S

Ws

BR

a

ARR1A

KCATCCW

1

isdmiieaarce

i

0d

Resuscitation 81 (2010) 568–575

Contents lists available at ScienceDirect

Resuscitation

journa l homepage: www.e lsev ier .com/ locate / resusc i ta t ion

imulation and education

ith or without an instructor, brief exposure to CPR training producesignificant attitude change�

onnie Lynch ∗, Eric L. EinspruchMC Research Corporation, 111 SW Columbia Street, Suite 1200, Portland, OR 97202, United States

r t i c l e i n f o

rticle history:eceived 23 September 2009eceived in revised form5 December 2009ccepted 23 December 2009

eywords:PRttitudesrainingonfidenceompetenceillingness

a b s t r a c t

Background: A common reason for bystanders’ failure to perform CPR in real or hypothetical situationsis their lack of confidence in themselves. CPR self-training, which uses learner-operated virtual mediarather than a live instructor, has not been assessed for its ability to influence learners’ attitudes towardperforming CPR in a real emergency. The aim of this study was to compare attitude effects associatedwith traditional, live instruction versus self-training or no instruction.Method: Data from 1069 lay learners were collected. Learners were assigned randomly to a traditionalinstructor-led course, a video-based self-training course, or a no-training control group. All learnerscompleted pre-training and post-training questionnaires that assessed competence, confidence, and will-ingness to perform CPR. Learners’ objective performance of CPR was also assessed, post-training, via arecording manikin.Results: ANOVA revealed that, in all 3 groups, all 3 attitudes changed significantly from pre- to post-questionnaire; further, the amount of attitude change did not differ reliably among the 3 groups (P < .05).

Of the objective measures, ventilation performance was the only one consistently and positively corre-lated with attitudes (P < .05). Despite focus group comments that suggested self-trained learners’ concernsabout the rudimentary nature of their training, these concerns did not manifest as a hindrance to positiveattitude change.Conclusions: Live training does not pose any measurable advantage for developing learners’ positiveattitudes. The counterintuitive finding that controls experienced similar levels of attitude change suggests

R tes

that mere exposure to CP

. Introduction

Increasing the number of CPR-trained lay responders is a log-cal step toward improving rates of survival from out-of-hospitaludden cardiac arrest. The recent introduction and empiricallyemonstrated efficacy1–3 of self-training kits make CPR trainingore readily accessible than ever before. However, participation

n training does not guarantee that an individual will respondn an emergency.4,5 The literature in both social psychology andmergency medicine suggests that failure to act is often morettributable to psychological factors than to practical factors such

s objective competence. Although potential rescuers may beepulsed by the presence of bodily fluids or other unpleasantharacteristics of the victim,6–8 or may fear contracting a dis-ase from mouth-to-mouth contact with the victim,7–10 the most

� A Spanish translated version of the abstract of this article appears as Appendixn the final online version at doi:10.1016/j.resuscitation.2009.12.022.∗ Corresponding author. Tel.: +1 503 223 8248; fax: +1 503 223 8399.

E-mail address: blynch@rmccorp.com (B. Lynch).

300-9572/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.resuscitation.2009.12.022

ting can have positive effects on attitudes.© 2010 Elsevier Ireland Ltd. All rights reserved.

commonly cited reason for reluctance to perform CPR in real orhypothetical situations involves bystanders’ lack of confidence orperceived competence for recognizing the emergency and respond-ing appropriately.5,8,9

Although most of the literature that relates confidence and per-ceived competence to helping behavior is now decades old, itsrelevance to CPR training has never been timelier. In one study11

that directly compared confidence and willingness to provideemergency aid, as confidence increased (via training and experi-ence), so did willingness. Several studies suggest that perceivedcompetence can be induced fairly easily, and that its increaseproduces corollary willingness to help and actual helping behav-ior. Midlarsky12 manipulated perceived competence and producedhigher rates of attempted helping (in this case, checking on some-one who was perceived to be choking). Kazdin and Bryan13 showedsimilar results for induced competence and blood donations. No

studies have investigated whether video-based training can suc-cessfully manipulate these characteristics.

