The characteristics of the included studies are presented in Multimedia Appendix 5. Most studies were conducted with trainees (n=40), primarily medical students (n=17) and nursing students (n=11). Study designs were predominantly randomized parallel-group trials (n=27), followed by quasi-experimental trials (n=12), randomized cross-over trials (n=4), and cluster randomized trials (n=3). The risks of bias for each study are presented in Multimedia Appendix 6. Although 74% (34/46) of the included studies were identified as randomized trials, only 30% (14/46) were rated as low risk of bias for random sequence generation, and 33% (15/46) were rated as low risk of bias for allocation concealment. For other dimensions of bias, low risk was observed in 35% (16/46) of studies for baseline outcome measurements, 37% (17/46) for baseline characteristics, 50% (23/46) for blinded outcome measurements, 50% (23/46) for contamination, 57% (26/46) for missing outcome data, and 80% (37/46) for selective outcome reporting. The PRISMA flowchart for our review is presented in Figure 1, and the PRISMA checklist can be found in Multimedia Appendix 7.

Table 1 presents the number of studies that were informed by a theory of motivation or model of motivational design. SDT and the ARCS model were the most commonly used theories, while 24 studies did not cite any theory. Five studies cited more than 1 theory of motivation. Among the 22 studies that used at least 1 theory, we judged the theory as informing the research questions in 20 (91%) studies, informing the experimental conditions in 15 (68%) studies, informing methods and measures in 17 (77%) studies, and informing the interpretation of results in 17 (77%) studies. Nine studies used theory to inform all 4 aspects of their research process [5,10,11,62,63,68,69,77,80].

Studies investigated motivational design strategies that targeted intrinsic value beliefs in 23 of the 46 (50%) studies, extrinsic value beliefs in 9 (20%) studies, competence and control beliefs in 6 (13%) studies, social connectedness in 4 (9%) studies, and autonomy in 2 (4%) studies. Ten (22%) studies targeted more than 1 construct; of these, 5 (11%) were informed by the ARCS model. Sixteen (35%) studies did not report targeting any specific motivational construct, instead aiming to enhance motivation in general.

While intrinsic value beliefs were the most commonly targeted construct, researchers drawing on a prominent theory or model (as listed in Table 1) tended to be more pluralistic in their foci. Specifically, studies that used a motivation theory or model targeted intrinsic value beliefs (n=11) at a similar level to extrinsic value beliefs (n=9) and, to a lesser extent, competence and control beliefs (n=6). By contrast, studies that did not use a theory or model focused solely on intrinsic value beliefs (n=10) compared to extrinsic value beliefs (n=0) and competence and control beliefs (n=0).

In this systematic review, we analyzed experimental comparison studies of online motivational design strategies in HPE. We aimed to identify which motivational constructs have been most frequently targeted in these studies and which remain understudied, offering insights into potential areas for future research.

A significant finding was that nearly one-third of the studies in our review did not specify which motivational constructs their design strategy was targeting, instead broadly aiming to enhance motivation. We argue that such research is of limited value to educators. Motivational design expertise relies on educators understanding how strategies work, specifically what constructs they influence and under what conditions they are most effective [106,107]. Studies that do not clarify which constructs a design strategy influences, either conceptually or empirically, cannot provide educators with the information needed to build expertise [16]. Therefore, we recommend that researchers explicitly define the motivational constructs their strategies aim to influence and test their impact on those constructs. This recommendation can be supported through the greater use of motivational theories, which were cited in fewer than half of the studies in our review. This lack of theory use is consistent with other reviews, such as those by Maheu-Cadotte et al [19] and Bajpai et al [29], who found similarly low levels of theory use in their reviews of serious games and digital education in HPE. Motivational theory should be used to inform the research questions, the design strategy, the outcome measures, and the interpretation of results. Excellent examples of theory use are present in our sample [5,11,80].

Among the studies that did specify targeted constructs, most focused on intrinsic value beliefs (eg, interest or enjoyment), compared to extrinsic value beliefs, competence and control beliefs, social connectedness, and autonomy. Accordingly, research in this area is disproportionately focused on ways to make online instruction more interesting and enjoyable. Given the volume of studies on design strategies targeting intrinsic value beliefs, we recommend that future research synthesize existing findings to identify the most effective strategies for enhancing interest and enjoyment and outline areas for future research.

