Recruitment and participants
Data collection took place from October 2019 to January 2020, a cross-sectional survey was conducted among colorectal cancer patients from four hospitals in Henan Province. This study was granted approval from the Ethical Review Board of Henan Provincial People’s Hospital, Zhengzhou, China. All the participants wrote informed consents before participating in the study. Inclusion criteria included, >18 years old and willing to participate in this study; 1~6 month after permanent enterostomy (which is referring to a colostomy in this study); Exclusion criteria included: patients who were not willing to participate; Diagnosed with mental illness and cognitive impairment. According to the Kendall criterion (i.e. 5–10-fold the number of items). A total of 16 items (7 items of general information, 4 dimensions of uncertainty, 2 dimensions of perceived stress, and 3 dimensions of self-care ability) were used in this study, with regard to 20% of the sampling error, 96-192 patients were needed, and 462 enterostomy patients participate in the study. Finally, 450 participants complete the research, with the recycling rate of 97.4%.
Measures
Uncertainty
The Mishel Uncertainty in Illness Scale (MUIS) was designed by Mishel[12], and localized by Xu Shulian[15], is a valid and reliable evaluation tool, which contains 32 items and was divided into four dimensions, including unpredictability (5 items), complexity (7 items), ambiguity(13 items), and lack of information (7 items), each item ranks by Likert 5-point scoring method. The total score of the MUIS ranges from 32 to 160, and was defined as 3 levels. The range of 32 to 74.7 was defined as low level, 74.8 to 117.4 as moderate level, 117.5 to 160 as high level, and higher scores indicating a higher degree of uncertainty in illness. The Cronbach’s alpha coefficient of the scale was 0.857.
Perceived stress
Perceived stress was measured by The Chinese Perceived Stress Scale (CPSS), which could assess the unpredictable, uncontrollable stress, and demonstrated to have a good reliability and validity. CPSS was 14 items and 2 subscales, each item was ranked by Likert 5-point scoring method, the total score range from 14 to 70, the higher score indicating greater awareness of stress. The Cronbach’s alpha coefficient of this scale was 0.891[16].
Self-care ability
Self-care ability was measured by self-care ability scale, which involved 45 items and three dimensions: self-care intention, self-care knowledge and self-care skills. The dimension of self-care intention has 12 items, the total score ranked 12~48, and the score≧29 indicated higher level of self-care intention. The dimension of self-care knowledge has 21 items, the total score ranked 0~21, the score≧13 indicated higher level of self-care knowledge. The dimension of self-care skills has 12 items, the total score ranked 0~12, the score≧7 indicated higher level of self-care skills[17]. The three dimensions (self-care intention, self-care knowledge and self-care skills) of this scale were all demonstrated to have a good reliability and validity, the Cronbach’s alpha coefficient of them were 0.831, 0.869, and 0.897, respectively.
Statistical analysis
SPSS version 21.0 (IBM Corporation, Armonk, NY, USA) and AMOS22.0 (IBM Corporation, Armonk, NY, USA) were used to conduct data analysis. Sociodemographic variables were analyzed by descriptive statistics. The relationships among uncertainty, perceived stress, and self-care ability were analyzed by parson correlation. Three multiple regression equations were conducted to check out whether uncertainty and perceived stress could affect the self-care ability remarkably. The study of Dale[18] and Goldblatt[19] had found that sex, age, educational level, and household income were closely related to the level of self-care ability. Therefore, sociodemographic variables (i.e. sex, age, educational level, family residence and household monthly income) were admitted in Model I to control for their effects on self-care ability. Then, Model II was conducted with uncertainty dimensions added based on Model I, Model III was established with perceived stress dimensions added based on Model II.
In order to verify the hypothetical relationship among uncertainty, perceived stress, and self-care ability, the structural equation model was conducted with the bootstrap method (5000 replicates) . The mediation model diagram was shown in Fig.1, in which uncertainty as an exogenous latent variable illustrating by four observed variables (i.e. unpredictability, complexity, ambiguity and lack of information), perceived stress as an exogenous latent variable illustrating by two observed variables (i.e. sense of uncontrollable and sense of nervous), self-care ability as an exogenous latent variable illustrating by three observed variables (i.e. self-care intention, self-care knowledge and self-care skills). Indices as: c2/df, the root mean square error of approximation (RMSEA), comparative fit index (CFI), goodness-of- fit index (GFI), Tacker-Lewis index (TLI) were used to determine whether the hypothesized model fit the observed data well.