A quasi-experimental crossover study on student engagement and performance in a gamified team-based learning module in physiology

Abstract

Purpose

Gamification is widely adopted in medical education, but recent systematic reviews reveal ambiguous effects on learning outcomes. It remains unclear whether engagement translates to learning or whether game mechanics impose extraneous cognitive load. This quasi-experimental crossover study investigated gamification's impact on student performance and engagement in Team-Based Learning.

Method

250 first-year MBBS students at a tertiary medical college (2024-2025) participated in two physiology modules (Central Nervous System, Endocrinology) using either traditional or gamified Group Readiness Assurance Tests (GRAT) in a crossover design. Primary outcomes were Individual Readiness (IRAT) and Application Exercise (AE) scores; secondary outcomes were engagement and satisfaction.

Results

Primary outcomes (IRAT and AE scores) did not differ significantly between formats. However, learning trajectory analysis revealed a potential mitigating trend: in Endocrinology, traditional groups showed significant IRAT-to-AE decline (–3.3 percentage points, p = 0.030, d = 0.19), whereas gamified groups' decline was non-significant (–1.9 pp, p = 0.124, d = 0.13).During the GRAT intervention phase, traditional teams outperformed gamified teams in Endocrinology (p = 0.006, d = 1.04), but not CNS (p = 0.43). Qualitative feedback was overwhelmingly positive (90% of responses), praising engagement and teamwork, though 13% cited time pressure as stressful.

Conclusions

Gamification may enhance engagement and may provide a motivational buffer against performance decline in challenging content. This suggest gamification is a context-sensitive design choice requiring careful calibration to balance motivational benefits against cognitive costs.

Introduction

Team-Based Learning (TBL) has become a cornerstone of active learning in health professions education. Its structured sequence of individual preparation, readiness assurance, and application fosters accountability and higher-order thinking [1].

A comprehensive umbrella review of 23 systematic reviews covering over 300 studies and 60,000 learners confirms TBL's effectiveness in improving knowledge acquisition, clinical reasoning, and teamwork skills across diverse medical curricula [2]. The Group Readiness Assurance Test (GRAT) is a critical phase where collaborative discussion reinforces concepts and corrects misconceptions [3]. During GRAT, teams work together to answer the same questions completed individually in the IRAT, receiving immediate feedback that creates teachable moments and promotes peer learning [1,4]. As educators seek to optimise active learning, gamification which is the “use of game design elements in non-game contexts” has emerged as a popular tool [5]. Proponents argue that mechanisms like points, competition, and rewards can increase learner motivation, participation, and enjoyment, with successful implementations in topics from anatomy to pharmacology [6–9].

Recent systematic and scoping reviews [10] document growing adoption of gamification in medical education, including applications in pharmacology [11,12], surgery [13,14], pathology [15], and basic science [16,17]. However, these reviews consistently report mixed effects on learning outcomes and highlight methodological limitations, including small sample sizes, lack of control groups, and inadequate theoretical frameworks [18–22]. The GRAT represents a particularly promising target for gamification for several reasons. First, it is inherently collaborative, aligning with game mechanics that reward teamwork and collective achievement [23]. Second, it provides natural opportunities for immediate feedback and visible progress core gamification elements [22]. Third, the competitive element between teams already exists in traditional TBL, making game mechanics a logical extension rather than an artificial addition [1,2]. Fourth, unlike the IRAT (where individual assessment integrity must be preserved) or application exercises (where deep analytical thinking is paramount), the GRAT balances knowledge reinforcement with team engagement, potentially accommodating game elements without compromising learning objectives [1,4].

However, the critical assumption that increased engagement automatically leads to improved learning remains largely untested. The literature on gamification presents a mixed and often contradictory picture [24–26]. A recent systematic review analysing gamification through the Structure of Observed Learning Outcomes (SOLO) taxonomy framework found that 78% of studies assessed only lower-order cognitive outcomes (knowledge recall, comprehension), with few examining higher-order skills like analysis and synthesis [19]. This raises questions about whether gamification enhances surface-level engagement at the expense of deep learning [19].

This ambiguity points to a crucial conceptual tension. On one hand, Self-Determination Theory (SDT) suggests gamification can fulfil core psychological needs for competence (skill mastery), autonomy (choice), and relatedness (teamwork), thereby boosting intrinsic motivation [27]. SDT distinguishes between autonomous motivation (personally endorsed, volitional) and controlled motivation (driven by external pressures). Recent applications in medical education demonstrate that autonomously motivated learning leads to better educational outcomes and enhanced well-being [28–30]. While on the other hand, Cognitive Load Theory (CLT) posits that learning is optimised when cognitive resources are focused on the learning task (germane load) [31]. CLT identifies three types of cognitive load: intrinsic load (inherent task complexity), extraneous load (non-essential processing imposed by poor design), and germane load (effortful schema construction) [32,33]. If game mechanics such as complex rules, extreme time pressure, or distracting visuals are used they may risk imposing a high extraneous cognitive load, which consumes mental resources that should be dedicated to learning. Recent work applying CLT to medical education emphasises that high element interactivity in clinical tasks makes learners particularly vulnerable to cognitive overload [33].

