Design
This study used a cluster randomized controlled study design with pretest and posttest assessments undertaken immediately before the intervention and 1 school year after the intervention, respectively.
The students were randomly assigned by Zhengfang software (Zhengfang Corp., West Lake District, Hangzhou, China) to two classes when they entered the university. Class 1 had 164 students with eight groups, while Class 2 had 131 students with seven groups. Each group consisted of approximately 20 students. Thus, the randomization in this study occurred at the class level. An administrator employed by the university, who had no information regarding the recruitment or data collection, conducted the random allocation. Two labels (numbered 1 and 2) were placed in opaque envelopes, and the administrator then requested one of the Medical Nursing teachers to selected one envelope at random. As “number 1” was selected, class 1 was to be the experimental class while class 2 was to be the control class.
Samples and setting
Cluster sampling was used and the sample comprised all the second-year undergraduate nursing students starting the Medical Nursing course in the fourth semester at the Nursing School of Fujian Medical University (Fuzhou, China). Students who changed majors or discontinued the course or had no desire to participate in the research were excluded in result analysis.
Intervention
Teaching arrangements
The Medical Nursing course was conducted for 1 academic year, in the fourth and fifth semesters, and it involved 144 teaching hours and 8 credits, which is the highest number of credits of all undergraduate nursing courses.
The learning objectives of the Medical Nursing course are: (i) to master the theoretical knowledge of this course; (ii) to improve students’ capability regarding holistic nursing; (iii) to equip students with the ability to analyze and solve problems as well as the ability of clinical reasoning and critical thinking; (iv) to cultivate nursing students’ professional ethics. Twelve teachers took part in teaching Medical Nursing. All of them underwent the same training regarding the specific processes of blended learning and offline case-centered learning. In addition, collective lesson preparation (involving all the teachers) was conducted before teaching each chapter in order to guarantee high-quality teaching.
The study was conducted between September 2018 and July 2019. The same teachers, textbook, and references were provided to both groups. All teaching materials of flipped classroom and cases of seminars for the control groups were discussed and decided in collective preparation meetings by the teaching team of Medical Nursing The experimental class underwent blended case-centered learning including online and offline learning, whereas the control class totally underwent offline case-centered learning (Table 1). Notably, whether in lectures, flipped classrooms, or seminars, the teaching method involved case-based learning. Although blinding of the students and teachers was not possible, the data analyst was blinded.
Experimental class
Apart from traditional face-to-face lectures in class, there were three stages in the implementation of the blended case-centered learning, comprising before, in, and after each flipped classroom (Figure 2).
Before the flipped classroom, all students were required to log in the Chaoxing platform (http://i.mooc.chaoxing.com/space/index.shtml) to watch the pre-recorded lecture videos and access the online forums and study materials. Besides, group cases study was also required and necessary. Students subsequently divided into 4 subgroups (approximately five students per subgroup) to discuss the cases and the questions that the teacher provided prior to the class, recording the whole process of their discussion and producing a report. Each flipped classroom involved one group, with approximately 20 students, and they were divided into four subgroups based on the pre-class group discussions. To assess the students’ mastery of the lesson prior to class, online quizzes, which lasted approximately 10 minutes, were performed at the beginning of the flipped classroom. Subsequently, the teacher provided detailed explanations of the quiz questions according to the students’ results, then the students reported what they had discussed prior to class. Next, other classmates provided supplementary answers and questions. The teacher commented on each report, summarized and guided the students to think and discuss. After the flipped classroom, the students were requested to submit homework to the platform, raise questions in the online forums or review teaching videos. All study data were recorded via the platform and could be accessed by the teachers.
Control class
The control class underwent offline case-centered learning via traditional face-to-face lectures and seminars throughout the two semesters (Figure 3). Before the seminar, the students were required to preview the textbook. Each seminar involved one group, with approximately 20 students. During each seminar, students formed 4 subgroups to discuss the cases under the guidance of the teacher. Subsequently, the teacher randomly selected subgroups to answer questions. Comments and explanations were provided by the teacher according to the answers. Additionally, the teacher guided the students to think, discuss, and summarize the knowledge points of the class. When class was almost over, a quiz was conducted to monitor the students’ mastery of the lesson. After the seminar, the students were required to finish the homework and summarize the key knowledge points and then submit the learning notes.
