Implementation of an Alternative Training Method for Cardiac Blood Collection in Mice

Abstract

Training personnel to work with animals presents a variety of challenges, both logistically and with regard to animal welfare. These issues make training an ideal opportunity to evaluate practices and to implement the 3R principles (refinement, replacement, and reduction). Cardiac blood collection from mice is a procedure that can compromise the 3Rs by requiring repeated practice and animal euthanasia. The development of a non-animal training model would promote the 3R principles. Our goals for the development of a new training model for cardiac blood collection from mice were to reduce the number of mice needed to achieve competency, improve our culture of care, and refine the training approach by improving competency. The training model was developed using commonly available materials. The total cost of the model was less than $15 USD per model. Two training curricula were conducted concurrently over a 5-mo period: 1) a curriculum in which trainees used the model before progressing to live mice and 2) the traditional curriculum, which used euthanized mice throughout. The measured variables included the total number of mice used, proportions of trainees who reached competency, the time needed to reach competency, method comprehension, quality of skill performance, trainer and trainee feedback, and training costs. The alternative group used at least 10 fewer mice per technician as compared with the traditionally trained group. The alternative group had a higher competency rate, with 82% (9 of 11 trainees) reaching competency compared with 60% (3 of 5 trainees) in the traditional group. Skill comprehension and quality were superior in the alternative group, as evidenced by fewer gross lesions at necropsy. Overall, personnel in the alternative group provided positive feedback with regard to the use of fewer mice, acquisition of both skill and confidence, and benefits for compassion fatigue. The use of this model is now our standard approach for training personnel in cardiac blood collection in mice. Our results demonstrate that the use of models in training curricula can enhance skill development and reduce the use of mice.

Introduction

The goal of training programs used in animal research is to teach skills and enhance competency. However, personnel training is continuously evolving. Training programs allow new employees to learn the skills they need to be successful in their roles and for existing employees to develop new skills and maintain competency over time. At many institutions, particularly in the United States, animals may be used to train research and staff personnel. These animals may be maintained in designated training colonies that are used for the sole purpose of training employees. They can be unneeded animals from purchases or in-house breeding programs, retired breeders, animals used in nonterminal research, or animals purchased specifically for use in training. Some of the risks and challenges to animal welfare that are associated with using animals for training purposes include lack of experience or skill of the learner, overuse of training animals, lack of consistent endpoints, need for space and resources to maintain the animals at the facility, and the development of compassion fatigue or job dissatisfaction due to pain, distress and euthanasia that animals experience when used for training.

The 3Rs are an important ethical framework for decision-making in animal-based research.²⁴ Alternatives, which are defined as any change from standard procedures, can be used for training, thereby replacing the use of animals, reducing the number of animals used, or refining techniques to minimize animal pain and distress.^22,24 Alternative training methods can teach skills such as handling, restraint, blood collection, catheterization, anesthesia, and injections.^14,26 The European Union directive specifies that animals can only be used during training after a trainee demonstrates a specified level of competency, and some animals, such as primates and dogs, can never be used solely for technical skill acquisition.⁶ During the initial stages of training, personnel mainly learn technical skills, such as how to hold and use a syringe. Learning these basic skills with alternative nonanimal training methods allows learners to practice and develop muscle memory at their own pace; this can improve competency before introducing the additional complications of using animals.^4,18,31 To reach competency, personnel often must practice on multiple animals. Thus, introducing alternative nonanimal methods to the initial stages of learning could significantly reduce the number of animals needed for training. Many learners prefer or have a more positive attitude toward learning on alternative models rather than on animals.^5,16

