Data were collected through a pregame survey, 30 minutes of gameplay, and a postgame survey; a common method to evaluate SGs [25]. Pregame and postgame opioid-related survey questions were adapted from various validated surveys and scales or created by the investigators [26]. Questions from a Wisconsin statewide survey collecting perceptions, awareness, and use of prescription medications in Wisconsin residents were adapted for adolescent use and assessed the participants’ knowledge of safe prescription opioid storage and disposal [14,27]. Questions assessing self-efficacy of prescription opioid use and other learning objectives from the SG were adapted from the MUSE (Medication Understanding and Use Self-Efficacy) scale and a survey assessing workplace safety and health knowledge in adolescents [28,29]. Questions measuring participants’ knowledge about naloxone (Narcan) were adapted from the Maryland Opioid Survey Summary Report [30]. Attention check questions were included in the pregame and postgame surveys to ensure the participants were thoroughly reading survey questions and to prevent straight lining. All surveys were hosted online through Qualtrics (Qualtrics LLC), and the data collected were securely stored through Qualtrics. Surveys were reviewed by Qualtrics staff and the study team to ensure functionality prior to distribution. This study was approved by the Institutional Review Board of University of Wisconsin-Madison.

The pregame survey consisted of 66 nonrandomized items, and 59 of the items were used to examine the participants’ baseline opioid knowledge, knowledge of safety, disposal and misuse, and the perceived effect an SG may have on opioid safety awareness. These 59 items consisted of questions with “Yes,” “No,” and “Don’t know” as answers or a 5-point Likert scale. Two items were also used as attention check questions to prevent straight lining by asking the participants to select a specific option within the 5-point Likert scale. Five demographics questions were asked to assess the participants’ characteristics, including race and ethnicity, grade in school, age, gender, and the number of children under the age of 18 living in their household. The online survey consisted of 12 pages with up to 10 items on each page. The pregame survey questions are included in Multimedia Appendix 1.

The postgame survey consisted of the same 59-item baseline opioid-related questions in the pregame survey and 2 attention check questions, as well as 25 additional questions asking about the participants’ perspectives on the MedSMART game. The postgame survey included a question asking the participants if they had ever been prescribed opioids by a doctor, and questions asking them to list the opioids they had been prescribed and any prescribed medications they perceived might be opioids. Additional personal health information was not collected in the survey per Qualtrics guidelines for surveys recruiting participants through their research panels. Response options included 5-point Likert scale questions, multiple choice questions, free response questions, and “yes,” “no,” and “unsure” answers. The online survey consisted of 17 pages with up to 10 items on each page, and the items were not randomized. The postgame survey is included in Multimedia Appendix 2.

The SG MedSMART: Adventures in PharmaCity was designed to educate adolescents about safe opioid use and management as well as enhancing the players' ability to make informed decisions about proper opioid use in real life. MedSMART: Adventures in PharmaCity was developed over the course of several months by a team of researchers, health care personnel, and game design experts [18]. The key elements of the game were defined, including goals, levels, mechanics, game flow, story, and characters. The game was only playable on a computer or tablet with keyboard access and had simple commands (all player actions were made using keys “A,” “D,” “W,” and spacebar). The game engine was built using Unity (Unity Technologies). The situations presented in the game were designed to be similar to scenarios adolescents may encounter in real life. The game provided immediate feedback to the player if they made a wrong decision, and the player could only progress in the game if they made the right decision and followed the correct sequence in the story line. After the prototype was built, the game was piloted with adolescents and pharmacy students to examine whether they liked the game, understood how to use it, and could navigate throughout the game [18,21]. Their feedback was incorporated into the final game design used in this study.

MedSMART: Adventures in PharmaCity is a task-guided game divided into 5 levels. Participants play as “Shan,” the anthropomorphized sheep, who has been prescribed opioids after breaking their arm. The player’s objective is to guide Shan to make the right choices regarding the proper use and care of opioids. The player navigates through the levels, each covering different opioid safety topics. Level 1, A Quiet Sunday Afternoon, is focused on teaching the player about the safe storage of opioids; the player is taught to lock up medication in a cabinet, so their friends do not take them (Figure 1). The second level, Monday Morning Bus Ride, enhances the game story line by adding a real-life scene about being in pain and forgetting there was an important assignment due that day (Figure 2). Level 3, A Persuasive Speech at School, focuses on teaching the player not to take others pain medication and the negative consequences of taking someone else’s medications; even if 2 people are prescribed a pain killer, that does not mean it is the same kind of medication or dose (Figure 3). Level 4, Bus Ride Home, focuses on teaching the player not to share their medication with others, the negative consequences of sharing medications with others, and what Narcan is used for (Figure 4). Level 5, Last Minute Chore, shows the player the correct way to dispose of an opioid medication and why they should not throw medications away or flush them down the toilet (Figure 5). The scenarios in each level were intended to be realistic and relevant to the adolescents’ daily life.

From October to November 2020, a nationally representative (quotas were set for race or ethnicity and gender based on the US census data) sample of adolescents was recruited via Qualtrics research panels. Parents of eligible adolescents were targeted. Eligible participants were adolescents aged 12 to 18 years who lived in the United States and could speak and understand English. After screening for eligibility, the parents were provided with a link to the online consent document and asked to provide consent for their child to participate in the pregame survey. The child was also provided with an online assent document and asked to provide assent. The parents were then directed to the pregame survey for their child to complete independently. Upon completion of the pregame survey, the parents of participants received compensation based on their predetermined agreement with Qualtrics. The parents were then provided with a link to the online consent document and asked to consent to their child’s participation in gameplay and a postgame survey. The children were also provided with a link to an online assent document and asked to provide assent for participation. Once consent and assent were confirmed, the parents received a link with instructions for their child to play the SG for 30 minutes and complete a postgame survey. Upon verification of gameplay and completion of the postgame survey, a US $10 amazon e-gift card was emailed to the parent of the adolescent for participation. All consent and assent documents detailed the study activities, time commitment, purpose of study, principal investigator’s information, and confidentiality and data security measures.

