This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Serious Games, is properly cited. The complete bibliographic information, a link to the original publication on http://games.jmir.org, as well as this copyright and license information must be included.
Physical activity games developed for a mobile phone platform are becoming increasingly popular, yet little is known about their content or inclusion of health behavior theory (HBT).
The objective of our study was to quantify elements of HBT in physical activity games developed for mobile phones and to assess the relationship between theoretical constructs and various app features.
We conducted an analysis of exercise and physical activity game apps in the Apple App Store in the fall of 2014. A total of 52 apps were identified and rated for inclusion of health behavior theoretical constructs using an established theory-based rubric. Each app was coded for 100 theoretical items, containing 5 questions for 20 different constructs. Possible total theory scores ranged from 0 to 100. Descriptive statistics and Spearman correlations were used to describe the HBT score and association with selected app features, respectively.
The average HBT score in the sample was 14.98 out of 100. One outlier, SuperBetter, scored higher than the other apps with a score of 76. Goal setting, self-monitoring, and self-reward were the most-reported constructs found in the sample. There was no association between either app price and theory score (
There are few content analyses of serious games for health, but a comparison between these findings and previous content analyses of non-game health apps indicates that physical activity mobile phone games demonstrate higher levels of behavior theory. The most common theoretical constructs found in this sample are known to be efficacious elements in physical activity interventions. It is unclear, however, whether app designers consciously design physical activity mobile phone games with specific constructs in mind; it may be that games lend themselves well to inclusion of theory and any constructs found in significant levels are coincidental. Health games developed for mobile phones could be potentially used in health interventions, but collaboration between app designers and behavioral specialists is crucial. Additionally, further research is needed to better characterize mobile phone health games and the relative importance of educational elements versus gamification elements in long-term behavior change.
Serious games, or games whose primary purpose is to educate rather than entertain [
Serious games for public health have typically been developed as video games [
Much research suggests that health interventions designed around health behavior theory (HBT) are more effective in changing behavior than those which are not [
As health professionals are increasingly using mobile phone apps in interventions to increase physical activity, research on the content of such apps is important. Although many content analyses for health and fitness apps have been recently conducted [
Our study was a content analysis of HBT contained in physical activity game apps selected from among the apps available in the iTunes App Store’s Health and Fitness category. Two graduate students trained in HBT coded the apps.
The sample was collected from the Apple App Store in the fall of 2014. Apps designed for iPhone use were chosen, because many similar app content analyses have used Apple’s App store for sample selection [
Search terms.
Physical activity terms | Gamification terms |
Dance | Game |
Exercise | Avatar |
Fitness | Reality game |
Run | Virtual |
Fit | Challenge |
Team | Race |
Train | Quest |
Trainer | Adventure |
Goal | Interactive |
Walk | Simulator |
Track | Augmented reality |
Tracker | Running |
Trek | Workout |
Health | Cycling |
Aerobics | Cardio |
Weight training |
|
We have formal training in public health and health behavior and adapted the search terms from a previous content analysis of health theory in fitness apps [
The first 500 most popular apps were chosen for each search term, as the app store does not sort by page number. Additionally, as adapted from Lister et al [
The detailed written descriptions for the first 500 apps that appeared in the search results under each topic were analyzed to assess whether each app met the criteria for a serious health game. The definition for serious game was taken from definitions provided by Michael and Chen [
The initial search revealed 86 apps that were originally selected as meeting the criteria. After reviewing all of the apps, 52 were identified for final inclusion. Apps were excluded that required special equipment (e.g., bikes, treadmills, pedometers, heart rate meters, GPS) (8), could not be located in the App Store upon subsequent searches (5), failed to operate (8), or upon further investigation did not meet the criteria for a physical activity game (13).
Each app selected for final inclusion was coded into an initial sampling rubric using Qualtrics online survey software. The coders downloaded each app to an iPad or iPhone and played each game for a minimum of 30 minutes or until completing one level to increase familiarity with the user interface and available functions. The coders then used a theory-based instrument to conduct the content analysis for each app.
The instrument and methodology used for coding was adapted from an instrument used by Cowan et al [
Each app was coded using a rubric with 100 theoretical items, containing five questions each for 20 different constructs, as used by Cowan et al [
We also coded for gamification elements. The specific gamification elements coded for were selected from an instrument used by Lister et al [
Coded data were imported from Qualtrics and analyzed using SAS Studio. To verify the level of interrater reliability, both coders independently coded 10 common apps, approximately 12% (10/86) of the original sample and 19% (10/52) of those retained for final inclusion. A Cohen’s kappa coefficient, a method commonly used in content analysis research, was calculated to measure interrater agreement (
Characteristics of the sample are shown in
App characteristics.
