a b s t r a c t

Background: We hypothesized that recent hands-on practice for cardiopulmonary resuscitation (CPR) would be strongly associated with a higher likelihood of self-efficacy in bystander CPR among laypersons according to age and gender group.

Methods: We used the National Korean Community Health Survey database of 228921 representatively sampled responders from 253 counties in 2012. Laypersons who had previous CPR training were eligible. Exposure vari- ables were having had CPR training with hands-on practice session with a manikin (Practical-CPR-Training) and CPR training within the last 2 years (Recent-CPR-Training). Primary outcome was self-efficacy in bystander CPR. Multivariable logistic regression analysis was performed. The final model with an interaction term was evaluated to compare the effects of CPR training across different age and gender groups.

Results: Of 62425 eligible respondents who have had CPR training, 20213 (32.4%) had Practical-CPR-Training. Adjusted odds ratios (AORs) for self-efficacy were 4.08 (3.78-4.41) in Practical-CPR-Training, 2.61 (2.50-2.73)

in male, 1.26 (1.16-1.36) in good self-rated health, 1.19 (1.10-1.29) in high school graduate, 1.19 (1.01-1.39)

in persons living with stroke patients in household, and 1.17 (1.10-1.24) in Recent-CPR-Training. In interaction models, Practical-CPR-Training showed higher self-efficacy in all age and gender groups, whereas Recent-CPR- Training was not associated with better self-efficacy in elderly group, male (AOR, 0.90 [0.69-1.18]) and female (AOR, 0.94 [0.72-1.23]).

Conclusion: Self-efficacy in bystander CPR was higher in person with recent CPR training with hands-on practice with a manikin.

(C) 2015

Introduction

Early bystander cardiopulmonary resuscitation (CPR) is vital to initiate the Chain of survival for out-of-hospital cardiac arrest (OHCA) and to im- prove neurologic outcomes [1,2]. Increasing rates of early provision of by- stander CPR have been observed in a number of communities through vigorous efforts of public advocacy, campaign, and training programs of CPR, leading to improved survival outcomes of OHCA patients [1-3].

Bystander CPR provision depends on the availability of a person competent to attempt CPR near the incident site of arrest. Therefore, the probability of bystander CPR would increase along with a higher

* Corresponding author at: Department of Emergency Medicine, Seoul National Univer- sity College of Medicine, 101 Daehak-Ro, Jongno-Gu, Seoul 110-744, Republic of Korea. Tel.: +82 2 2072 3257; fax: +822 741 7855.

E-mail addresses: [email protected] (Y.S. Ro), [email protected] (S.D. Shin), [email protected] (K.J. Song), [email protected] (S.O. Hong), [email protected] (Y.T. Kim), [email protected] (S.-I. Cho).

number of community members competent to perform CPR. Several strategies have been developed based on health behavior theories to en- courage laypersons to participate in CPR training program [4,5]. Fur- thermore, various CPR education programs from brief intervention to mass CPR training with hands-on practice are implemented for public awareness and competency [6,7].

Cardiopulmonary resuscitation self-efficacy is refers to one’s confi- dence in his or her ability to save one’s life with CPR and to perform by- stander CPR when in an emergency situation [4,8,9]. Public self-efficacy in CPR is known to be associated with bystander CPR rate in community as well as with OHCA Survival outcomes. Cardiopulmonary resuscitation self-efficacy can be determined by a variety of factors including CPR training method and quality, individual and household characteristics, cultural sensitivity, and medico-legal environment [10,11]. Internation- al guidelines recommend basic life support training to be repeated every 2 years and to include a hands-on practice session with a manikin, where the effectiveness of the training can be evaluated by measuring CPR self-efficacy in trainees [6].

http://dx.doi.org/10.1016/j.ajem.2015.12.001

0735-6757/(C) 2015

Improving laypersons’ self-efficacy in CPR performance is crucial to in- crease bystander CPR rate [6,12]. We hypothesized that recent CPR train- ing with hands-on practice would be strongly associated with a higher likelihood of having self-efficacy in bystander CPR provision among lay- persons, and the effects of CPR training would vary in magnitude across different age and gender groups. This study aims to assess whether CPR training is associated with self-efficacy in provision of bystander CPR and to compare the findings according to age and gender groups.

Methods

Study setting

Korea has 253 county health authorities serving a population of 50 million. The national CPR training program launched in the early 2000s and has been actively promoted ever since by the Korea Centers for Disease Control and Prevention in collaboration with many academ- ic societies. The emergency medical service division of the Ministry of Health and Welfare financially supports 16 provincial governments to conduct CPR campaigns and provide courses. The ministry collects in- formation on the number of annual trainees in each province since 2010; the total number of CPR trainees under government support was 186589 in 2010, 159455 in 2011, and 223952 in 2012. In addition to governmental efforts, hospitals, nongovernmental organizations, ac- ademic and scientific societies, fire departments, and Red Cross also pro- vide CPR training. Each provincial government has its own program and criteria developed according to the national standards for training sites and training instructors as recommended by the Korea Centers for Dis- ease Control and Prevention [13].