Survey questions were divided into 10 categories to represent key concepts and summarized into concept scores. Pregame and postgame surveys had Likert scores on a 1 through 5 scale and knowledge scores on a 0 to 100% scale; changes in concept scores, from presurvey to postsurvey, were calculated for each participant and were the primary efficacy outcomes. Concept score changes were described by overall mean (SD) and were stratified by gender, race, school, grade, and age. Differences in concept score changes were analyzed using the Kruskal-Wallis tests for categorical characteristics and correlation tests for continuous characteristics. Primary analysis included participants who accessed the game based on IP (located within the United States) and satisfied attention checks for both pregame and postgame surveys. Secondary analyses aimed to (1) assess the differences between participants included in primary analysis vs those who did not meet full attention criteria; and (2) assess the associations between the levels played and the length of play with concept scores. The participants who did not meet full attention criteria either did not access the game or did not self-report playing the game but met other attention criteria (ie, select “A great deal” when asked). No P value adjustments were made to account for inflated type 1 error rate. Significance was assessed at the α=.05 level. All statistical analyses were performed using R statistical software, version 4.0.5 (The R Foundation).

This paper describes the use of an SG designed to educate adolescents on opioid safety. The game was evaluated in terms of impact on the players’ pregame and post survey results for self-efficacy, perceived knowledge, misuse harm and behavior, behavioral intent, safe storage and disposal, opioid knowledge, and Narcan knowledge. The results indicate significant improvement in all areas except for Narcan knowledge and safe storage. However, not all remaining improvements were still significant when compared to a self-selecting control group of participants who either did not play the game or reported playing the wrong game. Those who accessed the game had better improvement in misuse behavior, perceived knowledge, and self-efficacy. Thus, the findings from this study suggest the potential effectiveness of the SG, MedSMART: Adventures in PharmaCity on improving opioid medication safety awareness among adolescents.

The findings from this study revealed female participants had better improvement for safe disposal, but males had greater improvement for misuse behavior. The participants who were non-White or Hispanic had higher improvements in perceived knowledge than White participants, and older participants were associated with greater improvements in opioid knowledge. There are limited studies that examine why these differences exist amongst gender, race, and age [31,32]. Those that examine these differences typically only study adults and may not apply a pre-post survey study design [33]. Racial and ethnic disparities in pain management are well documented in literature and may contribute to these differences as well [34-38]. Studies show that not only are non-White patients less likely to receive opioid prescriptions for certain conditions, but also that disparities exist in access to specialized care, such as mental health, addiction, or pain specialists [34-38]. Data from this study and current literature suggest that research focused on these aspects is warranted in the adolescent population.

Studies have indicated that SGs possess the potential for increasing students’ learning, motivation, and engagement as well as the ability to develop their minds and improve their learning efficiency [39-41]. Additionally, the Pediatric Pharmacy Advocacy Group recommends pharmacists educate adolescents about the proper administration, storage, and disposal of opioids [42]. Current literature suggests that reducing opioid prescribing by health care professionals, specifically dentists and surgeons, could substantially lower prescription opioid exposure in adolescents [11]. There is a need for initiatives that target high-volume prescribers and provide medication safety education for the patients.

An opportunity exists to implement MedSMART: Adventures in PharmaCity in multiple settings: (1) a school setting where students can play the game in a health course or a similar setting; (2) a doctor’s office where the adolescent is being prescribed the opioid and could play the game in the waiting room; and (3) at the pharmacy where the adolescent is picking up their opioid prescription. The addition of an SG in any of these settings may help proactively curb opioid misuse in adolescents.

This single-arm study did not have a randomized researcher-blinded control group. Although the participants who did not play the game or who reported playing the wrong game were compared to those who did play the game (Table 4), this was a self-selecting control group and not randomized. Efficacy outcomes relied on self-completion questionnaires, which may have been affected by participant dishonesty and reporting bias.

Additionally, although the sample demographics were similar to those of the US demographics, the majority of the participants were White. Thus, generalizability of study findings may be somewhat limited, as there are racial and ethnic differences in access to prescription opioids and pain management.

The participants could not ask for assistance navigating the game or clarification on survey questions because they were not monitored by research staff during data collection. Moreover, the participants’ attentiveness during gameplay was not observed. The study design relied on trusting the parent to not influence their child’s survey responses or gameplay. To ensure accuracy, future designs should confirm that adolescents are taking the surveys and playing the game and that parents are not affecting the responses. A focus group or interview could also be incorporated to a study’s design to understand what players learned from the game or to identify potential game improvements.

Health messaging surrounding opioid safety requires novel and engaging strategies to be effective in increasing knowledge, changing behavior, and preventing prescription opioid misuse. The SG MedSMART: Adventures in PharmaCity was designed to engage adolescents in real-life scenarios and provide them with the information needed to correctly use opioids. The findings from this study show that the SG significantly improved all concept scores except for Narcan knowledge and safe storage. Future research should examine how the game can improve adolescents’ understanding for safe opioid storage and the use of Narcan. MedSMART: Adventures in PharmaCity could be a useful and effective tool for preventive opioid misuse intervention programs and should replicate this study using a randomized controlled trial with a prepost survey study design. Further research is needed to assess potential benefits of dissemination and implementation of the game in health care or school settings.