The scale was developed based on a comprehensive literature review [1,6,10–13,16–23]. At first we identified studies regarding nurses’ knowledge, attitude and practice of nurses regarding disaster preparedness [3,4,6,11–13,21]. Then, the questionnaire items and content was developed on collaboration with health specialist including an emergency nurse, an emergency physician, two nurse instructors and a disaster manager. The initial version of questionnaire was sent to three professionals in the areas of nursing, disaster management and statistics and also an educational instructor who was expert in the design and validation of such tools. So, they could appraise the importance and relevance of items and content, the fluency of language and the measurement scales. They were also asked to provide additional suggestions to incorporate in the next version of questionnaire. Then, a pilot study was carried out with 20 hospital nurses who did not participate in the final study. When nurses filled the questionnaire, we conducted a focus group and interviewed them to check the comprehension and clarity of the questionnaire. They were asked to report any ambiguity regarding understanding the items and identify the questions that were most difficult to understand. They were also asked to suggest changes for ambiguous items. Doubtful items reported by participants and their suggestions were recorded by researchers for possible reconsideration. The resulting questionnaire composed of two parts. The first part was to collect the demographic characteristics of nurses’ including age, gender, years in experience, ward, education level and working hour per week. The second part consisted of 6 questions on knowledge, 11 questions on attitude and 6 questions on practice of nurses regarding disaster preparedness. The items were formulated so that participants could then evaluate their level of agreement with each item according to a three-point Likert scale ranging from 1 to 3. The total score of questionnaire ranged from 11 to 33 for attitude and 6 to 18 for each of knowledge and practice domains. Finally, once adjustments based on the nurses’ comments were made, a cross-sectional study was conducted in three public educational center affiliated to Ilam University of Medical Sciences, Ilam, Iran. All of employed nurses in selected hospitals were invited to participate in the study. Finally, 112 nurses accepted to be enrolled in the study.

The construct validity using exploratory factor analysis and Cronbach’s alpha are two statistical methods that we used for measuring construct validity and internal consistency of the questionnaire [24]. Factor analysis tests, the hypothesized structure of the content domain. The two factor-analytic methods used for evaluating structural validity were Explanatory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA). EFA is a statistical method used to identify the underlying relationships between measured variables and discover the latent variables (constructs or factors) that underlie the scale. On the other hand, CFA allows the researcher to test whether the hypothesized relationship between observed variables and their underlying latent constructs exists and also test alternative factor models to determine which provides the best fit [17]. To assess the questionnaire reliability (stability) a test-retest study was done with a convenience sample of 10 nurses on two occasions, with a two weeks interval.

Before the data collection, the study proposal was approved by the regional ethics committee of Ilam University of Medical Sciences. Next, researchers were referring to the three public hospitals from April to June 2015. After giving basic information, the study subjects were asked to participate in a session for data collection. All subjects who participated in the study gave informed consent.

Continuous variables were reported as the mean value ± Standard Deviation (SD). Cronbach’s alpha coefficient was used as an estimate of the internal consistency of the questionnaire. Bartlett’s test of sphericity was used to determine the degree of interrelations between variables for use in factor analysis. Kaiser-Meyer-Olkin Test (KMO) was used to determine sample adequacy. The KMO statistic ranged between 0 and 1. KMO value close to 1 indicates the sample efficiency and justifiability for factor analysis. Chi-square test, Comparative fit index (CFI) and root mean square error of approximation (RMSEA) were used as fit indices. CFI measures equal to or greater than 0.9 and RMSEA less than 0.05 displays the most acceptable fit index [25].

The study participants comprised of 112 nurses (50% male) with the mean age (SD) of 32.5 (±6.08) for male and 34.8 (±7.5) for female. Most of participants had a bachelor degree (85%) in nursing and 33% had 5–10 years of clinical experience. The values of KMO and Bartlett’s Test showed in [Table/Fig-1]. As shown in that table the KMO measure of sample adequacy was 0.761 and Bartlett’s test of sphericity was χ² = 346.03 (p< 0.0001). The mean of participants score in each of knowledge, attitude and practice are domains are showed in [Table/Fig-2]. Also, the correlation of questionnaire domains depicted in [Table/Fig-3]. As shown in [Table/Fig-3] each of domains was statistically correlated with other domains. Explanatory factor analysis using varimax rotation revealed seven main factors associated with the nurses knowledge, attitude and practice regarding disaster preparedness which included: 1- familiarity with disaster management [12–17]; 2- training using disaster simulation [5,10,11]; 3-educating others regarding disaster management [3,7,9]; 4- having plan for unpredicted disasters [6,8]; 5- taking part in disaster maneuvers [18,20,22]; 6- possible threats to disaster management [4], and 7-preparedness during disasters [2]. The first factor accounted for 17.7% of the total variance and the second factor accounted for an additional 10%, while the other factors contributed less to that variance [Table/Fig-4]. For this CFA model, the chi-square value was 1939 (p<0.05). RMSEA value was 0.058 which signify good fit index. The CFI value was 0.92 which meet the criteria (0.9 or higher) and indicate acceptable model fit.

Internal consistency and test-retest were used as estimators of reliability of questionnaire. Cronbach’s alpha was computed for each domain and total scale to measure internal consistency. As a whole, the questionnaire internal consistency using Cronbach’s alpha was 0.785, showing acceptable internal consistency. Cronbach’s alpha coefficient for each of questionnaire domains are showed in [Table/Fig-5]. Test-retest scores were strongly and significantly associated with a correlation of 0.82.

The first step in the preparedness of nurses regarding disaster management was to identify what nurses really know and believe about disaster management. The development and validation of a scale that measure nurses knowledge, attitude and practice regarding disaster preparedness is essential in the attempt to design and implement effective disaster preparedness education curricula and continuing education programs. So, such programs could be helpful in preparing nurses to respond effectively in these difficult situations. According to extensive review of literature this is one of first studies which developed a valid questionnaire that specifically measure nurses knowledge, attitude and practice regarding disaster management. The use of confirmatory factor analysis meets a high standard for validation of a new scale. This validation study revealed that the questionnaire had good internal consistency and reproducibility. The exploratory factorial analysis showed seven factors, in which the items weighed down on a given factor had some shared conceptual meaning and on the other side, the items in different factors measured different concepts. Also, high correlation between items in each of factors showed their congruence.

Regarding internal consistency, the lowest value which would be indicative of a test as having good internal consistency is 0.7 or greater. In our study Cronbach’s alpha measure reached the recommended level for clinical use and the test-retest statistic examination revealed good stability of the responses to the items of questionnaire over time. Exploratory factor analysis also identified seven factors accounting for 67.57 % of the variance. It has been recommended that item extraction should continue until the researcher reaches at least 60% variance [24,26]. So, the variance explained by these factors in our study can be considered appropriate. Al Khalaileh et al., also conducted a study to determine psychometric properties of Arabic version of Disaster Preparedness Evaluation Tool. They found three factors explained 64% of the variance which include knowledge, skills and post disaster management [16].

Despite the strength of this study, it also has some limitations. First, a convenience sample of nurses in three public educational center in west of Iran do not represent variation of all the country population. In relation to future research, the scale should be used with larger samples across other hospitals of Iranian provinces. Second, the ambiguity regarding the role of Iranian nurses in disaster management and their lack of competency for providing care in disaster situation might influence their responses to questionnaire items.

In conclusion, our newly developed scale to measure nurses’ knowledge, attitude and practice regarding disaster preparedness has adequate validity and reliability for use in surveys in Iran. Having baseline information regarding nurses preparedness status could help health policy makers to provide nurses training courses and put these training in academic curriculums.