This research was conducted in three steps. First, the original PRE-HIT instrument was translated into Persian version. Then a cross-sectional questionnaire survey was conducted to collect empirical data from the patients using the Persian version of the PRE-HIT instrument. In Step 3, exploratory and confirmatory factor analysis were conducted to test the structural validity of the instrument.
Step 1. Translation of the PRE-HIT instrument into the Persian version
The translation task was completed in three sub-steps: forward translation, face and content validation, and back translation.
At first, items were translated by one translator, a specialist in digital health. Translation considered cross-cultural and conceptual equivalence rather than linguistic equivalence for words and phrases to ensure the translated version is concise, simple and fit with Persian language and culture.
Face and content validity
The expert panel is consisted of four faculty members; two from nursing faculty, one expert in health information management and one from medical informatics. All were familiar with psychometric studies. The panel evaluated the face validity and content validity of the Persian version of the PRE-HIT instrument both qualitatively and quantitatively. Qualitative face validity was assessed by identification of problems and ambiguity in translation, and time required to answer a question. The suggestions of every expert was taken to change words to improve clarity or modify sentences to correct grammar error, or to simplify the expression without losing meaning, or using more appropriate words. Quantitative face validity was assessed using the Impact Score, which was calculated by the formula of frequency× importance for each item. The experts ranked each item on a 5 point Likert Scale ranging from very important (Score 5) to least important (Score 1). Frequency referred to the percentage of experts who gave an item a score of 4 or 5. Importance referred to the mean score of each item . An item would be kept if its Impact Score was larger than or equal to 1.5. Quantitative content validity was evaluated by content validity index (CVI) and content validity ratio (CVR). We used the CVI to examine the relevance of each item with the PRE-HIT construct. The expert panel used a 4-point Likert Sale to rate an item (1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, 4 = highly relevant). CVI score was calculated by the following formula. Items with the CVI score greater than or equal to 0.79 were retained .
CVI= number of experts giving a rating of “highly relevant” for an item / total number of experts
The necessity of the items in the PRE-HIT construct was calculated by the Lawshe test . For this, the expert panel scored an item by 3-point Likert Scale, ranging from essential, useful but not essential, and not necessary. The CVR score was calculated by the following formula. Items with the CVR greater than or equal to 0.49 were retained .
CVR = (Ne – N/2) / (N/2)
Where Ne is the number of experts identifying an item as “essential” and N is the total number of experts.
No cross-cultural and conceptual problems were found. All items achieved the impact scores and all items were equal to or greater than 1.5, the CVI and CVR scores above 0.79 and 0.49, respectively; therefore, their face and content validity were proved.
The Persian version of the questionnaire was translated back to English by an independent translator, who does not know the questionnaire. The translator was an expert in Health Informatics. Attention was paid to conceptual and cross-cultural equivalence. Afterwards, the translator and the research team discussed the English translation and reached agreement on its validity.
Step 2. Cross-sectional questionnaire survey
The design of the questionnaire
The questionnaire was comprised of two parts. The first part asked questions about demographic characteristics such as age, sex, level of education, and ownership of the International Computer Driving Licence (ICDL). The second part contained the 4-point Likert Scale questions asking about the PREHIT items.
Sample size calculation
Because factor analysis (FA) would be applied to investigate the psychometric properties of the PRE-HIT instrument. For valid FA, 5 to 10 samples are required to address a question item . As the PRE-HIT has 28 items, 280 questionnaire responses were required.
Inclusion and exclusion criteria
Patients who met the following inclusion criteria were recruited: 1) aged 18 years or over; 2) being conscious and not having serious complications such as mental disorders; and 3) able to read in Persian.
The doctors in the teaching hospitals in Cardiology, Dermatology, Gastroenterology, and Internal Medicine recruited the patients with chronic diseases at discharge and the inpatients with stable conditions. The questionnaires were handed to the patients directly by the researchers. The aim of the study, its voluntary nature, and assurance about anonymity of results in any resulted publications were orally explained by the researchers. Informed consent was sought before distributing the questionnaire. Data collection was conducted during March 1 to August 1, 2020.
Step 3. Data analysis
Exploratory and confirmatory factor analysis
To evaluate the construct validity, the exploratory factor analysis (EFA) was conducted in SPSS version 19. Due to the significant correlation between items, the Promax rotation was used to extract the latent factors. Eigenvalue ≥1 was used to identify the factors. Explained variance of each factor and cumulative explained variance for the entire survey were obtained. The Kaiser-Meyer-Olkin (KMO) index was checked for proportion of variance in the variables that might be caused by the underlying factors. Bartlett’s Test of Sphericity was conducted to check redundancy between the variables. If an item had a Communality value below 0.5, it would be deleted .
Confirmatory factor analysis (CFA) with maximum likelihood was applied to evaluate the goodness of fit of the extracted structure by EFA. The goodness of fit indices such as Comparative Fit Index (CFI ≥ 0.90), Tucker–Lewis Index (TLI ≥ 0.90), Root Mean Square Error of Approximation (RMSEA ≤ 0.06), Chi-square/Degree of Freedom (CMIN/DF ≤ 3), and Goodness of Fit Index (GFI ≥ 0.90) were checked . Also, factor loading for each item was examined. Analysis was performed in Amos version 19.
The convergent and discriminant validity and internal reliability
Convergent and discriminant validity are two aspects of construct validity. The convergent validity, evaluated through average variance extracted (AVE), and construct reliability (CR), ensures the relationship between two theoretically related factors of a construct. The CR and AVE for the factors of a construct should exceed 0.70 and 0.50 respectively. Discriminant validity, evaluated through maximum shared squared variance (MSV), and average shared square variance (ASV), ensures no relationship between two theoretically unrelated factors. For discriminant validity, the AVE value must be higher than two MSV and ASV values . Also, the internal reliability was assessed by Cronbach's alpha coefficient with value higher than 0.7 indicating acceptable level of reliability .
Comparison of the mean value of the factors between different groups
The criteria of sex, computer literacy and education level were used for demographic groupings. Shapiro-Wilk test was conducted to assess the normality of distribution of data. One-way ANOVA test was conducted to compare means of each factor in different education levels. Independent samples t-test and Mann-Whitney U test was conducted to compare different demographic group’s performance on each factor.