Study design and data source

This is a cross-sectional study using the Korean Community Health Survey (CHS) database, which is a nationwide, community-based, cross-sectional household-level survey, conducted by the Korea Centers for Disease Control and Prevention between September and November of 2012. The Korean CHS began in 2008 and has been conducted annu- ally by 253 county health authorities.

A total of 228 921 participants, who were members of representative households selected to participate in 253 counties, completed the survey in 2012. The participants were selected through probability proportional to size sampling and systematic sampling method. An average of 920 adults older than 19 years of age participated in the survey in each county. The survey was conducted as a face-to-face, computer-assisted interview and was administered by trained interviewers using structured survey papers with a strong Quality management program and survey protocol [14]. The survey includes 247 questionnaire items and aims to obtain health-related information on targeted acute and chronic diseases, health care utilization, health behavior, quality of life, socioenvironmental fac- tors, and basic demographic information of the respondents.

Study population

All respondents who had CPR training experience were enrolled in the study. Cardiopulmonary resuscitation training experience was defined as having participated in 1 or more CPR classes with composed cluster training for 40 minutes or longer.

Main outcomes and variable measurements

The primary end point was self-efficacy in provision of bystander CPR. We collected CPR training information in regard to having had CPR train- ing experience within the last 2 years (Recent-CPR-Training) and wheth- er the training included hands-on practice session with a manikin (Practical-CPR-Training). We collected information on age, gender, socio- economic position (economic activity, occupation, education level, and

marital status), medical history (diabetes mellitus, hypertension, ischemic heart disease [acute myocardial infarction or angina pectoris], stroke, and cancer), self-rated health status, and health care center visits for public health education programs regarding hypertension or diabetes or others, and household health information (any members in the household with ischemic heart disease, stroke, or cancer, number of generations living in the same household if multigenerational, household annual income, and residence in metropolitan city). Medical history and the ill household member living with the survey respondent were recorded as a dichoto- mous variable upon physician-confirmed diagnosis. Occupation was categorized as nonmanual worker, manual worker, and others; others included no occupation, students, house wife, and others. Economic activity is defined as having had any paid work for more than 1 hour in the last 1 week.

Statistical analysis

Descriptive analysis was performed to examine the distribution of study variables; the distributions of categorical variables are reported as percentages. Categorical variables were compared using the ?² test. We used Multivariable logistic regression models to determine the associations of CPR training, individual factors, and household factors with the study outcome. Adjusted odds ratios (AORs) with 95% confi- dence intervals (95% CIs) of CPR training (Recent-CPR-Training and Practical-CPR-Training) for the study outcome were calculated for all el- igible population after controlling for potential confounders. Then, all terms in the models were assessed for multicollinearity to evaluate high correlation between 2 or more predictor variables. No multicollinearity was detected in this analysis. To calculate the AOR (95% CIs) of Recent-CPR-Training and Practical-CPR-Training according to gender and age (young, 19-44 years old; middle aged, 45-64 years old; and elderly, 65-107 years old), we used interaction models with in- teraction terms (CPR training x age and CPR training x gender) as the final multivariable logistic regression models for the outcomes. All sta- tistical analyses were performed using SAS software, version 9.4 (SAS Institute, Inc, Cary, NC). P values were based on a 2-sided significance

level of .05.

Ethics statements

The study was approved by the Institutional Review Board of the Korea Centers for Disease Control and Prevention (IRB no. 2012- 07CON-01-2C).

Results

Recent CPR training with hands-on practice

Among 228 921 participants for the 2012 Korean CHS, 62 425 (27.3%) reported to having had CPR training experience. Demographics for all participants are reported in the Appendix Table.

Table 1 shows the demographic characteristics for all eligible sub- jects and the elderly according to Practical-CPR-Training. Of 62425 re- spondents with CPR training, 20 213 (32.4%) had Practical-CPR- Training, whereas 316 (15.1%) of 2088 elderly respondents experienced Practical-CPR-Training. Participants who were younger, male, nonman- ual workers, high school graduate or some college, never married, living with 2 generations, with high household annual income, had good self- rated health, and visited health care center for public health education programs were more likely to participate Practical-CPR-Training (all P b .01). Participants who had known diagnosis of comorbidities or lived with ill household member(s) were less likely to have attended Practical-CPR-Training.

Table 1

Demographics for all eligible population and the elderly stratified by CPR training with hands-on practice session within the last 2 years

	All eligible		P	Elderly (age 65-107 y)	P
	Total	Practical-CPR-Training		Practical-CPR-Training

