Predictors of cardiopulmonary resuscitation and automated external defibrillator skill retention

Predictors of cardiopulmonary resuscitation and automated external defibrillator skill retention Barbara Riegel, DNSc, RN, CS, FAAN,a,b Alice Birnbaum, MA,c Tom P. Aufderheide, MD,d Henry C. Thode, Jr, PhD,e Mark C. Henry, MD,e Lois Van Ottingham, BSN, RN,f Robert Swor, DO,g,h and the PAD Investigators Philadelphia, Pa, San Diego, Calif, Seattle, Wash, Milwaukee, Wis, Stony Brook, NY, and Royal Oak and Detroit, Mich

Background Few data exist regarding the retention of cardiopulmonary resuscitation (CPR) and automated external defibrillator (AED) skills over time in relationship to characteristics of lay volunteer responders, training, or risk of exposure to victims. The purpose of this study was to describe the characteristics associated with adequate CPR and AED skill retention. Methods and Results Skill retention was tested 3 to 18 months (mean 6.9 F 3.5 months) after initial training. Instructors judged adequacy of performance of essential CPR or AED skills and provided an overall assessment (adequate/inadequate), which was used as the outcome. Data on 7261 laypersons trained in CPR (4358 also received AED training) in 24 sites across the United States and Canada were available from the Public Access Defibrillation (PAD) Trial. Characteristics of the volunteers, classes, and facilities were evaluated as predictors of performance adequacy. Adjusting for site, intervention assignment (CPR-only or CPR + AED), and time since initial training, volunteer characteristics associated with adequate CPR performance were age (OR 0.78 per 10-year increment), male sex (OR 1.44), minority (OR 0.62), married (OR 1.35), prior emergency experience (OR 1.66), prior CPR class (OR 1.68), prior advanced training (OR 1.59), and extracurricular CPR training (OR 1.91) (all P b .05). Characteristics associated with AED performance included age (OR 0.69), college education (OR 1.34), and native language other than English (OR 0.51) (all P b .05). Conclusions Certain subgroups of lay volunteers may need targeted outreach programs in CPR and AED use, classes with longer training time, more practice, or more intense retraining to maintain their CPR and/or AED skills. (Am Heart J 2005;150:927- 32.)

Sudden out-of-hospital cardiac arrest (OOH-CA) remains a significant cause of death in spite of recent declines in overall mortality from cardiovascular disease.1 Existing methods of emergency resuscitation

From the aSchool of Nursing, University of Pennsylvania, Philadelphia, Pa, bSharp HealthCare, San Diego, Calif, cDepartment of Biostatistics, University of Washington, Seattle, Wash, dMedical College of Wisconsin, Milwaukee, Wis, eDepartment of Emergency Medicine, Stony Brook University, Stony Brook, NY, fDepartment of Biostatistics, PAD Clinical Trial Center, University of Washington, Seattle, Wash, g Department of Emergency Medicine, William Beaumont Hospital, Royal Oak, Mich, and hDepartment of Emergency Medicine, Wayne State University, Detroit, Mich. This study was supported by contract #N01-HC-95177 from the National Heart, Lung, and Blood Institute (Bethesda, Md), with additional support from the American Heart Association (Dallas, Tex); Guidant Corporation (Indianapolis, Ind); Medtronic, Inc (Minneapolis, Minn); Cardiac Science/Survivalink, Inc (Minneapolis, Minn); Medtronic Physio-Control (Redmond, Wash); Philips Medical Systems, Heartstream Operation (Seattle, Wash); and Laerdal Medical Corporation (Wappingers Falls, NY). Submitted July 14, 2004; accepted January 26, 2005. Reprint requests: Barbara Riegel, DNSc, RN, CS, FAAN, School of Nursing, University of Pennsylvania, 420 Guardian Dr, Philadelphia, PA 19104-6096. E-mail: [email protected] 0002-8703/$ - see front matter n 2005, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2005.01.042

