Automated External Defibrillators: Skill Maintenance Using Computer-assisted Learning

709 ACADEMIC EMERGENCY MEDICINE July 1998, Volume 5, Number 7

Automated External Defibrillators: Skill Maintenance Using Computer-assisted Learning JOHN M. JERIN, MS, BRUCE A. ANSELL, EMT-P, MARY PAT LARSEN, MS,

RICHARD 0. CUMMINS, MD

Abstract. Objective: To determine whether computer-assisted learning (CAL) can maintain the automated external defibrillation (AED) skills of emergency medical technicians (EMTs). Methods: The authors conducted a l-year prospective comparison of a n AED-skill training software program, running on desktop computers, with traditional instructor-led training. The subjects were experienced EMT-Ds (EMT-defibrillation), already trained in automated defibrillation ( n = 105) employed as full-time profes- sional EMT-Dkrefighters. Two of the 3 groups (groups A and C) in the study were assigned to use the CAL program for 6 months. The third group (group B) remained on the normal, instructor-led training regimen. Pre- and poststudy skill levels were measured using a skills performance test. Results: A secular trend of improved mean treatment scores was

observed across all 3 groups [mean rise of 0.49 point (p = 0.01), repeated-measures analysis of variance]. There were no differences between training groups in the increase in performance scores (p = 0.3). The 1- time cost of supplying t h e CAL program to the 105 EMT-Ds was $1,575, significantly less than the $3,240- per-year cost associated with instructor-led training. Conclusions: The authors observed satisfactory AED skill maintenance for experienced EMT-Ds using CAL to replace 2 of 4 quarterly instructor-led skills reviews. CAL has cost and convenience advantages over instructor-based skill maintenance and is an accept- able alternative. Key words: automated external defibrillation; computer-assisted learning; computer- based instruction; computer-based training; emergency medical technicians; education. ACADEMIC EMERGENCY MEDICINE 1998; 51709-717

UTOMATED external defibrillators (AEDs) A are a major advance to achieve early defibrillation,' and to strengthen the chain of survival in every community.2 Public safety personnel such as emergency medical technicians (EMTs),~ fire- f i g h t e r ~ , ~ and police p e r ~ o n n e l ~ . ~ can learn to use a n AED with just a few hours of training. The de- vices eliminate the need for training in rhythm rec- ognition and decision making about shock delivery.

From the King County EMS Division, Seattle-King County Department of Public Health (JMJ, BAA, MPL), the Division of Emergency Medicine, University of Washington (ROC), Jerin and Associates (JMJ), and the Center for Evaluation of Emer- gency Medical Services (ROC, MPL), Seattle, WA; and the Bellevue Fire Department, Bellevue, WA (BAA). Received November 3, 1997; revision received December 17, 1997; accepted January 2, 1998; updated January 26, 1998. Presented a t the American Heart Association Public Access De- fibrillation I1 conference, Washington, DC, April 1997. The computer simulation program was developed by John M. Jerin of Jerin and Associates. The Center for Evaluation of Emergency Medical Services receives both categorical and non- categorical grant support from a number of sources, including foundations, several defibrillator manufacturers, and other companies that manufacture products for EMS use. No specific financial support was provided from any defibrillator manufacturer for this program or project. Address for correspondence and reprints: John M. Jerin, MS, Division of Emergency Medical Services, Seattle-King County Department of Public Health, 999 3rd Avenue, Suite 700, Se- attle, WA 98104-4039. Fax: 206-296-4866; e-mail: jjerinaseanet .com

Despite their simplicity and ease of learning, emergency medical services (EMS) managers and direc- tors must not neglect to maintain AED skills and AED field performance.

Automatic defibrillation requires skills in both cognitive and psychomotor domains, including re- sponse to device prompts, recall of rules and protocols, and manipulation of the AED d e ~ i c e . ~ While most of these skills are learned during initial training, without periodic practice, skills may de- grade quickly.

The national guidelines published by the Amer- ican Heart Association (AHA) prescribe regular skill maintenance for AED-equipped agencies." These guidelines recommend refresher training for personnel every 90 days to maintain satisfactory skill levels. Refresher training allows EMT-Ds to practice using field equipment under the guidance of a n instructor. This refresher training, however, can be costly and difficult to coordinate because of instructor and personnel training costs, equipment needs, varied availability of EMT-Ds during working hours, and the necessity to compensate personnel for their training time when training is conducted during off-duty hours.

