Flipped classroom showdown: game-based vs. case-based learning in pharmacy education Authors: Mahtab Amini, Shahideh Amini, Maryam Alizadeh Published: 2026 ABSTRACT BackgroundActive learning methods like flipped classrooms, case-based learning (CBL), and game-based learning (GBL) are increasingly important in medical and pharmacy education. While studies suggest integrating these methods may improve outcomes, direct comparisons of CBL and GBL within flipped classrooms are limited, often focusing on small sample sizes and different student populations. This study compares the effectiveness of GBL and CBL in a flipped classroom for pharmacy education which are held completely virtual, aiming to assess learning outcomes and student satisfaction.MethodsParticipants were randomly assigned to the GBL or CBL group. In-class activities for both groups followed virtual flipped instruction classrooms on pharmacotherapy topics. Knowledge-based tests were used to assess learning outcomes, and a reliable and validated questionnaire was employed to measure student satisfaction. The basis for data analysis would include demographic data and the results of tests and questionnaires on satisfaction assessment. Data were analyzed via descriptive statistics methods (means, frequencies, percentages, standard deviations, and variances) and inferential analyses (Student’s t-test, analysis of variance, and correlation).ResultsThe 56 fourth-year PharmD students completed the study. Whereas both groups showed significant learning outcome gains (CBL T1-T3: ρ = 0.006, GBL: T1-T3 ρ = 0.001, T2-T3 ρ = 0.002), the scores for all tests were invariably higher in GBL; however, none of them are significant. Satisfaction with the learning method as also greater in the GBL group (3.82) than in the CBL group (3.61), particularly for students with lower GPAs in the GBL group (t = -0.412, ρ = 0.033). No statistically significant differences were observed for either the test scores or the satisfaction levels of either group; however, for the final test scores and progress rates, GBL had a slight-to-moderate effect size advantage (Cohen’s d = -0.36, -0.34 respectively).ConclusionsThis study demonstrated that both case-based learning (CBL) and game-based learning (GBL) within a virtual flipped classroom framework significantly enhanced student learning outcomes, without any statistically significant differences in test performance or student satisfaction between the two methods. Overall, both CBL and GBL prove to be equally effective in pharmacy education in a virtual environment.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12909-026-09754-y. Background Active learning is a type of educational paradigm in which learners are encouraged to participate actively by performing activities such as discussion, writing, manipulation of information, and even reflection of their actions in order to enhance their higher-order thinking skills [1–4]. According to the literature, active learning is any form of learning in which learners are actively engaged and encouraged to think about what they are doing during the process [5]. Some of these active learning techniques include flipped classroom, case-based learning (CBL), and game-based learning (GBL) techniques, and are widely used in medical education due to their effectiveness compared to passive lectures [1–3, 6–11]. In recent years, active learning has been widely used in pharmacy education as graduates from such programs are required to be better suited for the modern-day interprofessional healthcare environment [12, 13]. One of the advantages of active learning for students is that it encourages their engagement in the process and helps them develop critical thinking skills. Nevertheless, there are also some disadvantages associated with active learning, as it takes time and effort from educators and is not always accepted by students [2, 6, 9, 14–16]. Therefore, the choice of active learning strategies is based on the objective of learning and lesson objectives, as well as on the availability of teaching faculty resources [2, 4, 17, 18]. CBL is an instructional approach that is student-oriented, involves inquiry, and employs real-life clinical problems to get learners ready for professional practice. This strategy entails the application of theoretical concepts to actual situations [9, 19–21]. In contrast, GBL is an instructional method that requires learners to participate in a competitive activity with preset rules and game mechanics to achieve specific learning outcomes within an immersive and enjoyable environment [10, 14, 16, 22]. To address diverse educational needs, these two modalities were specifically selected for this study because they present different pedagogical spectrums: CBL focuses on analytical clinical application sometimes through a team-based collaboration [9, 19, 20, 23], while GBL emphasizes student engagement and motivation through play [15, 16, 22, 24]. As distance learning becomes a global norm, these strategies are increasingly transitioned to virtual settings [25]. Although virtual CBL has proven itself more effective for long-term results than lectures presented virtually, its adoption is not without challenges. These include the perception of a conflict between the ‘fun’ of games in a digital environment and the ‘serious business’ of learning. Proper implementation of virtual learning requires effective digital technology platforms along with a strong facilitator presence [9, 15, 16, 19, 26]. While the advantages offered by both these techniques are many, there are a few studies that directly compare GBL against CBL in the “in-class” aspect of the flipped classroom. Previous research may have