The protocol for traditional CPR classes asks instructors toaddress many of the psychological factors that foster confidence,perceived competence, and willingness to perform CPR: encour-

esusci

acmlspaa

tlowtomticicwh

beacowadtfpat

2

tstaFbe24wvtpptaRe

i3lwLa

were average ventilation volume, average compression depth, andpercent of compressions with full release (these were the onlymeasures available from the infant manikin, due to technical con-

B. Lynch, E.L. Einspruch / R

ging, assessing, and providing real-time feedback; addressingoncerns about exposure to disease risk; helping learners prepareentally for emergency situations; and ultimately pronouncing

earners competent to perform CPR. In video-based self-training,uch as the American Heart Association’s (AHA) CPR Anytime®

rogram, learners watch a brief video and practice on an inflat-ble manikin, with the whole process directed by a narrator whoddresses many of the same psychological factors.

Although an extensive review14 shows that education at a dis-ance can produce learning outcomes as good as or better thanive instruction, precious few studies address the psychologicalutcomes of learning in direct comparisons of distance learningith live instruction. Proponents of distance learning15,16 contend

hat learners tend to learn from a virtual instructor and a livene in similar ways. Two studies within the realm of emergencyedicine concur with this assertion. A virtual reality system for

eaching mass casualty triage to medical students produced signif-cant increases in confidence for all skills that were tested,17 and aomputer game for teaching CPR to high school students inducedmprovements in self-efficacy (a relatively general measure ofonfidence).18 But neither of these studies included comparisonsith live instruction, so it is not known whether the results wouldave been similar.

Ultimately, the effectiveness of self-training for increasingystander CPR requires understanding of how such training influ-nces learners’ attitudes and beliefs about performing CPR onnother person. This study therefore focuses on confidence, per-eived competence, and willingness to perform CPR as a functionf participating in 1 of 3 training modes: with a live instructor;ith a self-training kit; and without training (control group). We

lso investigate whether the accuracy of self-assessed competenceiffers for traditionally trained participants (who, according to pro-ocol for such classes, receive explicit, iterative, tailored feedbackrom the instructor on all CPR skills), versus self-trained partici-ants (who get feedback from the manikin on compression depthnd ventilation volume, and generic positive encouragement fromhe video narrator for other skills); versus control participants.

. Methods

The methods of random assignment and the details of theraining interventions have been described previously1 and areummarized briefly here. The data collection occurred in down-own Portland, Oregon, in 3 separate waves (W1, W2, and W3)s part of 3 separate, but related, studies. W1 occurred betweenebruary and May 2004 and included 224 learners; W2 occurredetween November 2004 and June 2005 and included 602 learn-rs; W3 occurred between January and February 2006 and included43 learners. In both W1 and W2, the learners were adults between0 and 70 years of age; in W3, which involved infant CPR, learnersere between 18 and 70 years of age. All learners were recruited

ia newspaper advertisements, flyers, and word of mouth fromhe participants. The only exclusion criteria other than age werearticipation in any CPR training within the previous 5 years androfessional status as a healthcare provider. The study conformedo the principles of the Declaration of Helsinki and the protocol waspproved by Portland State University’s Human Learners Researcheview Committee. Informed consent was obtained from all learn-rs.

Learners in W1 and W2 were exposed to one of the followingnterventions: an AHA Heartsaver course that lasted approximately

h; a CPR Anytime session (AHA’s 22-min video self-training for

aypersons [www.CPRAnytime.org, on 7/21/2009]; or a 10–40-minaiting period in the study space (without receiving training).

earners in W3 were exposed to one of the following interventions:1–1.5 h AHA Family and Friends course (infant portions only), or

tation 81 (2010) 568–575 569

a 30-min CPR Anytime for Infants session (analogous to the adultversion of CPR Anytime, but designed specifically for infant CPR).

All learners in W1, W2, and W3 completed a demographicsand attitudes pre-questionnaire (see Appendix) upon arrival in thestudy space, and a post-questionnaire immediately before testingon a recording manikin, being debriefed and departing. The post-questionnaire contained only the attitude items, but in a differentorder from the pre-questionnaire. The attitude items were iden-tical across the 3 waves, except that the infant version had oneadditional item related to choking.