Using a quasi-experimental crossover design, we evaluated a gamified GRAT on student performance and satisfaction during two physiology TBL sessions. Our primary aim was to move beyond a simple "does it work?" question and instead examine how gamification operates within a specific TBL phase, one chosen for its particular alignment with gamification principles that is exploring the complex interplay between engagement, game design and learning outcomes. We hypothesised that while gamification might enhance students' affective experience (engagement, enjoyment) its impact on cognitive outcomes (performance) would depend on the balance between motivational benefits and extraneous cognitive load imposed by game mechanics. This study therefore contributes a theoretically framed crossover evaluation that tests efficacy in a specific TBL context, identifies conditions under which motivational benefits and cognitive costs may simultaneously operate, and generates testable hypotheses for future research. Outcome measures were intentionally focused on immediate knowledge acquisition and engagement appropriate for establishing foundational evidence base within a single TBL cycle with performance-based clinical assessment and long-term retention identified as priorities for future investigation.

Materials and methods

Study design and setting

This quasi-experimental crossover study was conducted at the Department of Physiology, Kasturba Medical College, Manipal India, during the academic year 2024–2025. Ethical approval was obtained from the Institutional Ethics Committee of Kasturba Medical College and Kasturba Hospital, Manipal (IEC1:56/2025), and informed written consent was obtained from all participants. The study was conducted following the Declaration of Helsinki, 2013.

Participants

The study included all 250 Phase I MBBS students, representing the entire cohort. Students were assigned to one of two groups (Group A: n = 125, Group B: n = 125) based on their pre-existing batch allocation (Batch A vs. Batch B), which were formed by the medical college administration at the beginning of the academic year by random allocation. Participation was voluntary, and no academic incentives or penalties were associated with the study. All students had prior TBL experience, having completed one traditional session on blood physiology.

TBL session design and intervention

All students participated in two physiology TBL sessions covering:

• Session 1: Central Nervous System and Special Senses (CNS-SS)

• Session 2: Endocrinology

Each session followed the standard four-stage TBL sequence and lasted approximately 2 hours total:

1. Preparation: Students received reading material and learning objectives one week in advance.

2. Individual Readiness Assurance Test (IRAT): An individual MCQ-based quiz (20 questions) assessing pre-class preparation. (20 minutes)

3. Group Readiness Assurance Test (GRAT): A team version of the IRAT in either traditional or a gamified format. (30 minutes)

4. Application Exercises (AE): Case-based MCQs (15 questions) completed individually (20 minutes). The AE in this study was deliberately administered individually rather than as a group activity, to serve as a post-test of individual learning, with the IRAT as the baseline, thereby isolating the knowledge gain attributable to the intervening GRAT phase and eliminating the confounding influence of group dynamics on individual performance measurement.

A 10-minute transition interval was provided between activities to allow students to move between the individual format (IRAT and AE) and the team format (GRAT). Following the AE, a 30-minute debrief session was conducted during which facilitators reviewed each question with the correct answers and rationale; students also completed the feedback questionnaire during this period. This was done to assess individual learning outcomes as a post-test so to look for any increase in knowledge gained from baseline levels of IRAT. AE was administered individually rather than as a team activity.

The key intervention involved randomising the GRAT format (gamified vs. traditional) for one of the two sessions per group. Teams of 6-7 students were formed by instructors before the initial blood physiology TBL ensuring heterogeneous team composition. Team assignments remained constant throughout the study (across both Session 1 and Session 2) to control team dynamics and familiarity effects. Each batch of 125 students comprised 20 teams (5 teams of 7 students and 15 teams of 6 students), totalling 40 teams across both batches.

Question development and validation: IRAT and AE questions were developed by three faculty members with expertise in physiology education and aligned with course learning objectives. Questions were pilot tested with 125 students from the previous academic year and revised based on item analysis (difficulty index 0.3-0.7, discrimination index > 0.3). Questions were reviewed by an independent expert panel to ensure content validity and appropriate cognitive level. Bloom's taxonomy level was documented for each question: IRAT and GRAT questions targeted knowledge and comprehension (Bloom's levels 1–2), consistent with their readiness assurance function, while AE questions used case-based clinical scenarios targeting application and analysis (Bloom's levels 3–4), consistent with standard TBL methodology [1]. CNS and Endocrinology questions were matched for difficulty based on pilot testing.