Several alternative nonanimal teaching methods exist for skill training, including artificial models, mannequins, mechanical simulators, computer and virtual reality simulations, videos, self-experimentation, observational studies, in vitro cell lines and organotypic cultures, ethically sourced cadavers, and supervised practice.³¹ Models are good starting points for skill training; however, for skills involving rodents, only 6 models are commercially available for rats and only one for mice, and the skills that can be using these models are limited.^11,12 Some do-it-yourself models can be built by institutions at low cost based on published recipes and methodologies,^27,28 or institutions can develop their own models based on their needs. The fidelity of a model refers to the level of realism.¹⁷ Some low-fidelity models feel less real than the actual procedure but can still be useful in skill transfer.¹⁷ High-fidelity models are more realistic but are typically costly.¹⁷ The fidelity of a model is not linked to learning success; therefore, low-cost do-it-yourself models can also advance the 3Rs without significant financial investment.^9,14,17,29 Many veterinary, medical, and science education programs use models for skill learning. Previous reports indicate that models are as or more effective for skill training and are associated with fewer errors,^{2,3,7,15,18,21,25,30,31} suggesting that biomedical research facilities should adopt more alternative methods into their training programs.

The goal of this project was to develop and validate a non-animal model for training in cardiac blood collection in mice. The aims were to assess the efficacy of the novel alternative with regard to skill acquisition for new technical personnel compared with traditional live-animal training methods and to compare total animal numbers, cost, and personnel feedback between the 2 training methods. We hypothesized that the alternative model would reduce the number of animals needed for training, improve competency, skill comprehension and quality after training, produce positive personnel feedback, and reduce the cost of training.

Materials and Methods

The study protocol, 999-953-06, was approved by the Charles River Mattawan Institutional Animal Care and Use Committee.

Approval for employee participation and data use was gained via a consent statement in the feedback surveys and employee contractual agreements. The feedback obtained from participants was collected through training programs and was related to the model itself and, therefore, is exempt from research ethics board approval.

Animals.

Approximate equal numbers of male and female CD1 mice (n = 489, 8- to 9-wks-old at study start) were obtained from Charles River Laboratories (Wilmington, MA) and housed at the Charles River Mattawan facility, an AAALAC-accredited site. Mice weighed 25 to 35 g at arrival and were acclimated to the facility for 5 days before use. Males and females were initially housed in pairs by sex, but male mice that showed aggression were separated and housed individually. Mice were housed in Allentown polycarbonate cages (polycarbonate 63, Allentown, NJ) on aspen chip bedding (Nepco, Warrensburg, NY) and received Safe Harbor Mouse Retreats (Bio-Serv, Flemington, NJ) per standard operating procedures of the facility. The mice had unlimited access to Block Lab Diet (Certified Rodent Diet #5002, PMI Nutrition International, St. Louis, MO) and water via an automatic water system. Mice were housed under 12-h cycle of alternating light and dark cycles in rooms, from 0600 to 1800 for light and 1800 to 0600 for dark, and were maintained at 68 to 79 °F and 30 to 70% relative humidity. Randomization procedures were not used to select specific mice for use in training.

Alternative model design and construction.

Our motivation to develop an alternative training model arose based on the creation of the Cardiac Balloon, one of the translational training tools developed by the Center for Animal Resources and Education (CARE) team at Cornell University.²⁷ CARE team’s original training tool was modified by Charles River-Mattawan employees to further meet Charles River’s specific needs, which included an introductory model that displayed internal anatomy and landmark identification and an advanced model that more accurately represents a live animal and provides a more realistic feel that improved training and incorporated animal restraint.

Items necessary for the fabrication of the model were readily available and included a 3-in. hair clip​, a golf ball-sized amount of playdough, a natural latex water balloon​, facial tissue​, a food storage bag or piece of plastic, a 10-mL BD syringe, an Instech (Instech Laboratories, Plymouth Meeting, PA) plastic feeding tube (13 gauge × 150 mm), a 5-in. stuffed animal for the model “shell,” theatrical quality fake blood​ (manufactured by Kangaroo Manufacturing and purchased on Amazon), and 1 × 4-in. adhesive-backed fabric hook-and-loop fasteners material (Figure 1)​. The total cost for the creation of the model was approximately $15 (created in 2021). The introductory model is comprised of the hair clip with half of the tines removed (to simulate the thoracic and abdominal region), the water balloon filled with approximately 2 mL of fake blood and water nestled in a rolled facial tissue and placed within the tines of the hair clip, playdough placed over the region without the tines, and a cut-to-size piece of plastic that covered the entire model to provide cleanliness (Figure 2). This introductory model permitted trainees to identify critical anatomy and gain comfort with landmarks, equipment, and techniques necessary to complete the skill. The advanced model was created by simply inserting the introductory model inside a small stuffed animal shell and closing the back of the shell with no-sew fabric hook-and-loop fastener material (Figure 3). The aim of an advanced model was to provide a more realistic skill while also incorporating the element of animal restraint.