Measurements
Academic performance
The Medical Nursing assessment contains two parts, the final exam and the process assessment. The maximum total score for the course was 100. The total score was determined using the weights (Table 1).
The final exam was conducted one week after the course was finished. The total score of the final exam is 100, it consists of 50 multiple choice (50%), 5 explanation of nouns (15%), 3 short answer questions (15%), and 1 case analysis questions (20%). The examination questions are taken from the question bank of Medical Nursing according to the teaching outline, then reviewed by another two lecturers and finalized by a professor who was the leader of the teaching team. The content validity of the final examination was 0.9, marked by the two reviewing lectures, and the average difficulty coefficient and discrimination of this paper is about 0.6 and 0.5, respectively. The Spearman-Brown split-half reliability of the examination questions is 0.76.
The process assessment included the quizzes, online learning progress (for experimental class) or offline learning notes (for control class), performance and homework. There were 29 times quizzes in total and each quiz comprised ten multiple option questions for each flipped classroom or each seminar. The Spearman-Brown split-half reliability of the quizzes (290 questions in total) is 0.79. Online learning progress could be recorded and assessed automatically by the learning platform according to the preset weight coefficient, while offline learning notes and homework are mainly marked by teachers according to the grading criteria. Students’ performance was assessed according to their answers in class. The homework consisted of two experiment reports. The average difficulty coefficient of the quiz and homework is 0.7 and 0.5, respectively. The average difficulty coefficient of homework was assessed according the formula that is 1 minus the ratio of the average score of the homework over the full mark of the homework.
Assessment of the critical thinking ability of students
The California Critical Thinking Disposition Inventory (CTDI) has been specifically developed and used for nursing students [37]. It has been shown to be a valid instrument for assessing critical thinking ability among nursing students in different cultural contexts [38]. The population included in the present study consisted of Chinese students. Hence, the CTDI-Chinese Version (CTDI-CV), which was translated, modified, and validated by Chinese researchers, was more suitable than the original [39]. The CTDI-CV exhibits a good overall content validity index (0.89) and Cronbach’s alpha (0.90), indicating satisfactory content validity and internal consistency, respectively [39].
The CCTDI measures overall critical thinking disposition. There are 70 items in total, and it consists of seven subscales that measure the following dispositions: truth-seeking, open-mindedness, analyticity, systematicity, critical thinking self-confidence, inquisitiveness, and cognitive maturity. Items are scored with a 6-point Likert scale: 1= strongly disagree to 6= strongly agree, with a maximum total score of 420. The internal reliability coefficients (Cronbach’s alpha) in our study were 0.86 (pretest) and 0.87 (posttest), and the values for the subscales ranged from 0.73 to 0.82.
Ethical considerations
Approval was obtained from the Research Ethics Committee of Fujian Medical University. The participants were informed that they had the right not to participate and could withdraw from the study at any time. Written informed consent was obtained from all students who agreed to participate in the study. We informed the participants of the purpose, content, and extent of the study, and guaranteed that their responses were confidential.
Data collection procedure
The students in the two classes completed the questionnaire prior to and following the teaching experiment. The data collection procedure was explained to all participants, and information regarding the estimated time and number of contacts with participants was provided. Students who agreed to participate in the study were asked to provide demographic data and outcome data (critical thinking assessment) in a pretest assessment conducted during the meeting in which they were enrolled in the study. Outcome variables were measured again 1 school year after the intervention, as a posttest assessment. Additionally, we collected the final course grades of the participants. The data collection procedure primarily relied on a descriptive, noninvasive questionnaire on each student’s demographic characteristics and critical thinking ability.
Data analysis
All statistical analyses were conducted using Statistical Package for the Social Sciences (SPSS), version 22.0 (IBM Corp., Armonk, NY, USA). Mean and standard deviation were used to represent the normally distributed continuous data, while median (M) and quartiles (P25-P75) were used to represent the non-normally distributed continuous data. A Shapiro-Wilk test was first performed to verify the normal distribution of the academic performance data, and the Mann-Whitney U test was subsequently used. We used the two-tailed approach for the unpaired t test. In addition, ANCOVA analyses that were controlled for academic performance in seven professional basic courses in the previous academic year as confounders were performed to examine the effect on academic performance in Medical Nursing. ANCOVA analyses was also performed to examine the two teaching methods’ effect on critical thinking. The significance level was set at p<0.05.