Gamified GRAT implementation

For the gamified GRAT, each team used a single shared digital device (laptop or tablet) to encourage consensus-based decision-making. The activity was conducted using a web-based gamification platform (Interacty®). Technical support was available throughout the activity, and no technical disruptions occurred during data collection.

A total of 20 multiple-choice questions were administered in three sequential phases. The gamification framework incorporated the following mechanics:

• 1. Coin-based reward and penalty system

• Correct responses to each question in Phases 1, 2, and 3 were rewarded with one virtual “gold coin”.

• Incorrect responses resulted in the loss of one gold coin, serving as a penalty mechanism to discourage random guessing.

• 2. Bonus mechanisms

• At the completion of each phase, teams were provided a “spin-the-wheel” opportunity, introducing an element of randomness and strategic anticipation. Possible outcomes included:

• “Free Pass”: a one-time waiver allowing the team to avoid penalty for a single incorrect response in the subsequent phase.

• “Award Gold Coins”: a bonus of two additional gold coins added to the team’s total.

• “Steal Gold Coins”: the opportunity to transfer up to two gold coins from another team of their choice.

• 3. Time-pressure mechanics

• A time-sensitive scoring mechanism was incorporated, wherein faster responses were associated with higher coin retention, reinforcing prompt yet collaborative decision-making.

• Delayed submissions led to progressive coin deductions, operationalizing the concept of “time as money.”

This integration of reward, competition, and time pressure fostered active participation and collaboration within groups. The gamified approach thus preserved the core objectives of the GRAT while transforming the assessment into a competitive, interactive, and motivating exercise designed to promote higher levels of engagement, positive emotional response, and team collaboration. A short video of the functioning of the module has been shared as a link https://doi.org/10.6084/m9.figshare.30997195.

Due to the design parameters of the gamified platform, the GRAT utilised a single-attempt, binary scoring system. Teams deliberated to select a single consensus answer; correct responses were awarded the full coin value, while incorrect responses received zero coins. Immediate correct-answer feedback was displayed following the submission, but the system did not allow for subsequent attempts or partial credit.

Traditional GRAT

The traditional GRAT used the same questions, structure and timing as the gamified version but excluded all game mechanics. It was conducted as a conventional electronic quiz using Microsoft quiz with collaborative discussion.

Crossover sesign

The crossover design was implemented to control order and topic effects:

• Group A: Gamified GRAT in Session 1 (CNS-SS), Traditional GRAT in Session 2 (Endocrinology)

• Group B: Traditional GRAT in Session 1 (CNS-SS), Gamified GRAT in Session 2 (Endocrinology)

Outcome measures

• IRAT Scores

• Application Exercise Scores

• Student Engagement and Satisfaction measured via a validated 5-point Likert-scale questionnaire

• Open-ended Feedback on perceived experience, engagement, and team dynamics

GRAT scores were also recorded during the intervention phase to characterise team performance under each condition, though GRAT itself is primarily an educational intervention rather than an outcome assessment. Our primary learning-related analysis focuses on raw GRAT scores recorded identically for both formats, while net coin scores provide insight into how the gamification mechanics functioned as a motivational system. The IRAT-to-AE learning trajectory comparison was pre-specified as an exploratory secondary analysis; primary outcomes were absolute IRAT and AE scores compared between formats.

Feedback evaluation

The feedback questionnaire was developed based on Kirkpatrick’s four-level model, focusing on Level 1 (reaction) and Level 2 (learning) outcomes [34]. Development involved a literature review, expert panel assessment, content validity testing by three physiology education experts, and pilot testing with six postgraduates to confirm face validity. The final Likert-scale instrument covered domains such as Motivation to contribute, Teamwork helps understanding, excited to participate, staying engaged, Interactive/enjoyable, requires extra effort, improves understanding, retain better, encourages critical thinking, prepares for assessments, questions appropriately challenging or not, alongside open-ended questions for qualitative insights (Supplementary file 1).

Data collection and management

All sessions were facilitated by the same four instructors who were trained in TBL methodology. Instructors followed a standardised script for introducing each session and maintained consistent facilitation approaches across both formats to minimise instructor effects. All quantitative scores were recorded electronically during the sessions. Student attitudes and perceptions toward the gamified and non-gamified GRAT were collected immediately after each TBL session using surveys including Likert responses and open-ended comments which were transcribed verbatim for qualitative analysis. Data were de-identified and stored on a secure, password-protected file, with master datasets maintained in limited-access repository files and working files were managed in Microsoft Excel.

