0.999|
|Needle penetration of vein||0.22|
|US image of needle penetration of vein||0.29|
|Aspiration of vein contents||0.052|
|Passage of dilator into vein||0.86|
|Insertion of central venous catheter||0.26|
|US image of catheter insertion||0.78|
Past procedural training paradigms have followed the Halsteadian apprenticeship model, with a significant amount of time spent observing, practicing, and teaching the skill on actual patients, under the supervision of an expert mentor.8 This training model resulted in the universal mantra of “see one, do one, teach one” familiar to most physicians engaged in postgraduate medical education. However, this approach neglects the new realities of graduate medical education: limitations in resident duty hours imposed by the Accreditation Council for Graduate Medical Education (ACGME) and increasing concern for patient safety.9,10 More recent reviews of the development of expertise in a variety of domains, including medical specialties, suggest that a key component of psychomotor skills training is the actual number and nature of the repetitions, in association with feedback, which is defined as “deliberate practice.”1,2,11,12 Unfortunately, bedside clinical practice often lacks opportunities for immediate feedback, and repetition of a certain skill is usually not feasible, making deliberate practice impossible. Strategies to overcome these limitations are required to ensure optimal procedural skills training.13 To prepare for and supplement the clinical practice of procedures by trainees, educators must provide not only the learning opportunities conducive to deliberate practice, but also an effective training model. The key aspects that correlate with effective simulator-based learning include provision of feedback, repetitive practice, range of difficulty level, and capture of clinical variation.2,14,15 While our study asked participants to rate the cadaver-based model against their actual experiences in performing the procedure on patients, the cadaver-based model correlates well with these important elements of high-fidelity simulation and may serve as a superior model for some elements.
Providing feedback refers not only to the cognitive feedback during and after the simulation, but also to the psychomotor feedback provided to the learner during the procedure. Key procedural elements, such as the US image of the needle traversing subcutaneous tissue and entering the lumen of the vein, coupled with the tactile feel of the needle penetrating skin, subcutaneous tissue, and wall of the vein, followed by the appearance of venous blood in the syringe, make up the psychomotor feedback necessary to effectively learn central venous access. To provide this psychomotor feedback, procedural simulation models must include the key psychomotor tasks that make up a procedure and also provide visual and tactile feedback closely approximating the actual clinical practice. While fresh/frozen and lightly embalmed cadavers both provide favorable tissue haptics for emergency procedures, such as cricothyrotomy and tube thoracostomy, to the best of our knowledge, no reported model includes the key elements required for successful central venous access training: distinguishable intravenous and intraarterial contents.3,4 Cadaver models without intravenous and intraarterial infusions that distinguish arterial from venous circulation and reliably remain in the intravascular spaces may not provide the visual feedback necessary to confirm correct needle position or incorrect intraarterial needle position prior to insertion of the guide wire, dilator, and catheter. In preliminary investigations utilizing lightly embalmed cadavers, the ability to aspirate embalming fluid from the intravenous, intraarterial, and extravascular spaces could potentially result in misleading feedback and a poor training model.6,7
Our model provided psychomotor feedback closely simulating actual patient experience for six of the seven procedure elements measured. Although the intravascular infusions made up a key component of the model, the only element with a median rating of “3,” or “neutral,” on the five-point Likert scale was “aspiration of vein contents.” The high viscosity of the gel required evaluators to apply increased traction on the syringe plunger, greater than what is required to aspirate blood, most likely contributing to the lower rating (Figure 1). Future investigations will seek to determine a mixture of the water-based gel that will allow for easy aspiration while also resisting leakage from the intravascular space following needle, dilator, and catheter penetration of the vein. Another aspect of the gel infusion potentially affecting the ratings of the model is the presence of air bubbles in the gel, which result in US images that closely simulate venous air embolization rather than normal venous blood (Figure 2). Commercially available models for central venous access training also allow for scheduled, repetitive practice, but the psychomotor feedback of such models may not reflect patient experience. The elasticity and resistance to needle stick of plastic skin and vascular elements may not simulate actual patient experience as closely as a cadaver-based model, but no studies to date have compared the tissue haptics of a cadaver-based model to the commercial models. In contrast, repetitive practice is also possible in a live animal laboratory, with tissue haptics very similar to actual human patient experience. Animal labs unfortunately involve the negative aspects of the sacrifice of a living animal, and it is uncertain whether the learned procedure is readily translatable to human patient experience due to anatomic differences. Our model allows for an anatomically accurate biologic model without the need to sacrifice animals.
Capturing clinical variation and range of difficulty level in a procedural training model is essential to the development of both competence and, eventually, expert-level skills that are adaptable to changing clinical conditions. According to Ericsson,2 achievement of expert-level skill requires an “avoidance of automaticity.” For any psychomotor task, once the learner progresses through the “cognitive” and “associative” phases, his or her performance tends to becomes autonomous, resulting in an arrested stage of development. To achieve an acceptable level of competency in a procedure involving a range of anatomic variation, the learner must counteract automaticity by changing the conditions or altering the playing field, allowing the learner to develop higher levels of control of the skill through remaining in the cognitive and associative phases. For procedural simulation, repetition of a psychomotor task may lead to automaticity if no variation in technique is required by the model.2 Deliberate practice using high-fidelity simulators typically involves repetition of a psychomotor task with no variation in anatomy, potentially resulting in arrested development at less-than-expert skill levels. Furthermore, the lack of anatomic variation may also affect translation of skills learned through deliberate practice on simulators or task trainers to clinical practice. A study by Britt et al.16 suggests that lack of clinical/anatomic variability may result in less-than-optimal skills transfer. Of 15 residents receiving central line instruction utilizing the CentralLineMan (Simulab Corp., Seattle, WA) partial task trainer, only four succeeded in their initial central line placement on an actual patient.16 Our cadaver-based model, on the other hand, captures the anatomic variation inherent to clinical practice, as we typically use three to four cadavers per session. This allows for true deliberate practice and may prevent the development of automaticity, with a range of anatomic variations and difficulty levels for each practice session.