App name | HBT score | App name | HBT score |
SuperBetter | 76 | Rare Candy—Epic Habit and Goa | 12 |
Yoga Retreat | 37 | Rare candy free | 12 |
Zombies, Run! 5k Training | 35 | Wokamon | 12 |
iBelly Workout | 29 | Silk Road Walk | 12 |
The Walk | 29 | Runno | 11 |
Workout in a Bag—for kids | 29 | Box the Bag | 9 |
Yes, Drill Sergeant! | 28 | Block Sports | 9 |
RunAlice | 28 | Walky | 8 |
Walk it! | 24 | AR Basketball | 8 |
Zombies Run! | 24 | Battlesuit Runner Fitness | 8 |
Ninja Fitness Free | 24 | Jump Boy | 8 |
Streetquest—run a game | 23 | iBowl | 8 |
Walk n' Play | 22 | MotionMaze Trick or Treat | 7 |
Burn Your Fat with Me! | 21 | Hike the World—GPS Tracker | 6 |
PushUp Club Free | 20 | MotionMaze | 6 |
NFL Play 60 | 19 | Paranoid | 5 |
Habit Monster | 17 | treasure island GPS, | 5 |
Turfly | 17 | AR Soccer | 4 |
Daily Spartan | 17 | Keep Moving | 4 |
Walkr—Galaxy Adventure in You | 16 | Pygmalions Challenge | 1 |
FitQuest Lite | 16 | GPS Fun Lite | 1 |
RunZombieRun | 16 | Gigaputt | 1 |
7 Min Workout Zombie Survival | 15 | GPS Invaders | 0 |
HuntedApp | 14 | MotionMaze Holiday Adventure | 0 |
TapCloud | 13 | Mapventures | 0 |
Superhero Workout | 13 | TrezrHunt Free | 0 |
Exercise type in apps.
Forms of exercise | n | % |
Walking | 29 | 56 |
Running | 22 | 42 |
None/general movement | 17 | 33 |
Weight lifting/bodyweight exercises | 13 | 25 |
Other | 9 | 17 |
Stretching | 8 | 15 |
Jumping | 5 | 10 |
HBT score by construct (N=52).
HBT |
|
n (%) | Median | Mean | SD |
|
|
52 (100) | 12.5 | 14.98 | 12.92 |
|
|
|
|
|
|
|
General information | 20 (38) | 0 | 1.48 | 1.99 |
|
Self-monitoring | 30 (58) | 4 | 2.25 | 1.99 |
|
Stress management | 3 (6) | 0 | 0.23 | 0.94 |
|
Time management | 5 (10) | 0 | 0.33 | 1.08 |
|
Learning | 4 (8) | 0 | 0.27 | 1.03 |
|
|
|
|
|
|
|
Incentives | 13 (25) | 0 | 0.65 | 1.37 |
|
Barriers | 7 (13) | 0 | 0.44 | 1.26 |
|
Risks | 10 (19) | 0 | 0.62 | 1.39 |
|
Goal-setting | 36 (69) | 4 | 2.25 | 1.87 |
|
Self-reward | 23 (44) | 0 | 1.77 | 2.01 |
|
Readiness | 9 (17) | 0 | 0.5 | 1.18 |
|
Self-talk | 4 (8) | 0 | 0.31 | 1.08 |
|
Self-efficacy | 9 (17) | 0 | 0.63 | 1.46 |
|
Norms | 3 (6) | 0 | 0.17 | 0.79 |
|
|
|
|
|
|
|
Peer pressure | 18 (35) | 0 | 1.33 | 1.89 |
|
Modeling | 16 (31) | 0 | 0.83 | 1.50 |
|
Relapse prevention | 2 (4) | 0 | 0.10 | 0.57 |
|
Follow-up | 6 (12) | 0 | 0.38 | 1.09 |
|
Guilt | 4 (8) | 0 | 0.19 | 0.79 |
|
Stimulus control | 4 (8) | 0 | 0.25 | 0.95 |
There was no association between price and HBT (Spearman correlation coefficient
Price and HBT score.
Physical activity games by price.