	N (%)		Yes, n (%)	No, n (%)		Yes, n (%)	No, n (%)
Total	62425		20213	42212		316	1772
Age					b.01			b.01
Young (19-44 years old)	40869 (65.5)		14458 (71.5)	26411 (62.6)
Middle aged (45-64 years old)	19468 (31.2)		5439 (26.9)	14029 (33.2)
Elderly (65-107 years old)	2088 (3.3)		316 (1.6)	1772 (4.2)
Gender					b.01
Male	41475 (66.4)		13658 (67.6)	27817 (65.9)		199 (63.0)	1425 (80.4)
Economic activity
Yes	49078 (78.6)	15885 (78.6)		33193 (78.6)	.90	186 (58.9)	812 (45.8)	b.01
Occupation					b.01			b.01
Nonmanual worker	21932 (35.1)	7459 (36.9)		14473 (34.3)		39 (12.3)	129 (7.3)
Manual worker	27067 (43.4)	8397 (41.5)		18670 (44.2)		148 (46.8)	681 (38.4)
Others	13426 (21.5)	4357 (21.6)		9069 (21.5)		129 (40.8)	962 (54.3)
Education					b.01			.05
Bachelor’s degree or higher	19659 (31.5)	6218 (30.8)		13441 (31.8)		51 (16.1)	369 (20.8)
High school graduate or some college	36931 (59.2)	12407 (61.4)		24524 (58.1)		101 (32.0)	603 (34.0)
Less than high school graduate	5835 (9.3)	1588 (7.9)		4247 (10.1)		164 (51.9)	800 (45.1)
Marital status					b.01			.14
Married, living with spouse	40995 (65.7)	11969 (59.2)		29026 (68.8)		249 (78.8)	1477 (83.4)
Married, not living with spouse	3611 (5.8)	923 (4.6)		2688 (6.4)		65 (20.6)	287 (16.2)
Never married	17819 (28.5)	7321 (36.2)		10498 (24.9)		2 (0.6)	8 (0.5)
Medical history
Hypertension	6729 (10.8)	1793 (8.9)		4936 (11.7)	b.01	165 (52.2)	866 (48.9)	.27
Diabetes mellitus	2320 (3.7)	572 (2.8)		1748 (4.1)	b.01	50 (15.8)	316 (17.8)	.39
Ischemic heart disease	712 (1.1)	149 (0.7)		563 (1.3)	b.01	19 (6.0)	178 (10.0)	.02
Stroke	281 (0.5)	41 (0.2)		240 (0.6)	b.01	7 (2.2)	91 (5.1)	0.02
Cancer	1035 (1.7)	253 (1.3)		782 (1.9)	b.01	15 (4.7)	152 (8.6)	.02
Self-rated health status					b.01			b.01
Good	32527 (52.1)	11477 (56.8)		21050 (49.9)		133 (42.1)	603 (34.0)
Fair	25113 (40.2)	7577 (37.5)		17536 (41.5)		121 (38.3)	676 (38.1)
Poor	4785 (7.7)	1159 (5.7)		3626 (8.6)		62 (19.6)	493 (27.8)
Health care center visits for public health education programs
Yes	1461 (2.3)	562 (2.8)		899 (2.1)	b.01	33 (10.4)	108 (6.1)	b.01
Living with ill household member(s)
Ischemic heart disease	2274 (3.6)	663 (3.3)		1611 (3.8)	b.01	39 (12.3)	280 (15.8)	.12
Stroke	1355 (2.2)	408 (2.0)		947 (2.2)	.07	21 (6.6)	135 (7.6)	.55
Cancer	3411 (5.5)	1010 (5.0)		2401 (5.7)	b.01	32 (10.1)	224 (12.6)	.21
Multigenerational household					b.01			.13
Single generation	13527 (21.7)	4140 (20.5)		9387 (22.2)		218 (69.0)	1262 (71.2)
2 generations	42328 (67.8)	13903 (68.8)		28425 (67.3)		79 (25.0)	366 (20.7)
3 generations	6570 (10.5)	2170 (10.7)		4400 (10.4)		19 (6.0)	144 (8.1)
Household annual income, US $					b.01			.33
0-19999	8047 (12.9)	2364 (11.7)		5683 (13.5)		168 (53.2)	857 (48.4)
20000-39999	21127 (33.8)	6509 (32.2)		14618 (34.6)		85 (26.9)	555 (31.3)
40000-59999	13345 (21.4)	4465 (22.1)		8880 (21.0)		18 (5.7)	118 (6.7)
60000-	16686 (26.7)	5865 (29.0)		10821 (25.6)		32 (10.1)	152 (8.6)
Unknown	3220 (5.2)	1010 (5.0)		2210 (5.2)		13 (4.1)	90 (5.1)
Residential area					.05			.02
Metropolitan city	21758 (34.9)	6936 (34.3)		14822 (35.1)		120 (38.0)	559 (31.5)

Self-efficacy on provision of bystander CPR

Of 62425 respondents who experienced CPR training, 48701 (78.0%) re- ported self-efficacy in provision of bystander CPR. The individual and house- hold characteristics of the study population according to the CPR self-efficacy are shown in Table 2. Practical-CPR-Training had the highest AOR (95% CI) for self-efficacy in bystander CPR, which was 4.08 (3.78-4.41). Recent-CPR- Training (AOR, 1.17; 1.10-1.24; reference, CPR training before longer than 2 years), male (AOR, 2.61; 2.50-2.73; reference, female), middle age (AOR, 1.56; 1.39-1.75; reference, elderly), bachelor’s degree or higher ed- ucation level (AOR, 1.17; 1.07-1.27; reference, less than high school grad- uate), high school graduate of some college education level (AOR, 1.19; 1.10-1.29), and good self-rated health status (AOR, 1.26; 1.16-1.36; refer- ence, poor self-rated health) were more likely to have self-efficacy in pro- vision of bystander CPR. Stroke patients, never married, and married and living with spouse subjects were less likely to have self-efficacy. Among household factors, living with a stroke patient under 1 roof was positively associated with self-efficacy in bystander CPR (AOR, 1.19; 1.01-1.39).