involved comparison between these two methodologies [23]; however, such studies may involve postgraduate medical learners or subjects not related to pharmacy [19, 22]. Furthermore, much of the existing research on active learning in pharmacy focuses on the effectiveness of a single modality compared to lectures, rather than a direct comparison between two active learning methods in a complete virtual environment [6, 19, 27]. There is a distinct lack of research evaluating how these two strategies compare when delivered in a completely virtual, online-only environment [28, 29]. This choice of strategies aligns with the objectives of hypertension pharmacotherapy, which require both factual retention and clinical reasoning, and leverages our institutional capacity to develop standardized cases and custom digital tools. In order to fill this gap and assess which approach is better suited for the pharmacy program, this research seeks to conduct a comparison between GBL and CBL in terms of effectiveness and students’ satisfaction in a virtual flipped classroom environment among TUMS pharmacy students during the period of 2021–2023. This study will include the following objectives: The two groups were compared according to age, sex, and cumulative GPA.The performance of participants in three tests (Test 1, Test 2, and Test 3), satisfaction, and progress rates between tests were analyzed intra- and intergroups.The correlation of the cumulative GPA with test performance, the progress rates between tests, and satisfaction was measured for each group. The two groups were compared according to age, sex, and cumulative GPA. The two groups were compared according to age, sex, and cumulative GPA. The performance of participants in three tests (Test 1, Test 2, and Test 3), satisfaction, and progress rates between tests were analyzed intra- and intergroups. The performance of participants in three tests (Test 1, Test 2, and Test 3), satisfaction, and progress rates between tests were analyzed intra- and intergroups. The correlation of the cumulative GPA with test performance, the progress rates between tests, and satisfaction was measured for each group. The correlation of the cumulative GPA with test performance, the progress rates between tests, and satisfaction was measured for each group. Methods For this study, we have designed and used both a CBL-based flipped classroom and a GBL-based flipped classroom that was fully virtual. The procedures followed in the conduct of the study are shown in Fig. 1. Fig. 1The implementation steps of the study The implementation steps of the study Participants and sample size The selected population comprises fourth-year students pursuing their PharmD at TUMS who have already completed Pharmacology 1 and Pharmacotherapy 1. According to the PharmD syllabus in Iran, Pharmacology 1 entails learning about the pharmacological and therapeutic management of cardiovascular disease and conditions with special emphasis on hypertension by covering antihypertensive medication types, mechanism of action, pharmacokinetics/pharmacodynamics, clinical and laboratory data analysis, selection and application of drug therapies, dosing regimens, assessment of the effects of the drugs, as well as patient-based pharmacotherapeutics. The performance of the students will be evaluated through examination, case studies, and seminars. To minimize prior exposure bias, students who had enrolled in the drug information course or the voluntary pre-shift pharmacy training course were excluded from the study. The drug information course includes review of medications and pharmacy-related scenarios encountered in real practice settings, such as prescription interpretation, dispensing procedures, and theoretical evaluation of pharmacotherapy. In addition, the pre-shift training, which is typically undertaken after completion of the drug information course, is conducted in a real community pharmacy setting under the supervision of pharmacists and pharmacy technicians to familiarize students with pharmacy workflow, patient interactions, and professional responsibilities. Additionally, only the data of those who attended all the sessions and completed all the forms and tests were entered into the data analysis process. The sample size was calculated as 54 (27 per group) via nonprobability sampling, with a 5% significance level (α = 0.05) and 80% test power (β = 0.2), with the aim of detecting a 75% difference in standard deviation (σ = 0.75 × δ), following Chow S-C’s formula [18]. Participants were allocated using a computer-generated constrained randomization algorithm (Python software) designed to minimize differences between groups in sex distribution and cumulative GPA: the CBL group (29 students) and the GBL group (27 students). Intervention At the beginning of the study, a registration form was distributed to the target group to collect demographic information and present consent details. Test 1 was then sent to the participants, and once all participants completed it, the electronic educational content was provided as the self-directed portion of the flipped classroom. The participants were given one week to study the content. After the study period, Test 2 was administered. The participants were then randomly assigned to either the case-based learning (CBL) group or the game-based learning (GBL) group, ensuring a balanced distribution of cumulative GPA and gender. Both classes were conducted online. The first fifteen minutes were dedicated to a review session and an explanation of the class process. Next, the participants were randomly divided into subgroups of three to four members, each led by a facilitator. In the CBL group, each subgroup