Statistical analysis

Data was analysed using R software version 4.3, with statistical significance set at p < 0.05. Descriptive statistics were reported as means and standard deviations (SD) for continuous variables, medians with interquartile ranges (IQR) for Likert-scale items, and frequencies with percentages for categorical data. Normality of continuous score distributions was assessed using the Shapiro–Wilk test. Within-group comparisons of student scores across gamified and traditional GRAT conditions were performed using paired t-tests or Wilcoxon signed-rank tests. Between-group comparisons for each session were analysed using independent t-tests or Mann–Whitney U tests. A mixed-model repeated measures ANOVA was used to assess interaction effects between GRAT type (gamified vs. traditional) and session order (crossover sequence). Open-ended responses were analysed thematically using Braun and Clarke’s six-phase approach. Two researchers independently coded the data to enhance credibility and reduce bias. To ensure rigour, the authors-maintained reflexivity throughout the qualitative analysis, acknowledging that their dual roles as primary physiology faculty and researchers may have influenced thematic interpretation. This was mitigated by utilising independent coding by both the authors. A post-hoc power analysis was also done using G*Power (version 3.1). Codes were systematically organised into themes, which were iteratively refined for coherence and representativeness. Data integration followed a convergent parallel mixed-methods design, with separate analyses merged during interpretation via joint display tables to identify areas of convergence and expansion.

Results

Participant characteristics

A total of 250 Phase I MBBS students participated in the study. The final sample included 125 students in Group A (gamified first, traditional second) and 125 in Group B (traditional first, gamified second). No significant baseline differences were observed between groups in terms of prior academic performance in Group A and Group B with means (SD) being 46.8 (10.6) and 44.4 (10.9) respectively, (p-value = 0.08). The demographics of the study population are shown in Table 1.

Table 1.

Demographics of the study population.

Characteristic	Frequency (n)	Percentage (%)
Total Participants (N)	250	100%
Gender
Female	134	53.6%
Male	116	46.4%
Age (Years)
Mean (SD)	18.86 (0.83)	-
Range	17–25	-
Gaming Frequency
Monthly (1-3 times)	79	34.3%
Rarely (few times a semester)	110	47.8%
Never use any digital games	41	17.8%
Previous Gamification Experience
Yes	154	67.0%
No	76	33.0%
Preferred Game Types (Multi-select)
Problem-solving (e.g., Sudoku, Word games)	114	49.6%
Quick-reaction (e.g., Action, Sports)	93	40.4%
Strategic thinking (e.g., Chess, SimCity)	85	37.0%
Story-based adventure (e.g., RPGs)	67	29.1%
Quiz or Trivia	53	23.0%
Educational/Learning Apps	51	22.2%

IRAT session scores

There was no evidence of a difference in IRAT performance between the two groups across both modules as seen in Table 2.

Table 2.

Descriptive statistics and group comparisons of IRAT, GRAT and AE scores in Gamified and Traditional groups.

Measure	Group	n	Mean ± SD	p-value
CNS IRAT (out of 20)	Gamified	117	8.07 ± 2.79	0.610
	Traditional	116	8.25 ± 2.58
Endo IRAT (out of 20)	Gamified	122	9.07 ± 2.72	0.923
	Traditional	119	9.10 ± 2.95
CNS GRAT (out of 20)	Gamified	20	11.40 ± 2.80	0.43
	Traditional	20	12.15 ± 2.89
Endo GRAT (out of 20)	Gamified	20	12.05 ± 2.28	0.006*
	Traditional	20	14.30 ± 2.30
CNS AE (out of 15)	Gamified	117	7.05 ± 2.49	0.841
	Traditional	116	7.11 ± 2.13
Endo AE (out of 15)	Gamified	122	6.52 ± 2.51	0.561
	Traditional	119	6.33 ± 2.52

Abbreviations: CNS: Central Nervous system, IRAT: Individual Readiness Assurance Test, Endo: Endocrinology, GRAT: Group Readiness Assurance Test, AE: Application Exercise, SD: Standard Deviation.

^* p-value < 0.05 significant.

GRAT performance during the intervention phase: gamified and traditional

Raw GRAT scores (Collaborative Learning Performance):

In the CNS session, there was no significant difference between formats. However, in the Endocrinology session, traditional format teams significantly outperformed gamified format teams as shown in Table 2.

Game Mechanics Performance:

Despite comparable total coins earned between sessions (CNS median: 49, IQR 44-56; Endocrinology median: 50, IQR 43-58), net winnings differed substantially: 95% of CNS teams finished with negative net scores compared to 40% in Endocrinology teams. The "steal" mechanism was particularly punitive, allowing up to 22-coin losses when targeted by other teams versus a maximum 2-coin gain from bonus mechanisms. Critically, these net coin scores reflect game mechanics performance, not learning performance that the teams could answer questions correctly (contributing to raw GRAT scores) while accumulating negative net coins due to punitive game mechanics. To assess the distributional impact of the game mechanics, teams were stratified into four performance categories (Excellent, Good, Average, Poor) based on total coins earned and the timing of the teams completing the activity faster than the other teams. This analysis revealed a top-heavy distribution where only 7.5% of teams achieved 'Excellent' status, indicating that the game's reward structure was highly unequal. Detailed bonus mechanism statistics, performance category distributions, and steal mechanism analysis are provided in Supplementary File 2 including Supplementary Table 1, Supplementary Table 2 and Supplementary Table 3.