The factors associated with CPR self-efficacy among the elderly pop- ulation are shown in Table 3. Practical-CPR-Training had the highest AOR (95% CI) in the elderly, which was 4.66 (2.87-7.56). Male (AOR, 1.66; 1.28-2.15), bachelor’s degree or higher education level (AOR, 1.77; 1.30-2.42), and high school graduate or some college education level (AOR, 1.66; 1.31-2.10) were more likely to have self-efficacy, whereas stroke patient (AOR, 0.36; 0.16-0.80) and other occupation (AOR, 0.66; 0.53-0.83) were less likely.

Interaction analysis

We observed a wide range in the magnitude of CPR training effect on self-efficacy across age and gender groups. Adjusted odds ratios (95% CI) of Practical-CPR-Training for self-efficacy of bystander CPR were 4.44 (3.88-5.07) in middle-aged female and 3.50 (2.41-5.08) in elderly male. Practical-CPR-Training had significant associations with self- efficacy in all age and gender groups. Recent-CPR-Training was associat- ed with higher self-efficacy in young and middle-aged groups. However,

Table 2

Cardiopulmonary resuscitation training, individual, and household factors by CPR self-efficacy among all eligible population

			CPR self-efficacy
	Total		Yes	No
	N		n (%)	n (%)	AOR (95% CI)
Total	62425		48701 (78.0)	13724 (22.0)
CPR training
Recent-CPR-Training	26384		23123 (87.6)	3261 (12.4)	1.17 (1.10-1.24)
Practical-CPR-Training	20213		18559 (91.8)	1654 (8.2)	4.08 (3.78-4.41)
Age
Young (19-44 years old)	40869	31425 (76.9)		9444 (23.1)	1.08 (0.96-1.22)
Middle aged (45-64 years old)	19468	15768 (81.0)		3700 (19.0)	1.56 (1.39-1.75)
Elderly (65-107 years old)	2088	1508 (72.2)		580 (27.8)	1.00
Gender
Male	41475	34748 (83.8)		6727 (16.2)	2.61 (2.50-2.73)
Occupation
Nonmanual worker	21932	17190 (78.4)		4742 (21.6)	1.00 (0.95-1.06)
Manual worker	27067	21809 (80.6)		5258 (19.4)	1.00
Others	13426	9702 (72.3)		3724 (27.7)	0.90 (0.85-0.95)
Education
Bachelor’s degree or higher	19659	15363 (78.1)		4296 (21.9)	1.17 (1.07-1.27)
High school graduate or some college	36931	28895 (78.2)		8036 (21.8)	1.19 (1.10-1.29)
Less than high school graduate	5835	4443 (76.1)		1392 (23.9)	1.00
Marital status
Married, living with spouse	40995	31962 (78.0)		9033 (22.0)	0.91 (0.83-0.99)
Married, not living with spouse	3611	2770 (76.7)		841 (23.3)	1.00
Never married	17819	13969 (78.4)		3850 (21.6)	0.88 (0.80-0.97)
Medical history
Hypertension	6729	5308 (78.9)		1421 (21.1)	0.97 (0.90-1.04)
Diabetes mellitus	2320	1830 (78.9)		490 (21.1)	1.00 (0.89-1.11)
Ischemic heart disease	712	566 (79.5)		146 (20.5)	1.11 (0.88-1.40)
Stroke	281	201 (71.5)		80 (28.5)	0.66 (0.48-0.91)
Cancer	1035	774 (74.8)		261 (25.2)	1.03 (0.86-1.24)
Self-rated health status
Good	32527	26238 (80.7)		6289 (19.3)	1.26 (1.16-1.36)
Fair	25113	18892 (75.2)		6221 (24.8)	0.98 (0.91-1.06)
Poor	4785	3571 (74.6)		1214 (25.4)	1.00
Health care center visits for public health education programs
Yes	1461	1165 (79.7)		296 (20.3)	1.06 (0.93-1.22)
Living with ill household member(s)
Ischemic heart disease	2274	1789 (78.7)		485 (21.3)	1.04 (0.91-1.18)
Stroke	1355	1062 (78.4)		293 (21.6)	1.19 (1.01-1.39)
Cancer	3411	2648 (77.6)		763 (22.4)	1.01 (0.91-1.12)
Multigenerational household
Single generation	13527	10650 (78.7)		2877 (21.3)	1.00
2 generations	42328	32916 (77.8)		9412 (22.2)	0.95 (0.90-1.00)
3 generations	6570	5135 (78.2)		1435 (21.8)	0.95 (0.88-1.03)
Household annual income, US $
0-19999	8047	6254 (77.7)		1793 (22.3)	1.00
20000-39999	21127	16465 (77.9)		4662 (22.1)	0.94 (0.88-1.00)
40000-59999	13345	10481 (78.5)		2864 (21.5)	0.95 (0.88-1.03)
60000-	16686	13097 (78.5)		3589 (21.5)	0.94 (0.87-1.01)
Unknown	3220	2404 (74.7)		816 (25.3)	0.78 (0.71-0.87)

the AORs were not significant in both elderly male (0.90; 0.69-1.18) and female (0.94; 0.72-1.23) (Table 4).