was tasked with answering two scenarios, each consisting of four parts. The answers were submitted as a group. Once all the groups had completed their responses, a clinical pharmacy professor in the main class reviewed and discussed the scenarios. In the GBL group, the game was facilitated separately for each subgroup, with each participant playing individually and in turn according to the game’s rules. After an hour of gameplay, any questions the participants had regarding the educational content were addressed by the clinical pharmacy professor in the main class. Finally, the participants were asked to complete Test 3 and the Satisfaction Questionnaire. Content, cases, and game development CBL and GBL were chosen due to their complementarity in teaching through active learning methods: while CBL is characterized by reasoning, GBL is characterized by the stimulation of interest. Both groups had the same educational material concerning the treatment of hypertension, consisting of audio and video presentations, along with additional material, which was provided online before the classes. For the CBL group, learning activities were conducted through structured case-based discussions focused on selected clinical questions and therapeutic decision-making scenarios. For the GBL group, the same educational domains and learning objectives were incorporated into a board game format consisting of multiple-choice practical questions developed in Python 3.9 (all steps of game design and standardization are presented in Additional file 1). Figure 2 shows the main view of the board game. The GBL system presented questions randomly from a predefined question bank, and participants could select question difficulty levels during gameplay. Although the exact questions presented during the sessions were not always identical between CBL and GBL groups, all questions and scenarios were designed to assess the same hypertension pharmacotherapy learning objectives and were standardized with input from three clinical pharmacy specialists according to Kohen’s 12 tips for effective case design [30]. With the GBL group, the game enabled students to choose any of the three levels of difficulty; these included Level 1, which involved the recollection of facts (Knowledge); Level 2, which involved comprehending information (Comprehension); and Level 3, which involved solving problems related to their clinical practice (Application & Analysis). Pedagogically, the CBL scenarios were designed to ensure that all these levels were covered in a systematic manner so that students from both groups underwent similar cognitive demands. Fig. 2The view of the board game at the beginning of the game The view of the board game at the beginning of the game Data collection A knowledge-based test and a satisfaction questionnaire were used to measure learning outcomes and satisfaction. The knowledge-based test questions were prepared in compliance with Millan’s checklist [19] and comprehensively covered the educational content. The Satisfaction Questionnaire was developed by reviewing the literature about similar questionnaires that were used in flipped classrooms and game-based learning settings. Three databases—Scopus, PubMed, and Eric—were searched via a defined search strategy (Additional file 2) and inclusion/exclusion criteria (Additional file 3). After removing duplicates, articles were screened on the basis of title, abstract, and full text, and irrelevant studies were excluded. From the final set of articles, initial codes were extracted, classified into subcategories and categories, and used to develop the initial questionnaire items (Additional file 4). The final satisfaction questionnaire consisted of 18 items rated on a 5-point ordinal response scale ranging from ‘very low’ to ‘very high,’ along with three open-ended questions. Psychometric properties of tools The knowledge-based test Content validity, face validity, and reliability of the test were analyzed. The test was checked by fifteen experts in the field of clinical pharmacy for the content validity ratio using the Lawshe method, where all the items had values more than 0.49 [31]. For the calculation of CVI, the Waltz and Bausell formula was used, with all items having a value higher than 0.75 and an average score of 0.94 [32]. Facial validity was established through individual cognitive interviews with three clinical pharmacy experts and five fifth-year or above PharmD students at TUMS to evaluate the clarity of the meaning of the questions. The revisions were made on the basis of the received feedback. The reliability of the test was evaluated with twenty PharmD students in their fifth year or above at TUMS via the Kuder‒Richardson KR20 method [33]. Additionally, the reliability was recalculated in three stages following the completion of the test. The finalized test includes 26 multiple-choice questions with 0.7 acceptable reliability. Satisfaction questionnaire To confirm the validity and reliability of the satisfaction questionnaire, both content validity and face validity were assessed. The questionnaire was sent to twelve third-year PharmD students at TUMS. In this evaluation, students assessed the content validity ratio (CVR) via the Lawshe table, with all the questions scoring over 0.56 [31]. The content validity index (CVI) was calculated according to the Waltz and Bausell formula, with all the questions scoring over 0.75 and an average score of 0.93 [32]. To establish face validity, feedback from five third-year PharmD students at TUMS was examined through cognitive interviews regarding the meaning and clarity of the questions. The research team reviewed these suggestions and made necessary changes to