Study design
Reporting of the two cross-sectional studies adheres to GRRAS guidelines (23). After the content validity had been confirmed during the translation and adaption, two cross-sectional studies were conducted in a tertiary hospital in Hubei, China. The convergent validity, construct validity, internal consistency, and homogeneity were evaluated both in the first and second studies. The test-retest reliability was assessed only in the first study. The two studies were undertaken from January-February in 2021.
Sample Size And Participants
In two cross-sectional studies, participants were included if they had a registered nurse qualification certificate and at least 12 months of work experience by convenience sampling. The construct validity was evaluated by exploratory factor analysis (EFA) in study one and by confirmatory factor analysis (CFA) in study two.
The sample size focused on EFA in the first study. According to the principles for EFA minimum sample size (5:1 ratio of participants to variables) (24), and considering the possibility of 10% invalid answers, a sample size of 187 participants was more than adequate for EFA. The PSNSC-SC was distributed to 216 clinical nurses.
The sample size focused on CFA in the second study. The sample source of CFA is different from EFA, and the sample size is at least 200 larger than for EFA (25). Therefore, another 373 emergency nurses completed the PCNSC-SC.
Instruments
Demographic characteristics information questionnaire
A socio-demographic questionnaire was developed through literature review and expert consultation and included 11 items: age, gender, ethnicity, educational level, professional title, marital status, area of employment, employment in years, religious beliefs, PC training, and experience in caring for dying patients.
Palliative Care Nursing Self-competence Scale
The PCNSC is composed of 34 items and eight dimensions, namely, physical needs: pain and other symptoms (eight items), psychological needs (four items), spiritual needs (four items), needs related to functional status (three items), ethical and legal issues (four items), interprofessional collaboration and communication (three items), personal and professional issues related to nursing care (four items) and end-of-life care (four items). It is scored using an 11-point Likert scale. The score for each item ranges from 0 (not competent at all) to 10 (highly competent) and the total score is 340. Higher values reflect a higher level of perceived self-competence. The Item-level Content Validity Index (S-CVI) was 0.95 which was first determined by Canadian clinical nurses (17).
With the author’s permission to use the original scale, the translation of the PCNSC was based on Brislin's Translation procedure (26). The scale was translated from English into Chinese by two independent native Chinese translators and back-translated by two bilingual English translators on the research team. The preliminary translated version was submitted in two rounds to a committee of six experts to review and evaluate the semantic equivalence and conceptual equivalence (27). This committee consisted of three college faculty with PhDs in nursing and three hospital nurses with bachelor’s degrees and extensive work experience. These experts first provided comments and suggestions regarding whether the items could be understood and applied in the Chinese context using a Likert scale with responses from 1 to 4 (1 = not relevant at all, 4 = very relevant). Pilot testing was carried out among a group of 25 eligible clinical nurses according to the principle that the sample size of the pilot testing accounted for 1/10 of the sample size of the formal investigation (28). Then, The pre-final Chinese version was developed (Supplementary 1).
Data Collection
A survey was carried out using Wenjuanxing (https://www.wjx.cn), an online survey system. The researchers distributed a link to the survey via WeChat (http://weixin.qq.com) and explained the aim of this survey. Moreover, enable the IP address restriction function to ensure that each IP address can fill in only one survey. For quality control purposes, the questionnaire could not be submitted if there were missing items. Finally, two researchers downloaded and checked the data.
Data analysis
IBM-SPSS statistical version 24 and AMOS version 23 (29) were used for data analysis. Descriptive statistics were used to profile the general demographic data.
The content validity was calculated by the committee of six experts using a four-point Likert scale (1 = not relevant at all, 4 = very relevant). While the Interrater Agreement (IR) level is above 0.7, the Content Validity Index can be calculated involving the content validity at the average scale level (S-CVI/ Ave), the content validity index of universal agreement (S-CVI/UA), and the content validity index of each item (I-CVI).
The construct validity was evaluated by EFA in the first study and by CFA in the second study. A Kaiser-Meyer-Olkin (KMO) test value above 0.7 is acceptable, and Bartlett’s test of sphericity was significant (p < 0.001), supporting EFA and CFA (25). The principal component extraction method was used for EFA, and obtained factors were orthogonally rotated through the varimax method. The criteria of factor extraction is followed by eigenvalues greater than 1.0, and the criteria of item retention are set by a factor loading cut-off greater than 0.4. The maximum likelihood method was used for CFA. The fit indices include the chi-square degree of freedom (χ2/df), root means square error of approximation (RMSEA), the goodness of fit index (GFI), adjusted goodness of fit index (AGFI), comparative fit index (CFI), Tucker-Lewis index (TLI), incremental fit index (IFI) and normal fit index (NFI).
For the reliability was determined by homogeneity, internal consistency, and test-retest reliability. The homogeneity was obtained by calculating the correlation coefficient (r) between the scores of each item and the total score. There was a significant association while the r was more than 0.4 (p < 0.05). The internal consistency was evaluated by Cronbach’s alpha coefficient for the scale and each dimension. Furthermore, the PCNSC was filled by 30 participants in a two-week interval in the first study to evaluate the test-retest reliability. The scores acquired in these two tests were compared by the Pearson correlation (Pearson r).
Ethical Considerations
The studies were authorized by the Academic Ethics Committee of ***blinded to review***. All participants gave informed consent to the survey. The studies were conducted with authorization from Dr. Desbiens following ethical guidelines for methodological research.