Discussion

This analysis of a nationwide community survey identified signifi- cant associations between CPR training method and self-efficacy in the provision of bystander CPR among adult laypersons as well as the elder- ly. Greater self-efficacy in bystander CPR was observed in persons who had CPR training with hands-on practice session within the last 2 years, and the effect of the CPR training did not vary across different age and gender groups. Recent CPR training was associated with higher self-efficacy in young and middle-aged populations, but not in elderly population. We observed differences in several individual and house- hold factors including comorbidities and socioeconomic position be- tween laypersons who were received different CPR training courses in terms of hands-on practice.

Repeated high-quality CPR trainings with hands-on practice or simula- tion sessions increase posttraining CPR self-efficacy [6,15]. In this study, re- cent CPR training with hands-on practice showed 4 times higher odds for self-efficacy in bystander CPR among those who had previous CPR training. The association was coincident among the elderly population. Cardiopul- monary resuscitation campaign and advocacy may vary in contents, methods, tools, or interval of repetition, which would impact individual’s self-efficacy in provision of bystander CPR differently. Having had recent CPR training is reported to result in better bystander CPR quality [12], and hands-on practice or simulation sessions enhanced laypersons to remem- ber the skills and procedures for a long time [16]. Previous studies reported that simple CPR advocacy and campaign can alone be effective for increas- ing bystander CPR rates [7,17,18]; however, the simple training did not yield high Quality of CPR in absence of dispatcher assistance [19,20]. Performing CPR in real-life emergencies is an extremely rare event for lay- persons; therefore, training methods beyond simple CPR must be advocat- ed such as inclusion of hands-on practice session with a manikin to increase the layperson’s self-efficacy or high-quality bystander CPR [6].

Table 3

Cardiopulmonary resuscitation training, individual, and household factors by CPR self-efficacy among the elderly population

			CPR self-efficacy
	Total		Yes	No
	N		n (%)	n (%)	AOR (95% CI)
Total	2088		1508 (72.2)	580 (27.8)
CPR training
Recent-CPR-Training	474		376 (79.3)	98 (20.7)	0.77 (0.54-1.11)
Practical-CPR-Training	316		279 (88.3)	37 (11.7)	4.66 (2.87-7.56)
Gender
Male	1624		1214 (74.8)	410 (25.2)	1.66 (1.28-2.15)
Occupation
Nonmanual worker	168		136 (81.0)	32 (19.0)	0.93 (0.59-1.46)
Manual worker	829		633 (76.4)	196 (23.6)	1.00
Others	1091		739 (67.7)	352 (32.3)	0.66 (0.53-0.83)
Education
Bachelor’s degree or higher	420		325 (77.4)	95 (22.6)	1.77 (1.30-2.42)
High school graduate or some college	704		542 (77.0)	162 (23.0)	1.66 (1.31-2.10)
Less than high school graduate	964		641 (66.5)	323 (33.5)	1.00
Marital status
Married, living with spouse	1726	1268 (73.5)		458 (26.5)	1.03 (0.78-1.36)
Married, not living with spouse	352	234 (66.5)		118 (33.5)	1.00
Never married	10	6 (60.0)		4 (40.0)	0.59 (0.15-2.30)
Medical history
Hypertension	1031	742 (72.0)		289 (28.0)	1.03 (0.84-1.27)
Diabetes mellitus	366	261 (71.3)		105 (28.7)	0.99 (0.75-1.29)
Ischemic heart disease	197	142 (72.1)		55 (27.9)	0.83 (0.47-1.45)
Stroke	98	59 (60.2)		39 (39.8)	0.36 (0.16-0.80)
Cancer	167	114 (68.3)		53 (31.7)	0.83 (0.45-1.54)
Self-rated health status
Good	736	573 (77.9)		163 (22.1)	1.22 (0.92-1.62)
Fair	797	559 (70.1)		238 (29.9)	0.93 (0.72-1.19)
Poor	555	376 (67.7)		179 (32.3)	1.00
Health care center visits for public health education programs
Yes	141	104 (73.8)		37 (26.2)	1.12 (0.74-1.69)
Living with ill household member(s)
Ischemic heart disease	319	237 (74.3)		82 (25.7)	1.45 (0.92-2.29)
Stroke	156	106 (67.9)		50 (32.1)	1.76 (0.87-3.53)
Cancer	256	181 (70.7)		75 (29.3)	1.08 (0.64-1.81)
Multigenerational household
Single generation	1480	1082 (73.1)		398 (26.9)	1.00
2 generations	445	324 (72.8)		121 (27.2)	1.02 (0.79-1.32)
3 generations	163	102 (62.6)		61 (37.4)	0.73 (0.49-1.07)
Household annual income, US $
0-19999	1025	736 (71.8)		289 (28.2)	1.00
20000-39999	640	480 (75.0)		160 (25.0)	1.02 (0.79-1.30)
40000-59999	136	95 (69.9)		41 (30.1)	0.80 (0.52-1.24)
60000-	184	132 (71.7)		52 (28.3)	0.88 (0.58-1.34)
Unknown	103	65 (63.1)		38 (36.9)	0.66 (0.42-1.03)

Higher public self-efficacy in CPR in communities can exert impres- sive impacts on the bystander CPR rate and clinical outcomes [3]. Self- efficacy is one of the factors that directly influence health behavior such as performing CPR, based on health behavior theories [4,5].