finalize the questionnaire. After the participants completed this questionnaire, its reliability was measured by the Cronbach’s alpha coefficient. The final satisfaction questionnaire includes 18 five-point ordinal response scale ranging from very low (= 1 score) to very high (= 5 score) plus one extra choice defining that they do not have any idea regarding the question(= 0 score) questions and three open-ended questions [34]. The reliability of the satisfaction questionnaire was evaluated separately for both groups after completion. Cronbach’s alpha was calculated as 0.92 for the CBL group and 0.94 for the GBL group, which, above 0.90, indicates a very good level of reliability. Statistical analysis Knowledge Score was defined as the sum of correct answers to 26 multiple-choice questions (range 0–26). While individual questions were dichotomous, the Knowledge Score was regarded as a continuous variable for purposes of analysis. Satisfaction Score was defined as the average of 18 Likert scale items (0–5). While individual items are inherently ordinal in nature, the Satisfaction Score was regarded as a continuous variable because of the high reliability between items. SPSS software, version 16.0 (SPSS Inc., Chicago, IL, USA), was used in data analysis. Results of descriptive statistics are presented as mean values ± standard deviation or frequencies and percentages. Independent samples t test was applied to measure differences in quantitative parameters between the CBL and GBL groups with regard to their age, cumulative GPA, test scores, progress rate, and satisfaction scores. Chi-square test was employed to find out sex distribution differences between two groups. Test scores intragroup comparison during three testing periods was done using repeated-measures ANOVA analysis, and LSD post hoc analysis was applied if necessary. Correlation analysis was based on Pearson’s method. Results Participants and their characteristics Fifty-six students from the fourth-year pharmacy doctorate of TUMS participated in this study after the elimination of those who did not meet the inclusion criteria. The demographic information of the participants was collected with a registration form. There were no statistically significant differences in the number, sex distribution (χ² (1) = 1.03, p = 0.31), or distribution of cumulative GPA between the CBL and GBL groups (Table 1). Table 1Descriptive characteristics of the study groups, means ± SDs or N (%), and their comparisonsCharacteristicsCBLGBLt Test aN/ mean% or ±N/mean% or ±Participants number2951.8%2748.2%Female1966%2178%Male1034%622%Mean of age (years)21.48± 0.5721.56± 0.890.716Cumulated GPA* (from 20 points)16.53± 1.1716.20± 1.240.302* Cumulated GPA stands for grade point averagea. The independent t test was used for the mean cumulative GPA and the mean age with a 95% confidence interval and p < 0.05. The chi-square test was used for sex distribution. A p-value < 0.05 was considered statistically significant Table 1Descriptive characteristics of the study groups, means ± SDs or N (%), and their comparisonsCharacteristicsCBLGBLt Test aN/ mean% or ±N/mean% or ±Participants number2951.8%2748.2%Female1966%2178%Male1034%622%Mean of age (years)21.48± 0.5721.56± 0.890.716Cumulated GPA* (from 20 points)16.53± 1.1716.20± 1.240.302* Cumulated GPA stands for grade point averagea. The independent t test was used for the mean cumulative GPA and the mean age with a 95% confidence interval and p < 0.05. The chi-square test was used for sex distribution. A p-value < 0.05 was considered statistically significant Descriptive characteristics of the study groups, means ± SDs or N (%), and their comparisons * Cumulated GPA stands for grade point average a. The independent t test was used for the mean cumulative GPA and the mean age with a 95% confidence interval and p < 0.05. The chi-square test was used for sex distribution. A p-value < 0.05 was considered statistically significant Learning outcomes and satisfaction The participants completed the tests in three stages. The intragroup comparison revealed significant differences between Test 1 and Test 3 in the CBL group and between Test 1 and Test 3, as well as between Test 2 and Test 3 in the GBL group (Table 2). Table 2LSD post hoc test resultsGroupijMean differencesStdSig.CBLTest 1Test 2-1.9661.0500.072Test 3-2.9660.9970.006Test 2Test 3-1.0000.6860.156GBLTest 1Test 2-1.8151.1850.138Test 3-4.0371.0220.001Test 2Test 3-2.2220.6630.002The distribution normality of tests and progress rates are approved by the Kolmogorov‒Smirnov test. Therefore, parametric tests were used to compare the results of the CBL and GBL groups. After the rejection of the assumption of homogeneity of variances by Mauchly's sphericity test (Huynh–Feldt correction for the CBL group and Greenhouse–Geisser correction for the GBL group) and indicating significant differences among the test means by intragroup ANOVA comparing the means of Tests 1, 2, and 3 (the CBL group’s F, Sig = 5.322, 0.011, and the GBL group’s F, Sig = 8.492, 0.002), the LSD post hoc test was conducted with p < 0.05 Table 2LSD post hoc test resultsGroupijMean differencesStdSig.CBLTest 1Test 2-1.9661.0500.072Test 3-2.9660.9970.006Test 2Test 3-1.0000.6860.156GBLTest 1Test 2-1.8151.1850.138Test 3-4.0371.0220.001Test 2Test 3-2.2220.6630.002The distribution normality of tests and progress rates are approved by the Kolmogorov‒Smirnov test. Therefore, parametric tests were used to compare the results of the CBL and GBL groups. After the rejection of the assumption of homogeneity of variances by Mauchly's sphericity test (Huynh–Feldt correction for the CBL group and Greenhouse–Geisser correction for the GBL group) and indicating significant