are inadequate because of time delays inherent in the arrival of a trained responder with defibrillation capabilities to the side of an OOH-CA victim. In the Public Access Defibrillation (PAD) Trial, volunteer nonmedical responders (lay volunteers without a traditional responsibility to respond to a medical emergency) were trained to perform cardiopulmonary resuscitation (CPR)–related and automated external defibrillator (AED)–related skills and to respond to OOH-CA. The design and rationale for the PAD Trial have been described in detail elsewhere.2 Briefly, the study recruited public and residential facilities or bcommunity unitsQ (eg, office buildings, shopping centers, gated communities) that were expected to have a relatively high frequency of cardiac arrests. In some locations, multiple facilities combined to form a community unit to achieve the PAD Trial’s cardiac arrest risk criteria. The volunteers were employees or residents in these facilities. The community units were randomized to receive CPR- only or CPR + AED training. All volunteers were trained to access the emergency medical system and begin CPR.

American Heart Journal November 2005

928 Riegel et al

Table I. Volunteer characteristics

Volunteer characteristics Mean age (y) Male sex (%) Minority status (%) Hispanic origin (%)T Married or has significant other (%) Education: at least some college (%) Possesses driver’s license (%) Able to operate a computer (%) Native language other than English (%) Back problems (%) Lung disease (%) Heart disease (%) Vision problems (%) Arthritis (%) Experience aiding in emergency (%) Experience aiding in seizure (%)T CPR class before PAD (%) First aid before PAD (%)y Advanced training before PAD (%)T Non-PAD CPR training after PAD initial training (%) Non-PAD AED training after PAD initial training (%) Initial PAD trial training in CPR + AED intervention (%)

Returned for skill retention testing (n = 7261)

Failed to attend testing (n = 3162)

Dropped out of trial (n = 3552)

Testing not scheduled at facility (n = 2900)

40.73 (SD 13.82) 53.6 18.7 7.3 58.7 73.0 95.9 86.9 6.0 8.3 0.5 0.9 2.7 5.5 51.6 36.8 58.4 51.1 10.5 6.5 1.6 58.4

39.12 (SD 14.50) 54.6 22.3 10.7 52.2 72.9 94.5 85.8 6.7 8.3 0.6 1.1 3.8 6.6 45.8 29.3 53.2 45.8 9.8 Unknown Unknown 60.1

35.16 (SD 14.08) 55.4 24.8 10.6 43.5 70.4 92.1 86.8 7.4 8.5 0.8 0.6 4.1 4.7 48.4 33.8 56.5 51.3 12.1 Unknown Unknown 56.3

39.97 (SD 14.21) 52.2 21.0 7.7 59.1 70.7 95.4 85.9 3.8 8.4 0.6 0.9 3.5 6.3 45.6 32.0 47.0 40.3 10.8 Unknown Unknown 47.2

TApproximately 10% of responses were missing on the volunteer questionnaire. yApproximately 20% of responses were missing on the volunteer questionnaire.

Volunteers in the CPR + AED facilities were also trained to defibrillate OOH-CA victims using an AED.3 A total of 1260 distinct facilities from 24 sites in the United States and Canada participated in the trial. Of concern is the realization that OOH-CA events are relatively infrequent (2-3/1000 population per year)4,5 and that lay responders will rarely, if ever, be called upon to respond to an arrest and use their skills. Therefore, important questions addressed by the PAD Trial were whether volunteers would adequately retain the skills of CPR and AED use and how the volunteer, training, and facility characteristics were related to the retention of these skills. Previous work by Todd et al6 has suggested that there are high rates of skill deterioration after initial CPR training. This decline in skills is apparent for both lay persons and medical professionals.7-9 A body of literature is developing for AED use, documenting ease of use by elders, children, and other lay groups.10-12 However, few data exist regarding skill retention over time in relationship to characteristics of responders, their training, or their environment. Knowledge of these relationships would allow instructors to tailor the training approach for specific subgroups of individuals interested in learning and using CPR and AED skills. The specific objective of this analysis was to describe volunteer demographic, training, and facility characteristics associated with adequate CPR and AED skill retention 3 to 18 months after initial training.

