Introduction
Disaster refers to situations where the impact of natural and man-made phenomena exceed the capacity of the hazard-affected community to cope with it, thus leading to the destruction of normal conditions, the loss of life, and damage to property and the environment. Reference Usuzawa, Telan and Kawano1 According to the International Disaster Database, from 2007 to 2012, 1931 natural disasters occurred in 117 countries, causing 610768 deaths, 1 billion injuries, and economic losses to the tune of 75 trillion dollars. Reference Albrecht2,Reference Vaillancourt and Haavisto3
Disaster preparedness is defined as “activities and measures taken in advance to ensure effective responses to disasters, including the issuance of early warnings and the emergency evacuation from threatened areas.” Reference Said and Chiang4 The well-prepared individuals and teams could take timely and effective measures to reduce the damage caused by disasters. Nurses play an indispensable role during disasters and are often the first responders in disaster relief operations. Reference Rokkas, Cornell and Steenkamp5 When a disaster occurs, nurses should be able to evaluate and meet patients’ needs, utilize available resources and collaborate with other professionals to deliver emergency care. Disaster nursing refers to the systematic and flexible application of knowledge and skills with the characteristics of the nursing discipline, and cooperates with other professional fields to carry out activities aimed at mitigating the hazards of disasters to human life and health. The International Council of Nursing (ICN) established a framework for disaster nursing core competencies in 2009, which comprised 4 areas that corresponded to different stages of a disaster, namely prevention, preparedness, response, and recovery. Reference Al Thobaity, Plummer and Williams6 Therefore, it is necessary to evaluate the disaster preparedness of nurses before preparing and implementing education and training programs on disaster nursing.
Background
Disasters have taken place frequently and have caused serious damage in mainland China. According to the China Earthquake Administration (CEA), over 100 earthquakes of magnitude 5 or above occurred in China between 2016 and 2019. Reference Xu, Liu and Deng7 The Chinese government enacted the Response Plan for disasters, and prioritized sending the China Emergency Medical Team (CEMT), as well as doctors and nurses to designated areas after a disaster. Reference Hao, Li and Zheng8 Emergency nurses are the largest group of disaster relief providers in mainland China. They operate as frontline health care providers, educators, mental health counsellors, and triage officers. Disaster relief work under dangerous and hard conditions requires emergency nurses with good health, psychological qualities, adaptability, and professional capacities. Reference Li, Turale, Stone and Petrini9 Disaster nursing started late in mainland China, and the discipline system was not complete. The Disaster Nursing Committee of the Chinese Nursing Association was established in 2009; It gradually carried out disaster-related lectures and short-term courses. However, the demand of disaster nursing training for nurses is high and the overall disaster preparedness level is low at present. Some studies have reported that emergency nurses showed a low-to-moderate level of disaster preparedness in mainland China. Reference Usher, Mills and West10 They said that they had not received systematic disaster nursing training and felt underprepared. Therefore, it is necessary to develop a suitable training program for emergency nurses in mainland China in order to help them respond to disaster better.
Although there are a few tools to measure nurses’ disaster preparedness in English, most lack reliability, and validity tests and are not widely used. Reference Tavan, Menati, Azadi, Sayehmiri and Sahebi11-Reference Wisniewski, Dennik-Champion and Peltier13 The Disaster Preparedness Evaluation Tool© (DPET) is 1 of the most widely used scales to assess disaster preparedness and has good psychometric properties. Reference Tichy, Bond, Beckstrand and Heise14 Moreover, the DPET measures the knowledge, skills, awareness, and management ability of disaster preparedness in the 3 stages before, during, and after a disaster, with the broadest coverage, and meets the definition of disaster preparedness. Since Bond and Tichy’s initial study in 2009, the DPET has been used to evaluate different nursing populations around the world, including in Taiwan, China, Reference Chen, Chou, Liao, Ho and Chung15 Saudi Arabia, Reference Al-Khalaileh, Bond, Beckstrand and Al-Talafha16 Japan, Reference Öztekin, Larson, Akahoshi and Öztekin17 Jordan, Reference Al-Khalaileh, Bond and Alasad18 Indonesia, Reference Sangkala and Gerdtz19 and America. Reference King, Spritzer and Al-Azzeh20
Although a Taiwan’s version of the DPET (DPET-C) has been adapted in Taiwan, China, a mainland China version is still necessary, considering the differences between the 2 places. First, the political backgrounds are different because of the implementation of the ‘1 country, 2 systems’ policy. Second, there are cultural differences between the 2 places, resulting in different understanding and language habits. Third, the 2 studies focus on different population groups. The DPET-C was based on public health nurses while our study is based on emergency nurses. Fourth, both places differ in terms of terrain and climate, and each is vulnerable to different types of disasters and has different requirements when it comes to disaster preparedness. Reference Lin, Li and Shieh21 In addition, researchers introduced the DPET into mainland China and translated it into Chinese in 2014. Reference Li and Sheng22 However, this version was only a direct translation of the DPET, without a validity test, cultural adaptation and reliability. Therefore, it is necessary to translate and adapt a version of the DPET for mainland China.