The testing procedure followed the same protocol in all 3 waves.The test scenario occurred within 1.5 h of the learner’s training.Learners entered the testing room individually, where they encoun-tered an examiner and normal office furnishings, with a LaerdalResusci AnneTM (or Laerdal Resusci BabyTM, for infant CPR) record-ing manikin on the floor. The manikin was connected to LaerdalPC SkillReportingTM software via a laptop computer. The examinerrecited a standardized script to the learner before the assessment.During the 3-min assessment period, the sensored manikin pro-vided data on volume of ventilations, depth of compressions, andrate of compressions. Data on hand placement for compressionswere available only with adult manikins, and were not included asan outcome variable in this study.

In all 3 waves of the study, self-trained learners were selected atrandom for participation in focus groups to elicit additional, quali-tative data. A total of 100 learners participated: 53 in W1, 36 in W2,and 11 in W3.

Experimental conditions and tasks for all 3 waves are summa-rized in Fig. 1.

The primary outcomes of interest were: changes in pre- topost-intervention attitudes, and relationship of self-assessed com-petence to objective measures of competence.

2.1. Analyses

Demographic data (age, gender, ethnicity, education, and his-tory of heart disease) were summarized using descriptive statistics.

Two sets of 4 reliability analyses were conducted on the pre-training questionnaire. Cronbach’s alpha was examined for itemsgrouped by their 3 originally hypothesized scales (confidence, per-ceived competence, and willingness) and for all items together.Items were dropped from each hypothesized subscale until alphawas maximized. This analysis was conducted once for the combinedgroup of learners in W1 and W2 and once again for the learners inW3 study.

A factor analysis determined whether the items loaded on thescales as originally hypothesized. This analysis was repeated for the2 groups of learners, as per the reliability analysis.

The Wilcoxon signed-rank test was used to test, for eachtraining group, whether attitudes changed from pre-training topost-training. ANOVA assessed effects of demographic variables onattitudes.

Accuracy of self-assessed competence was assessed in 2 ways.First, scores for each scale were correlated with key objective mea-sures. For W1 and W2 studies these measures were percent ofventilations with adequate volume, percent of compressions withadequate depth, and average compression rate.1 For W3, measures

straints). Second, for each scale the distribution of scores was

1 For both ventilation and compression performance, attempts that produced avolume or force at or above the recommended level were considered “adequate.”

570 B. Lynch, E.L. Einspruch / Resuscitation 81 (2010) 568–575

assoc

dt

evdotvtrcf

3

pitmo

fieifiapWf

Wtc

the attitudinal factor that changed most (see Table 4). ANOVA, withBonferroni corrections for multiple comparisons (P < .002), revealedthat the only demographic factor that differentially affected atti-tudes was gender: males were more confident overall than females

Table 1Demographic characteristics of the study samples.

Item Percent of respondents

W1 and W2(N = 822)

W3a (N = 243)

GenderFemale 55.7 49.8Male 44.3 50.2

Age49 years of age or less 35.9 72.350–59 years of age 39.3 22.360 years of age or more 24.8 5.5

Ethnic groupWhite 84.8 65.3Non-white 15.2 34.7

EducationHigh school or less 21.5 38.1Some college 39.0 37.7Bachelor’s degree or higher 39.5 24.1

History of heart diseaseDo you or does anyone in your family havea history of heart disease or heartdisorder?

48.1 29.7

Have you or has anyone in your familyever had a heart attack?

45.0 –

Fig. 1. Experimental groups and

ivided into terciles, and MANOVA compared performance amonghe 3 groups.

The focus groups provided useful qualitative data about learn-rs’ subjective experiences in self-training, and the ways in whicharious components of the video may have affected attitudes. Fouriscussion topics were analyzed: participants’ 1-word descriptionsf the overall training; their impressions of the video’s narrator;heir discussions of any confusion they experienced during theideo; and their sense of feeling ready to perform CPR after self-raining. A simple content analysis was performed to categorizeesponses as positive, negative, or neutral/ambivalent. The mostommonly occurring categories of response and some exemplarsrom each category are included.

. Results

Table 1 shows the demographic characteristics of the study sam-les. Males and females were about equally represented. Learners

n W1 and W2 were generally older, more educated, and more likelyo be White than those in W3. Learners in W1 and W2 were also

ore likely to have a personal or familial history of heart diseaser heart disorder.