Methods The PAD Trial was approved by the institutional review board of the coordinating center and by institutional review boards of participating sites. Initial training in both the CPR-only and the CPR + AED arms of the trial emphasized recognizing OOH-CA, accessing the local emergency medical system, and performing CPR. In addition, volunteers in the CPR + AED arm were trained to operate 1 of 3 AED models used in the trial. Certified basic life support instructors taught the initial training classes. Training was standardized and consistent with the American Heart Association HeartSaver ABC and HeartSaver AED programs, although other curricula were acceptable, provided that the criteria below were met. Initial training guidelines specified 3 to 4 hours in course length, with a recommended student to instructor ratio of 4:1 and no more than 12 students per class. Training was as realistic as possible, with case studies and practice scenarios. Watching a video was recommended. Instructor lecture was minimized (not N45 minutes total for lecture and demonstration). Individual trainees were required to have at least 20 (optimally, 30) minutes of skill practice time. In retesting sessions, volunteers were individually tested on component skills and on overall CPR or AED performance. The PAD Trial investigators agreed, a priori, upon 5 essential CPR skills (assess responsiveness, access 911, provide ventilation adequate to make the chest rise, place hands properly on the chest, and compress the chest an adequate depth) and 5 essential AED skills (bare the chest for electrode placement, place electrodes correctly, clear self, verbally clear the area, and shock within 90 seconds of the arrival of the AED). In

American Heart Journal Volume 150, Number 5

Riegel et al 929

Table II. Independent predictors of adequate CPR and AED skill performance Univariate predictors of skill retention CPR skills ‘‘Adequate’’

Volunteer characteristics Age (per 10-y increment) Male sex Minority Ethnicity (Hispanic) Married (or has significant other) Education (at least some college) Native language other than English Back problems Lung disease Heart disease Vision problems Arthritis Experience aiding in emergency Experience aiding in seizure CPR class (before PAD) First aid (before PAD) Advanced training (before PAD) Non-PAD CPR training after PAD initial training Non-PAD AED training after PAD initial training Initial training characteristics Student/teacher ratio Student/manikin ratio Student/practice time ratio Student/AED ratio Facility characteristics Residential (vs public) facility Estimated risk of cardiac arrest

Independent predictors of skill retentionT

AED skills ‘‘Adequate’’

CPR skills ‘‘Adequate’’

AED skills ‘‘Adequate’’

Odds ratio

P

Odds ratio

P

Odds ratio

P

Odds ratio

P

0.76 1.49 0.63

.001 .001 .001 NS .043 .001 .001 .017 .049 .056 .016 .001 .001 .001 .001 .001 .001 .001 .002

0.68

.001 NS NS NS NS .001 .001 NS NS NS NS NS .002 NS .002 NS NS NS NS

0.78 1.44 0.62

.001 .001 .001 NS .001 NS NS NS NS NS NS NS .001 NS .001 NS .009 .001 NS

0.69

.001 NS NS NS NS .049 .014 NS NS NS NS NS .067 NS .058 NS NS NS NS

1.15 1.30 0.59 0.76 0.44 0.52 0.62 0.58 1.97 1.78 2.04 1.96 2.76 2.04 2.93

0.62

1.53 0.44

1.47 1.55

1.35

1.66 1.68 1.59 1.91

1.34 0.51

1.30 1.35

NS NS NS NA

NS NS NS NS

NS NS NS NA

NS NS NS NS

.009 NS

NS NS

NS NS

NS NS

TLogistic GEE model adjusting for site, time since initial training, and randomization assignment (CPR-only or CPR + AED) of the participating facility.