The dramatic rise i n the number of personal computers used by emergency agencies supports the use of computers for some AED training. Com- puter-assisted learning (CAL) eliminates the prob-

710 AEDCAL Jerin et al. AED COMPUTER-ASSISTED LEARNING

lem of scheduling and conducting training sessions with an instructor. With CAL, EMT-Ds may have more opportunities for high-fidelity practice without increasing the demand on human instructors.

We developed an AED simulation software program to allow EMT-Ds to practice AED treatment protocols on a personal computer. The program provides guided practice without direct supervi- sion by an instructor. The purpose of this 1-year, prospective, controlled project was to determine whether skills review and practice with a CAL AED program are effective in maintaining the skills of experienced EMT-Ds.

METHODS

Setting. The fire department of the city of Belle- vue, WA, was selected as the study fire department. The Bellevue Fire Department is one of 44 fire departments in King County, WA, and serves an urban population of 130,000 with 9 fire stations and 2 medic units. The Bellevue Fire Department was selected because it has a mature EMT-D program (>lo years old) that requires all active duty, firefighter EMTs to be trained in the use of AEDs. More than 70% of the department’s calls are for EMS; the remaining calls are for fire-related inci- dents. During this 1-year project, 105 firefighter/ EMT-Ds, with an average of 5.6 years of EMT-D experience, participated. AEDs were used 54 times in the field during the study. One of the authors (BAA) is the EMT-D program coordinator for the Bellevue Fire Department. He is responsible for the training and skills maintenance of the EMT-D program and was able to closely supervise the project.

Design. All EMT-Ds in the fire department are assigned to 1 of 3 separate 24-hour-duty shifts. This provided 3 convenient study groups (A, B, and C) that did not overlap in training. All groups received a 1-hour initial practical skill review. Each EMT-D was then evaluated by 1 of 4 trained

TABLE 1. Refresher Training Schedule by Method of Training

evaluators on performance and speed according to predetermined evaluation criteria (discussed be- low). The evaluators were all trained and moni- tored for consistency in their testing methods during a training session held before the study began. During this training session, the evaluators were asked to rate videotaped AED performances using a checklist. The group discussed the results after each performance. Evaluations continued until in- terevaluator rating was consistent.

Groups A and C received a brief introduction to the computer program. This training session cov- ered log-on procedures and interaction instruc- tions. For 6 months, group A (continuous-schedule computer group) used the computer program in place of quarterly, instructor-ledhands-on skill review. During this time group A had unlimited access to the computer program. Computer-based performance tests were conducted at 3 and 6 months. Group A was restricted from computer use at 6 months and was evaluated at 9 and 12 months by conventional instructor-led evaluations.

Group B (control group) received instructor-led reviews once every 3 months, which is standard for AED refresher training. Skill evaluations for the control group were held immediately after the reviews in order to document skill level at a consistent point in the 90-day review cycle. This evaluation point was chosen due to the difficulty of assembling the group more than once during the review period. Since the review cycles continued indefinitely, timing of the evaluation point was not critical so long as it was held at a consistent interval. Group B received no information about the computer simulation program.

Group C (alternating-schedule computer group) received alternate computer use and instructor-led reviews every 3 months. All groups were re-evaluated at the end of the 12-month study period by 1 of the same 4 instructors. These instructors did not conduct any of the hands-on refresher training sessions held during the study.

Table 1 shows the schedule for the training of

Group Months 1-3 Months 4-6 Months 7-9 Months 10-12 ~~ ~

Group A (continuous) Computer Computer Instructor Group B (control) Instructor Instructor Instructor Group C (alternating) Computer Instructor Computer

Instructor Instructor Instructor

TABLE 2. Testing Schedule by Method of Testing

Group Initial Test 3-month Test 6-month Test 9-month Test 12-month Test

Group A (continuous) Instructor Computer Computer Instructor Instructor Group B (control) Instructor Instructor Instructor Instructor Instructor Group C (alternating) Instructor Computer Instructor Computer Instructor

711 ACADEMIC EMERGENCY MEDICINE July 1998. Volume 5. Number 7

the 3 groups of EMT-Ds, by months of the study. Groups A and C were given regular instructor-led reviews during the 3-month period in which they were not using the computer program. We believed that there is a need for occasional hands-on practice with the device to help retain familiarity with handling the AED and its components. Therefore, these groups completed 2 instructor-led reviews during the study. This design allowed us to exam- ine the effects that a realistic practice regimen using both computer and instructor-led methods would have over the study period. Table 2 shows the schedule for the testing of the 3 groups.