Table 4

Interaction analysis of CPR training on self-efficacy across different age and gender groups

Male Female

	AOR	95% CI		AOR	95% CI
Practical-CPR-Training Total	3.95	3.59-4.34	4.23		3.84-4.65
Young (19-44 years old)	3.91	3.55-4.31	4.16		3.74-4.61
Middle aged (45-64 years old)	4.18	3.62-4.82	4.44		3.88-5.07
Elderly (65-107 years old) Recent-CPR-Training Total	3.50 2.84	2.41-5.08 2.67-3.02	3.72 3.14		2.56-5.40 2.95-3.36
Young (19-44 years old)	1.15	1.06-1.24	1.20		1.10-1.31
Middle aged (45-64 years old)	1.17	1.05-1.31	1.23		1.10-1.37
Elderly (65-107 years old)	0.90	0.69-1.18	0.94		0.72-1.23

Adjusted for gender, age, occupation, education level, marital status, medical history, self- rated health status, health care center visits, living with ill household member, multigen- erational household, annual Household income, and interaction terms (CPR training x age and CPR training x gender).

Cardiopulmonary resuscitation self-efficacy interacts with attitudinal beliefs, social norms, and perceived benefits or risks, where all of afore- mentioned factors are influenced by CPR training method and quality, individual and household characteristics, cultural sensitivity, and medico-legal environment [8,10,11,21]. These findings suggest that strategies for improving bystander CPR should include comprehensive and systematic approaches to increase an individual’s intention to per- form bystander CPR [5].

For the elderly population, CPR training with hands-on practice in- creased self-efficacy in bystander CPR, whereas recent CPR training did not. Out-of-hospital cardiac arrest commonly occurs in elderly and at home, which in hand increases the chance of an elderly confronting an emergency situation having to perform bystander CPR to a patient with OHCA. Therefore, customized training programs and methods in- cluding hands-on practice session for the elderly population are needed to enhance self-efficacy.

In previous qualitative studies, disability and perceived physical lim-

itation were accounted as the barriers to providing CPR as well as at- tending CPR training [9]. In this study, stroke patients were less likely to have self-efficacy in bystander CPR, whereas people who lived with stroke patients under the same roof were more likely to have self- efficacy. Individual perceived risk of cardiac arrest would influence

one’s willingness to attend CPR training [4,5] and to have self-efficacy, especially when the risk was visualized through incidences such as a

Appendix Table. Demographics for all respondents stratified by previous CPR training

family member with disabilities and rehabilitation process. However, the proportions of CPR training and recent CPR training with hands-on practice were lower in people having comorbidities or living with Any CPR training experience
household members having comorbidities including ischemic heart dis- Total, N (%)		Yes, n (%)	No, n (%)
ease and cancer, except stroke. This result is in parallel with previous Total	228921	62425	166496
studies [22] despite specifically targeted CPR training program was of- Age fered for family members of high-risk patients of cardiac arrest such Young (19-44 years old)	85060 (37.2)	40869 (65.5)	44191 (26.5)
as ischemic heart disease [23]. Targeted CPR training and advocacy Middle aged (45-64 years old)	85757 (37.5)	19468 (31.2)	66289 (39.8)
Elderly (65-107 years old)	58104 (25.4)	2088 (3.3)	56016 (33.6)
should be more highlighted for the high-risk groups, and various tools Gender to visualize the risk would impact individual perceived risk. Male	102898 (44.9)	41475 (66.4)	61423 (36.9)
Higher education level was one of factors that influenced self- Economic activity
efficacy in provision of bystander CPR and was associated with recent Yes	145971 (63.8)	49078 (78.6)	96893 (58.2)
Occupation