Table 2 details the mean scores and standard deviations of thenal items in the 3 attitude scales, and the internal consistency ofach scale. Items that were on the questionnaire, but ultimately notncluded in the scales, are displayed for completeness. The final con-dence scale comprised 4 items, with alpha = .699 for W1 and W2nd alpha = .631 for W3. The final perceived competence scale com-rised 2 items, with alpha = .808 for W1 and W2 and alpha = .748 for3. The final willingness scale comprised 2 items, with alpha = .905

or W1 and W2 and alpha = .926 for W3.Changes in attitudes toward CPR are detailed in Table 3 for

1 and W2. Significant increases in confidence, perceived compe-ence, and willingness occurred for all 3 training groups. Perceivedompetence was initially lowest and showed the greatest gain of

iated tasks for W1, W2, and W3.

the 3 attitudes. Similarly, the improvement in attitudes from pre-training to post-training in W3 was significant and competence was

Have you been told that you are at highrisk for a heart attack?

8.6 –

Has anyone with whom you live been toldthey are at high risk for heart attack?

10.5 –

a Ellipses indicate items that were not contained in the W3 questionnaire.

B. Lynch, E.L. Einspruch / Resuscitation 81 (2010) 568–575 571

Table 2Item-specific results of pre-training attitudes toward CPRa,b.

Scale (with correlation for each study)/item W1 and W2 (N = 822) W3 (N = 243)

Mean SD Mean SD

Confidence (alpha = .699, .631)I feel confident that I could perform CPR in a real emergency. 2.75 1.25 2.60 1.27In a real emergency I think I would be too nervous or scared to do CPR.c 3.78 0.99 3.61 1.09I feel confident that I can perform the skills of CPR on a manikin. 3.55 1.14 3.57 1.22I would probably do more harm than good if I tried to perform CPR in a real emergency.c 3.60 1.00 3.48 1.00

Competence (alpha = .808, .748)I know how to do [infant] CPR. 2.36 1.07 1.99 1.97I have the knowledge to teach another person how to do [infant] CPR. 1.77 0.95 1.73 .98CPR should only be taught by someone who has the proper certification.c – – – –There are many things I have to remember in order to do CPR effectively.c – – – –CPR is difficult to learn.c – – – –It is a good idea for people who know how to do CPR to teach others how to do it. – – – –

Willingness (alpha = .905, .926)I would be willing to perform CPR on a friend or family member [infant] in a real emergency. 4.06 1.18 3.69 1.28I would be willing to perform CPR on a stranger [infant] in a real emergency. 3.68 1.23 3.56 1.29I am willing to teach what I know about CPR to my friends or family. – – – –

a Scale: 1 = strongly disagree, 2 = disagree, 3 = unsure, 4 = agree, 5 = strongly agree.b Ellipses indicate items that were on the questionnaire but not included in the scale.c Item was reverse coded.

Table 3Change in attitudes toward CPR pre-training and post-training for subjects in W1and W2.

Scale/group Pre-training Post-training

N Mean SD N Mean SD

ConfidenceNo-training control 288 3.40 .75 292 4.32*** .52Self-training 368 3.44 .82 368 4.26*** .58Traditional 157 3.39 .82 158 4.26*** .58

CompetenceNo-training control 284 2.06 .90 285 4.08*** .75Self-training 360 2.11 .97 361 4.16*** .72Traditional 151 1.98 .85 152 4.05*** .78

WillingnessNo-training control 290 3.89 1.14 290 4.51*** .50

iecs

tcv

TC

Table 5Correlation between post-training attitudes toward CPR and CPR performancea.

Group/scale Correlation of attitude and objectivemanikin measure

Vent HP CD ACR

Subject in W1 and W2 studiesConfidence

All .106* .032 .136** .109**

No-training control .136* .108 .207*** .150**

Self-training .040 .046 −.048 .079Traditional .176 −.110 .167 .061

CompetenceAll .007 −.062 .076 −.048No-training control .027 .077 .127* .029Self-training −.016 −.122 −.039 −.094Traditional .169 .014 −.130 −.086

WillingnessAll .026 −.000 .132** .035No-training control .102 .110 .207*** .119*

Self-training −.013 .053 −.029 −.034Traditional −.043 −.118 −.107 −.002

Group/scale Correlation of attitude and objectivemanikin measure

Self-training 368 3.85 1.16 365 4.48*** .56Traditional 157 3.88 1.17 154 4.47*** .60

*** P < .001.

n between-group comparisons (P < .001). There was no significantffect in within-subjects comparisons of confidence, nor in anyomparisons involving the perceived competence or willingnesscales (all P values > .002).