Application exercise session scores

For CNS and endocrinology application exercises, the gamified group compared with the traditional group showed no significant differences, as shown in Table 2.

Learning trajectory: IRAT to AE gains

In the CNS topic, both gamified and traditional GRAT led to significant gains from IRAT to AE (~ + 6 percentage points, moderate effect sizes) as shown in Table 3. In contrast, during the endocrinology topic, scores declined from IRAT to AE, with a small but significant decrease in the traditional GRAT (–3.3, p = 0.030) and a non-significant decrease in the gamified GRAT (–1.9, p = 0.124) as shown in Table 3.

Table 3.

Comparison of IRAT–AE Performance Across Sessions and GRAT Formats.

Session	GRAT Format	N	IRAT Mean % (SD)	AE Mean % (SD)	Mean Difference (pp)	p-value
CNS	Gamified	117	40.3 (13.9)	46.6 (17.1)	+6.4	<0.001*
	Traditional	116	41.2 (12.9)	47.0 (14.8)	+6.2	<0.001*
Endo	Gamified	122	45.3 (13.6)	43.4 (16.7)	–1.9	0.124
	Traditional	119	45.5 (14.8)	42.2 (16.8)	–3.3	0.030*

Abbreviations: pp = percentage points; CNS: Central Nervous System, Endo: Endocrinology, IRAT = Individual Readiness Assurance Test; AE = Application Exercise; GRAT = Group Readiness Assurance Test. CNS session → both gamified and traditional GRAT led to significant gains from IRAT → AE (~ + 6 pp). Endocrinology session → both formats showed a decline in AE relative to IRAT; the decline was significant in traditional but not significant in gamified GRAT.

^* p-value < 0.05 significant

Order and carryover effects mixed-design ANOVA: session × format effects

A two-way mixed-model (repeated-measures) ANOVA was conducted with Session (CNS vs. Endocrinology) as the within-subjects factor and Format/Order (gamified-first vs. traditional-first) as the between-subjects factor.

The analysis revealed a significant main effect of Session, F(1, 224) = 10.78, p = .001, η²G = .017, indicating a difference in performance between the CNS and Endocrinology sessions. There was no significant main effect of Format/Order, F(1, 224) = 0.167, p = .683, η²G = .001, suggesting that the sequence in which students experienced gamified versus traditional GRAT did not influence overall performance. The Session × Format/Order interaction was also non-significant, F(1, 224) = 0.216, p = .642, η²G < .001, indicating that differences between sessions were consistent irrespective of the order in which the formats were experienced.

Student engagement and satisfaction (likert data)

Table 4 presents the mean Likert scores across four key domains: Engagement & Motivation, Learning Efficacy, Team Dynamics, and Instructional Design & Usability. Post-session scores were consistently higher than pre-session baseline scores, indicating a positive reception to the TBL format regardless of the specific intervention.

Table 4.

Theme-wise comparison of student perceptions (mean scores) before and after the CNS and Endocrinology TBL sessions on the basis of Likert responses.

	Traditional group (CNS)		Gamified group (CNS)		Traditional group (Endo)		Gamified group (Endo)
Domains (emerged)	Pre session	Post session	Pre session	Post session	Pre session	Post session	Pre session	Post session
1. Engagement & Motivation	3.80	3.90	3.89	4.21	3.67	3.94	3.92	4.08
2. Learning Efficacy	3.86	3.93	3.88	4.26	3.72	4.00	3.97	4.13
3. Team Dynamics	3.89	4.07	3.96	4.32	3.80	4.08	4.01	4.17
4. Instructional Design & Usability	3.84	3.93	3.89	4.03	3.67	3.97	3.94	3.95
5. Exploring gamification vs traditional TBL	N/A	N/A	N/A	3.88	N/A	N/A	N/A	3.73

Scores max. out of 5, Abbreviations: CNS: Central Nervous System, Endo: Endocrinology.

CNS Session: The Gamified group demonstrated notably higher scores compared to the Traditional group across all four comparative domains. The most distinct differences were observed in Team Dynamics (Gamified: 4.32 vs. Traditional: 4.07) and Learning Efficacy (Gamified: 4.26 vs. Traditional: 3.93), suggesting that students found the gamified format particularly effective for collaboration and understanding complex CNS concepts.