A disproportionate focus on enhancing intrinsic value beliefs aligns with an increased uptake of SDT in HPE, as documented in our studies and other reviews [24,108]. SDT emphasizes the role of intrinsic motivation—which is grounded in feelings of interest and enjoyment—in effective learning [55]. However, we found that studies using SDT were often pluralistic in the constructs they targeted, suggesting a more nuanced approach than studies without a theoretical basis. A theoretical perspective, whether based on SDT or another theory, may help researchers avoid equating motivation solely with enjoyment and interest, thus neglecting other facets of motivation, such as confidence and relatedness, despite evidence suggesting that these constructs may be particularly at risk when learning online [7,8]. Supporting this perspective, we found that studies informed by the ARCS model—which explicitly states the importance of supporting learners’ attention, relevance, confidence, and satisfaction—were most likely to report targeting multiple motivational constructs. We recommend that studies test design strategies targeting a broader range of motivational constructs to expand the set of design strategies that educators can choose from (eg, confidence-enhancing strategies or relatedness-enhancing strategies). For example, though serious games are often framed as ways to enhance interest and enjoyment, they may also be configured to support feelings of practical relevance or boost confidence [24]. Researchers could build on the serious games literature by investigating ways to design serious games to support feelings of extrinsic value, confidence, social connectedness, and autonomy.

We encourage researchers to study ways of motivating learners in established online modalities (eg, asynchronous modules or webinars) and by using emerging technologies such as virtual reality and artificial intelligence. For example, artificial intelligence chatbots have the potential to provide personalized coaching and feedback during learning [109,110]. Providing such support and scaffolding instruction in a learner’s zone of proximal development may foster a sense of autonomy and confidence. As research on the motivational design of emerging online modalities is still in its infancy, future studies could investigate how to design emerging technology-enabled instruction to optimize learner motivation.

The risk of bias was a concern across many of the included studies. To ensure that future research can make more defendable claims regarding the effects of design strategies, researchers should clearly specify procedures for random sequence generation and allocation concealment, which are often missing from published papers. They should also capture relevant variables at baseline, blind assessors to condition, and attempt to limit attrition and contamination [27].

Several limitations are worth noting. We did not include any synonyms for the word “motivation” (eg, “engagement” or “satisfaction”) or motivational constructs (eg, “value,” “relevance,” or “confidence”) in our search terms because we believed these terms would greatly increase the number of nonrelevant studies in our search results. We assumed that studies using synonyms for “motivation” or referencing motivational constructs would also use the word “motivation” and thus would be retrieved in our searches. Consequently, we may have missed some otherwise eligible studies that exclusively referenced concepts that are related to, or treated as synonymous to, motivation (eg, engagement) or motivational constructs (eg, confidence). We also chose to exclude studies written in a language other than English, which may have resulted in missed studies.

We decided to focus our review on experimental studies because they provide a critical source of evidence regarding the effectiveness of design strategies. We acknowledge that many different kinds of studies can generate evidence to support educators’ motivational design efforts when producing online learning [31,111]. For example, qualitative studies can help us understand how learners make meaning of instructional designs in context [112], and single-group studies can investigate the factors influencing engagement with motivational design strategies [113]. It may be that studies leveraging nonexperimental designs demonstrate a different distribution of foci regarding motivational constructs. We recommend that a breadth of methodologies, including but not limited to experimental comparison studies, be used to investigate novel motivational design strategies in the future.

Finally, our review focused on online instruction in HPE, and it is unclear whether the trends we observed apply to other types of HPE, such as in-person simulation. While the trend toward enhancing interest and enjoyment may also be present in other HPE contexts—such as through the gamification of in-person instruction [114-116]—we cannot make definitive claims about the generalizability of our results to other types of HPE. Conducting similar reviews in other areas of HPE may be a focus of future research.

A key challenge for educators when teaching online involves keeping learners motivated. To address this challenge, educators need access to motivational design strategies that target a range of motivational constructs. The existing research provides an important starting point, but there is much work to be done. Researchers can use our findings to guide future primary and secondary research that generates a more robust evidence base for educators wishing to motivate their learners.