addition, instructors provided a summary assessment (the outcome measure) of whether the CPR performed was adequate (ie, would have perfused) and/or whether the AED skills were adequate (ie, would have delivered a shock approximately through the heart). After testing, volunteers were retrained to proficiency. Potential predictors of volunteer skill performance were proposed a priori based on previously published11,13,14 or established variables likely to influence the acquisition and retention of CPR or AED skills. The volunteer characteristics considered were age, sex, marital status, education, ethnicity, native language other than English, and minority group. Minority was defined as any volunteer who indicated at least one non-white racial category of black, Asian/Pacific islander, Native American/First Nations, or other. Additional characteristics considered for exploratory analysis were possession of a driver’s license, ability to operate a computer, history of physical limitations (back problems, lung disease, heart disease, poor vision even with corrective lenses, and arthritis), experience aiding in an emergency (or seizure) before the PAD Trial, resuscitation training before the PAD

Trial (CPR, first aid, or advanced training), and extracurricular training during the PAD Trial (non-PAD CPR or AED class). Initial training course characteristics assessed included student/teacher ratios as well as student/practice time ratios, student/manikins ratios, and student/AEDs ratios. Facility characteristics assessed were residential versus public facility and the risk of cardiac arrest (estimated from the facility’s history of OOH-CAs onsite and/or exposure to individuals older than 50 years).2

Statistical analysis By protocol, testing sessions were recommended to be scheduled 3 to 6 months after initial training; however, in practice, these sessions occurred anywhere from b1 month to N24 months after initial training. This analysis was limited to volunteers who were tested 3 to 18 months (mean 6.9 F 3.5 months) after their initial training (86% of all volunteers tested). To account for correlation among volunteers at a given facility, associations of volunteer, facility, and training characteristics with overall skill performance (adequate/inadequate) were evaluated using logistic generalized estimating equations

American Heart Journal November 2005

930 Riegel et al

random effects models. Models were adjusted for site and for time since initial training. In the CPR analysis, intervention assignment was considered to adjust for any AED training impact on CPR performance. Stata version 6.0 (College Station, Tex) and SPSS version 10.0 (Chicago, Ill) were used for data analysis. A 2-sided P value b.05 was considered statistically significant.

Results A total of 19 398 volunteers received initial training by January 1, 2003. Of these, 1406 were excluded from analyses for protocol violations (eg, testing in groups rather than individually), 1117 attended skills retention testing early (b3 months after initial training) or late (z18 months after initial training), and 9614 were ineligible, unavailable, or unwilling to attend testing (based on information available, approximately 19% dropped out of the trial because of job transfers, job turnover, etc; 12% were volunteers at facilities that had difficulty scheduling testing; 7% were not active in the trial when testing occurred; and 62% failed to attend testing sessions offered at their facilities). Demographic characteristics of those who returned for testing and those who did not were generally similar. The average age of volunteers who returned for testing was 41 years, versus 35 years for volunteers who dropped out of the trial. Among volunteers who returned for testing, 59% were married, versus 44% among those who dropped out of the trial. Fifty-eight percent of volunteers who returned for skill retention testing had been initially trained in the CPR + AED intervention, versus 47% of volunteers at facilities where no testing was scheduled. A total of 7261 volunteers trained to perform CPR (4358 of whom were also trained to operate an AED) was included in the analysis assessing potential predictors of skills retention. Table I provides descriptive information about the volunteers. Table II summarizes associations between volunteer characteristics and skill retention. Characteristics univariately associated with badequateQ CPR skill performance included male sex, married/significant other, education beyond high school, extracurricular CPR or AED training after the initial PAD Trial training, experience aiding in an emergency or seizure before the PAD Trial, and training such as CPR class, first aid, or advanced training before the PAD Trial. Characteristics including older age, minority race, native language other than English, and having a physical disability such as back problems, lung disease, vision problems, or arthritis were associated with binadequateQ CPR skill retention (all P b .05) (Table II). In a backward-stepwise generalized estimating equations logistic analysis, characteristics remaining in the model after adjusting for site, time since initial training, and initial training assignment to CPR-only or CPR + AED included age (OR 0.78 per 10-year increment, P b .001), male sex

Table III. Characteristics of participating facilities and initial training classes CPR-only Practice time (min) per student median [IQR] Student/teacher ratio median [IQR] Student/manikin ratio median [IQR] Student/AED ratio median [IQR]

CPR + AED

11.89

[7.50-17.86]

14.38

[9.72-21.13]

4.50

[3.58-5.56]

4.58

[3.54-5.50]

2.50

[1.40-4.00]

3.00

[1.60-4.00]

–

–

3.73

[2.82-4.67]

IQR, interquartile range.

