Table of Contents
The current methodological study was carried out on female breast cancer survivors referred to hospitals affiliated with Mashhad University of Medical Sciences from May to November 2022.
Study sample
The sample used in this study was selected through convenience sampling. Women with breast cancer had the following inclusion criteria: a history of at least two years of breast cancer, having no recurrent or metastatic disease, receiving treatment in one of the stages of surgery, chemotherapy, or radiotherapy, consenting to participate in the study, and having the ability to answer the questionnaire items. The exclusion criteria were no mental disorder, not using psychiatric medications, and reluctance to participate in the study. According to Kellar and Kelvin (2012), the minimum sample size to perform factor analysis is five to ten times higher than the number of items in the intended instrument [12, 13]. Thus, the Persian version of the HLI-BCS contains 20 items: 420 patients with breast cancer participated in two stages, including 220 samples for confirmatory factor analysis (CFA) and 200 patients meeting the inclusion criteria, to assess the convergent validity between the Persian versions of the HLI-BCS and the 12-item Short Form Health Survey (SF-12).
Study instruments
The Demographic Information Questionnaire
Data were collected using the demographic information form, Persian version of the HLI-BCS, and SF-12. The demographic information form included age, marital status, education level, employment status, economic status, health insurance, type of treatment, cancer stage, and duration of breast cancer.
The Healthy Lifestyle Instrument for Breast Cancer Survivors (HLI-BCS)
HLI-BCS was developed by Wang et al in 2015 [11]. This instrument contains 20 items in five areas: dietary habits (items 1–5), environment and physiology (items 6–8), health responsibility and stress management (items 9–13), social and interpersonal relations (items 14–16,) and spiritual growth (items 17–20). Through the information obtained by filling out the instrument, the lifestyles of breast cancer survivors can be evaluated in different dimensions, and plans can be made to improve their lifestyles and quality of life. Of the instrument items, 13 had negative expressions and seven had positive expressions. The scoring of the items was based on a 5-point Likert scale from 1 to 5 [1 (never), 2 (relatively), 3 (sometimes), 4 (often), and 5 (always)]. The total score on the numerical instrument ranged from 20 to 100. In the psychometrics of the original version of the instrument, face validity and content validity were confirmed by 10 patients with at least five years of disease and five oncologists with ten years of experience working with patients with breast cancer. CFA has confirmed the instrument’s construct validity on 230 breast cancer survivors in the Taiwanese population, and the mean Cronbach’s alpha for the instrument’s dimensions was 0.8 [10].
The Iranian version of 12-item Short Form Health Survey (SF-12)
The SF-12 is the short form of the 36-item Short Form Health Survey (SF-36), designed by Ware et al. in 1996. For the reliability of the questionnaire, the calculated Cronbach’s alphas for the physical and mental dimensions were 0.89 and 0.76, respectively, showing the favorable reliability of the questionnaire items [14]. Montazeri et al. [15] published the Iranian version of SF-12 containing 12 items in two physical and mental health dimensions. The present questionnaire investigated the quality of life regarding the general perception of one’s health, physical performance, physical health, emotional problems, physical pain, social performance, vitality, and perceived mental health. The scores obtained by the participants showed good status (37–48), moderate status (25–36), and poor status (12–24). The reliability of the Iranian version in physical and mental dimensions was reported as 0.73 and 0.72, respectively [15].
Translation
The translation was performed using a forward–backward translation protocol from English to Persian using the World Health Organization (WHO, 2016) method [16]. Two English-to-Persian translators were asked to independently translate the questionnaire. A Persian-to-English translator was then asked to re-translate the Persian questionnaire into English. It was then sent to five experts in Persian language and literature, English language, instrument design, and nursing for final approval to ensure the precise transfer of concepts and proper translation of the translated instrument. After receiving these comments, an initial Persian version was prepared.
Analysis methods
Face and Content validity
A cognitive interview and pretest were performed with the target group to ensure translation validity. Accordingly, ten patients with breast cancer who met the criteria to fill out the questionnaire were interviewed individually regarding the clarity of every single item and the precise measurement of the real variable using the Persian instrument. The target group was asked to express their opinions and suggested words to better understand each item or to replace words [17].
A group of 15 experts calculated the content validity ratio (CVR) and content validity index (CVI) to assess the relevance, clarity, and simplicity of the items. The group of experts consisted of faculty members from the nursing department, oncologists, oncology nurses, oncology surgeons, clinical nutritionists, psychiatrists, and social workers. When the number of experts was 15, the minimum acceptable CVR was 0.49, based on Lawshe’s table [18]. The minimum acceptable CVI value for each item is 0.7 [19].
Construct validity
CFA was performed using the maximum likelihood estimation method and the most common goodness-of-fit indices for 220 participants to evaluate the structural factors of the Persian version of the HLI-BCS. Model fit indices were evaluated according to (2 < χ2/df < 5), the root mean square error of approximation (RMSEA < 0.08), the comparative fit index (CFI > 0.9), the goodness of fit index (GFI > 0.9), parsimonious normed fit index (PNFI > 0.5), and the Tucker–Lewis index/non-normed fit index (TLI/NNFI > 0.9). Items with factor loadings less than 0.4 were removed from the model [20, 21].
Convergent validity
The correlation between the results of the Persian version of the HLI-BCS and the SF-12 was compared to evaluate the convergent validity of the questionnaire. The Pearson’s correlation coefficient was used for the convergent validity test. Pearson correlation assumes that the variables are parametric; both variables should be quantitative and interval/ratio, and the data should have a normal distribution. In this study, the normality of the data was determined based on a kurtosis between ± 7 and skewness between ± 3. Therefore, a Pearson’s correlation coefficient of ≥ 0.40 was considered acceptable (17).
Reliability
Internal consistency was examined using Cronbach’s alpha to check instrument reliability. Cronbach’s alpha ranged from 0 to 1. Higher values indicate higher reliability, and a minimum level of 0.7 is recommended for alpha. In addition, the test–retest method and intraclass correlation coefficient (ICC) were used to determine reliability, and 30 patients who met the criteria for completing the instrument responded to the questions of the Persian version within two weeks. The ICC rate is between zero and one, the reliability coefficient of 0.6 is acceptable, and reliability coefficients of 0.8 and higher denote the instrument’s great stability [17, 22].
Multivariate normality and outliers
Univariate and multivariate distributions were assessed for skewness and kurtosis, and Mardia’s coefficient for outliers. One of the signs of deviation from the normal distribution was a Mardia’s coefficient > 8. Multivariate outliers were evaluated by assessing the Mahalanobis distance. Items with a Mahalanobis distance of p < 0.001 were considered as multivariate outliers. All statistical analyses were performed using IBM SPSS Statistics version 24 and IBM SPSS Amos version 25. A significance level of p < 0.05 was considered for all statistical tests [23].