Study design and participants
This observational descriptive and analytical study was carried out in all counties of Hamadan province, located in the west of Iran, in 2019. These counties includes Hamadan (the capital of Hamadan province), Malayer, Tuyserkan, Nahavand, Razan, Bahar, Kabudarahang, Asadabad, and Famenin. Based on the previous study[19], the estimated sample size was 600 households. Cluster sampling was used for this study and we used the design effect of 1.5 plus 10 percent attrition. Subsequently, the final sample size was calculated at 1000 households. The data were collected from February to July 2019. From each county, a university graduate person was recruited and trained for data collection. The supervision and training were done by the first author. The verbal informed consent was obtained from all participants before the data gathering. The participants were first provided a description of the study and they were informed that the participation in the study was voluntary, and all study data were anonymous and confidential. Then, if they gave verbal informed consent, they would participate in the study and fill out the anonymous questionnaires. A person aged 18 or above was randomly selected from each household and answered the questions. For illiterate people, questionnaires were filled out through interviewing them. The inclusion criteria were being at least 18 years old and being able to answer the questions. The exclusion criteria were an incomplete questionnaire.
Participants have been selected by multi-stage cluster random sampling. First, stratified sampling was used for each county based on its urban and rural populations. Then, in urban and rural areas, a list of urban or rural health centers was listed and one health center was randomly selected in each county. After that, from the list of all households covered by the selected health center, one household was selected by simple random sampling and sampling started taking the clockwise direction of the selected household and continued until the required sample was collected. For selecting the sample of the urban population of Hamadan County, we selected one health center from each district by simple random sampling (in Hamadan city, there are four districts). In the next stage, from the list of covered households, one household was randomly selected and the sampling was started taking the clockwise direction until the required sample in each district was collected.
Measurements
The questionnaire used for data collection comprises four domains including 1) demographics, 2) earthquake preparedness 3) awareness on earthquake response, and 4) predictor of earthquake preparedness based on the HBM. Earthquake preparedness was response variable.
1) Demographics included age, sex, occupation, education, economic status, family size, number of individuals over 60 years old and under 16, earthquake experience, homeownership, marital status, and having a person with a disease that needs medication at their home.
2) We measured earthquake preparedness by an earthquake preparedness checklist[22]. This checklist was developed and validated by Spittal et al., in 2006. It consists of 23 questions with yes or no answers. The questions are about: having a working torch (flashlight), a first aid kit, a working battery radio, a working fire extinguisher, etc[22]. We adapted this checklist by adding two items according to the context of the study. These two questions were: 1) do you know the necessary contact numbers such as fire station, police, and emergency so that you will be able to call them if needed?; 2) are you familiar with the phrase, ''Drop, Cover, and Hold"? Also, we adapted it with some minor changes. We added “ have learned first aid” to “have purchased first aid kit” statement. We added “and extra cloths and blankets” at the end of ”put aside extra plastic bags and toilet paper for use as an emergency toilet” statement. We replaced “roof” with “my way” in “ensuring that the roof will probably not collapse in an earthquake. We added some examples to “take some steps at work” statement such as attending an earthquake preparedness class and having fire insurance. The content validity of the Persian checklist was tested by 10 experts. We calculated CVI and CVR equal to 0.92 and 0.95, respectively. Also, the face validity and reliability of this checklist were examined in a pilot study on 40 adults. According to their recommendations, minor revisions were made to increase the transparency and understandability of the statements. Likewise, the reliability of this checklist was measured by internal consistency (Chronbach α = 0.858). The total score of this checklist was ranging from 0 to 25 and the higher score reflects more preparedness.
3) The awareness on earthquake response questionnaire included seven questions with true/false answers (In an earthquake: you should get down close to the ground; you should get under a big piece of furniture such as a desk or other covers; you should hold on to a firm object until the end of the shaking; you should stand in a doorway; If you are indoors during an earthquake, you must exit the building; If you are in bed during an earthquake, you should stay there and cover your head with a pillow; next to pillars of buildings and interior wall corners are the safe areas). One point was given for each correct answer. Therefore, the total score of this domain was seven points.
4) The adapted questionnaire of earthquake preparedness based on the HBM was used. The original questionnaire has been established and validated by Inal et al[1] in Turkey. The forward and backward translation method was used for translating the original questionnaire. According to the experts' opinions, some minor changes were made to adapt the items of the questionnaire for the study population in the present study. Thereby, three questions were added to the questions of the cues to action (Radio and TV encourage me to prepare for disasters, I usually seek information about disaster preparedness from Radio and TV, and I usually obtain information about disaster preparedness from health providers). Besides, one question was added to the questions of perceived benefits (preparedness for disaster will reduce financial losses and injuries). Then, the content validity of the questionnaire was assessed by a panel of experts including 10 Health specialists in the field of health in disasters, health education, health promotion, and safety promotion (CVR= 0.92 & CVI=0.85). Next, the face validity and reliability of the questionnaire were measured in a pilot study on 40 people over 18 years old. The reliability was calculated by using internal consistency. One question from the perceived severity (emergency and the experience of disasters does not change my life) and one question from self-efficacy (I cannot create an emergency plan with my neighbors) was excluded based on the results of Cronbach's alpha. In Iran, neighbors don’t share their plans; therefore, it was logical to exclude these items. Finally, the questionnaire consisted of 33 questions, including perceived severity (2 questions, α=0.709), perceived susceptibility (6 questions, α=0.664), perceived benefits (4 questions, α=0.758), perceived barriers (6 questions, α=0.822), self-efficacy (7 questions, α=0.677), cues to action (8 questions, α=0.683), and total questions (33 questions, α=0.809). All of the items were assessed by a 5-point Likert scale ranging from ‘completely disagree’ (one point) to ‘completely agree’ (5 points). Some items were scored reversely.
Statistical analysis
We used the analysis of variance (ANOVA) and independent t-test to determine the relationship between variables. Besides, the multivariate linear regression model was used to determine the predictors of household earthquake preparedness. The Stata 14.2 software was used to analyze the data.