differences among the test means by intragroup ANOVA comparing the means of Tests 1, 2, and 3 (the CBL group’s F, Sig = 5.322, 0.011, and the GBL group’s F, Sig = 8.492, 0.002), the LSD post hoc test was conducted with p < 0.05 LSD post hoc test results The distribution normality of tests and progress rates are approved by the Kolmogorov‒Smirnov test. Therefore, parametric tests were used to compare the results of the CBL and GBL groups. After the rejection of the assumption of homogeneity of variances by Mauchly's sphericity test (Huynh–Feldt correction for the CBL group and Greenhouse–Geisser correction for the GBL group) and indicating significant differences among the test means by intragroup ANOVA comparing the means of Tests 1, 2, and 3 (the CBL group’s F, Sig = 5.322, 0.011, and the GBL group’s F, Sig = 8.492, 0.002), the LSD post hoc test was conducted with p < 0.05 In all tests, the mean score of the GBL group was greater than that of the CBL group, but the difference was not significant. Although the progress rates from Test 1 to Test 2 were higher in the CBL group than in the GBL group, in the latter, the progress rates between Test 2 and Test 3 and between Test 1 and Test 3 were higher than those in the former (Table 3). Table 3Descriptive statistics, comparison, and effect size of the study groups' outcomesTestsDescriptive StatisticsEquality of MeansCohen's daCBLGBLLevene's Test for Equality of Variancest test for Equality of Means (95% CI)Mean± SDMean± SDFSigtSigTest114.312.9814.73.2080.4250.517-0.4760.636-0.13Test216.284.5616.525.0180.2150.645-0.1900.850-0.05Test317.284.4418.743.6960.2520.618-1.3370.187-0.36T1toT21.975.651.816.1580.1240.7270.0950.9240.03T2toT31.003.692.223.4460.1740.678-1.2780.207-0.34T1toT32.975.374.045.3100.0430.837-0.7500.456-0.20Satisfaction3.610.8083.820.8780.0530.819-0.9290.357-0.25a. Effect sizes for pairwise comparisons were calculated via Cohen’s d, which is defined as the difference between the means of two groups divided by their pooled standard deviation [35] Table 3Descriptive statistics, comparison, and effect size of the study groups' outcomesTestsDescriptive StatisticsEquality of MeansCohen's daCBLGBLLevene's Test for Equality of Variancest test for Equality of Means (95% CI)Mean± SDMean± SDFSigtSigTest114.312.9814.73.2080.4250.517-0.4760.636-0.13Test216.284.5616.525.0180.2150.645-0.1900.850-0.05Test317.284.4418.743.6960.2520.618-1.3370.187-0.36T1toT21.975.651.816.1580.1240.7270.0950.9240.03T2toT31.003.692.223.4460.1740.678-1.2780.207-0.34T1toT32.975.374.045.3100.0430.837-0.7500.456-0.20Satisfaction3.610.8083.820.8780.0530.819-0.9290.357-0.25a. Effect sizes for pairwise comparisons were calculated via Cohen’s d, which is defined as the difference between the means of two groups divided by their pooled standard deviation [35] Descriptive statistics, comparison, and effect size of the study groups' outcomes a. Effect sizes for pairwise comparisons were calculated via Cohen’s d, which is defined as the difference between the means of two groups divided by their pooled standard deviation [35] The differences in participants’ test results, progress rates, and satisfaction between the two groups were not statistically significant according to the independent-samples t test (Table 3). However, the effect size suggests a small-to-medium effect, which may have practical implications, especially in larger-scale studies, such as Test 3 (Cohen’s d = -0.36) and Test 2 to Test 3 (Cohen’s d = -0.34), which favor GBL over CBL (Table 3). The questionnaire comprised 18 Likert-scale items. The first six questions and question 18 addressed course objectives, educational content, and the questions and scenarios presented in both CBL and GBL classes. Overall, items 7 to 17 showed higher mean scores in the GBL group than in the CBL group, except for questions 8, 9, and 17. Questions 8 and 9 were related to intragroup discussions, whereas question 17 was reverse-coded and assessed the perception of time passing, such that lower scores indicated a more favorable outcome. Item-level comparisons using independent-samples t-tests revealed significantly higher scores in the GBL group for question 7 (“participation in activities contributed to learning”; p = 0.049) and question 12 (“the method met expectations compared with other teaching methods”; p = 0.011). However, these findings should be interpreted cautiously because multiple comparisons were performed (Table 4). Table 4Satisfaction questionnaire item scores, percentages of high/very high score responses, and between-group comparisonsSatisfaction QuestionnaireCBLGBLAverage Score (from 5)The total percentage of the high and very high LikertAverage Score (from 5)The total percentage of the high and very high Likertp-value a1. To what extent were the objectives of this course explained to you?4.1775.9%4.1277.8%0.9132. To what extent was the educational content consistent with the course objectives?3.8669.0%4.0077.8%0.6063. To what extent was the time allocated for this course appropriate?2.7927.6%3.0425.9%0.3444. To what extent was the technical quality of the course content satisfactory? (The technical aspect refers to the sound and image quality of the educational videos, the textbook and the content of the class, the possibility of accessing them, the technical infrastructure used in this course, etc.)3.7665.5%3.6955.6%0.9555. To what extent was the pre-class content helpful for understanding the material and preparing for CBL/GBL?3.5955.2%3.5066.7%0.9306. To what extent were the CBL/GBL questions and scenarios practical?3.9069.0%4.0077.8%0.6427. To what extent did participation in CBL/GBL activities contribute to your learning?3.3844.8%3.8870.4%0.0498. To