Purpose
The purpose of our study was to translate and adapt the DPET in order to create a version for mainland China (DPET-MC), and to validate its psychometric properties.
Methods
Design
The translation and validation processes were performed in 3 phases. In phase 1, the DPET was translated from English into Chinese through a translation and back-translation process. In phase 2, a sample of 30 emergency nurses was pre-tested to establish the preliminary reliability of the DPET-MC. In phase 3, a psychometric evaluation of the DPET-MC was performed (Figure 1).
Figure 1. Translation and validation of the Disaster Preparedness Evaluation Tool (DPET).
The Original DPET Questionnaire
The DPET was constructed by Bond and Tichy (2007) to assess the disaster preparedness of nurse practitioners (NPs) in America. Reference Tichy, Bond, Beckstrand and Heise14 The DPET was developed based on the ICN framework of disaster nursing competencies by a respected academic institution. The instrument includes 68 items and 2 sections to measure NPs’ perceptions of their disaster preparedness. Section 1, titled ‘Introduction,’ contains 47 Likert-type questions. The first 25 items discuss pre-disaster preparedness and the items are grouped into 3 categories, namely knowledge, disaster skills, and personal preparedness. The next 16 items discuss disaster response and the items are grouped into 2 categories, namely knowledge and patient management. The last 6 items discuss disaster recovery and the items are grouped into 2 categories, namely knowledge and management. Each item is rated on a 6-point Likert scale wherein 1 = ‘strongly disagree,’ 2 = ‘disagree,’ 3 = ‘partly disagree,’ 4 = ‘partly agree,’ 5 = ‘agree,’ and 6 = ‘strongly agree.’ Adding the score for each item produces a total score that can range from 47 to 282, with higher scores indicating a greater degree of disaster preparedness. The DPET has adequate internal consistency reliability. The Cronbach’s alpha coefficients are 0.91, 0.93, 0.93, and 0.91 for the overall scale and 3 dimensions, respectively. Section 2, titled ‘Demographic data’ consists of 21 open-ended questions.
Phase 1: Translation and back-translation
The original version was translated into Chinese in keeping with the ‘Process of Cross-Cultural Adaptation Guidelines.’ Reference Beaton, Bombardier, Guillemin and Ferraz23 The translation was carried out by an emergency nurse and a medical teacher and the back-translation was carried out by a professional translator who had never seen the index before. A multidisciplinary committee panel reviewed the 3 versions of the DPET including the original, the translated, and the back-translated versions, consolidated the 3 versions, and developed a pre-test version for the pre-test stage. In order to evaluate the comprehensibility and universality of the final version, a face validation was conducted wherein 10 Chinese women, aged 25 to 45 years with normal literacy and comprehension levels, and without any medical background, were included to complete the questionnaire and undergo an interview. Comprehension problems, verbal errors, and unclear expressions were recorded and suitable modifications were made.
Phase 2: Pre-test
A pre-test was undertaken using the pre-test version of the DPET questionnaire with a convenience sample of 30 emergency nurses (sub-sample 1) in a third-grade hospital in Guangdong, China. Reference Perneger, Courvoisier, Hudelson and Gayet-Ageron24 The internal consistency and test-retest reliabilities were analyzed using Cronbach’s alpha coefficients and internal correlation coefficients (ICC).
Phase 3: Psychometric Evaluation
Study Design
A total of 2 cross-sectional studies were undertaken following the STROBE statement. Reference von Elm, Altman and Egger25 The former examined the construct validity of the DPET while the latter tested the model fit, internal consistency reliability and split-half reliability of the DPET-MC.
Participants and Settings
In this study, researchers selected 26 public hospitals in Guangdong province according to the convenient principle, surveyed nurses in emergency departments, and obtained 2 convenience samples. A sample of 633 emergency nurses was recruited in the first study between May and August, 2018. A sample of 205 emergency nurses was recruited in the second study in April 2019. The inclusive criteria were age (≥ 18 years) and willingness and ability to provide informed consent to participate in the study. Non-registered nurses and student nurses were excluded.