(OR 1.44, P b .001), minority status (OR 0.62, P b .001), married/significant other (OR 1.35, P b .001), experience aiding in an emergency (OR 1.66, P b .001), CPR class before PAD initial training (OR 1.68, P b .001), advanced training before PAD initial training (OR 1.59, P = .009), and extracurricular CPR training after PAD initial training (OR 1.91, P b .001) (Table II). Characteristics univariately associated with badequateQ AED skill performance included education beyond high school, experience aiding in an emergency, and having taken a CPR class before PAD initial training. Older age and native language other than English were associated with binadequateQ AED skill performance (all P b .05). In the logistic analysis, characteristics remaining in the model included age (OR 0.69 per 10-year increment, P b .001), education beyond high school (OR 1.34, P = .049), and native language other than English (OR 0.51, P = .014). Trends were noted for volunteers with experience aiding in an emergency (OR 1.30, P = .067) and CPR class before PAD initial training (OR 1.35, P = .058) (Table II). Neither the facility’s estimated risk of cardiac arrest nor the characteristics of the volunteers’ initial training classes were predictive of AED or CPR skill performance (Table II). Table III provides descriptive information about the facilities and the PAD Trial classes.

Discussion Many studies have reported that retention of resuscitation skills degrades over time for lay responders.6,9-12,15-23 What remains unknown is whether skill degradation is uniform across subgroups of lay responders. In the PAD Trial, when a standardized training program was provided to a large sample of lay volunteers, several characteristics were associated with adequacy of skill performance measured at 3 to 18 months (mean 6.9 months) after initial training. Two themes were identified in the analyses. First, certain student characteristics (eg, older age, no college education, minority status, native language other than

American Heart Journal Volume 150, Number 5

English) can be used to identify volunteers who may need particular attention (eg, more time during training, specific skills targeted, or more frequent refresher training). The second theme is that repetition promotes skill retention. Volunteers who had prior training in CPR or other first aid, supplementary training, or previous experience assisting in an emergency event all performed relatively better during testing, but this may also simply reflect a selection bias. Many of the individual characteristics predictive of poor performance were anticipated. The relatively small stature of women could be anticipated to interfere with the physical effort required to compress a chest during CPR. Whereas CPR requires physical skill performance, use of an AED is primarily a cognitive task and no sex difference in performance was found. Problem-solving skills are acquired during higher education, so it is not surprising that volunteers with some college education performed better on AED testing. Younger volunteers demonstrated skill retention in both CPR and AED testings in greater proportions than did older volunteers, perhaps because short-term or working memory is known to decline with age, especially in stressful situations such as testing.24 Being married or having a significant other was associated with better CPR performance, perhaps because the thought of saving a loved one motivated the volunteer to learn. However, other characteristics also deserve comment. The characteristic native language other than English was associated with inadequate AED performance (but not CPR) in the multivariate model. This finding may be explained by the fact that the AEDs used in this study use voice prompts that were spoken in English. Being in a minority group was associated with inadequate CPR performance. This apparent effect may reflect the biases of a largely white instructor pool or trainee sensitivity to being judged by a nonminority instructor. Inadequate resuscitation performance could also reflect shortcomings with regard to cultural sensitivity in the course materials or cultural beliefs about death that influence receptivity to learning and performing CPR.25 Further research may help clarify these issues and identify specific groups that would benefit from targeted outreach programs. The observation that CPR and AED skill retention was better in volunteers who had some other experiences (eg, prior CPR class) on which to build is not really surprising because repetition is a tool for acquiring knowledge and retaining skills. This finding has important implications for public policy. If children learned basic life support in elementary school, for example, and received periodic retraining throughout their elementary and secondary education, later performance might be improved. In the PAD Trial, initial training was 3 to 4 hours in length. Although most