their family would be beneficial especially in combination with contin-

with greater Willingness to perform CPR and ability to use automatic ex- Bachelor’s degree or higher		40832 (17.8)	19659 (31.5)	21173 (12.7)
High school graduate or		98200 (42.9)	36931 (59.2)	61269 (36.8)
ternal defibrillators properly [16,22]. some college
In Korea, public CPR education was led by national initiatives, which Less than high school graduate		89889 (39.3)	5835 (9.3)	84054 (50.5)
resulted in increased bystander CPR rates from 2.5% (between 2006 and Marital status 2010) [24] to 38.1% (between 2012 and 2013). Through the Korean CHS, Married, living with spouse 158618 (69.3)			40995 (65.7)	117623 (70.6)
Married, not living with spouse		37302 (16.3)	3611 (5.8)	33691 (20.2)
however, we found that the proportion of CPR trainees was 27.3% Never married		33001 (14.4)	17819 (28.5)	15182 (9.1)
(Appendix Table) and the proportion of CPR trainees who recently Medical history attended hands-on practice was only one-thirds of the total trained Hypertension		53049 (23.2)	6729 (10.8)	46320 (27.8)
population. Extensive and vigorous public campaigns should be pur- Diabetes mellitus		19958 (8.7)	2320 (3.7)	17638 (10.6)
sued including mandatory CPR education in elementary schools, man- Ischemic heart disease		6630 (2.9)	712 (1.1)	5918 (3.6)
Stroke		4475 (2.0)	281 (0.5)	4194 (2.5)
datory CPR education when acquiring a driver’s license, mass CPR Cancer		7759 (3.4)	1035 (1.7)	6724 (4.0)
training, and media advocacy. In addition, continuous medical educa- Self-rated health status tion with hands-on practice every 2 years should be recommended. Fur- Good		88928 (38.8)	32527 (52.1)	56401 (33.9)
thermore, targeted CPR education program for high-risk population and Fair		91472 (40.0)	25113 (40.2)	66359 (39.9)
Poor		48521 (21.2)	4785 (7.7)	43736 (26.3)
uous monitoring of CPR training experience and self-efficacy.	health education programs
	Yes	6296 (2.8)	1461 (2.3)	4835 (2.9)
	Living with ill household

Health care center visits for public

Study limitations

member(s) Ischemic heart disease	13670 (6.0)	2274 (3.6)	11396 (6.8)
Stroke	9260 (4.0)	1355 (2.2)	7905 (4.7)
Cancer	16371 (7.2)	3411 (5.5)	12960 (7.8)
Multigenerational household

The study has several limitations. First, we used self-reports to mea- sure self-efficacy in bystander CPR. Measuring the level of self-efficacy

in the context of a particular behavior has been shown to correlate	Single generation	86315 (37.7)	13527 (21.7)	72788 (43.7)
well with actually performing the behavior [8]. Nevertheless, self-	2 generations	118727 (51.9)	42328 (67.8)	76399 (45.9)
efficacy was not objectively measured and survey questionnaire may be interpreted differently by the respondents, which can result in bias. Second, pediatric population was not included in this study, and ex-	3 generations Household annual income, U 0-19999 20000-39999	23879 (10.4) S $ 70625 (30.9) 69424 (30.3)	6570 (10.5) 8047 (12.9) 21127 (33.8)	17309 (10.4) 62578 (37.6) 48297 (29.0)
treme elderly with cognitive impairment might not have been able to	40000-59999	34985 (15.3)	13345 (21.4)	21640 (13.0)
participate in this survey. This selection bias could have yielded under-	60000-	40607 (17.7)	16686 (26.7)	23921 (14.4)

estimation or overestimation.

Unknown	13280 (5.8)	3220 (5.2)	10060 (6.0)
Residential area
Metropolitan city	67852 (29.6)	21758 (34.9)	46094 (27.7)

Conclusions

Self-efficacy in bystander CPR was higher in persons with recent CPR training with hands-on practice with a manikin among laypersons as

CPR training

CPR training within the last 2 years

CPR training with hands-on practice

26384 (11.5) 26384 (42.3) 0 (0.0)

20213 (8.8) 20213 (32.4) 0 (0.0)

well as the elderly. Public health strategies to engage laypersons to par- ticipate in CPR training using manikins, continue medical education program, and provide targeted training for high-risk groups may help to increase self-efficacy and to enhance bystander CPR.

CPR training with hands-on practice in this study. In the elderly, the ed-	Nonmanual worker	42883 (18.7)	21932 (35.1)	20951 (12.6)
ucation level was strongly associated with self-efficacy after adjusting	Manual worker	102903 (45.0)	27067 (43.4)	75836 (45.5)
for other socioeconomic position such as occupation and household in- come. Previous studies found that high education level is associated	Others Education	83135 (36.3)	13426 (21.5)	69709 (41.9)

Funding acknowledgements

None.

Disclosures

None.

Abbreviation: CPR, cardiopulmonary resuscitation.

References

1. Sasson C, Meischke H, Abella BS, Berg RA, Bobrow BJ, Chan PS, et al. Increasing car- diopulmonary resuscitation provision in communities with low bystander cardio- pulmonary resuscitation rates: a science advisory from the American Heart Association for healthcare providers, policymakers, public health departments, and community leaders. Circulation 2013;127(12):1342-50.
2. Hasselqvist-Ax I, Riva G, Herlitz J, Rosenqvist M, Hollenberg J, Nordberg P, et al. Early cardiopulmonary resuscitation in out-of-hospital cardiac arrest. N Engl J Med 2015; 372(24):2307-15.
3. Abella BS, Aufderheide TP, Eigel B, Hickey RW, Longstreth Jr WT, Nadkarni V, et al. Reducing barriers for implementation of bystander-initiated cardiopulmonary re- suscitation: a scientific statement from the American Heart Association for

   healthcare providers, policymakers, and community leaders regarding the effective- ness of cardiopulmonary resuscitation. Circulation 2008;117(5):704-9.

   Sasson C, Haukoos JS, Eigel B, Magid DJ. The HANDDS program: a systematic ap- proach for addressing disparities in the provision of bystander cardiopulmonary re- suscitation. Acad Emerg Med 2014;21(9):1042-9.