Endocrinology Session: A similar positive trend for gamification was observed, though the margins between groups were narrower. The Gamified group reported higher scores for Engagement & Motivation (4.08 vs. 3.94) and Learning Efficacy (4.13 vs. 4.00). However, perceptions of Instructional Design & Usability were comparable between the two formats (Gamified: 3.95 vs. Traditional: 3.97).

Post-hoc power analysis

We conducted a post-hoc power analysis to evaluate the adequacy of our achieved sample size. To detect a medium effect size (Cohen's d = 0.50) in individual performance outcomes (IRAT and AE) with an alpha of 0.05 and 80% power, a minimum of 128 students was required. Our cohort of 250 students comfortably exceeded this threshold. For team-level outcomes (GRAT), detecting a large effect size (d = 0.80) under the same parameters required 52 teams. While our sample of 40 teams fell slightly below this generalised threshold, it was highly sufficiently powered to detect the large effect size (d = 1.04) observed in the Endocrinology module, which required a minimum of only 32 teams.

Qualitative feedback: thematic analysis

The analysis of the 657 responses revealed four key themes related to the efficacy of the TBL sessions and the impact of the Gamification elements. Following Braun & Clarke’s six steps, four higher-order themes emerged that consistently explain students’ perceptions of the gamified GRAT within TBL [35,36]. Percentages below indicate the proportion of responses that referenced each theme (responses could reference multiple themes). Analysis of the open-ended responses revealed four key themes reflecting students’ experiences with the gamified TBL approach.

Theme 1: Learning Enhancement and Clinical Application.

Theme 2: Collaborative Process and Team Dynamics.

Theme 3: Challenge and Effort (Cognitive Load).

Theme 4: Specific Feedback on Gamification Elements.

Most students highlighted the gamified approach as more engaging, enjoyable, and motivating compared to traditional formats. Gamification was seen to foster teamwork and collaboration while adding elements of fun through rewards and competition. However, some students noted challenges such as time pressure and occasional technical issues with the reward mechanisms. A small proportion of students continued to prefer traditional TBL or felt neutral between the two methods. Several students expressed interest in extending gamification to other topics, suggesting broad acceptance of the approach. An expanded thematic analysis with themes and specific student quotes is shown in Supplementary file 3.

Discussion

Our study yielded a nuanced and paradoxical set of results that challenge simplistic assumptions about gamification in medical education. Rather than applying gamification arbitrarily, the GRAT was specifically chosen as the target phase because of its inherent alignment with game mechanics, collaborative structure, immediate feedback, and existing inter-team competition making this a theoretically justified context for testing gamification efficacy. The primary learning outcomes IRAT and AE scores were comparable across formats when analysed as absolute scores. However, learning trajectory analysis revealed a potential mitigating trend: gamification appeared to buffer against performance decline in challenging content in Endocrinology, the traditional format showed a statistically significant but small IRAT-to-AE decline (p = 0.030), whereas the gamified format's decline was smaller and non-significant (p = 0.124). While this pattern is consistent with a potential mitigating effect of gamification, the difference between these two trajectories is modest and this exploratory finding should be interpreted cautiously it is hypothesis-generating rather than confirmatory, and requires replication in future studies with larger samples and direct motivational measures.

Most gamification work in medical education focuses on review games, quizzes or simulation-style activities [10,37–40]; comparatively fewer studies embed game mechanics into core TBL phases [41–43]. Our findings demonstrate the feasibility of integrating gamification directly into GRAT, within routine teaching. Consistent with prior cautions, engagement gains do not guarantee score gains. In our data, traditional teams outperformed gamified teams in Endocrinology (moderate effect size), whereas differences in CNS were not significant, and AE outcomes remained comparable. This nuanced profile mirrors reports that topic complexity and instructional design can moderate the impact of gamification. Our results therefore support the view that gamification is a context-sensitive design choice rather than a universally performance-enhancing add-on [44]. While standard TBL benchmarks suggest iRAT and GRAT scores of 60–75% and 85–90% respectively [45], our reported averages were lower. This reflects the standard grading scales and assessment rigour of our specific educational context, where a score of 60–70% indicates proficiency. Additionally, students' initial unfamiliarity with the intensive pre-class preparation required for TBL may have contributed to these baseline scores.

The dual pathways model: reconciling SDT and CLT

To reconcile these divergent findings, we propose a dual-pathway model (Figure 1) as a post-hoc, explanatory framework not an empirically confirmed mechanism that conceptualises the simultaneous motivational and cognitive effects of the gamified GRAT and generates hypotheses for future research with direct measures of motivation and cognitive load.

Figure 1.

A conceptual model illustrating the cognitive and motivational pathways of gamified TBL.