Figure 1.

Materials used in the creation of the mouse cardiac training model: a 3-in. hair clip, a golf ball-sized amount of playdough, water balloon, facial tissue, food storage bag or piece of plastic, syringe and catheter, 5-in. stuffed animal for model ‘shell,’ fake blood, and no-sew fabric hook-and-loop fasteners material.

Figure 2.

Introductory level mouse cardiac training model used to learn anatomy, landmarks, approach, instrument handling, and overall technique associated with the skill.

Figure 3.

Advanced mouse cardiac training model used to expand previous knowledge and add animal restraint training and a more realistic level of complexity to the training.

Design of the training curriculum.

The key learning objectives for trainees for cardiac blood collection from mice included:

• 1.Positioning the animal for a ventral approach to the heart
• 2.Identifying anatomic landmarks and orienting the needle relative to the landmarks
• 3.Stabilizing the mouse’s body appropriately to minimize the degree to which the heart shifts as the needle advances toward it
• 4.Repositioning the needle, without removing it entirely from the thoracic cavity, in the event the first attempt is unsuccessful
• 5.Feeling the sensation of the needle penetrating the “heart”
• 6.Learning the depth to which the needle should be inserted for sample collection
• 7.Applying appropriate back pressure to the syringe plunger to facilitate blood collection without collapsing the heart
• 8.Stabilizing the needle and syringe while drawing back on the plunger for blood collection

Two individuals who were familiar and experienced with the procedure performed the training for this study. To maintain consistency in the training message and class conduct, a trainer’s guide was created to describe the overall flow of the class, the necessary equipment for each training curriculum and corresponding day of training, the message to be shared, the key learning objectives, and competency standards. The training session style was predetermined but was not revealed to trainees. Trainees were enrolled using our standard training request process and were scheduled based on trainee schedules and availability. The employment duration of the trainees ranged from approximately 6 mo to 5 y. None of the trainees enrolled in this study were previously competent in the skill of cardiac blood collection but may have had blood collection experience by other routes or in other species.

Two training curricula, the alternative model group (total participants = 11) and the traditional group (total participants = 5), were conducted monthly at concurrent times to allow for close comparison. Both groups received up to 5 days of training, with a maximum of one 3-hour training session per day. The alternative model group used the model on training day 1, followed by same training procedures used for the traditional group on days 2 to 5 (Table 1). Day 1 for the alternative group started with a video demonstration of the skill on an actual mouse; an overview of the necessary equipment, anatomy, and landmarks; a trainer demonstration of the skill on the model; and then unlimited practice (up to the 3-h class duration) on the models. Day 1 for the traditional group consisted of the same discussions and used up to 10 mice per trainee in the practice session. The competency criteria included demonstrating the ability to draw 0.7 to 1.0 mL of blood on 10 mice, immediately after euthanasia, within a 20-minute time period. Each trainee was permitted to have an assistant who could euthanize the next mouse (via CO₂ inhalation) while they were collecting a sample. The majority of trainees were competent in performing euthanasia, but were trained in this skill as necessary. Once a trainee achieved competency, no further practice was required. Participants who did not reach competency by day 5 were required to retake the training. Two participants (one from each curriculum) did not achieve competence and had to retake the training. The results of the second training sessions of these participants are not included in the data.

Table 1.

