Principal Findings
This study describes a novel BUS approach to test feasibility, data availability, and most common apps among a young adolescent population. The inclusion of the BUS approach within a Web-based survey adds the capacity to assess mobile smartphone screen time that is accessible for both participants and researchers in a highly technological environment. Battery use is automatically monitored by the operating systems of most smartphones [21], reducing both the bias of self-report as well as burden on participants to record real-time smartphone and app use. Compared with typical EMA approaches, this methodology can improve the capacity to deliver real-time data as well as allow for data collection from larger, Web-based samples [22].
The feasibility of our approach was mixed with both strengths and challenges. Among the participants who did upload a screenshot, the images provided clear data that could be viewed, stored in a deidentified manner, and categorized for further analysis. Our challenges in feasibility may be explained by the rate of phone ownership in our young adolescent sample, which is lower than recent reports of 56% smartphone ownership among youth aged 8-18 years [23]. Among adolescents with their own phone, approximately half were willing or able to upload a screenshot. It is possible that a study population of older adolescents and adults may have higher phone ownership and better understanding and capacity to upload a screenshot. An additional explanation for this challenge may be the survey platform Qualtrics, which allows users to take surveys on both mobile phones and desktop computers. Participants who used a desktop computer to complete the survey may have experienced an additional burden in uploading a screenshot from a different device, which may have contributed to the overall feasibility of this study. Compensation of US $13 was determined by Qualtrics based on the initial suggested sample size for the larger study. Based on our results, low monetary compensation may be a factor in willingness to complete the screenshot task and should be considered when incorporating the BUS approach into future studies.
Over half of the screenshots analyzed contained complete data including the display of “onscreen” versus “background” time per app. One possible reason for cases of incomplete data may be that participants had an older iOS operating system, as the battery usage feature is only available on iOS 9 and newer operating systems [24]. This should be taken into consideration in future studies utilizing this methodology in iPhones. In addition, this study relied on a single screenshot, which may not have captured total app usage. To strengthen overall data availability, future studies might require as many screenshots as necessary to provide the full range of apps used by a participant.
While our study illustrates both strengths and challenges to the proposed research protocol, it serves as a valuable starting point for considering how to advance data collection methods to understand mobile smartphone screen time and media use in iPhones. Studies have concluded that smartphone apps can be beneficial in monitoring and evaluating patients [25-27] as well as increasing adherence to medical interventions [28,29]. The BUS approach offers the ability to take these studies further in understanding real-time use and overall time spent on apps as a factor that could contribute to health outcomes. This methodology offers a framework that could be adapted to Android phones, which offer varying displays of battery use per app and time spent per app. When data and display of battery use are available, a BUS approach may still be utilized as a data collection process. Instructions for the screenshot would need to be adjusted to reflect the varying battery use displays of Androids, and further exploration of these displays should be used to inform these instructions. By expanding the BUS approach to Android devices, future studies would also allow for greater sampling of socioeconomic status, with lower socioeconomic status individuals and families more likely to own Android devices.
The BUS approach may also be combined with a variety of research methodologies including Web-based cross-sectional surveys, as was done in this study, or used as a tool for monitoring smartphone app use longitudinally. In a previous pilot study conducted using BUS, older adolescent iPhone users were asked to submit weekly screenshots of battery use for 9 weeks, with a 94% retention rate over 5 weeks and 60% retention rate over 9 weeks [30]. A further advantage of the BUS approach is the ability to collect comprehensive data related to mobile smartphone screen time without the need for an additional app or programming. Thus, this methodology is accessible for researchers without the means for software or app development.
Limitations
There are limitations to this feasibility study. The sample size of screenshots analyzed is limited and is not representative of the larger adolescent population. The sample overall may not be representative of the US racial or ethnic makeup for youth aged under 18 years because current US census records differ from our study’s demographics and estimate this population to be 51% white, 15% black or African American, and 25% Latino people [31]. In addition, the battery use page only displays phone use while an iPhone is not charging and may not account for the full time spent using a smartphone device. Furthermore, adolescents may have access to other phones, including their parents’, which may underestimate overall screen time. Finally, in this feasibility study, only iPhone screenshots were evaluated. Results may not generalize to all smartphones or mobile devices with battery use tracking. We did not test the BUS tracking method against other methods of tracking app use to assess its accuracy.