The computer program was a simulation exercise that provided rehearsal of the steps in the defibrillation process. This exercise was complete when the EMT-D demonstrated the ability to complete 3 computer-generated scenarios without error. The instructor-led review consisted of a small- group skill demonstration with subsequent practice using an AED and a CPIUdefibrillation training mannequin. The instructor-led training sessions ended with a hands-on skill evaluation.

Monitors were trained to use the software program and answer common questions. They moni- tored use of the program and helped familiarize EMT-Ds with the program. The program’s data files were inspected bimonthly to ensure that unauthorized personnel were not using the program and contaminating the results of the study.

As an evaluation of educational methodology, this study was exempt from requirements for in- formed consent according to the criteria set by the University of Washington Human Subjects Review Board.

AED Defibrillation Skills. Operation of the AED requires only 4 steps: power on, attachment of the device to the patient, analysis of the rhythm by the AED, and delivery of the shocks when indicated. The number and sequence of these steps, however, become more complex when integrated into full field resuscitations. With actual resuscitation at- tempts, personnel must add patient assessment, basic life support, and scene control. Table 3 shows the minimum critical actions required to deliver field defibrillation integrated with CPR (steps may vary depending on the model of AED).

In addition, most of these steps contain multi- ple “substeps.” For example, “open the airway” requires knowledge of proper hand positioning and movements; and even “attach defibrillation pads to AED cables” requires opening a sealed foil packet, removing 2 pads, locating the pad terminal posts, and properly snapping the cable connectors to the posts.

All EMT-Ds are required to meet a quarterly (every 3 months) skill maintenance requirement. This consists of a brief review of recent cases and

TABLE 3. Minimum Critical Outcomes to Deliver Field Defibrillation Integrated with CPR

Assess responsiveness Open the airway Check breathing Provide 2 ventilations Check the pulse Begin chest compressions Turn AED power on Attach defibrillation pads to AED cables Attach pads to patient’s chest Verbalize information about the patient and the care provided Clear the patient Press analysis button Clear the patient if shock indicated Begin CPR if no shock indicated Press the shock button if shock indicated 6Observe for proper shock delivery Clear the patient for reanalysis Repeat above sequence, including updated verbal reports

a scenario-based hands-on review of the AED protocols and equipment. These are the instructor-led reviews that CAL replaced for groups A and C. The EMT-Ds are presented with 2 scenarios: a patient with persistent ventricular fibrillation who requires a t least 3 analysis periods and shocks, and a patient in a nonshockable rhythm. These quarterly review sessions require a t least 1 hour of personnel time, and 1 instructor per 6 EMT-Ds. All EMT-Ds are “trained to criteria,” which means they must meet all performance criteria. Two of the 4 quarterly reviews were replaced by CAL for groups A and C.

The authors developed an AED quality assurance skill evaluation form that is used in initial training, during the quarterly skills maintenance sessions, and for postevent review of all clinical uses of the AEDs. The major dimensions that are evaluated are manual skill performance, protocol adherence, and safety. In addition, it is required that the time interval from power on to shock delivery must be more than 60 seconds. The evaluation form assigns numeric treatment and speed scores. Figure 1 reproduces the skill evaluation form used in the King County EMS system. Since the subjects in our study were experienced EMT-Ds whose CPR skills are well-maintained, the skill evaluations did not dissect CPR skill performance in detail.

The CAL Program. One of the authors (JMJ) developed the computer simulation program to run under the Apple Macintosh operating system. The program was programmed with Allegiant Super- Card (Allegiant Technologies, San Diego, CAI au- thoring software. The program incorporated Apple User Interface guidelines to ensure uniformity with other Macintosh software.