Training curricula for the alternative model and traditionally trained groups outlined by day

Training group	Training day
	Day 1	Day 2	Day 3	Day 4	Day 5*
Alternative (n = 11)	Employ non-animal model for introduction to skill Unlimited number of collection attempts per trainee within the 3-h class time limit 0 mice used per trainee	Use euthanized mice to practice technique 10 mice used per trainee	Practice day using euthanized mice Time constraint of 6 mice in 12 min 10 mice used per trainee (4 additional mice without time constraint)	Competency attempt using euthanized mice Up to 2 competency attempts per trainee 10 mice used for each competency attempt	Competency attempt using euthanized mice Up to 2 attempts per trainee 10 mice used for each competency attempt
Traditional (n = 5)	Use euthanized mice for introduction to skill 10 mice used per trainee

Training occurred monthly at concurring times, with each group receiving up to 5 d of training, with 3-h training sessions per day. An equal number of both males and females were used each day to address size and sex differences.

^*Only occurred if proficiency was not reached on day 4.

Before and after completion of training, a brief feedback survey was emailed to trainees to gather their impressions of the alternative training models, comfort with the skill, and compassion stress elicited during the training. Trainers also received a survey after completion of the study to assess the same elements and to also obtain their perspectives on skill acquisition and understanding related to trainees and each training curriculum. All electronic responses from both groups were anonymized and analyzed dependent on role.

Blood collection and clinical pathology.

Blood was collected by cardiac puncture as described after CO₂ euthanasia and euthanasia death confirmation. Fasting was not necessary before blood collection. Collections were performed using a 23-gauge × 3/4-in. needle and 1-mL syringe. Approximately 0.7- to 1.0-mL samples were collected from each mouse into Greiner Bio-One MiniCollect 0.5-mL K2EDTA tubes (Monroe, NC). Samples were inspected visually for clot formation and were stored at ambient temperatures until sample submission. Remaining blood from each sample was processed to plasma by centrifuging at 3,000 × g for 10 min and then visually inspected for hemolysis. The visual charts used in the inspection were generated from a chemistry analyzer for automated hemolytic indices (Beckman Coulter AU5800, Indianapolis, IN). Grading was based on a photometric test for the semiquantitative assessment of lipemia, icterus, and hemolysis with scores ranging from 0 to 4 based on the following values: < 50 mg/dL Hgb, score of zero; 50 to 99 mg/dL Hgb, score of one; 100 to 199 mg/dL Hgb, score of two; 200 to 299 mg/dL Hgb, score of three; and > 300 mg/dL Hgb, score of four. The remaining plasma and other blood components were discarded after completion of these evaluations.

Postmortem evaluations.

Focused gross necropsies were performed on at least the first 5 mice per sex from each trainee’s initial proficiency attempt (n = 122; alternative group = 87, traditional group = 35) by evaluators who were blind to the group’s training curriculum. Evaluations included the chest cavity, mediastinum, pleura, heart, and lungs of each mouse submitted. The mice were examined for signs of hemorrhage or other findings potentially associated with blood collection. If hemorrhage was seen in the thoracic cavity, on the heart, or in the remnant of the pericardial sac, then semiquantitative scoring was used. For the thoracic scores, 0 = no hemorrhage, 1 = hemorrhage up to 0.2 mL, and 2 = hemorrhage > 0.2 mL. For the heart and pericardium scores, 0 = no hemorrhage, 1 = hemorrhage on 0 to 50% of the circumference of the heart/pericardial sac, and 2 = hemorrhage on > 50% circumference of the heart/pericardial sac. Thoracic cavity and cardiac scores were conducted independently and not compared.

Statistical analysis.

Statistical analyses were completed using R Studio (2020; Vienna, Austria. https://R-project.org/). Data were assessed for normality using visual inspection of the quantile-quantile plot. Data were not normally distributed, and due to the small sample size, the nonparametric test Mann-Whitney U was used to compare the total number of mice used and the competency rates and attempts between training groups (traditional group compared with alternative group). The Cochran-Mantel-Haenszel test was used to determine the association between the training method and observed postmortem lesions in mice, with sex included as a factor.

