Global Journal of Medical Research, J: Dentistry and Otolaryngology, Volume 23 Issue 1

items in the knowledge dimension, 13 items in the belief dimension, and 22 items in the behavior dimension. e) Reliability and validity test of the questionnaire i. Research objects Convenience sampling method was used to select nurses in the oncology department of five tertiary hospitals in Jiangsu Province as the research objects from March to April 2022. Inclusion criteria: Qualified as a nurse practitioner and engaged in front-line clinical work; working time in the oncology department ≥ 1 year; voluntary participation in this study. Exclusion criteria: rotation, advanced study, practice nurses; those who are not at work due to illness, affairs, maternity leave, etc. According to the sample size of 5 to 10 times the number of items [9], the minimum sample size is 240 cases, and considering the dropout rate of 10%, this study finally included 270 subjects. The questionnaires were collected by 2 uniformly trained research team members. The data is collected in the form of questionnaire stars, and the purpose, meaning and precautions of this questionnaire survey are introduced to the research subjects with a unified guide language; in order to avoid omissions, all items are set as mandatory items; in order to avoid invalid questionnaires, the research object with the same user and IP address can only be filled in once. After the questionnaire was collected, it was exported to excel, checked by two people, and the unqualified data were deleted. ii. Project Analysis a. Discrimination analysis method The critical ratio decision value (CR) was used to test the discriminative degree and discriminating ability of the questionnaire. The questionnaire total scores of the research subjects were sorted from high to low, and the top 27% of the total scores were in the high group, and the last 27% were in the low group. The differences in the scores of each item between the two groups were compared, and the items with CR<3 and no statistically significant difference were deleted [10]. b. Correlation coefficient analysis method By calculating the correlation coefficient of each item with the overall questionnaire and the scores of each dimension, the representativeness of each item is reflected, and the homogeneity of each item with the overall questionnaire and each dimension is judged. Items with a correlation coefficient < 0.40 with the overall questionnaire or the dimension to which it belongs are deleted [11]. c. Internal consistency reliability analysis method Calculate the Cronbach's α coefficient of the overall questionnaire and each dimension, and then calculate the Cronbach's α coefficient of the overall questionnaire and each dimension after deleting each item. If the Cronbach's α coefficient of the questionnaire increases after removing an item, delete the item [9]. iii. Validity analysis a. Construct validity The construct validity of the questionnaire was tested by exploratory factor analysis and confirmatory factor analysis. Exploratory factor analysis: It is suitable for sampling appropriateness value (KMO) > 0.6 and the Bartlett sphericity test has a statistically significant difference (P < 0.05). Contribution rate> 40%; use the orthogonal rotation to maximize the variance to obtain the component matrix, and delete the entries with factor loading values < 0.40 [12]. Confirmatory factor analysis: using the maximum likelihood method for analysis; using the ratio of chi-square degrees of freedom (x2/df), root mean square error of approximation (RMSEA), incremental fit index (IFI), comparative fit index (CFI) , Parsimony Adjustment Fit Index (PCFI), Goodness of Fit Index (GFI), Normative Fit Index (PNFI) and other results to analyze the rationality of the questionnaire structure; the reference standards for each index are x2/df<5.0, RMSEA <0.10, IFI>0.90, CFI>0.90, PCFI>0.50, GFI>0.90, PNFI>0.50 [13]. b. Content Validity The 15 experts who originally participated in the Delphi letter inquiries were invited to evaluate the content validity of the revised questionnaire, using the Likert 4-point scoring method, with "1-4 points" indicating "very irrelevant", "irrelevant", "relevant", "very relevant". ". The content validity of the questionnaire was tested by the item-level content validity index (I-CVI) and the scale-level mean content validity index (S-CVI). It is generally believed that I-CVI>0.78 and S-CVI>0.9 indicate good content validity [14]. iv. Reliability Analysis The Cronbach's α coefficient was used to analyze the internal consistency reliability of the overall questionnaire and each dimension, and the Cronbach's α coefficient was generally required to be >0.80; the questionnaire filling results of 50 oncology nurses were re-collected after 2 weeks, and the correlation between the two questionnaire scores was tested. The test-retest reliability of the questionnaire generally requires a test- retest reliability > 0.70 [11]. f) Statistical methods Double check and input data, SPSS 23.0 and Amos 23.0 were used for statistical analysis. Use mean and standard deviation, frequency and composition ratio to describe the general data of the research object; use two independent sample t test, Pearson correlation coefficient and Cronbach's alpha coefficient method to analyze items and screen items of the questionnaire; use exploratory factor analysis and confirmatory Factor analysis was used to test the construct validity of the questionnaire; I-CVI and S-CVI were used to test the 11 Year Global Journal of Medical Research Volume XXIII Issue I Version I ( D ) 2023 J © 2023 Global Journals Development of Knowledge-Attitude-Practice Questionnaire on Oral Nutrition Supplement among Nurses in Oncology Department and its Reliability and Validity Test