Design
This is observational research that utilized the cross-sectional survey method and was divided into two steps: (1) translation of the English version of the PCSS into Chinese and (2) validation of the Chinese version of the PCSS in outpatients following total hip replacement between June and August in 2020.
Data collection
Based on the requirements for confirmatory factor analysis (CFA), we needed enough outpatients post-total hip replacement to complete this survey[16]. The participants did not have known cognitive impairments or reading problems.
Sample size
The sample-to-item ratio should be above 5:1. In practice, the ideal ratio often achieves 10:1[17]. In this study, the number of samples met the ideal conditions.According to Chinese law, ethical approval was indicated for this study.
Participants
Patients
The Hospital Research Ethics Committee approved the study protocol, and patients provided informed consent before participation. At the beginning of this research, the respondents were informed about the related rights and obligations to participants. The researchers used uniform language to explain the main points of this survey. All questionnaires were independently completed by outpatients. The privacy of the participants was protected for the entire duration of the study. A survey was conducted between June and August 2020, involving outpatients who underwent total hip replacement in the orthopedic trauma department. We randomly selected patients in the database.
The inclusion criteria were as follows: (1) outpatients status post unilateral primary total hip replacement who had no complications; (2) conscious and had comprehension ability sufficiently; (3) age > 60 years old; (4) Barthel Index[18] (BI) > 60; and (5) written informed consent before the survey.
The exclusion criteria were as follows: (1) nervous system diseases; and (2) inability to communicate with others, such as deafness and aphasia.
According to the Chinese law, Tianjin Hospital Medical Ethics Committee gave approval for this research (TJYY-2020-YLS-043).
The survey was self-administered. The researchers explained the aim of this study, participants’ rights and obligations and obtained consent from the participants prior. Participants’ privacy was protected for the entire duration of the study, and the participants must be complete the scales in 10 minutes. Finishing the first survey, participants were asked if they were willing to join the second (administered at approximately two weeks) and the third (administered at approximately two months) surveys later.
Translation of the Chinese version of the PCSS
The original English version of the PCSS comprises 16 items. The original version of the PCSS was published in English [15]. In the original version of the PCSS, all items were measured by a 5-point scale, from 1 (not at all confident) to 5 (completely confident). Higher scores reflect better self-efficacy on communication. The scale started with the following: ‘‘Think about the encounter you just had with your doctor. . .mark to what extent you feel capable of doing the following things’’[15]. According to Bandura’s suggestions [19], the initial guidance is to establish an appropriate mentality, which the patient should have when assessing their beliefs about their communication ability. Patients were then asked to judge their ability to act, not their potential [16]. Each item of the original version PCSS was translated into the initial Chinese version by two researchers respectively who have medical education backgrounds. Then, the initial version was translated back to English by two other researchers and reviewed by one of the original authors. The final Chinese version of the PCSS comprises 16 items and measures confidence in communication with doctors with the 5 Likert scale from 1 = “not at all confident” to 5 = “completely confident”.
Other scale for validation
The perceived efficacy of patient-physician interactions questionnaire (PEPPI-10) is used to measure patients’ self-efficacy in communicating with health-care professionals such as doctors or nurses[20]. There are two versions of the original PEPPI: 10-items scale and 5-items scale. They are all the 11 Likert scale (between 0 = no confidence and 10 = very high confidence). Higher score indicates that the individual has a better ability for patient-physician interaction. The Cronbach’s α coefficient of the original version PEPPI-10 is 0.91, which demonstrates a reliable theoretical basis in study. Meanwhile, the Chinese version PEPPI-10 has good validity and reliability in people with knee osteoarthritis[21].
Statistical analysis
SPSS 19.0 software (IBM, 2010) and LISREL 8.7 (Science and software international, Lincoln wood, IL, USA) were used for statistical analysis. After collecting the questionnaires, we analyzed the missing data and frequency of the scales. Then, we performed confirmatory factor analysis (CFA) to test the structural validity using LISREL 8.70. Then, we checked the distribution characteristics of scores from the Chinese version of the PCSS, tested the normality of the total score and determined the possible lower and upper limit effects, evaluating whether more than 15% of the participants scored the worst or best on the Chinese version PCSS, which represented the lower and upper limit effects [22]. CFA was used to test whether scoring of the Chinese version of the PCSS was suitable for three-factor modeling. In addition, the nonstandard fitting index (NNFI), comparative fitting index (CFI) and standard root mean square residual (SRMR) were used to assess the model fit. NNFI and CFI values ≥ 0.90 and SRMR values ≤ 0.08 were considered sufficient model fit [23,24]. In this study, we used Cronbach’s α coefficient to examine the internal consistency of the Chinese version of the PCSS. Cronbach’s α coefficient represents the average split half reliability coefficient of all items, which is the most common effectiveness measurement index in validation studies[25]. The Cronbach’s α coefficient score ranges from 0 to 1, with 0 meaning that there is no correlation of the items in the scale and 1 meaning complete correlation of all items in the scale. Many studies have proven that Cronbach’s α coefficient is a reliable index for measuring the internal consistency of a scale [26]. When the Cronbach's α coefficient exceeds 0.7, the scale has good internal consistency and can be used in clinical research [27]. Cohen's kappa was used to calculate the test-retest reliability in the validated study. The value of Cohen’s kappa has been used to examine the level of test-retested consistency [28]: < 0 represents inconsistency; 0.0-0.20 represents slight consistency; 0.21-0.40 represents fair agreement; 0.41-0.60 represents moderate agreement; 0.61-0.80 represents basically consistent; and 0.81-1.0 represents completely consistent.
We used the Bayesian network function of SPSS Modeler 18.0. This neural network builds the model by learning the potential correlation between the independent variable (the third score of the Chinese version of the PCSS) and the dependent variable. Then, the model results are verified by comparing the predicted values with the actual values. In such applications, the neural network system is better than the traditional computer, which solves the problem according to a set of instructions[29,30].