Descriptive data are presented as raw means, standard deviations, and incidence rates. Statistical comparisons with a P value less than or equal 0.05 are considered statistically significant. Hematologic findings were not analyzed statistically due to the low number of samples.

Results

Number of mice used.

Over the entire course of training, fewer mice were used in the alternative group than in the traditional group (W = 0.5, P = 0.002). The average number of mice used per trainee in the traditional group was 46 (min = 40, max = 49). For the alternative group, the average number of mice used per trainee was 33 (min = 30, max = 40). At least 10 fewer mice were used per trainee in the alternative model training curriculum because no live mice were used on the first day. Animal use was also lower in the alternative group due to the faster achievement of competency, but this reduction was somewhat variable (see below).

Time to reach competency.

For the traditional group, 60% (3 out of 5) of trainees passed the competency evaluations before the training period ended. The average number of attempts at competency (if gained) was 2.6. For the alternative group, 82% (9 out of 11) of trainees passed the competency evaluation after an average of 1.8 attempts. These differences were not statistically different (passed competency: P = 0.407; attempts: P = 0.226).

Sample quality and tissue damage.

Hematology.

Blood samples were evaluated for hemolysis and clot formation. The samples were generally of high quality, with no notable differences between training methods. Both types of training resulted in minimal hemolysis and clot formation, with only 2 clotted samples (both from the alternative group), 7 hemolyzed samples (five from the alternative group and two from the traditional group), and two unidentified hemolyzed samples. The severity of hemolysis was not significantly different between training types. None of the hemolysis scores were higher than a grade of 1 or 2. In our facility, a hemolysis score of three is viewed as an analytical concern and requires drawing another sample.

Pathology.

Mice of either sex had significantly fewer postmortem lesions of the liver (P = 0.009) and pericardium (P = 0.025) when bled by trainees in the alternative group as compared with the traditional group; however, no group differences were found with regard to the heart. Table 2 lists the lesions and percentage of these findings seen with each training curriculum.

Table 2.

Percentage of lesions noted during gross necropsy, arranged into the associated training curriculum (traditional group compared with alternative group)

Groups	Pathologic lesion (% of mice evaluated)
	Red cardiac fluid	Red thoracic fluid	Tear(s) in the pericardium	Tear(s) in the diaphragm	Tear in a liver lobe	Hemorrhage or blood clot associated with the liver
Traditional (35 mice)	Small amount: 38%	Small amount: 5%	Focal tears: 14%	Focal tears: 14%	11%	3%
	Large amount: 8%	Large amount: 3%	Multiple tears: 4%	Multiple tears: 3%
Alternative (89 mice)	Small amount: 26%	Small amount: 6%	Focal tears: 6%	Focal tears: 11%	1%	1%
	Large amount: 2%	Large amount: 0%	Multiple tears: 2%	Multiple tears: 1%

Only a subset of animals from each group were submitted for gross pathologic evaluations. The number of animals submitted correlates to the number of trainees within each group and the number of trainees who attempted competency within each group.

Training method feedback.

Feedback received in the surveys indicated that trainers and trainees saw value in using the alternative model and supported its use in the skill training. The trainees (regardless of training method) listed the benefits of alternative models as reduced animal use, improved understanding of the procedure, increased opportunity to practice, and improved animal welfare. Responses indicated that more technicians in the alternative model group than the traditional group felt confident performing the skill. At study completion, both trainers and trainees were positive about the alternative training model and suggested that it become the standard for this training.

Cost savings.

The mouse cardiac bleed training model resulted in a cost saving of approximately $45 USD/trainee and as much as $360 USD/training session in animal costs for the average 8-trainee session. Each model cost $15 USD to create and can be reused multiple times. The cost of the model was accounted for in the cost savings calculations. If direct labor and supply costs associated with animal maintenance are added, the estimated cost savings would be as much as $735/training session. At our facility, training for this skill occurs approximately 9 times per year, and thus, the projected annual savings is as much as $6,615 USD.