The motivational pathway (self-determination theory)

Our qualitative data are consistent with the gamified GRAT having activated the SDT motivational pathway: students explicitly praised teamwork and collaboration (Relatedness, 10.9% of coded responses), the coin-reward system provided immediate performance feedback (Competence), and 33.5% of responses emphasised enjoyment and fun (Engagement), collectively indicating enhanced intrinsic motivation [23,27,28]. This pattern in qualitative data is consistent with an SDT-driven motivational pathway and may help contextualise the potential mitigating trend observed in Endocrinology, though it cannot independently confirm a learning benefit. The qualitative reports of sustained effort and engagement are consistent with SDT predictions [27,30], but require verification through direct motivational measurement in future studies.

The cognitive pathway (cognitive load theory)

However, this motivational benefit came at a cognitive cost.

Our gamification mechanics likely imposed high extraneous load [31–33] the non-productive processing that competes with germane schema construction through multiple simultaneous demands: Time pressure, Punitive mechanics: Split attention, simultaneously tracked coins, timer, bonus spins, and physiology content. This aligns with recent CLT-based findings in medical education demonstrating that learners engaged in high element-interactivity tasks are particularly vulnerable to extraneous load [32,33]. Prior studies in flipped classrooms [46] and obstetrics and gynaecologist education [47] have shown that excessive instructional complexity and distraction impair learning efficiency in novice learners. In our study, the added extraneous load during the gamified Endocrinology GRAT likely exceeded students’ working memory capacity, explaining why gamified teams performed significantly worse during the intervention phase despite higher engagement. In contrast, traditional teams could allocate their full cognitive resources to germane processing and schema construction.

The dual-pathway model also explains the null CNS finding despite qualitative indications of enhanced motivation (90% positive feedback), motivational benefits were offset by cognitive costs from the punitive steal mechanism (95% negative net scores) and extreme time pressure (100% negative net scores), which imposed extraneous cognitive load during collaborative discussions, interfering with knowledge construction.

This dual-pathway analysis suggests that net outcomes depend on the relative balance of these two forces and are design-modifiable and content-dependent. Taken together, these findings reveal a critical coexistence: gamification simultaneously generated motivational benefits evidenced by higher engagement scores and positive qualitative feedback and cognitive costs evidenced by impaired collaborative performance in Endocrinology operating under the same bundled intervention.

Consistent with prior work, including Mallick and Waheed’s LUDO-based intervention [15], we observed that gamification promoted teamwork, strategic role allocation, communication and adaptive problem-solving. Teams dynamically leveraged individual strengths, supported peers during technical difficulties, and responded strategically to competitive mechanics, including inter-team coin stealing. However, while such collaborative complexity may mirror real-world demands, CLT emphasises that novice learners benefit from reduced extraneous load and structured guidance, with complexity and time pressure introduced later for near-transfer practice once schemas are established [31–33].

Our mixed-methods design was instrumental in revealing these dynamics. Quantitative Likert-scale data showed minimal pre–post changes, likely reflecting high baseline expectations, brief exposure, and items referencing TBL in general rather than format-specific experiences. In contrast, thematic analysis of open-ended responses uncovered tensions that Likert scales could not capture. While students praised enjoyment, engagement, and teamwork, they simultaneously described time pressure as stressful and game mechanics as distracting. This divergence between affective endorsement and cognitive efficiency is critical. The finding that 95% of teams in CNS finished with net-negative scores illustrates how a mechanic intended to motivate can become punitive, disproportionately penalising struggling teams and potentially undermining learning. Beyond the generation of extraneous cognitive load, the punitive mechanics of the gamified GRAT introduced a significant demotivating factor. The qualitative data highlights a strong perception of unfairness, aligning with the quantitative finding that a majority of teams were negatively impacted by these specific rules. When gamified penalties decouple academic effort from the anticipated reward, they undermine the core tenets of Self-Determination Theory specifically, the students' sense of competence and autonomy. Consequently, rather than stimulating engagement, these punitive elements likely induced frustration and impaired the intrinsic motivation to perform. This indicates a fundamental structural flaw in the game rules utilised in this iteration, demonstrating that gamification must carefully balance competitive elements with psychological safety. As this study did not utilise direct psychometric measures of motivation or cognitive load, we propose the following dual-pathway model as a post-hoc, explanatory framework to generate hypotheses for future research, rather than as an empirically confirmed mechanism.