Riegel et al 931

employers of volunteers in this study were receptive to providing the time necessary to acquire and maintain lifesaving skills, earlier exposure to these skills could decrease the time required to train volunteers. Early uniform learning by the populace of CPR and AED use may also increase the currently low rate of bystander CPR.26 Comparatively, fewer characteristics were associated with AED skill retention than with CPR skill retention. This finding is consistent with the relative simplicity of the psychomotor skills required for competent AED use (and the presence of AED voice prompts). Contrary to AED use, CPR performance is physically demanding and requires that volunteers recall the sequence of steps without prompting. The characteristics found to be associated with AED and CPR skill retention may assist instructors in individualizing training and retraining programs to meet the needs of trainees. Current guidelines emphasize practice over lecture. Training program developers and instructors may need to develop and evaluate different techniques for students with any of the characteristics associated with lower skill retention. For example, elderly volunteers in residential settings—logical target groups for CPR and AED training—may need a more intensive course, more practice time, and/or more frequent refresher training than younger groups.13 Interestingly, the training characteristics such as ratios of students to instructors, manikins, and practice time were not significant predictors of testing success in either CPR or AED use. It may be that there was not enough variation in the PAD Trial class resources (or intrinsically in resuscitation classes) to detect the influence of different approaches on performance. Given the interquartile ranges observed in the PAD Trial classes, one would expect little effect on outcome (see Table III). The facility characteristics tested were not associated with CPR or AED skill retention. This finding suggests that basic life support training can be implemented in a widespread fashion without concern that the setting will influence skill retention. Some limitations of this study deserve comment. A single instructor rated the skill adequacy of each volunteer at the facility without a second opinion regarding performance. Interrater reliability was not assessed and it is possible that some instructors were lenient and others were rigorous in their evaluations. The analyses were adjusted for site to address this potential effect. Although individual skill components of CPR and AED use were graded, this analysis used the instructor’s overall evaluation of performance (adequate/inadequate) as the gold standard. This evaluation could have been biased in some fashion such as by instructors’ desire to have bgoodQ outcomes or by negative stereotypes held about older students, women, those with disabilities, and minorities. More than half of the volunteers were in the

932 Riegel et al

CPR + AED arm perhaps because the novelty of working with AEDs may have caused a greater success of volunteer recruitment in the intervention arm. Finally, there is some potential for selection bias; a larger proportion of volunteers who returned for testing were married than were those who dropped out of the trial; volunteers who returned for testing were younger, on average, than volunteers who dropped out of the trial; and a larger proportion of volunteers who returned for testing were initially trained in the CPR + AED intervention than were those at facilities that did not schedule testing. Otherwise, there were no substantive differences across characteristics measured. In summary, this study identifies volunteer characteristics associated with CPR and AED skill retention. These results should be useful to further improve skill retention after CPR and AED courses, especially with volunteers needing extra time and special attention. Additional research is warranted to test innovative teaching methods for these special populations of lay volunteers.