Discussion

Our goals with regard to testing our alternative approach to training personnel to perform cardiac blood collection in mice were to assess its effectiveness in helping personnel to achieve technical competency, reduce the numbers of mice needed for training, and improve our culture of care. The results demonstrate the success of the alternative approach at achieving these goals. Personnel trained with traditional methods had a competency rate of 60% and personnel trained with the alternative method had an 82% competency rate, with fewer attempts to gain competency and fewer lesions and injuries observed postmortem in the alternative group. The alternative group also returned positive feedback about the model, including the suggestion that the alternative model be adopted as the standard training practice at the facility. Other positive outcomes of the study include the use of fewer mice and a reduction in the overall training costs.

Our results are similar to those of other studies examining the effectiveness of alternative training methods.^1,8,31 Those studies also reported benefits such as the use of fewer animals for training, limited availability of models, additional opportunity for practice, better understanding of the procedure, improved proficiency, reduced costs, and better animal welfare^1,8,31 Cadavers have often been used for training in animal science and medical programs, but the availability, storage, and costs of cadavers are prohibitive,² and real animals are still needed to provide cadavers, making them less effective for reducing overall animal use. Artificial models are available commercially for rats and mice^11,12 and can be made using do-it-yourself guides.^27,28 These models can aid in training of handling, restraint, needle and syringe handling, oral gavage, injections, blood collection, euthanasia, biosecurity practices, and surgical techniques.^10-12,27,28 In our study, participants suggested that the artificial model be permanently incorporated into the training program. Feedback also showed support for developing additional alternative approaches for training in other skills.

The use of our artificial mouse for training personnel to perform cardiac blood collection successfully addressed the 3Rs of replacing animals, reducing the total number of animals used during training, and refining the technique to minimize harm.^19,24 Each person trained with the alternative model used at least 10 fewer mice in the training curriculum. Use of the non-animal model for training improved trainee performance, as evidenced by fewer postmortem lesions on the liver and pericardium and fewer cardiac tissue injuries. A review of the effectiveness of alternative teaching methods found that 90% of studies had reported equivalent or superior knowledge and skill acquisition as compared with traditional training using live animals.³¹ Higher confidence and fewer errors have also been reported in studies evaluating learning in veterinary schools when comparing traditional live animal teaching methods and alternatives for learning gynecologic skills,¹⁸ ovariohysterectomy,³⁰ surgery,²⁹ and common procedures such as blood collection.⁴

The use of adjunctive nonanimal models may result in less compassion stress and an overall improved culture of care for technicians due to reductions in animal use and the need for euthanasia.²⁰ Culture of care programs should aim for continuous improvement in animal care and welfare, the wellbeing of personnel, scientific quality, and openness and transparency.²³ Based on feedback provided after training, our model allowed trainees to achieve better understanding and skill performance, and they felt more comfortable and confident performing the skill. This conclusion is consistent with previous studies showing better learning and more positive attitudes when learners can work with non-animal models during the initial stages of training.³¹ Comparing the number of animals used in training with and without use of models also provides openness and transparency when communicating with stakeholders and the public.

Trainers may also benefit from the use of non-animal teaching methods. Trainers are more likely to report compassion fatigue, likely due to the secondary trauma of training personnel to perform procedures that may be painful or harmful, observing trainees make mistakes that may harm the animals, and witnessing emotional responses from personnel as a result of procedures such as euthanasia.^13,20 Performing euthanasia, particularly physical euthanasia (that is, cervical dislocation), can contribute to compassion fatigue.^13,20 We also found cost savings in association with our implementation of our approach. Those savings could potentially be used to fund further culture of care and 3R initiatives.

In conclusion, the use of low-cost, do-it-yourself non-animal models can be effective for addressing the 3Rs in research training programs and should be incorporated more readily across institutions. The model we developed for cardiac blood collection from mice eliminated the use of mice on the initial day of training, reduced the number of mice used for training, and refined the overall technique. As evidenced by the positive feedback from personnel, introducing non-animal models during training could mitigate compassion stress and improve job satisfaction due to increased skill competency, thus improving welfare for both animals and personnel. Use of our model also reduced the overall costs of the training program for this skill.