After a brief log-on sequence, the program pre-

712 AEDCAL Jerin e t al. AED COMPUTER-ASSISTED LEARNING

sents a graphical scene of a patient, an AED unit, and 12 control buttons (Fig. 2). A message area gives initial dispatch information and performance feedback. The user initiates actions, such as per-

forming CPR or delivering defibrillatory shocks, by selecting 1 of the 12 control buttons with a mouse pointing device. The program evaluates each user action against a set of rules based on the AHA

AED Quality Assurance Report

2. Technician I I . I I I I I Technician's Name

3. Session: __ I . Initial Class 1. Three month CE 3. Six month CE 4. Nine month CE 5 . Twelve month class 6. Actual field event

Perforrnancflrotocol

1. Verified cardiac arrest and begins CPR = .5

5. Proper identification of self, department, and patient = .5

5. Identifies placement of defibrillation electrodes = .5

14.'Defibrillator 1 = LP 250 2 = L P 3 0 0 3 = HS 3000 4 = F M 160 5 = Other

r-7 7. Clears patient prior to every rhythm assessment = 2 U B. Clears patient prior to every shock (MUST PASS) = 3

3 Performs correct sequencing of protocols = 1

10. Checks pulse when no shock advised = .5

I I . Provides updated information oat least once = .5

15. Technician Type 1 = E M T 2 = First Responder 3 = Other

12. Directs resuscitation of patient = .5

13. BLS care in accordance to AHA = I

16. Total Treatment Score (Ten points possible)

PFM - BVM -

Speed

1o:oo :Power on) -- --

17. Time rhythm appears on the screen

30 seconds or less = 4 90 seconds or greater, or 3 1 to 50 seconds = 3 patient connected 51 to 70 seconds = 2 before power on = 0 7 1 to 90 seconds = 1

18. Time to first shock --.--

19. Total Speed Score

-. (4 points possible)

Zonunents:

Fiaure 1. Automated external defibrillator (AED) quality assurance report.

ACADEMIC EMERGENCY MEDICINE July 1998, Volume 5, Number 7 713

Figure 2. The computer simulation program.

AED algorithm and King County protocols. The program increases the user’s score for each correct action. Incorrect actions result in a deduction from the user’s score and an appropriate feedback message.

The program collects data for each user including the total number of scenarios, date and time of use, user errors, initial shock speed, elapsed time, and score for each scenario. Users may view the top 20 practice scores on a display accessed from the main menu. This display encourages competi- tion for the top score and increases interest in the program. The program contains both test and practice modes, which are identical except that scores are confidential in the test mode.

The user can select an advice option from a pull-down menu that provides clues about which action to take next. The AED displayed in the program emulates an existing manufactured AED, which is the model used by the EMT-Ds in the study fire department. The program includes appropriate audio and visual prompts.

The program randomly presents patients who may or may not require defibrillatory shocks. Pa- tient presentations vary in terms of cardiac rhythm, vital signs, symptoms, gender, age, and environmental circumstances. The program spon- taneously generates troubleshooting situations that require the user to solve common problems such as battery failure and electrode malfunction.

The program was placed in 9 fire stations with 1 computer each. Crews typically consist of 3 to 5 persons working 24-hour shifts. EMT-Ds in the study groups used the computer program while on duty between emergency calls and other tasks. The EMT-Ds had unlimited use of the program during

their assigned months. They were required to run at least 3 scenarios during their assigned period and had t o obtain a passing score.

This program was originally developed for the purposes of this study. Subsequent to this research, the program was modified (running under the Microsoft Windows operating system) and has been made available for distribution by a major defibrillator manufacturer. Only after documented success with the program was shown, was the program considered for commercial application. The rights to a modified version of the program are owned by one investigator (JMJ).

Measurements. The purpose of our study was to determine whether computer-assisted skill maintenance and instructor-led skill maintenance methods were equivalent. We looked for a significant effect of training method in either speed or performance of defibrillation protocols. Perfor- mance was measured a t the beginning and end of the study with identical hands-on practical tests with speed and treatment proficiency scores. The treatment score varied between 0 and 10 points; the speed score varied between 0 and 4 points. The higher the score, the better the performance. Ex- amination of score differences for normal distribution determined analysis method.