Table 5 details the correlations between the attitude scales andhe objective measures. A few statistically significant, but weak,orrelations emerged. The percentage of ventilations with adequateolume was correlated with confidence (r = .106, P < .05) in W1 and

able 4hange in attitudes toward CPR pre-training and post-training for subjects in W3.

Scale/group Pre-training Post-training

N Mean SD N Mean SD

ConfidenceSelf-training 117 3.30 .77 116 4.29*** .51Traditional 123 3.33 .85 125 4.42*** .540

CompetenceSelf-training 113 1.89 .91 114 4.28*** .64Traditional 121 1.82 .84 122 4.42*** .60

WillingnessSelf-training 116 3.71 1.23 115 4.49*** .50Traditional 122 3.55 1.25 125 4.52*** .50

*** P < .001.

Vent ACD CFR

Subjects in W3 studyConfidence

All .189** −.025 .020Self-training .247* .060 .040Traditional .125 −.031 −.035

CompetenceAll .144* .027 .043Self-training .135 .123 .055Traditional .154 −.006 .006

WillingnessAll −.027 .082 .043Self-training −.019 .185 .030Traditional .081 .031 .046

a Vent, percent ventilations with adequate volume; HP, percent compressionswith proper hand placement; CD, percent compressions with adequate depth; ACR,average compression rate; Vent, average ventilation volume; ACD, average com-pression depth; CFR, percent compressions with full release.

* P < .05.** P < .01.

*** P < .001.

572 B. Lynch, E.L. Einspruch / Resuscitation 81 (2010) 568–575

Table 6Comparison of CPR performance by post-training attitude levels for subjects in W1 and W2.

Scale Objective measure asrecorded by manikin

Percent ventilations withadequate volume

Percent compressionswith proper handplacement

Percent compressionswith adequate depth

Average compressionrate

Mean SD Mean SD Mean SD Mean SD

Confidencea

Low (N = 285) 55.64 40.12 84.93 26.49 63.15 41.89 92.93 22.91Medium (N = 299) 64.75 39.32 84.92 26.58 63.87 41.66 97.21 22.61High (N = 229) 70.46 38.05 85.50 26.64 68.98 41.00 95.46 22.59

F = 9.44 F = 0.04 F = 1.44 F = 2.61P < .001 P = .962 P = .237 P = .074

Competenceb

Low (N = 325) 61.41 40.03 85.12 25.43 64.81 41.34 95.98 23.05Medium (N = 289) 64.61 39.32 84.92 27.63 65.28 52.10 96.14 22.80High (N = 181) 64.86 39.43 85.31 27.02 66.90 40.59 91.94 24.65

F = 0.67 F = 0.01 F = .015 F = 2.18P = .514 P = .987 P = .859 P = .114

Willingnessc

Low (N = 479) 58.58 40.58 84.35 27.22 62.18 41.73 92.73 23.23Medium (N = 91) 69.45 36.73 87.08 22.27 66.70 42.56 98.70 22.66High (N = 245) 69.53 37.97 85.31 27.17 69.56 40.83 97.74 23.25

F = 7.60 F = 0.43 F = 2.65 F = 5.17P = .001 P = .649 P = .071 P = .006

WP(Ht

agaT

TC

a Wilk’s lambda = .97, F = 2.90, P = .003.b Wilk’s lambda = .99, F = 0.87, P = .543.c Wilk’s lambda = .97, F = 2.81, P = .004.

2 and with confidence (r = .189, P < .01) and competence (r = .144,< .05) in W3. Compression depth was correlated with confidence

r = .136, P < .01) and willingness (r = .132, P < .01) in W1 and W2.and placement for compressions was not correlated with the atti-

ude scales.To further explore the relationship between the attitude scales

nd objective measures, MANOVA was used to compare the 3roups on their tercile scores within each scale. The results for W1nd W2 are shown in Table 6 and the results for W3 are shown inable 7. These results mirror those shown in Table 5. For W1 and

able 7omparison of CPR performance by post-training attitude levels for subjects in W3.

