This was a cross-sectional and descriptive-analytic study conducted among active health volunteers referring to health centers in Qazvin in 1399. Multi-stage random sampling was used so that the list of health centers in Qazvin was prepared and then it was divided into two parts of north and south. Then, two health centers were randomly selected in each section. Finally, health volunteers were randomly selected through a lottery in each health center.
According to the results of the pilot study in 30 health volunteers (r = 0.15 for the correlation between the adoption of walking behaviors to prevent osteoporosis and health literacy) as well as the sample size table for correlation studies, the minimum sample size was estimated to be 175 . The sample size was estimated to be 263 considering DESIGN EFFECT = 1.5. Finally, to compensate for the probable loss of the samples, 290 people were enrolled in the study.
The inclusion criteria of the study were reading and writing literacy, Iranian citizenship, 18 to 65 years old, active as a health volunteer during the study, active presence in weekly or monthly volunteers' meetings in Health Centers, and informed written consent. Exclusion criteria were no unwillingness to continue the study and incomplete questionnaires.
The data collection instrument consisted of three parts: a) demographic and background information questionnaire including items about age, marital status, education level, address, weight, height, number of delivery, number of lactation, number of family members, and monthly family income.
B) To measure health literacy and its skills, the health literacy questionnaire for the urban population between 18–65 years old (HELIA) was used . This questionnaire included 5 main skills (reading, access, understanding, evaluation and decision making, and use of health information) and 33 items. The scoring scale was based on 5 -score Likert, so that 5,4,3,2 and 1 were given to quite easy, easy, not easy and not hard, hard, and completely hard in reading items respectively. In 4 other health literacy skills, 5,4,3,2, and 1 were given to always, quite often, sometimes, rarely, and never respectively. To score the questionnaire, the raw score for each individual in each skill was obtained by the sum of his scores. Then, to convert this score to a zero to one hundred range, the raw scores minus the minimum possible raw score were divided by the maximum possible score minus the minimum possible score. Finally, to calculate the total score, scores of all skills (based on a range of zero to 100) were added and it was divided by the number of skills (number 5), so that 0 to 50, 50.1 to 66, 66.1 to 84, 84.1 to 100 were considered as inadequate health literacy, adequate health literacy, sufficient health literacy, and excellent health literacy respectively. The questionnaire had a favorable validity and acceptable reliability (The alpha Cronbach between 72–89%) . In the present study, it was first pilot studied in 30 of the health volunteers and the alpha Cronbach coefficient were 0.85, 0.82, 0.79, 0.81, 0.76, and 0.79 for reading, access, understanding and perception, evaluation, the decision making and application of health information, and the whole questionnaire respectively.)
(C) Part three included measuring the adoption of walking behaviors to prevent osteoporosis. This section included 7 items and it measured the time spent on walking over the past week based on the given guanines. To score this part, zero, one, two, and three were given to no walking, light walking, average walking, and heavy walking respectively. Thus, the range of scores was between zero o 21. The validity of this part was more than 0.7 in the study of John and colleagues, and its reliability was above 0.79 by Cronbach's alpha coefficient (29, 26). it was also pilot studied in 30 of the health volunteers and its alpha Cronbach coefficient was 0.81. The rate of preventive behaviors was classified into three levels of poor (scores less than 50% of the total score), moderate (scores of 75 − 50% of the total score), and good (scores above 75% of the total score) . Also, the preventive behaviors were classified into two levels of poor (scores less than 50% of the total score) and good (scores between 100 − 50% of the total score) [25 and 31], and they were used in the logistic regression in the study.
The research number was received from the Deputy of Research and Technology of Qazvin University of Medical Sciences (Ethics code: ir.qums.rec.1398.380) in coordination with selected health centers. The purpose of the study was explained for the health volunteers to get their written consent. The questionnaires were self-reported, and all health volunteers were asked to complete the questionnaires honestly. They were also assured that all the information requested would be secret and without the names of the individuals. Data were analyzed using descriptive statistics and logistic regression in SPSS version 23, and the significant level was less than 0.05.