Repetitive practice is a key feature of effective procedural simulation, but is only feasible if the training model is durable. Durability is especially important if multiple trainees are to sequentially participate in the simulated procedure. Preliminary investigations of intravascular infusions demonstrated the propensity for saline and tap water to rapidly diffuse out of the intravascular space prior to performance of the simulated procedure.6,7 Water-based gel not only remained in the intravascular space following the initial infusion, but also resisted leakage from the multiple vessel punctures resulting from the simulations. During our study, the femoral vein sustained 14 documented punctures, with good retention of the infused gel. While a high-viscosity substance ensures intravascular retention, it also limits easy aspiration. The optimal intravascular substance would remain intravascular after multiple-needle, dilator, and catheter penetration of the vein wall, while still allowing for easy aspiration. We plan to address this limitation of the model with future investigations of serial gel:water mixtures to improve intravascular retention and the ease of infusion and eliminate air bubbles in the infused substance.
Further affecting overall durability of the model is the relatively fixed duration of use for each cadaver. Despite other reports of a more limited time frame for use, our method allows for 3 to 4 months of use for each lightly embalmed cadaver. The defined time frame requires advance scheduling of laboratory sessions to ensure availability. Other limitations of the lightly embalmed cadaver with intravascular infusion model include adequate facilities, such as cold storage and a well-ventilated laboratory, and cadaver-related limitations, such as lack of spontaneous circulation to aid in identification of anatomic landmarks (e.g., femoral pulse for localization of the femoral vein), and the previously described air bubbles within the infused gel.
Some general limitations of the lightly embalmed cadaver model include the recurrent cost of cadavers for preparation and storage, as well as the expertise required for cadaver embalming. The laboratory facility where the procedure training sessions occur also requires close proximity to refrigerated storage facilities, controlled entry, and an adequate ventilation system, limiting the ability to perform simulation in situ, such as in the emergency department or other patient care areas in the hospital.
Specific limitations of our study protocol include the use of past clinical performance of central venous access on actual patients for comparison of the model, as this experience is most certainly variable between evaluators and further depends on individual memory rather than a more consistent “criterion standard.” In addition, the study is limited by the relatively small numbers of evaluators, the single anatomic site investigated, and lack of assessment of guide wire passage. Future studies are planned to address the psychomotor elements associated with guide wire passage as well as other anatomic sites. Whether our results for femoral vein access translate into workable models for internal jugular and subclavian central venous access is unknown and warrants future investigation.
Last, learner preferences and personal beliefs may limit the use of the cadaver-based model. Some medical graduates may prefer not to use cadavers to learn procedural skills for personal reasons; however, we found that medical students interviewing for EM residency programs preferred to learn procedures on cadavers to learning on mannequins.17
For most of the key psychomotor elements of central venous access, the lightly embalmed cadaver with intravascular water-soluble gel infusion provided a procedural model that closely simulated actual clinician experience with patients. This cadaver provides an ultrasound-guided central venous access training model that allows for deliberate practice in a setting that preserves patient safety and provides anatomic variability.
The authors thank Paul Becker for his assistance in the preparation of the cadavers.
- 1EricssonKA, CharnessN, HoffmanRR, FeltovichPJ (eds.). Cambridge Handbook of Expertise and Expert Performance. New York, NY: Cambridge University Press, 2006, pp 457–88.
- 4A fresh frozen cadaver procedure laboratory. MedEdPORTAL; 2008. Available at: http://services.aamc.org/30/mededportal/servlet/s/segment/mededportal/?subid=794. Accessed Aug 7, 2009., , , , .
- 6Comparison of intravascular substances in a central venous access simulation model utilizing lightly embalmed cadavers. Council of Residency Directors (CORD) Academic Assembly, Orlando, FL, March 2007., , , , , .
- 7Enhancement of teaching central venous access using viscous lubricant in lightly embalmed cadavers. American College of Clinical Pathology meeting, Chicago, IL, 2006., ,
- 8The fruition of Halsted’s concept of surgical training. Surgery. 1952; 23:518–27..
- 9Accreditation Council for Graduate Medical Education. ACGME Duty Hours Language. Available at: http://www.acgme.org/acWebsite/RRC_110/110_guidelines.asp. Accessed Aug 7, 2009.
- 10KohnLT, CorriganJM, DonaldsonMS (eds.). To Err Is Human: Building a Safer Healthcare System. Washington, DC: National Academies Press, 1999.
- 16Central line simulation: a new training algorithm. Am Surg. 2007; 73:682–3., , .
- 17Self reported emergency medicine residency applicant attitudes towards a procedural cadaver laboratory curriculum. West J Emerg Med. 2008; 9:141–5., , , .