what extent were you able to participate in classroom and group discussions during CBL/GBL activities?4.1079.3%3.4659.3%0.0739. To what extent did participation in classroom and group discussions improve your learning and understanding?3.9072.4%3.4651.9%0.26610. To what extent would you like to participate again in similar courses in the future?3.7969.0%4.1974.1%0.20111. To what extent do you support using flipped classroom combined with CBL/GBL in the pharmacy curriculum?3.6258.6%4.2381.5%0.05712. To what extent did the flipped classroom combined with CBL/GBL meet your expectations compared to other teaching methods you have experienced?3.1748.3%4.0881.5%0.01113. To what extent did you find the flipped classroom combined with CBL/GBL attractive and innovative?3.6962.1%3.9674.1%0.39014. To what extent did the flipped classroom combined with CBL/GBL increase your learning motivation?3.5262.1%4.0470.4%0.11315. To what extent did the flipped classroom combined with CBL/GBL improve your concentration during class?3.4158.6%3.8563.0%0.21516. To what extent did the flipped classroom combined with CBL/GBL enhance your enjoyment of the learning experience?3.7258.6%4.3181.5%0.06517. To what extent did you the time pass quickly during the flipped classroom combined with CBL/GBL sessions?2.8337.9%2.6333.3%0.90018. To what extent were you satisfied with this course overall?3.6969.0%4.0485.2%0.072Total3.6160.2%3.8067.1%a. Between-group comparisons for individual questionnaire items were performed using independent-samples t-tests. A p-value < 0.05 was considered statistically significant. Effect size analysis using Cohen’s d showed that most questionnaire items favored the GBL group, with several items demonstrating small-to-moderate effect sizes. In contrast, only questions 8 and 9 showed small effect sizes favoring the CBL group, whereas the remaining CBL-favoring items demonstrated negligible effects. Because 18 individual item comparisons were performed, Bonferroni correction for multiple comparisons yielded an adjusted significance threshold of p < 0.0028 (0.05/18) Table 4Satisfaction questionnaire item scores, percentages of high/very high score responses, and between-group comparisonsSatisfaction QuestionnaireCBLGBLAverage Score (from 5)The total percentage of the high and very high LikertAverage Score (from 5)The total percentage of the high and very high Likertp-value a1. To what extent were the objectives of this course explained to you?4.1775.9%4.1277.8%0.9132. To what extent was the educational content consistent with the course objectives?3.8669.0%4.0077.8%0.6063. To what extent was the time allocated for this course appropriate?2.7927.6%3.0425.9%0.3444. To what extent was the technical quality of the course content satisfactory? (The technical aspect refers to the sound and image quality of the educational videos, the textbook and the content of the class, the possibility of accessing them, the technical infrastructure used in this course, etc.)3.7665.5%3.6955.6%0.9555. To what extent was the pre-class content helpful for understanding the material and preparing for CBL/GBL?3.5955.2%3.5066.7%0.9306. To what extent were the CBL/GBL questions and scenarios practical?3.9069.0%4.0077.8%0.6427. To what extent did participation in CBL/GBL activities contribute to your learning?3.3844.8%3.8870.4%0.0498. To what extent were you able to participate in classroom and group discussions during CBL/GBL activities?4.1079.3%3.4659.3%0.0739. To what extent did participation in classroom and group discussions improve your learning and understanding?3.9072.4%3.4651.9%0.26610. To what extent would you like to participate again in similar courses in the future?3.7969.0%4.1974.1%0.20111. To what extent do you support using flipped classroom combined with CBL/GBL in the pharmacy curriculum?3.6258.6%4.2381.5%0.05712. To what extent did the flipped classroom combined with CBL/GBL meet your expectations compared to other teaching methods you have experienced?3.1748.3%4.0881.5%0.01113. To what extent did you find the flipped classroom combined with CBL/GBL attractive and innovative?3.6962.1%3.9674.1%0.39014. To what extent did the flipped classroom combined with CBL/GBL increase your learning motivation?3.5262.1%4.0470.4%0.11315. To what extent did the flipped classroom combined with CBL/GBL improve your concentration during class?3.4158.6%3.8563.0%0.21516. To what extent did the flipped classroom combined with CBL/GBL enhance your enjoyment of the learning experience?3.7258.6%4.3181.5%0.06517. To what extent did you the time pass quickly during the flipped classroom combined with CBL/GBL sessions?2.8337.9%2.6333.3%0.90018. To what extent were you satisfied with this course overall?3.6969.0%4.0485.2%0.072Total3.6160.2%3.8067.1%a. Between-group comparisons for individual questionnaire items were performed using independent-samples t-tests. A p-value < 0.05 was considered statistically significant. Effect size analysis using Cohen’s d showed that most questionnaire items favored the GBL group, with several items demonstrating small-to-moderate effect sizes. In contrast, only questions 8 and 9 showed small effect sizes favoring the CBL group, whereas the remaining CBL-favoring items demonstrated negligible effects. Because 18 individual item comparisons were performed, Bonferroni correction for multiple comparisons yielded an adjusted significance threshold of p < 0.0028 (0.05/18) Satisfaction questionnaire item scores, percentages of high/very high score responses, and between-group comparisons a. Between-group comparisons for individual questionnaire items were performed using independent-samples t-tests. A p-value < 0.05 was considered statistically significant. Effect