References 1. Zheng CJ, Giles WGM. Sudden cardiac death in the United States, 1989-1998. Circulation 2001;104:2158 - 63. 2. The PAD Investigators. The Public Access Defibrillation (PAD) Trial study design and rationale. Resuscitation 2003;56:135 - 47. 3. Hazinski MF. Saving lives through public access defibrillation. Emerg Med Serv 2001;30:56 - 60. 4. Han BH, Meyer PM, Wright FA, et al. Incidence of cardiac arrest: a neglected factor in evaluating survival rates. N Engl J Med 1993;329:600 - 6. 5. Becker L, Eisenberg M, Fahrenbruch C, et al. Public locations of cardiac arrest. Implications for public access defibrillation. Circulation 1998;97:2106 - 9. 6. Todd KH, Heron SL, Thompson M, et al. Simple CPR: a randomized, controlled trial of video self-instructional cardiopulmonary resuscitation training in an African American church congregation. Ann Emerg Med 1999;34:730 - 7. 7. Nichol G, Detsky AS, Stiell IG, et al. Effectiveness of emergency medical services for victims of out of hospital cardiac arrest: a meta-analysis. Ann Emerg Med 1996;27:700 - 10. 8. Lewis FH, Kee CC, Minick MP. Revisiting CPR knowledge and skills among registered nurses. J Contin Educ Nurs 1993;24: 174 - 9.

American Heart Journal November 2005

9. Wenzel V, Lehmkuhl P, Kubilis PS, et al. Poor correlation of mouthto-mouth ventilation skills after basic life support training and 6 months later. Resuscitation 1997;35:129 - 34. 10. Ecker R, Rea T, Meischke H, et al. Dispatcher assistance and automated external defibrillator performance among elders. Acad Emerg Med 2001;8:968 - 73. 11. Cummins RO, Schubach J, Litwin P, et al. Training lay persons to use automatic external defibrillators: success of initial training and one-year retention of skills. Am J Emerg Med 1989;7:143 - 9. 12. Lawson L, March J. Automated external defibrillation by very young, untrained children. Prehosp Emerg Care 2002;6:295 - 8. 13. Swor R, Farr L, Jackson R, et al. A randomized controlled trial of chest compression only CPR for older adults: a pilot study. Resuscitation 2003;58:177 - 85. 14. Iwashyna TJ, Christakis NA, Becker LB. Neighborhoods matter: a population-based study of provision of cardiopulmonary resuscitation. Ann Emerg Med 1999;34:459 - 68. 15. Riegel B. Training nontraditional responders to use automated external defibrillators. Am J Crit Care 1998;7:402 - 10. 16. Vanderschmidt H, Burnap TK, Thwaites JK. Evaluation of a cardiopulmonary resuscitation course for secondary schools retention study. Med Care 1976;14:181 - 4. 17. Kaye W, Rallis SF, Mancini ME, et al. The problem of poor retention of cardiopulmonary resuscitation skills may lie with the instructor, not the learner or the curriculum. Resuscitation 1991;21:67 - 87. 18. Mancini ME, Kaye W. Resuscitation training: a time for reassessment. J Cardiovasc Nurs 1996;10:71 - 84. 19. Fossel M, Kiskaddon RT, Sternbach GL. Retention of cardiopulmonary resuscitation skills by medical students. J Med Educ 1983;58:568 - 75. 20. Gombeski Jr WR, Effron DM, Ramirez AG, et al. Impact on retention: comparison of two CPR training programs. Am J Public Health 1982;72:849 - 52. 21. Van Kerschaver E, Delooz HH, Moens GF. The effectiveness of repeated cardiopulmonary resuscitation training in a school population. Resuscitation 1989;17:211 - 22. 22. Nyman J, Sihvonen M. Cardiopulmonary resuscitation skills in nurses and nursing students. Resuscitation 2000;47:179 - 84. 23. Gass D, Curry L. Physicians and nurses’ retention of knowledge and skill after training in cardiopulmonary resuscitation. Can J Cardiol 1983;128:550 - 1. 24. Logie RH, Maylor EA, Sergio DS, et al. Working memory in eventand time-based prospective memory tasks: effects of secondary demand and age. Eur J Cogn Psychol 2004;16:441 - 56. 25. Cone DC, Richardson LD, Todd KH, et al. Health care disparities in emergency medicine. Acad Emerg Med 2003;10:1176 - 83. 26. Becker L, Ostrander M, Barrett J, et al. Outcomes of CPR in a large metropolitan area — where are the survivors? [see comments]. Ann Emerg Med 1991;20:355 - 61.