Introduction

While a number of questionnaire measuring the various dimensions of bystander cardiopulmonary resuscitation (CPR) [1,2,3,4] and AED [2,3,4,5,6,7] have been published, there are few questionnaires that combine both bystander CPR and automated external defibrillator (AED) measurement [8, 9]. We reported the development and construct validation (exploratory factor analysis, EFA and confirmatory factor analysis, CFA) of a bilingual (English and local Malay language) questionnaire that comprehensively measures four objectives: (1) the perception of AED placement strategy, (2) the perception on the importance of bystander CPR and AED, (3) the perception on the confidence and willingness to apply these two lifesaving interventions as well as (4) the fears and concerns in applying these two interventions.

Main text

Methods

Participants

For EFA, 184 administrative employees from Universiti Malaysia Sarawak (UNIMAS) (who were participants of CPR and AED workshops), responded to the questionnaire. Healthcare employees such as doctors, nurses and paramedic staff were excluded from this study. The mean age of the participants was 37.6 years (standard deviations ± 6.85); and 100 of them (54.3%) were male participants. The number of participants who responded exceeded the estimated sample size according to Costello and Osborne [10] (i.e., minimum of 5 responses per item). As there were 22 items to be validated, the estimated sample size was 110 participants. For CFA, 100 medical doctors from the emergency department of Sarawak General Hospital (age range from 28 years to 32 years old) and 122 final year medical students from UNIMAS (age range from 23 to 25 years old) and who had been trained in CPR and AED, responded to the questionnaire. Convenient sampling was applied in recruiting the participants.

Materials

The EFA was performed in Statistical Package for the Social Sciences (SPSS) software using principal axis factoring as the extraction method. For CFA, reflective measurement modelling using partial least square was performed using SMART-PLS software. For translation of the questionnaire from English language into the local Malay language, the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) principles of good translation practice for cultural adaption was adopted as the working framework [11]. According to the ISPOR principles, a good translation has the following steps: (1) forward translation, (2) reconciliation (3) backward translation (4) harmonization [11]. In the forward translation stage, two authors who are fluent in both English and Malay language (SYL, DNK) independently translated the questionnaire into Malay language. After completion of the initial Malay language translation, these 2 authors of this paper then discussed and resolved any discrepancy of the translations (reconciliation). Subsequently, an independent language expert who is fluent in both languages were invited to translate the Malay language version back to English language (backward translation). Finally the original English version and the backward English version would be compared to check for significant discrepancy (“harmonization”).

Procedures

Ethical approval was obtained from the Medical Research and Ethics Committee, Ministry Of Health Malaysia and the study was registered under the Malaysian National Medical Research Register (NMRR, website URL: www.nmrr.gov.my) with the research number of NMRR-16-696-39041. Written informed consent was obtained from all participants prior to their participation in this study.

With regards to the development and construct validation of the questionnaire, a preliminary version of the checklist was first constructed based on previous works [1, 3,4,5,6,7,8,9, 12] as well as opinions from four authors of this paper (KSC, SYL, NS and DNK). A modified Delphi method via iterative email communications and face-to-face discussion sessions were conducted. The purpose of these discussions was to come up with a list of the pertinent items purported to measure (1) the perception of the importance of bystander CPR and public access AED, (2) the perception of the confidence and willingness of bystanders to apply these two lifesaving interventions as well as (3) fears and concerns of bystander that may deter their willingness and (4) the perception of the placement strategies of these AED.

Once the preliminary list of items was determined, the participants were asked to rank the items in a Likert scale of four, ranging from “1 = strongly disagree” to “4 = strongly agree”. An initial run of EFA was performed in order to determine the number of factors to be fixed (eigenvalue > 1). After fixing the number of factors, re-run of EFA was then performed to determine the factor loadings of the items as well as to identify items that may need to be removed. Promax oblique rotation was again used. In the pattern matrix, factor loading with cut-off value of < 0.4 was used as the criteria to determine whether an item was to be removed or not [13]. The communality value, which indicates convergent validity of the items, was set at 0.25. Finally, the Cronbach’s alpha coefficients (with > 0.6 cut-off value) were then checked to evaluate the degree of internal consistency of the items in each construct or factor [13].

With regards to CFA, convergent validity, internal consistency, and discriminant validity were determined using partial least square (PLS). For internal consistency, the composite reliability was determined, whereas for convergent validity, factor loadings and average variance extracted (AVE) were determined [14]. In this regard, factor loading of > 0.70 is used as the cut-off point; whereas for item with factor loading between 0.4 and 0.7, the effect of its removal on the overall AVE would be considered. If the removal of the item improves the AVE of the factor, the item would be removed unless the item is a priori determined by the authors to be of critical importance in terms of content validity [14, 15].

With regards to the translation process, two of the authors of this paper (SYL, DNK) independently translated the original English version into the targeted Malay language (“forward translation”) version. These two authors are proficient in both English language and Malay language. SYL is a medical doctor working in the emergency department of Sarawak General Hospital while DNK is a nursing educator from UNIMAS. These 2 authors then compared their versions of the translation with the aim of merging their versions into a single forward translation version (“reconciliation” stage). The backward translation was performed by an independent translator who is proficient in both English and Malay languages and who has vast experience in doing journalistic translation work in two languages. After completion of the backward translation, three authors (SYL, DNK, NS and KSC) then compared the back translated English version with the original English version to check for significant discrepancy (“harmonization”). Any significant discrepancy would be discussed, and revised if deemed necessary. In the unlikely event where there is discrepancy in which the authors could not amicably resolve, an independent language expert who is fluent in both English and Malay languages would be called in.