Data Analusis. Normally distributed differences were tested with repeated-measures analysis of variance (ANOVA)g to determine overall training effect and differences between the 3 training methods, A, B, and C. Non-normally distributed differences were examined with the Wilcoxon signed rank test to determine the overall training effect


12 1

0

0

N i 37 35 u 33 35 Y

A 0 rn pre C

Training Method O P O S i

Figure 3. Box and whisker plot shows median (line in box), interquartile range, 25th and 75th percentiles (tops and bottoms of boxes), lowest expected values for normal distribution (whiskers), and outliers observa- tions > 1.5 standard deviations from the mean (circles). Distributions of treatment group test scores pre- and posttraining are shown. Differences among the 3 groups A, B, and C were not statistically significant.

and the Kruskal-Wallis (xZ)lo to determine the effect of training method. Although change in performance due to overall training was not the end- point of primary interest, without an overall statistically significant pre-post change in performance, statistical tools would be unable to detect differences between the 3 study groups. A study- wide type I error (rejecting the null hypothesis when the null is true) level was set at p = 0.05. The same type I error was used for post-hoc cal- culations of observed power.

RESULTS

One hundred five subjects enrolled in the study were divided into 3 groups: A (37 subjects), B (33 subjects), and C (35 subjects). Final speed scores were missing for 16 subjects: A (8 subjects), B (4 subjects) and C (4 subjects), leaving 89 subjects for speed score analysis. Final treatment scores were missing for 3 subjects: A (2 subjects) and C (1 sub- ject), leaving 102 analyzable subjects.

Treatment scores, analyzed with repeated-measures ANOVA, increased over the study period with a mean overall increase of 0.49 points on a 10-point test (p = 0.01). This modest difference was not considered clinically significant. No effect due to training method (Fig. 3) was observed (p = 0.3). The observed power of the repeated-measures

ANOVA test to detect a difference of 1 test point in mean score was modest (0.281.” Neither EMT years of experience (p = 0.66) nor field use of defibrillators (p = 0.97) affected treatment scores in any measurable way. Table 4 shows the mean treatment scores and the speed scores for the 3 groups.

Speed scores, which varied between 2 and 4 for all subjects, showed little variation. Medians and interquartile ranges” (25th to 75th percentiles) were identical (equal to 3) for all 3 study groups, both before and after training (Fig. 4). Since the distribution of speed score differences was dis- tinctly non-normal, we analyzed these scores with non-parametric methods (Wilcoxon signed rank, Kruskal-Wallis 1-way test, and xz statistic), none of which was significant.

During the period of the study there were 54 cardiac arrests in which the EMT-Ds used the AEDs. These field events were equally distributed among the personnel in the 3 study groups so that “practice” on actual field events was unlikely to af- fect results. Review of performance on these 54 events revealed excellent performance regardless of group. Three subjects of the 105 EMT-Ds who started the study dropped out due to disability or other reasons.

DISCUSSION

Our study found no evidence that replacing 2 of the quarterly hands-on, instructor-led AED training sessions with CAL would result in adverse effects. We chose a conservative design and replaced

TABLE 4. Mean Treatment Scores for Skill Evaluation ~~

Group A Group B Group C (Continuous (Control (Alternating Computer) Group) Computer)

Scores (n = 37) ( n = 33) ( n = 35)

Time zero Treatment more 8.4 Speed score 2.97

3 months Treatment score Speed score

6 months Treatment score Speed score

9 months Treatment score 9.2 Speed score 2.81

12 months Treatment score 8.5 Speed score 2.76

8.5 8.4 2.94 2.94

8.5 2.88

8.6 9.3 2.91 2.77

8.8 3.0

9.1 9.2 3.07 2.88

ACADEMIC EMERGENCY MEDICINE July 1998, Volume 5. Number 7 715

v) 2 . 5

cn 8 c 2.0 ‘ 8

only 2 of the instructor-led sessions. One study group (group C) went 6 months between instructor-led training sessions; and 1 group (A) went 9 months. We compared all groups using the same skill and speed evaluation criteria. We observed no differences among these groups when compared with a control group of instructor-led skills review sessions. It is quite possible tha t these EMT-Ds could have gone a full year between hands-on, instructor-led sessions. We propose that EMT-Ds in- itially trained to predetermined criteria in the use of AEDs can maintain their AED skills over a 1- year period through the combined use of computer- assisted practice and instructor-led practice.