Scale Objective measure as recorded by manikin

Average ventilation volume

Mean SD

Confidencea

Low (N = 83) 25.42 18.96Medium (N = 81) 24.00 19.47High (N = 46) 30.91 18.22

F = 2.02P = .135

Competenceb

Low (N = 71) 21.92 19.66Medium (N = 78) 28.81 18.48High (N = 56) 28.59 18.30

F = 3.02P = .051

Willingnessc

Low (N = 76) 27.88 18.64Medium (N = 66) 23.24 18.81High (N = 66) 26.64 20.11

F = 1.09P = .340

a Wilk’s lambda = .98, F = .74, P = .612.b Wilk’s lambda = .97, F = 1.05, P = .389.c Wilk’s lambda = .97, F = .89, P = .504.

W2, those with higher scores on the confidence scale had a greaterpercentage of ventilations with adequate volume (F = 9.44, P < .001),and those with higher scores on the willingness scale had a greaterpercentage of ventilations with adequate volume (F = 7.60, P = .001)and a higher average compression rate (F = 5.17, P = .006). For W3,those with higher scores on the competence scale tended to have a

greater percentage of ventilations with adequate volume (F = 3.02,P = .051).

Focus group participants’ 1-word descriptions of their overallexperience in CPR self-training were strongly positive: of the 48

Average compression depth Percent compressions withfull release

Mean SD Mean SD

20.88 7.17 60.83 30.4220.12 7.62 61.33 32.0021.15 8.76 60.72 30.31F = 0.32 F = 0.01P = .724 P = .992

20.34 8.09 59.97 33.0920.85 7.73 59.62 30.4920.84 7.27 62.30 28.72F = 0.10 F = 0.14P = .906 P = .872

20.63 6.77 61.96 29.4619.80 7.99 57.86 31.6821.55 8.47 63.14 31.59F = 0.84 F = 0.54P = .433 P = .586

esusci

rm(“

tcpra“hcw

lwtmtTwwnfhh

iOpswwmirtbi

4

tawfilatcevsmpa

wdtsme

Deck for statistical assistance, and Lori Sievert for administrativesupport.

B. Lynch, E.L. Einspruch / R

ecorded descriptions, 40 were categorized as positive (e.g., “infor-ative,” “eye-opening,” “interesting,” “practical”); 4 as neutral

e.g., “rigorous,” “adequate”); and 4 as negative (e.g., “repetitive,”simplistic”).

When asked to discuss the narrator, participants’ commentsended to emphasize aspects of performance, competence, andlarity (30 comments) or personality (20 comments), rather thanhysical characteristics and appearance (2 comments). The nar-ator was regarded very positively (48 positive comments, suchs “clear,” “calming,” “not distracting,” “good,” “competent,” andenthusiastic”); negative comments were rare (4 comments: “badair,” “lacked personality,” too loud and energetic, and “almost tooheerful”). Three neutral comments concurred that the narratoras “not irritating.”

When asked to describe points of confusion in the video, theargest category of responses (with 29 of 67 comments) had to do

ith practical aspects of using the kit (e.g., unsure whether to playhe video before assembling other kit materials; how to inflate the

anikin; whether it was permissible to pause the video; whereo place the manikin in relation to the video monitor or oneself).he next largest category of responses (22 of 67 comments) dealtith confusion about why certain topics were not covered (e.g.,hat to do if alone, what to do about a choking adult, and whyo pulse-checking was advised). Only 8 comments expressed con-

usion about the skills of CPR themselves (e.g., uncertainty aboutand placement or depth for compressions, or about how to tilt theead and give ventilations).

Participants were also asked to comment on whether the train-ng had given them the skills to perform CPR in a real situation.f the 28 comments given, 12 were classified as unambiguouslyositive and 11 of these emphasized either the efficiency of thehort format or the amount of practice offered. Thirteen commentsere classified as positive, but with reservations. Common themesithin these comments were: the training was good, but thereust be more to it; or it was better than nothing, but more train-

ng was probably needed. Three slightly negative comments wereecorded: two participants felt that the presence of a live instruc-or to answer questions and otherwise provide support would haveeen an improvement, and one participant believed that the learn-

ng would be forgotten with time.