4. Panchal AR, Fishman J, Camp-Rogers T, Starodub R, Merchant RM. An “Intention-Focused”

   paradigm for improving bystander CPR performance. Resuscitation 2015;88:48-51.

   Bhanji F, Mancini ME, Sinz E, Rodgers DL, McNeil MA, Hoadley TA, et al. Part 16: ed- ucation, implementation, and teams: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;122(18 Suppl. 3):S920-33.

5. Del Rios M, Aldeen A, Campbell T, Demertsidis E, Heinert S, Sinchi J, et al. Save-a-life at the ballpark: 10-min spectator training achieves proficiency in cardiac arrest re- sponse. Resuscitation 2015;87:21-5.
6. Vaillancourt C, Kasaboski A, Charette M, Islam R, Osmond M, Wells GA, et al. Barriers and facilitators to CPR training and performing CPR in an older popula- tion most likely to witness cardiac arrest: a national survey. Resuscitation 2013; 84(12):1747-52.
7. Vaillancourt C, Charette M, Kasaboski A, Brehaut JC, Osmond M, Wells GA, et al. Bar- riers and facilitators to CPR knowledge transfer in an older population most likely to witness cardiac arrest: a theory-informed interview approach. Emerg Med J 2014; 31(9):700-5.
8. Godfred R, Huszti E, Fly D, Nichol G. A randomized trial of video self-instruction in cardiopulmonary resuscitation for lay persons. Scand J Trauma Resusc Emerg Med 2013;21:36.
9. Batcheller AM, Brennan RT, Braslow A, Urrutia A, Kaye W. Cardiopulmonary resusci- tation performance of subjects over forty is better following half-hour video self- instruction compared to traditional four-hour classroom training. Resuscitation 2000;43(2):101-10.
10. Takei Y, Nishi T, Matsubara H, Hashimoto M, Inaba H. Factors associated with quality of bystander CPR: the presence of multiple rescuers and bystander-initiated CPR without instruction. Resuscitation 2014;85(4):492-8.
11. Lee MJ, Hwang SO, Cha KC, Cho GC, Yang HJ, Rho TH. Influence of nationwide policy on citizens’ awareness and willingness to perform bystander cardiopulmonary re- suscitation. Resuscitation 2013;84(7):889-94.
12. Oh DH, Kim SA, Lee HY, Seo JY, Choi BY, Nam JH. Prevalence and correlates of depres- sive symptoms in Korean adults: results of a 2009 Korean community health survey. J Korean Med Sci 2013;28(1):128-35.
13. Cho GC, Sohn YD, Kang KH, Lee WW, Lim KS, Kim W, et al. The effect of basic life sup- port education on laypersons’ willingness in performing bystander hands only car- diopulmonary resuscitation. Resuscitation 2010;81(6):691-4.
14. Swor R, Khan I, Domeier R, Honeycutt L, Chu K, Compton S. CPR training and CPR performance: do CPR-trained bystanders perform CPR? Acad Emerg Med 2006; 13(6):596-601.
15. Bobrow BJ, Vadeboncoeur TF, Spaite DW, Potts J, Denninghoff K, Chikani V, et al. The effectiveness of ultrabrief and brief educational videos for training lay responders in hands-only cardiopulmonary resuscitation: implications for the future of citizen car- diopulmonary resuscitation training. Circ Cardiovasc Qual Outcomes 2011;4(2): 220-6.
16. Berg RA, Hemphill R, Abella BS, Aufderheide TP, Cave DM, Hazinski MF, et al. Part 5: adult basic life support: 2010 American Heart Association Guidelines for Cardiopul- monary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;122(18 Suppl. 3):S685-705.
17. Birkenes TS, Myklebust H, Neset A, Kramer-Johansen J. Quality of CPR performed by trained bystanders with optimized pre-arrival instructions. Resuscitation 2014; 85(1):124-30.
18. Lee JS, Jeon WC, Ahn JH, Cho YJ, Jung YS, Kim GW. The effect of a cellular-phone video demonstration to improve the quality of dispatcher-assisted chest compression-only cardiopulmonary resuscitation as compared with audio coaching. Resuscitation 2011;82(1):64-8.
19. King R, Heisler M, Sayre MR, Colbert SH, Bond-Zielinski C, Rabe M, et al. Identifica- tion of factors integral to designing community-based CPR interventions for high- risk neighborhood residents. Prehosp Emerg Care 2015;19(2):308-12.
20. Weiner SG, Kapadia T, Fayanju O, Goetz JD. socioeconomic disparities in the knowl- edge of basic life support techniques. Resuscitation 2010;81(12):1652-6.
21. Chu KH, May CR, Clark MJ, Breeze KM. CPR training in households of patients with chest pain. Resuscitation 2003;57(3):257-68.
22. Ro YS, Shin SD, Song KJ, Lee EJ, Kim JY, Ahn KO, et al. A trend in epidemiology and out- comes of out-of-hospital cardiac arrest by urbanization level: a nationwide observa- tional study from 2006 to 2010 in South Korea. Resuscitation 2013;84(5):547-57.