Recent years have witnessed an explosion of “multimedia” educational and training products that are based, to various degrees, on interactive, computer-based technology.12- l 5 Surprisingly few educators or manufacturers have objectively evaluated the effectiveness of these products with EMS personnel.12J- Most articles are review articles without original data.12J9 We could locate only a few publications in peer-reviewed journals of’ a n evaluation of interactive videodisc instruction for EMS skills.20.21 Several doctoral theses are quoted widely in other articles to support the concept of interactive learning, but for unclear reasons these dissertations have not become available in peer- reviewed journals.21,22

Despite the paucity of published evaluations of CAL, we think there a re some learning tasks in which these approaches are appropriate. AED skill review lends itself well to CAL because AED protocols involve a finite set of actions and outcomes and a well-defined course of treatment. The tasks are simple enough to capture with validity in multimedia, but complex enough to require practice and review.

We are not proposing that CAL can duplicate the psychomotor skills of mannequin-based CPR and defibrillator attachment. Initial defibrillation training requires hands-on, practical experience that duplicates the skills and actions required during real field resuscitations. We see few alterna- tives to closely supervised, hands-on initial training, even with continued efforts by manufacturers to make AEDs a s simple and intuitive to operate as possible.

The ma.jor advantages of CAL revolve around cost. convenience, and time management. EMT-Ds can self-schedule their skill reviews a t convenient times, most often during quiet, on-shift hours. The expenses associated with instructor salaries, em- ployee compensation, and facility fees are reduced significantly.

The cost for instructor-led reviews for all EMT- Ds a t the Bellevue Fire Department for one 3- month period was $810 (27 hours @ $30.00 per

4.5

* * * * I I

3.5 1 30j-- -- --

* * * * * *

N P 37 29 32 30 15 31

A 0 C W p r e

Training Method o p o s t

Finure 4. Box and whisker plot shows median (line in box). Since the interquartile range, 25th and 75th percentiles (tops and bottoms of boxes), and lowest expected values for normal distribution (whiskers) are identical to the median, they all appear as one thick line at 3 points. Asterisks represent extreme values >3 standard deviations from the mean. Distributions of group speed test scores pre- and posttraining show little variation in speed scores among all participants.

hour) to cover instructor salaries. The cost to in- stall the program on the department’s 9 computers would have been about $1,575 for purchasing cop- ies of the program (based on the retail price of sim- ilar programs), including $225 for training tech- nical support staff.

There were no other direct costs associated with instructor-led reviews-such a s overtime pay for EMT-Ds-since the reviews were done while EMT- Ds were on duty, Other potential costs tha t some organizations may encounter for instructor-led reviews are: maintenance for equipment (e.g., CPR mannequins), classroom space, and EMT-D overtime pay for training. Organizations must also con- sider scheduling problems associated with the instructor-led review method because EMT-Ds often must leave the review prematurely to respond to a n emergency dispatch. This disrupts the class and forces the instructor to return at another time or stay much longer to finish the training. Other cost considerations include: computer hardware, hardware maintenance, a n d computer experience of personnel. Staff who are unfamiliar with computers may need brief training on use of the computer program.

Over time the computer review method be- comes much less expensive than the instructor-led method because instructor salaries do not have to


be paid for each review. The cost for implementing the computer program was approximately $1,575 for the Bellevue Fire Department. The cost of delivering instructor-led training 4 times a year is approximately $3,240. An agency the size of the Bellevue Fire Department can expect to save $1,620 per year after the first year by replacing 2 instructor-led training sessions with computer reviews.

Once developed, the software program costs for the CAL are minor. However, some EMS systems may not be able to absorb even modest costs related to CAL. One suggestion is that defibrillator manufacturers could bundle the software costs into the price of new defibrillators and supply the teaching program with every defibrillator.

LIMITATIONS AND FUTURE QUESTIONS

There are several alternative explanations for why we observed almost the exact same skill and speed performances among the 3 study groups. First, the existing level of training and experience of the Bellevue Fire Department EMT-D personnel (5.6 years), with ~ 3 5 0 field events using the AEDs, may contribute to the consistent level of performance during this study period.