Feasibility of TBL implementation in routine schedules

The time-intensive nature of TBL warrants practical consideration. Each session spanned approximately two hours nearly double the conventional lecture slot. Integrating such sessions without displacing existing content poses a logistical challenge. However, scheduling TBL within existing two-hour laboratory or tutorial blocks can mitigate timetable disruption. Importantly, TBL consolidates preparation, peer learning, and assessment into a single session, potentially replacing separate tutorial and assessment slots and thus remaining time-efficient across the broader curriculum. Once materials and question banks are established, faculty preparation burden diminishes, improving long-term sustainability. In this study, four trained faculty facilitated each session, and the gamification platform required only a standard internet connection and shared devices, minimising infrastructure demands. Furthermore, because these TBL modules are designed as end-of-system consolidation exercises, they require a frequency of only about one session per month, making them highly feasible to maintain without overwhelming the standard academic schedule.

Limitations

Limitations include the single institution setting and Likert items oriented to TBL generally rather than format-specific constructs; and subjectivity in coding open-ended responses (mitigated by transparent theme reporting but still a consideration). Although a post hoc power analysis confirmed adequate statistical power at the individual student level (N = 250), our team-level sample size for the GRAT (N = 20 per format) limited our statistical power to detecting only large effect sizes. Additionally, this study evaluated cohort-level outcomes and did not stratify the data to assess the specific impact of gamified versus traditional formats on slow learners or students with lower baseline academic performance. A limitation of this study's gamified design was the absence of a partial credit mechanism during the GRAT. Standard TBL best practices advocate for iterative testing formats, which allow teams to make subsequent attempts for partial credit if their initial answer is incorrect. A significant limitation is the absence of a validated instrument to directly measure cognitive load. Consequently, assertions regarding extraneous cognitive load and working memory overload are inferential, drawn primarily from qualitative student feedback rather than objective psychometric data. Additionally, outcomes were measured using immediate, MCQ-based assessments; the study did not evaluate long-term knowledge retention or utilise performance-based clinical assessments. Furthermore, the gamified GRAT was deployed as a bundled intervention incorporating simultaneous elements of reward, penalty, time pressure, and competition. Therefore, it is not possible to isolate the independent impact of any single game design choice on student outcomes or cognitive load. Methodological limitations of the crossover design include limited exposure (only two sessions), inherent differences in module difficulty (CNS versus Endocrinology), potential student maturation between sessions, and the possibility of carryover effects, all of which limit strict causal inference. The gamified system also exhibited high variance between sessions, as evidenced by 95% of teams finishing with negative net winnings in the CNS & SS session compared to only 40% in the Endocrinology session. This shift likely reflects a learning curve where students adapted their strategies to avoid punitive mechanics like the 'Steal Gold' feature, which Supplementary file 2 demonstrates was the primary driver of negative scores.

Future directions

Future research should prioritise direct measurement of the dual-pathway mechanisms using validated instruments the Leppink et al [48]. Cognitive Load Scale for extraneous load and the Intrinsic Motivation Inventory [49] for autonomous motivation administered concurrently during gamified and traditional sessions. Dismantling the bundled intervention to isolate the independent effects of individual game mechanics (time pressure, reward structures, competitive elements) would clarify which design choices drive motivational benefits and which impose cognitive costs. Long-term retention should be examined through summative assessments and delayed recall testing to determine whether engagement translates beyond the immediate session. Finally, multi-institutional studies with diverse learner populations and exploration of gamification across other TBL phases would establish generalisability and identify where game mechanics deliver the greatest learning benefit relative to their cognitive cost.

Conclusion

This study suggests that gamification is a feasible and engaging approach for enhancing motivation and teamwork within TBL. However, it provides empirical support for the view that engagement is not automatically a proxy for learning and that gamification operates as a double-edged, design-sensitive intervention whose net effect depends critically on the balance between motivational benefits and cognitive costs. When game mechanics are well-calibrated to content complexity and learner readiness, they may invigorate collaboration without compromising learning; when poorly designed, they risk imposing extraneous cognitive load that undermines the very performance they aim to support. Thoughtful, theory-driven alignment of gamification with content complexity, timing and learner readiness is therefore essential to ensure that the pursuit of engagement ultimately serves, rather than subverts, the primary goal of learning in medical education. Future research with direct measures of cognitive load and motivation is now well-positioned to test the mechanisms proposed here.

Supplementary Material

Supplementary_file_2 clean.docx

Supplementary_file_2 clean.docx

Supplementary file 3.docx

Supplementary file 3.docx

Supplementary_file_1 clean.docx

Supplementary_file_1 clean.docx

Disclosure statement

No potential conflict of interest was reported by the author(s).

Funding

None.

Data availability statement

The data analysed/generated is available with the corresponding author and will be made available on reasonable request.

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/10872981.2026.2715862.

Ethical approval

Ethical approval was obtained from the Institutional Ethics Committee of Kasturba Medical College and Kasturba Hospital, Manipal (IEC1:56/2025). The study was conducted following the Declaration of Helsinki, 2013.

Disclaimers

None.

Previous presentations

None.