The sample size was determined by the requirements of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). For EFA (sub-sample 2), the sample size was estimated to be at least 470 (for the 47-item version of the questionnaire), according to the recommendations that a minimum participant to variable ratio of 10:1 should be achieved. Reference Gaskin and Happell26 For CFA (sub-sample 3); We used the guideline that for maximum likelihood estimation (MLR), at least 200 participants were required. Reference Cappelleri, Bushmakin, Harness and Mamolo27
Instruments
The first study used the DPET tool with the original 47 Likert-type questions. The second study used the DPET-MC tool after determining the construct validity by performing EFA. Furthermore, 6 questions on demographic data (age, gender, education, hospitals, years of work and previous experience of disaster relief) were filled by all participants.
Procedure and Data Collection
Both studies were based on an online questionnaire survey which followed every item of the ‘CHERRIES’ guidelines rendered via a software named ‘Questionnaire Star’ (Changsha Ranxing Information Technology Limited Liability Company, Changsha, China). The questionnaire had 2 sections. Section A focused on demographic and clinical background questions and Section B focused on the Likert-type questions. The questionnaire was distributed online through links and Quick Response (QR) codes, which were accessible without a password and took approximately 10 minutes to complete. To ensure active participation, researchers contacted head nurses of each hospital individually, asking them to help distribute questionnaires in their WeChat groups and monitor the whole process. Nurses who agreed to participate in the study were expected to open the questionnaire and submit it only after completing all questions. Participants would receive a small gift after completing the questionnaire. To avoid alter-time access in completing the questionnaire, an IP number was used for access restriction. The online survey would be available for 1 week. After the deadline, the system would close automatically and participants were unable to fill in and submit questionnaires. Completed questionnaires were collected automatically and were exported to an Excel file manually.
Statistical Analysis
Data were analyzed using Excel®, SPSS® Software (version 23) (IBM SPSS Inc., Chicago, Illinois), AMOS® Software (version 23) (IBM SPSS Inc., Chicago, IL), and Vista® Software (version 64) (Microsoft Corporation, Redmond, WA).
First, descriptive statistics were drawn up to examine the data distributions. Second, EFA was used to establish the construct validity of the DPET tool by performing principal components analysis (PCA) with varimax rotation. The appropriateness of the factor analysis was evaluated using the Kaiser-Meyer-Olkin (KMO) measure, where a score is considered good if it is higher than 0.90 and excellent if it is greater than 0.95. Bartlett’s test was used to test the sphericity and determinant of the correlation matrix (P < 0.05). The eigenvalues-greater-than-1 rule and the scree plot can determine the factor number. Reference Cattell28 Items were excluded when they met the following criteria: (1) item loading < 0.50 on 1 factor or (2) item loading > 0.40 on 2 factors. Reference Al-Balhan, Khabbache, Watfa, Re, Zerbetto and Bragazzi29 Parallel analysis (PA) verified the accuracy of the factors extracted. Reference O’Connor30 The number of factors was equal to the number of sample eigenvalues exceeding the corresponding mean eigenvalues and the 95th eigenvalues obtained from the random data correlation matrices. Reference Weng and Cheng31
Third, CFA was performed using Maximum Likelihood Estimate (MLE) after EFA to examine the structural model fit of the DPET-MC tool. The following criteria were used to evaluate the fit indexes: Chi square statistic/ degrees of freedom (x2 /DF) < 3.0; root-mean-square error of approximation (RMSEA) < 0.08; the incremental fit index (IFI), the Tucker-Lewis index (TLI), and the comparative fit index (CFI) > 0.90. Reference Bentler32,Reference Schermelleh-Engel, Kerwer and Klein33
Fourth, internal consistency reliability was performed to assess the coordinated degree among dimensions and items in the scale. A Cronbach’s alpha higher than 0.70 is considered acceptable and higher than 0.90 is considered good. Reference Streiner34,Reference Scale Reliability35 Split-half reliability was used to test the degree of consistency within items. Items were grouped using odd and even-half method. In general, a split-half reliability coefficient higher than 0.75 was acceptable and lower than 0.40 was poor. Reference Zhang, Zhou and Ge36
Results
Phase 1
After the translation and back-translation process, a pre-test version of the DPET was produced. All the translators made modifications and cultural adjustments during discussions. No difficulties were encountered during the translation and adaptation process because the translators were experienced and there was a Taiwanese version to act as a reference. Similarly, the remaining steps in the process proceeded without any challenges. The results of face validity indicated that the scale was acceptable and comprehensive.