Price (US$) | Apps |
0 | 34 |
0.99 | 6 |
1.99 | 5 |
2.99 | 4 |
3.99 | 1 |
4.99 | 2 |
The number of elements of gamification was not associated with HBT score (Spearman correlation coefficient
All of the elements of gamification were present in the sample, except for real-world prizes. The most common gamification elements in the sample were fantasy environment (96%, 50/52), whereas storyline was present in half of the sample (50%, 26/52). Rankings or standings (19%, 10/52) and leaderboards (29%, 15/50) were the least commonly utilized feature of gamification in the sample (
Gamification elements (N=52).
Gamification elements | n (%) |
Storyline | 26 (50) |
Fantasy environment | 50 (96) |
Competition | 31 (60) |
Possibility of failure | 43 (83) |
Leaderboards | 15 (29) |
Clear expectations | 49 (94) |
Score | 38 (73) |
Ranking or standing | 10 (19) |
Levels | 24 (46) |
Real world prizes | 0 (0) |
Gamification and HBT score.
The purpose of our study was to determine the presence of HBT in physical activity games developed for a mobile phone platform. This study also analyzed (1) the prevalence of specific health behavior constructs, (2) the association between price and presence of HBT, and (3) the association between elements of gamification and HBT in these same apps.
The presence of HBT in this sample varied and many of the apps contained low levels of HBT, but HBT levels were higher on average than HBT levels in analyses of non-game health apps. The average HBT score of this sample was 14.98 out of 100, with SuperBetter, an outlier, yielding the highest HBT score at 76. Excluding SuperBetter, the next highest HBT score was 37, and the average HBT score became 13.78. In other content analyses of non-game mobile phone apps that utilized the same coding rubric as this study, the average HBT score was lower; in a sample of physical activity apps, Cowan et al [
SuperBetter, an app used for achieving nonspecific health goals, was an outlier with a much higher HBT score than the sample average. SuperBetter was unique in its heavy inclusion of educational elements, including individually tailored assistance; it required feedback on not only whether users completed each exercise, but also how well users completed each exercise, as well as tips for improvement [
SuperBetter stands in stark contrast to physical activity app games focused more on entertainment with few educational elements, such as GPS Invaders (
Despite the differences in content, both SuperBetter and GPS Invaders are considered serious games as determined by the inclusion criteria of this study. It should also be noted that the coding rubric and inclusion criteria utilized in this study emphasized the importance of educational content, in conjunction with the definition of serious games as primarily intended to educate, rather than entertain [
The most prevalent health behavior constructs (after gamification elements) included goal setting, self-monitoring, and self-reward. In a review of mobile apps utilized in health interventions, Payne et al [
Paid apps were no more likely to include elements of HBT than free apps. However, the sample of paid apps was small and these results should be interpreted with this limitation in mind. West et al [
There was also no significant association between elements of gamification and presence of other HBT constructs—that is, having more game elements did not increase the overall HBT score. Researchers disagree on definitions of serious games, and the number and type of gamification elements needed to merit the classification of a serious game vary [
The findings of our study are significant for practical use in public health, especially as mobile apps are being increasingly utilized in health interventions [
SuperBetter.
GPS Invaders.
The findings of our study should be interpreted in the context of some limitations. First, the coders only used the mobile phone for either 1 level or 30 minutes to code for HBT elements. It is possible that some HBT components were missed by limiting use to this time frame, though unlikely; a recent study demonstrates that while mobile phone and app use is increasing, average app session length has stayed constant at about 5.7 minutes [
The coders analyzed only app descriptions to determine if each app qualified as a health game, and it is possible that some games were missed due to inadequate descriptions. The coders attempted to compensate for this limitation by selecting a sample (10 apps) that appeared in the search but did not appear to meet the description of a game via the description; the coders found that none of these games fit the definition of a serious game, so the likelihood that apps were overlooked due to weaknesses in the description appears low. Finally, it should be noted that the definition of a serious game is not consistent across research; a number of legitimate serious exercise games exist that do not fulfill the criteria we proposed. Legitimate exergames (according to other researchers) may have been excluded from our sample. In this particular study, we were more interested in games emphasizing education, though content analyses of serious games with different criteria would be interesting for future research.
Physical activity health games developed for mobile phones are a potentially viable option for health interventions, though further research and development of such games should continue. Further research should be conducted to determine whether these health games are efficacious in health interventions, and the extent to which educational and gamification elements impact efficacy should be further assessed as well. Collaboration between app designers and behavioral specialists is also crucial to help promote lasting behavior change. Investigations into whether serious app games for health are more conducive to inclusion of HBT and whether they universally contain more elements of HBT is valuable in order to assess whether such games can improve individual and community health in the long-term.
None declared.