size analysis using Cohen’s d showed that most questionnaire items favored the GBL group, with several items demonstrating small-to-moderate effect sizes. In contrast, only questions 8 and 9 showed small effect sizes favoring the CBL group, whereas the remaining CBL-favoring items demonstrated negligible effects. Because 18 individual item comparisons were performed, Bonferroni correction for multiple comparisons yielded an adjusted significance threshold of p < 0.0028 (0.05/18) In addition, there were three open-ended questions on strengths, weaknesses, and suggestions for future improvements. The respondents felt that CBL was effective in providing realistic and relevant cases and promoted teamwork; however, working virtually was not as effective as learning face-to-face, and the time was limited to discussing the cases. On the other hand, GBL received praise for competitive and innovative learning experiences and more interesting and creative content. GBL has received the following criticisms: many inconsistencies in formula have been noted, and in many instances, it relies too much on luck. Hence, all the groups preferred face-to-face contact because of better group dynamics and improved communication. The recommendations involved the clarification of questions, various technical issues, and overall learning experiences. The associations among the variables in each group were computed via the Pearson correlation coefficient, as shown in Table 5. The same analysis in the CBL group revealed that there is a positive significant relationship between the cumulative GPA and Test 3 by Pearson correlation [27] = 0.372; p = 0.047 and a positive significant relationship between the cumulative GPA and the progress rate from Test 1 to Test 3 by Pearson correlation [27] = 0.474; p = 0.009. In the GBL group, on the other hand, no significant relationship was detected between test results and progress rates and cumulative GPA; only a moderate negative correlation was found between satisfaction scores and cumulative GPA, r [25] = -0.412, p = 0.033. Table 5Pearson correlation of variables within study groupsVariablesCBLGBL12Pearson CorrelationSig. (2-tailed)Strength of relationship aPearson CorrelationSig. (2-tailed)Strength of relationship aCumulated GPATest1-0.3010.1120.1760.380Cumulated GPATest20.2000.2990.2690.176Cumulated GPATest30.3720.047Moderate0.0530.792Cumulated GPAT1toT20.3200.0910.1270.527Cumulated GPAT2toT30.2000.298-0.3340.089Cumulated GPAT1toT30.4740.009Moderate-0.0690.731Cumulated GPASatisfaction-0.0910.638-0.4120.033ModerateSatisfactionTest10.0850.663-0.2950.135SatisfactionTest2-0.1830.343-0.0460.819SatisfactionTest3-0.0550.7780.2050.306SatisfactionT1toT2-0.1920.3190.1160.564SatisfactionT2toT30.1600.4080.2870.147SatisfactionT1toT3-0.0920.6340.3210.103a. Strength of relationship in Pearson correlation is classified by Dancey C.P., Reidy J. Pearson Education; 2007. Statistics without Maths for Psychology Table 5Pearson correlation of variables within study groupsVariablesCBLGBL12Pearson CorrelationSig. (2-tailed)Strength of relationship aPearson CorrelationSig. (2-tailed)Strength of relationship aCumulated GPATest1-0.3010.1120.1760.380Cumulated GPATest20.2000.2990.2690.176Cumulated GPATest30.3720.047Moderate0.0530.792Cumulated GPAT1toT20.3200.0910.1270.527Cumulated GPAT2toT30.2000.298-0.3340.089Cumulated GPAT1toT30.4740.009Moderate-0.0690.731Cumulated GPASatisfaction-0.0910.638-0.4120.033ModerateSatisfactionTest10.0850.663-0.2950.135SatisfactionTest2-0.1830.343-0.0460.819SatisfactionTest3-0.0550.7780.2050.306SatisfactionT1toT2-0.1920.3190.1160.564SatisfactionT2toT30.1600.4080.2870.147SatisfactionT1toT3-0.0920.6340.3210.103a. Strength of relationship in Pearson correlation is classified by Dancey C.P., Reidy J. Pearson Education; 2007. Statistics without Maths for Psychology Pearson correlation of variables within study groups a. Strength of relationship in Pearson correlation is classified by Dancey C.P., Reidy J. Pearson Education; 2007. Statistics without Maths for Psychology Discussion For understanding the outcomes, it is necessary to consider the difference between the two intervention stages: Test 1 to Test 2 is indicative of the impact of self-study, whereas Test 2 to Test 3 represents the impact of CBL/GBL in class. Overall impact of the flipped classroom approach can be seen through Test 1 to Test 3. In this study, we found that both case-based learning (CBL) and game-based learning (GBL) facilitated significant improvements in knowledge test scores over time. Distinct patterns of improvement were observed within each group. Specifically, the CBL group demonstrated significant improvement between Test 1 and Test 3, whereas the GBL group showed significant improvements from Test 1 to Test 3 and from Test 2 to Test 3. These within-group improvements are consistent with previous evidence presenting that both approaches can effectively enhance student performance over time. However, we did not observe statistically significant differences in test scores between the two instructional methods, suggesting both instructional approaches, including the specifically designed game used in the GBL arm, were similarly effective in improving short-term learning outcomes in this cohort. This observation further supports the notion that key educational “active ingredients,” such as cognitive interactivity and feedback, can be effectively delivered through both well-designed games and robust case-based activities [10, 23, 36–40]. Additionally, we found that there was no statistically significant difference between the satisfaction levels of students in both groups, unlike what has been reported previously about greater satisfaction