Phase 2
The pre-test scale was completed by 30 emergency nurses (sub-sample 1, overall response rate = 96.7%) in a third-grade hospital in Guangdong, China. Participants were invited to complete the questionnaires 2 weeks later (overall response rate = 96.7%). The Cronbach’s alpha coefficient was 0.99 and the ICC was 0.60.
Phase 3
Demographic and Clinical Background Information
In the first study, 633 emergency nurses were recruited (response rate = 100%). Participants’ ages ranged from 18 to 55 years. Most of them were female (n = 515; 81.4%). Most had attended university (n = 406; 64.14%). Most features of the second study were similar to those of the first. The participants’ demographic information was summarized in Table 1.
Table 1. Demographic and clinical background information of the analyzed participants
Note:
a the pilot testing phase
b the EFA phase
c the CFA phase.
Exploratory Factor Analysis and Parallel Analysis
The KMO value (0.972) and Bartlett’s test of sphericity (x 2 = 24551.88, P < 0.001) suggested that the data were appropriate for PCA. The PCA with varimax rotation and the scree plot extracted a 5-factor solution that explained 64.06% of the total variance. Finally, 5 factors with 34 items were extracted: pre-disaster management (10 items), post-disaster skills and knowledge (7 items), pre-disaster knowledge (8 items), knowledge and skills in the workplace (6 items), and pre-disaster awareness (3 items). Factor loadings, item communalities, and the variance explained by the factors are shown in Table 2. A total of 5 factors were extracted using PA, too.
Table 2. Exploratory factor analysis with factor loadings for the DPET scale
Note: Factor loadings exceeding 0.50 on one factor are in boldface (except for exceeding 0.40 on two factors).
Confirmatory Factor Analysis
CFA was performed to validate the 5-factor structure after determining the construct validity with sub-sample 3. A second-order model including 5 factors was measured using 34 items (Figure 2). The goodness of fit of the model was good: x 2/ DF = 1.978, RMSEA = 0.071, IFI = 0.907, TLI = 0.90, CFI = 0.90. The standardised pathway loading variance and number of items for each factor are listed as follows: 0.44-0.86 for 10 items (factor 1), 0.75-0.88 for 7 items (factor 2), 0.58-0.83 for 8 items (factor 3), 0.51-0.88 for 6 items (factor 4), and 0.53-0.66 for 3 items (factor 5).
Figure 2. The factor structure of Disaster Preparedness Evaluation Tool – Mainland China version (DPET-MC).
Note: Factor: disaster preparedness; Factor 1: pre-disaster management; Factor 2: post-disaster skills and knowledge; Factor 3: pre-disaster knowledge; Factor 4: knowledge and skills in the workplace; Factor 5: pre-disaster awareness; I: item; e: residual error.
Reliability
The Cronbach’s alpha for the total scale was 0.97, and the coefficients for the sub-scales ranged from 0.79 to 0.96. The split-half reliability for the total scale was 0.97, and the coefficients for the subscales ranged from 0.67 to 0.95 (Table 3).
Table 3. Internal consistency reliability and split-half reliability of the DPET-CM scale
Discussion
In this study, we obtained a DPET-MC scale with 5 dimensions and 34 items that indicated good psychometric properties. This scale may be a reliable and valid instrument to assess disaster preparedness among emergency nurses in mainland China and can provide guidelines for constructing training programs and assignments.