Second, our skill evaluation form may not be sensitive enough to detect subtle skill differences. We think, however, that the skill form does eval- uate all critical skills required for effective, safe, and timely shock delivery for people in cardiac arrest from ventricular fibrillation. We believe that more detailed skill evaluation is not necessary to ensure satisfactory field performance.

Third, we could not isolate totally the 2 study groups from ''cross-training" during the study period. While we requested that they use only the computer simulation to maintain skills during the study period, they may have chosen to improve their scores by separate, hands-on practice. How- ever, this was unlikely because there was very lim- ited access to the training equipment necessary to do this kind of practice; in fact, the opposite effect was common in that several control group mem- bers were curious about and wanted to use the computer simulation. Subjects from the control group were prevented from using the software because program data files were checked bimonthly by researchers to ensure that only authorized personnel logged on.

Fourth, AED use may be so easy and intuitive that it makes little difference, once learned, what method is selected for skill maintenance. We think, however, that an opposite phenomenon is more likely. That is, AED field use by emergency personnel actually is deceptively difficult, and effective skill maintenance is critical. EMS personnel face

many difficult challenges during field resuscitations, including scene control, equipment operation, and team coordination. Experiences in our EMT-D program, as well as other reports, reveal that EMT-Ds experience performance problems even after successful initial training.

The observed power for the repeated-measures ANOVA test (0.28) to detect a difference in 1 test point, was too low to rule out type I1 error (not rejecting the null hypothesis of no effect when, in fact, an effect exists). Figure 3 shows an increase in median treatment score by 1 point among groups B and C, but group A shows no increase. Repeating this experiment with greater power might show whether this difference in median scores between group A and the other 2 groups is truly significant. A more rigorous test, or a test with more rigorous scoring criteria, might prevent scores from bunching a t the top score (in this case, 10). This may give greater discrimination at the higher scores. Testing for a larger difference in scores and using fewer study groups would make this study considerably more powerful.

Further research examining the application of CAL in AED training would be in order. Issues to address would include: How much hands-on review can be replaced with CAL? What role does CAL have in initial AED training? Will responders maintain an interest in using CAL over a long period of time? What role does it have in training nontraditional responders and laypersons (public access defibrillation)?

CONCLUSIONS

There appears to be no skill degradation when using CAL to replace 2 of 4 quarterly, instructor-led reviews for AED refresher training. EMS agencies will likely realize cost and convenience benefits from computerized AED refresher training. At the same time, they will be able to maintain estab- lished skill levels.

Ensuring adequate skill performance from nontraditional responders in public access defibrillation programs will be problematic, and CAL may be able to address some of the issues associated with refresher and defibrillator maintenance training. This will be a fertile area for future re- ~ e a r c h . ~ ~ ~ ~ ~ Many of the subjects in this study ex- pressed positive comments about the program. They especially valued the immediate feedback, which helped them to remember AED protocols.

The effectiveness of any computer-based instructional program depends more on proper design and application than on any inherent advantages of the medium. Poorly designed instruction will not likely be effective, regardless of how it is delivered.

ACADEMIC EMERGENCY MEDICINE July 1998, Volume 5, Number 7 717

The authors gratefully express their appreciation to the Fire- fighter EMT-Ds of the City of Bellevue Fire Department, and in particular to Chief Peter R. Lucarelli who has demonstrated imagination and leadership in bringing improved EMS care to the citizens of Bellevue, Washington, and throughout the EMS system. J im Scappini. EMT-D, Louis Gonzales, EMT-P, and Eric M. Dulberg. DrPH, provided valuable advice and service in the conduct of this project,

References

1. Cummins R. From concept to standard-of-care? Review of the clinical experience with automated external defibrillators. Ann Emerg Med. 1989; 18:1269-75. 2. Cummins RO, Ornato JP, Thies W, et al. Improving survival from cardiac arrest: the “chain of survival” concept. Circula- tion. 1991; 83:1832-47. 3. Cummins RO. EMT defibrillation: national guidelines for implementation. Am J Emerg Med. 1987; 5:254-7. 4. Graves J, Austin D, Cummins R. Rapidzap! Automated De- fibrillation. Englewood Cliffs, NJ: Prentice-Hall, 1989. 6. Mosesso VN Jr, Davis EA, Carlin RE, e t al. Use of automated defibrillators by police first-responders for treatment of out-of-hospital cardiac arrests [abstract]. Ann Emerg Med. 1993; 22:920. 6. White R, Asplin B, Bugliosi T, Hankins D. High discharge survival rate after out-of-hospital ventricular fibrillation with rapid defibrillation by police and paramedics. Ann Emerg Med.