. Discussion and conclusions

The results converge to support the conclusion that CPR self-raining can produce the same type and level of attitude changes does a traditional, instructor-led course. A secondary findingas that participants’ own assessments of their competence, con-dence, and willingness to perform adult CPR tended to be weakly

inked to objective measures of compression depth and rate –nd, to a lesser extent – to ventilation performance; however,his correlation was almost entirely attributable to the untrainedontrols. One possible explanation for this finding is that thencouragement participants receive in training, whether throughideo self-instruction or a traditional class, makes them less sen-itive to their actual performance. Conversely, controls may beore likely to realize that, without the benefit of training, their

erformance is poor; therefore, they may more easily adjust theirttitudes to reflect this reality.

Two of the results are counterintuitive: controls, who merelyere tested and not trained, did express improved levels of confi-

ence; and in spite of self-trained participants’ expressed concernshat the training they were receiving might have been missingome important elements, these participants exhibited just asuch positive attitude change as did traditionally trained learn-

rs.

tation 81 (2010) 568–575 573

Crider’s11 results may help explain these findings. In Crider’sstudy, mere exposure to an emergency, either as victim or res-cuer, increased both confidence and willingness to help in asimilar emergency. In our study, control participants completedpre-questionnaires and then were confronted with a simulatedemergency and asked to do whatever they felt was best for thevictim. Nearly everyone in this group attempted CPR. If Crider’s11

findings apply more generally to the malleability of attitudes withmere exposure to emergencies, the test session may have beenenough to convince participants that they could perform – or atleast try to perform – CPR in a real emergency. The fact that no feed-back on performance was given by examiners in the test sessionsmay also have helped participants maintain a focus on the impor-tance of merely doing something, rather than doing all the skillscorrectly. Discovering whether this speculative explanation is thecorrect one, however, is perhaps less important than understandingthe implications of our and Crider’s11 findings. Learning programsthat provide brief exposure to simulated emergency situations,along with brief training on how to handle such emergencies, maybe the most efficient and effective combination for ensuring thatlay bystanders feel confident, competent and willing to help in areal emergency.

The study has 3 limitations worth noting. First, because theoriginal and primary purpose of data collection in W1 and W2was to determine the efficacy of this self-training program forolder learners, most of our participants were over 40 years of age.Although similar results were seen with the younger learners inW3, the overall sample is still skewed toward older learners. Sec-ond, our participants were paid for their involvement in the studyand thus may have felt some obligation to behave in ways thatthey thought would please us. However, this explanation wouldnot predict a positive attitude change in controls. Third, only asingle type of self-training program was investigated, so it is notknown whether the same results would obtain with other pro-grams.

As the literature has shown, positive attitudes toward a par-ticular helping behavior, whether these attitudes are instilledvia instruction or mere exposure, can increase the likelihood ofengaging in the behavior. Consistent with the assertions of virtuallearning scholars over the past few decades, the medium appearsto be less important than the message: CPR self-training can rein-force the positive attitudes that improve the likelihood of learners’stepping forward in an emergency.

Conflict of interest statement

The study was funded by the American Heart Association andLaerdal Medical. B. Lynch has served as an independent consultantto the American Heart Association. E. Einspruch has no conflicts ofinterest to declare.

The funders contributed input at the design stage of the study,but did not participate in or otherwise influence study at any otherstage.

Acknowledgments

The authors gratefully acknowledge M.E. Mancini and an anony-mous reviewer for helpful comments on the manuscript, Dennis

Appendix A. Infant CPR study participant questionnaire

See Tables A1 and A2.

574 B. Lynch, E.L. Einspruch / Resuscitation 81 (2010) 568–575

Table A1Please provide information about yourself by completing the following items. Please do not leave any items blank. If you have a question or concern about any of the items,ask the study coordinator.

B. Lynch, E.L. Einspruch / Resuscitation 81 (2010) 568–575 575

Table A2Please CIRCLE the response to indicate how much you agree or disagree with each of the following statements. Circle only one number and do not circle between numbers.

