Participants and Procedures
This cross-sectional study was carried out in 2017 in the city of Nahavand, located in the western part of Iran in Hamedan province with a population of 72000, among rural women. It should be noted that villages in Iran are covered by cities on the basis of geographical divisions. So that, if a researcher wants to perform a study on villages, he/she should at the first select the considered cities.
This city was selected using random digits method from the list of all cities in the west of Iran. There are 43 cities in the west of Iran. These cities locate in Kermanshah province (14 cities), Kurdistan province (10 cities) and Hamedan province (10 cities) (16). In one hand by attention that people living in cities located in the west of Iran have a similar cultural, economic, social, and somewhat common language and also they have a same climate and sunlight and on the other hand by attention to limited resources in the hands of the current study researchers, it was decided to only one of the mentioned cities become entered to the study.
Another important issue is that rural women are usually in front of sun to perform household affairs. In other words, many of affairs near and outside houses are performed using women as their duty description.
Rural population refers to people living in rural areas as defined by national statistical offices. A rural area is a geographic one that is located outside towns and cities. Villages are often located in rural areas. In other word, all population, housing, and territory not included within an urban area form villages (17). Through cluster sampling method, 4 villages were selected from Nahavand city, randomly. Then, using the documents of health centers located in the villages, the women were selected through random sampling method. All demographic information of Iranian rural population is recorded in health centers. Rural health centers provide this information through annually census by their employees. This operation is supervised by district health authorities. This census help planning and developing primary health care in rural areas (18).
The lowest sample volume by attention to the previous studies (11, 14), considering the maximum standard deviation of 5.4, acceptable error of 0.7, confidence interval of 95% and using n=z2s2/d2 formula, was estimated 230 persons. In one hand, on the basis of Kock et al. study on minimum sample size estimation in least squares‐based structural equation modelling (PLS‐SEM), the minimum and adequate sample size is 160. On the other hand, the population of women with inclusion criteria in the 4 selected villages were 1628 persons. The eligible women were selected through simple randomized sampling and proportional with the villages population. So that, 243 persons were entered to the study and the questionnaire distributed among them. Lastly, the participated women returned 230 fully completed questionnaires (19).
The written informed consent form was collected from the participants. This form included the items of study purpose, expected duration of the subject's participation, a description of the procedures, risks or discomforts and benefits, confidentiality, and a statement regarding voluntary participation and freely to leave out the study at any time (20). If one of the selected persons was not willing to participate the study, another person was invited to participates.
Inclusion criteria included rural women with at least minimum literacy, age higher than 18 years old, and no diagnosis of skin cancer. The exclusion criteria included were not continuously present at the training sessions and tendency to leave out during of the study. The training sessions regarding the importance of study, how to answer the questions, freely to leave out the study and so on were held for each participant separately which lasted about 20 minutes.
The study instrument included a standard questionnaire for skin cancer based on PMT which have 2 sections of socio-demographic variables and PMT theoretical constructs (21). The participants were interviewed by one of the research team members at their homes.
The socio-demographics variables included age, gender, marital status (single/ married/ widow), education level (illiterate/ elementary/ secondary/ high school/ diploma/ collage degree), job (household/ worker/ employee/ self-employment), number of hours working under sun, history of sunburn, number of family members and family monthly income level. Existence of a cancer patient in the participants or their relatives was asked.
The second part of the questionnaire included questions measuring PMT theoretical constructs including perceived vulnerability (e.g., If I have been exposed to sunlight for a long time, my skin will be damaged) (4 items), perceived severity (e.g., Skin cancer is not too concerning) (3 items), perceived rewards (e.g., It’s a pleasure to be under the sunlight) (3 items), perceived fear (e.g., I feel bad about skin cancer) (3 items), perceived response (e.g., If I use hat and sunglasses, I reduce the risk of skin cancer) (3 items), perceived costs (e.g., It’s time-consuming to wear a hat and sunglasses) (6 items), perceived self-efficacy (e.g., I can prevent skin cancer) (5 items) and protection motivation (e.g., I decided to be less exposed to sunlight) (5 items) and also skin cancer preventing behaviors (8 items). The responses in the theory constructs were scored using 5-points Likert scale ranging from 1(strongly disagree) to 5 (strongly agree). The responses in the behavior assessment questions were scored ranged from 0 (never) to 4 (always). Some questions were scored reverse.
Validity and reliability
To confirm face validity of the instrument, 10 experts reviewed the level of difficulty, the extent of inappropriateness, phrase ambiguity and failure in the meaning of words and lastly they presented their corrections.
To assess content validity, a panel of experts consisting 10 university professors in the area of health education were asked to assess the questions quantitatively and qualitatively. In the qualitative method, the experts were asked to assess the instrument on the basis of grammar compliance criteria, using the right words, putting the items in the right place, and scoring. Finally, their feedbacks, which were mainly related to the wording and phrasing of the items, were used to revise the instrument.
In the quantitative method, content validity ratio (CVR) and content validity index (CVI) were confirmed. To do this, 15 experts were requested to state their views for each item on the basis of 3 parts spectrum of “it is necessary”, “it is useful but not necessary” and “it is not necessary”. By attention that the number of experts were 15, so CVR amount on the basis of Lawshe table should be 0.49 to its content validity become confirmed. As CVR for all questions was higher than 0.49, so content validity was confirmed.
To assess CVI, the experts reviewed each item on the basis of relevance, simplicity and clarity. The results were applied in the questionnaire. The questionnaire reliability was assessed through Cronbach Alpha on 40 rural women with similar demographic characteristics with the study population. The questionnaire Cronbach Alpha was higher than 70%.
Path analysis was used to assesses PMT and predict preventive behavior of skin cancer. The used indices were Chi2 which its insignificant amount indicates theoretical fitness with the data, the ratio of chi2 to degree of freedom in which the amount lower than 3 is preferred, and comparative fit index (CFI), goodness of fit index (GFI), adjusted goodness of fit index (AGFI), normed fit index (NFI) which amounts higher than 0.9 is favorable for all these items. Regarding root mean square error of approximation (RMSEA) and root mean square of residuals (RMSR), the amounts lower than 0.05 is very good and 0.08 is acceptable (22).
The collected data were analyzed using SPSS 22 and LISREL8.8 through intraclass correlation coefficient, maximum likelihood method and correlation matrix. The linear structural relations model (LISREL) was used to determine whether the data fit the model.