The translation process was conducted rigorously following the ‘Process of Cross-Cultural Adaptation Guidelines’ to maintain equivalence. The results of face validity and the pre-test promoted the comprehensibility and credibility of the scale. The EFA extracted a 5-factor solution with 34 items, which accounted for 64.06% of the total variance, including (1) pre-disaster management, (2) post-disaster skills and knowledge, (3) pre-disaster knowledge, (4) knowledge and skills in the workplace, and (5) pre-disaster awareness. The 5 dimensions of the DPET-C were: post-disaster management, skills, knowledge of self-preparation in a disaster, knowledge to respond in the community and knowledge to respond in the workplace. The 5-factor solution of the DPET-MC scale was consistent with the DPET-C scale for Taiwan and China. Reference Chen, Chou, Liao, Ho and Chung15 Since both scales were applicable to Chinese nurses, they did have some common features. However, there were also some differences such as the number of items, item splitting, and the contents of the dimensions. The reasons for this may be the differences in the economic and medical environments, political policies, language habits (Taiwan uses traditional characters, while mainland China uses simplified characters), and applicable groups (the study on Taiwan targeted public health nurses, while this study on mainland China was tailored towards emergency nurses). Reference Lee, Chaou and Lin38 The DPET-C scale indicated that nurses should be able to participate in emergency planning and prevent the threat of biological and chemical weapons and nuclear radiation. However, according to the DPET-MC scale, nurses were more likely to be equipped with disaster knowledge, skills, management and coordination. The reasons for the differences might be: (1) there were few biological and chemical threats in mainland China, so relevant training and education were insufficient; (2) emergency departments of Chinese hospitals attached more importance to comprehensive first-aid skills, such as wound classification, cardiopulmonary resuscitation (CPR) and fracture fixation; (3) the rescue of biological and chemical weapons mainly relied on professional medical institutions, such as the Chinese Centre for Disease Control and Prevention; (4) the prevention of nuclear radiation was mainly in the charge of specialized military hospitals, which were not involved in this investigation. Meanwhile, the 5-factor solution in the DPET-MC scale was different from the Arabic version of the DPET, which extracted 3 factors, namely knowledge, skills, and post-disaster management. Reference Al-Khalaileh, Bond, Beckstrand and Al-Talafha16 The reasons for this might be the differences in culture and perceptions between the 2 different countries. Furthermore, the DPET-MC scale did not fit the factor structure proposed by Heather C King in the US version Reference King, Spritzer and Al-Azzeh20 which included 3 factors, namely disaster knowledge, disaster skills, and personal preparedness. The discrepancy might be the result of the different characteristics of the participants in both studies, as our study focused on emergency nurses and their study focused on health care personnel.
The reliability of the scale was good, as indicated by the excellent results of internal consistency reliability and split-half reliability. The Cronbach’s alpha was 0.97 and the split-half reliability coefficient was 0.97, which were higher than the acceptable coefficients of 0.70 and 0.75. The results of reliability indicated the items in the scale and subscales that measured the same trait. Reference Loven, Yngman-Uhlin and Hjortswang40
This study had some practical advantages. The DPET-MC scale had fewer items when compared to the original scale, which can save time and energy for emergency nurses in a fast-paced work environment. The EFA, PA, and CFA were used to identify the validity of the scale and draw up a reliable solution that could provide guidelines for educators and managers to construct pre-disaster training and post-disaster assignment. Reference Xia, Li, Chen, Jin and Zhang41
This study had a few limitations as well. First, the scale was translated and validated within Guangdong province in China. Thus, it is unclear whether the statistical results can be replicated in other regions in China. Second, the self-reporting method has the potential to lead to subjective results when nurses assess their capabilities, rather than objective observations and evaluations by a third person’s unbiased perspective. Third, emergency nurses in this study were mainly from third-grade hospitals and could not represent other grade hospitals. Therefore, the promotion of this tool is limited. Future studies should ensure appropriate proportions in the collection of research subjects. Fourth, most of the emergency nurses included in this study were female, which could not represent male nurses as well. Future research should take into account the representativeness of the sample. Fifth, in this study, the convenience sampling was used to obtain an un-representative sample. Future studies should consider more objective sampling methods, such as stratified sampling. Sixth, professional titles of emergency nurses should be included in the description of demographic data, because different professional titles might lead to different levels of disaster preparedness.
Conclusion
The DPET-MC scale was translated and validated using various methods with the involvement of teachers, experts, and emergency nurses in mainland China. The results indicated that the DPET-MC scale was a reliable and valid instrument to assess disaster preparedness of emergency nurses. Educators can identify the prevailing level of disaster preparedness among emergency nurses and were able to develop appropriate educational programs to enhance it. Managers in hospitals can measure emergency nurses’ disaster preparedness to provide evidence for pre-disaster targeted training and post-disaster assignments.
Funding Statement
This research was funded by the program of Key Platform and Scientific Research Projects of Guangdong Education Department [grant number: 2016KQNCX016], and by Elite Youth Education Program, Guangzhou University of Chinese Medicine, Guangdong, China [grant number: QNYC20170201].
Ethical Standardss
The School and all the hospitals approved the study. Participants provided their informed consent, and the information received from them was treated as confidential. 37 Professors Bond and Tichy authorized the researchers via email to use and translate the DPET into mainland China version.