7. Kellermann AL, Hackman BB, Dobyns P, Frazier C, Nail L. Engineering excellence: options to enhance firefighter compli- ance with standing orders for first-responder defibrillation. Ann Emerg Med. 1993; 22:1269-75. 8. ACLS Subcommittee, Emergency Cardiac Care Committee (Cummins R, Billi J, eds). Instructors’ Manual for AHA Ad- vanced Cardiac Life Support Course. Early Defibrillation Course. Dallas: American Heart Association, 1994. 9. Johnson R, Wichern D. Applied Multivariate Statistical Analysis. Englewood Cliffs, NJ: Prentice Hall, 1982. 10. Daniel WW Applied Nonparametric Statistics. Boston: Houghton-Millin, 1978, pp 135-40,200-6. 11. Narusis MJ. SPSS for Windows, Base System User’s Guide, Release 6.0. Chicago: SPSS, Inc., 1993, pp 185-7. 12. Parrish K. Utilizing interactive videodisc to teach and test

1996; 28:480-5.

advanced medical psychomotor and cognitive skills. J Med Educ Technol. 1990; 1(1):9-112. 13. Friedman C. Anatomy of the clinical simulation. Acad Med. 1995; 70:205-9. 14. Huckabay LMD, Anderson N, Holm DM, Lee J . Cognitive, affective and transfer of learning consequences of computer- assisted instruction. Nurs Res. 1979; 28:228-32. 16. Hon D. Interactive training in cardiopulmonary resuscitation. Byte. 1982; 7(6):108-38. 16. Lyness A. Effectiveness of interactive video to teach CPR theory and skills. In: Simonson M, Treirner M, eds. Proceedings of selected research paper presentations at the 1985 annual convention of the Association of Educational Communications and Technology. Ames, IA: Iowa State University, 1985, pp

17. Edwards M, Hannah K An examination of the use of interactive videodisc cardiopulmonary resuscitation instruction for the lay community. Comput Nurs. 1985; 3:250-2. 18. Ebner D, Mahoney J J , Lippert H, Bradt N. A study to eval- uate the instructional effectiveness and efficiency of the Ac- tronics CPR Learning System for small group cooperative learning. Unpublished thesis. Bethesda, MD: Medical Technol- ogy Research Group, Uniformed Services University of the Health Sciences, 1988. 19. Moser D, Coleman S. Recommendations for improving cardiopulmonary skills retention. Heart Lung. 1992; 21:372-80. 20. Heller M, Stoy W, Shuman L, Wolfe H, Zavada C. Effec- tiveness of interactive videodisc instruction for the continuing education of paramedics. Prehosp Disaster Med. 1994; 9:165- 71. 21. Stoy W. The effectiveness of a n interactive videodisc system for instruction of paramedic and fourth year medical students in cognitive and psychomotor skills required for advanced airway management. Unpublished doctoral disserta- tion. Pittsburgh: University of Pittsburgh, 1991. 22. Aukerman M. Effectiveness of a n interactive video ap- proach for CPR recertification of registered nurses. Unpub- lished thesis. Pittsburgh: University of Pittsburgh, 1986. 23. Weisfeldt ML, Kerber RE, McGoldrick RP, e t al., and the Automated External Defibrillation Task Force. American Heart Association Report on the Public Access Defibrillation Conference, December 8- 10, 1994. Circulation. 1995; 92:2740-7. 24. Weisfeldt ML, Kerber RE, McGoldrick RP, e t al. Public access defibrillation: a statement for healthcare professionals from the American Heart Association Task Force on Automated External Defibrillation. Circulation. 1995; 92:2763.

462-78.

Documents

Automated External Defibrillators: Skill Maintenance Using Computer-assisted Learning