Strongly agree Agree Not sure Disagree Strongly disagree

I know how to do infant CPR. 5 4 3 2 1I have the knowledge to show another person how to do infant CPR. 5 4 3 2 1Infant CPR should only be taught by someone who has proper certification. 5 4 3 2 1I feel confident that I could perform infant CPR in a real emergency. 5 4 3 2 1I would be willing to perform CPR on a friend or family member’s infant in a real emergency. 5 4 3 2 1I would be willing to perform CPR on a stranger’s infant in a real emergency. 5 4 3 2 1In a real emergency, I think I would be too nervous or scared to do infant CPR. 5 4 3 2 1I am willing to share what I know about infant CPR with my friends or family. 5 4 3 2 1There are many things I have to remember in order to do infant CPR effectively. 5 4 3 2 1I feel confident that I can perform the skills of infant CPR on a manikin. 5 4 3 2 1Infant CPR is difficult to learn. 5 4 3 2 1I would probably do more harm than good if I tried to perform Infant CPR in a real emergency. 5 4 3 2 1

do it.

T

R

1

1

1

1

1

1

11996.

It’s a good idea for people who know how to do infant CPR to show others how toI know what to do to help an infant who is choking.

HANK YOU! The study coordinator will collect this form.

eferences

1. Lynch B, Einspruch EL, Nichol G, Becker LB, Aufderheide TP, Idris A. Effective-ness of a 30-min CPR self-instruction program for lay responders: a controlledrandomized study. Resuscitation 2005;67:31–43.

2. Wik L, Thowsen J, Steen PA. An automated voice advisory manikin system fortraining in basic life support without an instructor. A novel approach to CPRtraining. Resuscitation 2001;50:167–72.

3. Batcheller AM, Brennan RT, Braslow A, Urrutia A, Kaye W. Cardiopulmonaryresuscitation performance of learners over forty is better following half-hourvideo self-instruction compared to traditional four-hour classroom training.Resuscitation 2000;43:101–10.

4. Shotland RL, Heinold WD. Bystander response to arterial bleeding: helping skills,the decision-making process, and differentiating the helping response. J Pers SocPsychol 1985;49:347–56.

5. Swor R, Khan I, Domeier R, Honeycutt L, Chu K, Compton S. CPR training andCPR performance: do CPR-trained bystanders perform CPR? Acad Emerg Med2006:596–601.

6. McCormack AP, Damon SK, Eisenberg MS. Disagreeable physical characteristics

affecting bystander CPR. Ann Emerg Med 1989;18:283–5.

7. Lester CA, Donnelly PD, Assar D. Lay CPR trainees: retraining, confidenceand willingness to attempt resuscitation 4 years after training. Resuscitation2000;45:77–82.

8. Smith KL, Cameron PA, Meyer AD, McR, McNeil JJ. Is the public equipped to actin out of hospital cardiac emergencies? Emerg Med J 2003;20:85–7.

1

1

5 4 3 2 15 4 3 2 1

9. Pane GA, Salness KA. A survey of participants in a mass CPR training course. AnnEmerg Med 1987;16:1112–6.

0. Villaincourt C, Stiell IG, Wells GA. Understanding and improving low bystanderCPR rates: a systematic review of the literature. CJEM 2008;10:51–65.

1. Crider D. A study of the attitudes and behaviors of potential rescuers and theirwillingness to respond in an emergency situation. Unpublished doctoral thesis;1996. Pennsylvania State University.

2. Midlarsky E. Aiding under stress: the effects of competence, dependency, visi-bility, and fatalism. J Pers 1971;39:132–49.

3. Kazdin AE, Bryan JH. Competence and volunteering. J Exp Soc Psychol1971;7:87–97.

4. Bernard RM, Abrami PC, Lou Y, et al. How does distance education compare withclassroom instruction? A meta-analysis of the empirical literature. Rev Educ Res2004;74:379–439.

5. Wang N, Johnson WL, Mayer RE, Rizzo P, Shaw E, Collins H. The politenesseffect: pedagogical agents and learning outcomes. Int J Hum Comput Stud2008;66:98–112.

6. Reeves B, Nass C. The media equation. New York: Cambridge University Press;

7. Vincent DS, Sherstyuk A, Burgess L, Connolly KK. Teaching mass casualty triageskills using immersive three-dimensional virtual reality. Acad Emerg Med2008;15:1160–5.

8. Youngblood P, Hedman L, Creutzfeld J, et al. Virtual worlds for teaching the newCPR to high school students. Stud Health Technol Inform 2007;125:515–9.