This was a quasi-experimental interventional study conducted on the staff of selected dentistry centers in Tehran in 2017. To determine the sample size, according to a similar study in Iran [26] in which the mean score of hand hygiene behavior after the intervention in the intervention and control groups were 20.42±2.66 and 19.54±2.90, respectively, considering 95% confidence interval and statistical power of 80%, the sample size was estimated to be 53, which was increased to 64 considering the 20% of the participants might leave the study. Participants were selected by multi-stage sampling. First, two centers were selected from the dentistry centers in Tehran by simple random method. In the second stage, one center was randomly allocated to the intervention and the other one as the control group. In the third step, the participants were randomly allocated to both groups by equal distribution. All general dentists and specialists, oral and dental health care staff, and dentists' assistants working in those centers who worked at least for one continuous year entered the study with written consent. None left the study.
Since there was not a valid questionnaire, one was developed by the researchers which included 59 items in 9 sections using the sources and reference books and the opinion of technical professors. The first part was the questions related to demographic and general characteristics [8 items] and the second part was the items related to knowledge assessment [7 items] whose score ranged from zero to 35. The third part was related to measuring the constructs of the health belief model, which included the structures of perceived susceptibility [5 items], perceived severity [4 items], perceived benefits [5 items], perceived barriers [5 items], and self-efficacy t [9 items] for which a Likert scale with five options [strongly agree, agree, have no opinion, disagree and strongly disagree] was used. The range of the scores for perceived severity, perceived benefit, perceived barrier, self-efficacy scores were between 4 and 20, 5 and 25, 5 and 25, and 9 and 45 respectively. The items related to measuring cues to action [3 items] were based on a Likert scale with 5 options [very much, much, little, very little, and not at all] which was between 3 and 15. The fourth part was the behavior assessment items [13 items] based on a Likert scale with 4 options "always, most of the time, rarely and never" which ranged from 13 to 52.
To evaluate the validity of the content, a panel consisting of 10 professors in the field of health education and promotion, dentists, infectious disease specialists, microbiologists, and epidemiologists, examined the validity of the content in qualitative and quantitative ways. Qualitatively, experts were asked to review the tool based on the criteria of grammar, use of appropriate words, placing the items in the right place, and proper scoring, and they provided the necessary feedback. For the content validity, the content validity ratio (CVR) and content validity index (CVI) were determined quantitatively. Experts were consulted to determine the necessity or non-necessity of each item to determine the CVR. The CVR values above 62 % were considered acceptable [27]. At this stage, one item of knowledge assessment, one for perceived intensity, two for action guide, and one for self-efficacy were removed from the questionnaire. In the CVI review, experts evaluated each item in terms of relevance, clarity, and simplicity, and values above 0.79 were considered acceptable [27]. None of the items were omitted in the CVI review, and the content validity ratio to the total number of items was 0.95.
In the second stage, a modified questionnaire was given to 10 members of the target group to assess its face validity. According to the suggestions of the target group and their understanding of the simplicity and ability to understand the items, the final necessary changes were made to the questionnaire. It is worth noting that these individuals were not included in the study.
To evaluate the reliability of the questionnaire, Cronbach's alpha coefficient was used to determine the internal consistency of the subscales of the Health Belief Model Questionnaire and the values of 0.70 and above were accepted [27]. Therefore, the questionnaire was completed by 15 participants who met the criteria of the present study and had similar demographic characteristics to the study population. Cronbach's alpha for knowledge, perceived susceptibility, perceived severity, perceived benefit, perceived barriers, cues to action, self-efficacy, and behavior were 0.85, 0.75, 0.73, 0.91, 75 0.0, 0.86, and 0.87, respectively.
Also, to measure reliability over time, the questionnaire was completed again after 2 weeks by 15 participants from the target population, and the correlation coefficient and ICC [intra-class correlation coefficient] were measured, which was 0.94 for knowledge, 0.96 for perceived susceptibility, 0.97 for perceived severity, 0.97 for perceived benefits, 0.85 for perceived barriers, 0.87 for cues to action, 0.96 for self-efficacy, and 0.89 for behavior assessment. It is worth noting that these individuals were not included in the study. Moreover, after careful studying reliable and relevant sources, the initial educational content was designed based on the health belief model and the data from the pre-test results based on which the most effective structures predicting hand health behavior were identified. In fact, perceived barriers and knowledge significantly predicted health behavior [28], and the lesson plan and teaching content were revised and adjusted based on all structures and with emphasis on these effective structures. Finally, it was given to the panel of experts to evaluate and give their comments. After correcting some parts, general goals, specific goals, and related behavioral goals were formulated based on the training program of each session. The training was held in three 90-minute sessions.
The questionnaires were completed as a self-report in 30 minutes. After explaining the study and its objectives, all participants were asked to complete the questionnaire with complete honesty, and they were assured that all the information requested in the questionnaire would be used confidentially. Before the study, written informed consent was obtained from all participants. The questionnaires were completed at the employees' workplace. Also, before starting the study, the ethics code with the number IR.SBMU.PHNS.REC.1396.3 was received from the research ethics committee of Shahid Beheshti University of Medical Sciences. Then, the data were analyzed using SPSS by independent t-test, Mann-Whitney, Chi-square, and repeated measures analysis of variance. In addition, the significance level was considered less than 0.05.