associated with GBL compared to CBL. This could be attributed to the superior quality of the CBL method used in our research or the application of stricter statistical analyses, since other research works have reported satisfaction differences without adjusting for multiple comparisons [10, 23, 36, 39]. The results of the exploratory analysis of the association between cumulative GPA and learning outcomes revealed different trends in each group. For the CBL group, the correlation between cumulative GPA and Test 3 results, and the increase from Test 1 to Test 3 proved positive. This result is in accordance with earlier studies, which indicated that CBL is highly likely to work well for students having a strong theoretical background since this method presupposes application of the theoretical material to real clinical cases [9, 20, 21]. However, no significant connection between GPA and tests’ results was found in the GBL group, which supports claims that GBL learning can be independent from traditional assessment methods of academic progress [16, 22, 23, 36]. Nonetheless, the results should be interpreted carefully due to the limited sample size. Similar to this, exploratory sub-analyses showed that there was a possible inverse relationship between GPA and satisfaction among the GBL group, since there is an existing moderate inverse relationship between the GPA and the scores for the satisfaction of the students. One possible reason could be the “play paradox,” wherein academically inclined learners view play-based activities as less effective compared to traditional instructional strategies [25]. However, since these observations were inconsistent and did not occur across all analyses performed, these results are more hypothesis-generating in nature [25, 41]. Overall, our results correlate with those reported in other similar works, showing learning results comparable for the two teaching strategies in question. Both CBL and GBL have already proven to be effective means for fostering student engagement and acquiring knowledge in active learning settings. We expand upon the current body of evidence by conducting a direct comparison of CBL and GBL within the context of flipped classroom instruction in pharmacy [42, 43]. However, it is important to take into account the heterogeneity associated with various GBL approaches. Game-based learning includes various instructional designs, game mechanisms, and learner interactions. As the game applied within the current investigation is designed for this specific course, any comparison with other GBL interventions discussed in the academic literature requires caution. There are several limitations that should be taken into consideration when discussing our results. Firstly, our sample was relatively small (n = 56), which constrained our ability to detect anything other than very large effect sizes using statistical analysis. This implies the possibility that there may have been significant differences between CBL and GBL, yet we failed to identify those differences because of Type II errors (false negatives). Therefore, non-significant statistical results do not necessarily mean that the two groups performed equally well. Secondly, while both interventions shared the same learning objectives, differences in question delivery in GBL and scenarios, in CBL, and interaction style between the CBL and GBL formats may have influenced the educational experience. Another limitation lies in the fact that the CBL approach was conducted in a collaborative group format, while GBL was primarily individual-based. As such, any differences seen might relate not only to the educational modality but to social interaction and peer learning between groups as well. Thirdly, multiple-choice tests might not be able to provide enough evidence on higher-order cognitive functions, and classification of the cognitive level of test items according to Bloom’s Taxonomy was not performed. Finally, the current study was conducted during the COVID-19 pandemic still affecting routing educational process in TUMS in a virtual learning environment. In future studies, it is recommended that more participants are involved, in-person implementation is performed, and long-term knowledge retention is investigated. In addition, future research is needed to explore the cost-efficiency and scalability of game-based learning in pharmacy curriculum implementation and to explore the interaction between individual differences and instructional techniques. Conclusion In this study, we found that both case-based learning (CBL) and game-based learning (GBL), implemented within a flipped classroom framework, were associated with significant improvements in student learning outcomes over time. However, no statistically significant differences were identified between the two instructional approaches with respect to test performance or overall student satisfaction. Although certain numerical differences and subgroup patterns were observed, these findings were not statistically robust and should therefore be interpreted cautiously. Overall, our findings support the use of both CBL and GBL as effective student-centered instructional approaches in pharmacy education, with neither method demonstrating clear superiority under the conditions of this study. Further research using larger sample sizes and more rigorously controlled study designs is warranted to better understand how different instructional strategies may influence learning outcomes across diverse student populations. Supplementary Information Additional file 1. Additional file 1. Additional file 2. Additional file 2. Additional file 3. Additional file 3. Additional file 4. Additional file 4.