Sample and data collection
The study participants were randomly selected from all health care centers of Kashan county, using stratified random sampling with proportional allocation. First, the entire population was divided into five categories (strata) based on the geographical location and livelihood level. Then, a random sample from each stratum (subgroup) was taken on the basis of health care coverage, from which the subgroup was randomly selected using a random number table. To address this, the medical record information of pregnant women referred to health care centers from April to July 2020 was screened using random number table. Women were screened for their eligibility and invited to participate in the study via phone call for completing the questionnaire. The questionnaire was completed by three trained interviewers. Interview was used for data collection. The interview took place in one face-to-face interview session. To encourage the participants, they were given an online exercise program during pregnancy. The sample size was considered according to other studies [
8] and the sample size formula:
\( n=\frac{\left({z}_{1-\frac{\propto }{2}}\right)2\times p\left(1-p\right)}{d^2} \)) where α: 5%, d: 6%, p: 50%. Finally, a total of 294 pregnant women (with considering 10% for drop sample) were selected for the study.
Data collection tools included a researcher-made questionnaire including questions about demographic information (13 items), midwifery information (13 items), and awareness (7 items, such as “herbal medicine can affect the fetus”). Awareness questions were scored on a true or false basis. The total score ranged from 0 to 7, with a higher score indicating better awareness. The reliability of the scale was measured by Cronbach alpha and it was satisfactory (0.81).
The second part of the questionnaire was designed based on the theory of planned behavior including: 1- Attitudes towards the use of medicinal herbs during pregnancy and post-delivery (7 items, e.g., in my opinion, the use of herbal medicines is .... absolutely useful / absolutely harmful, the total score ranged from 7 to 49, with a higher score indicating positive attitudes); The reliability of the scale was measured by Cronbach alpha and it was satisfactory (0.89).
2- Subjective norms (4 items, e.g., most people who are important to mee suggest that I can use herbal medicine and traditional medicine .... Strongly agree/strongly disagree, the total score ranged from 4 to 16, with a higher score indicating being more influenced by people to practice behavior); The reliability of the scale was measured by Cronbach alpha and it was satisfactory (0.82).
3- Perceived behavioral control (2 items, choice of perceived behavioral control to reasoned action, e.g., despite the special conditions of pregnancy (nausea, vomiting), I can refrain from taking herbal and traditional medicines (the total score ranged from 2 to 4, with a higher score indicating more perceived behavioral control). The reliability of the scale was measured by Cronbach alpha and it was satisfactory (0.74).
4- Intention (2 items, e.g., I plan to take herbal medicine if I experience health problems during pregnancy, strongly agree/strongly disagree, the total score ranged from 2 to 4, with higher scores indicating greater intention). The reliability of the scale was measured by Cronbach alpha and it was satisfactory (0.80).
5- Behavior performance (22 items, five-point Likert scale including always, often, sometimes, rarely, never, the total score ranged from 22 to 110, with higher scores indicating more tendency toward herbal medicines during pregnancy). The reliability of the scale was measured by Cronbach alpha and it was satisfactory (0.89).
The validity of the questionnaire was determined using content validity ratio (CVR) and content validity index (CVI). To establish content validity, 10 academicians including experts on health education, public health, midwifery, and obstetrics were asked to categorize the items into three categories of “necessary,” “beneficial but not necessary,” and “not necessary. According to the Lawche’s table, the items with CVR > 0.62 were remained in the questionnaire. For CVI estimation, the experts were asked to score each item based on relevance and clarity using a 4-point Likert type scale. Items were regarded as clear and relevant if they obtained values ≥0.79. Finally, after two rounds, the content validity ratio (CVR) and content validity index (CVI) were calculated, which were both equal to one.