Results

With regards to the development and construct validation of the questionnaire, the Kaiser–Meyer–Olkin measure of sampling adequacy was 0.79 indicating sampling adequacy for EFA. The p-value for Bartlett’s test of sphericity was < 0.001 indicating that there are worthwhile correlations among the items based on the correlation matrix. There are five factors with initial Eigenvalue > 1 (also demonstrated in scree plot, see Fig. 1). The re-run of EFA using promax oblique rotation showed that the communalities of all items > 0.25. Pattern matrix analysis showed that all items were loaded into the factors with factor loading > 0.4 (see Table 1 for the detailed factor loadings after the initial run of EFA). The Cronbach’s alpha value for Factor 1 was 0.955 with 6 items loaded into it, suggesting that there are redundant items. When checked for redundancy, it was noted that item “The directions that point to the location of the AED are clear” carries very similar meaning with item “The signage that shows the location of the AED is clear”. Hence, item “The directions that point to the location of the AED are clear” was removed. After removal, the internal consistency of Factor 1 is still good with Cronbach’s alpha of 0.942. The internal consistency for Factor 2, Factor 3, Factor 4 and Factor 5 were also good with Cronbach’s alpha of 0.855, 0.787, 0.914 and 0.893 respectively. The corrected item-total correlation of all items ranged from 0.41 to 0.89. No negative correlation was noted. A re-run of the EFA after deletion of item “The directions that point to the location of the AED are clear” was subsequently performed and showed that the factor loadings are still good.

Fig. 1
figure 1

Screen plot shows 5 factors with Eigenvalue > 1

Full size image
Table 1 The factor loadings with pattern coefficient values after the initial run of exploratory factor analysis
Full size table

Based on the items that are loaded into the 5 factors, Factor 1 is labelled as “perception of AED placement strategies”, Factor 2 as “perception of importance of CPR and AED”, Factor 3 as “concerns of injuring victims during CPR and AED”, Factor 4 as “concerns of legality in performing CPR and AED” and Factor 5 as “confidence and willingness to perform CPR and AED”.

With regards to the CFA, the factor loadings of all items ranged between 0.48 and 0.98, except for “formal training” where the factor loading was 0.40. This item was subsequently removed as its removal improved the AVE of the factor from 0.30 to 0.40. The AVEs of the other factors ranged from 0.55 to 0.94. In terms of internal consistency, the composite reliability of all factors ranged from 0.69 to 0.97. In terms of the discriminant validity, there was no significant cross loading, the AVEs of all factors were higher than that of other factors according to Fornell and Lacker’s criterion [14] and the confidence interval for heterotrait-monotrait ratio (HTMT) for all items did not include the value of 1.0. The final structural model developed from PLS and its measurement results are detailed in Additional file 1: Figure S1. The forward translation version (after the reconciliation stage) as well as the backward translation version by an independent translator (after the harmonization stage) are tabulated in Additional file 2: Table S1. The final version of the validated questionnaire is shown in Table 2.

Table 2 The finalized version of the bilingual questionnaire to measure the perception and confidence towards cardiopulmonary resuscitation and automated external defibrillator training and placement strategy
Full size table

Discussion

The objective (1) perception of placement strategies of public access AED, was captured in Factor 1, labelled as “perception of AED placement strategies”; objective (2) perception on the importance of bystander CPR and public access AED is captured in Factor 2, labelled as “perception of importance of CPR and AED”; objective (3) perception on the confidence and willingness to apply these two lifesaving interventions is captured in Factor 5, labelled as “confidence and willingness to perform CPR and AED” and objective (4) the fears and concerns in applying these two interventions is captured in Factor 3 “concerns of infection and injuries during CPR and AED” and Factor 4 “concerns of legality in performing CPR and AED”.

To ensure a successful implementation of public access AED program, the preparedness of trained bystanders (with positive attitude and confidence) is as important as the placement of the AEDs itself [16, 17]. Whitney-Cashio et al. [10] suggested that AEDs should be placed in highly visible locations (“visibility”) that can be easily accessible (even after working hours) and with the “direction” and “signage” to access the AED should an emergency arises. Besides accessibility and visibility, AEDs should have clear instruction on how to use it and be placed in a secure area (e.g. with surveillance cameras) to minimize the risk of the AEDs being stolen (“security”). The corresponding item to measure each of these criterion in Factor 1 “perception of AED placement strategies” is listed in Additional file 3: Table S2.

Limitations

The items in this questionnaire were mainly constructed and culled from published papers. In other words, there could have been other valid dimensions that also measures a specific factor or construct but are missed in our questionnaire. This is especially so in the construct of “concerns of injuring victims during CPR and AED”. In this construct, we merely measured the concerns or fears of injuring victims and contracting infectious diseases that may deter the initiation of bystander CPR and the use of AED. There may have been other fears and concerns that are valid and relevant but are not captured in this questionnaire.