designing for engaging experiences in mobile social- health...

8
32 PERVASIVE computing Published by the IEEE CS n 1536-1268/13/$31.00 © 2013 IEEE UNDERSTANDING AND CHANGING BEHAVIOR Designing for Engaging Experiences in Mobile Social-Health Support Systems O ver the last few years, there has been significant academic and commercial interest in using mobile platforms to fos- ter improved health. Although most of the published research in this area has focused on a system’s efficacy in promoting health (see the sidebar), the kinds of experiences afforded by these systems are also important, 1 particularly with regard to system adoption and continued use. This question of how to design for sustained engagement in the context of supporting or encouraging health and well-being has been relatively underexplored. To address this, we per- formed two studies using Virtual Environ- ments for Raised Awareness (VERA), a mobile social-health application we developed and then deployed for two years. VERA lets users share health-related activities and decisions with one another in an open-ended fashion—that is, it doesn’t focus specifically on one area, such as physical exercise or nutrition, and it doesn’t prescribe specific actions for users to take with respect to their health decisions. Rather, it lets the users determine what’s healthy, providing a contrasting approach to existing systems that sense and analyze only a specific set of human behaviors. 2 Our analysis of usage data from the two studies, along with user interviews, revealed lessons learned that serve as informal design suggestions for facilitating user engagement in the context of mobile-health applications. These lessons address common technology con- cerns—such as the application’s visible interface and the infrastructure behind it. However, they also suggest new strategies for deploying mo- bile social-health applications within real-world cohorts and for recognizing relevant external social contexts that govern social interactions within the mobile platform. Virtual Environments for Raised Awareness We followed a sociotechnical systems approach in designing VERA—that is, we assumed the system design included not only the applica- tion’s interface and architecture but also the various social contexts in which the application is embedded. Thus, the two studies presented here, while involving few changes to the tech- nology itself, explore different social configura- tions among the system and user groups. This How do you design for sustained engagement in the context of supporting or encouraging health and well-being? Two studies of a mobile social-health application reveal how different aspects of the sociotechnical design affect user engagement with the system. Eric P.S. Baumer, Vera Khovanskaya, Phil Adams, John P. Pollak, Stephen Voida, and Geri Gay Cornell University

Upload: others

Post on 04-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

32 PERVASIVE computing Published by the IEEE CS n 1536-1268/13/$31.00 © 2013 IEEE

U n d e r s ta n d i n g a n d C h a n g i n g B e h av i o r

designing for engaging experiences in Mobile social-health support systems

Over the last few years, there has been significant academic and commercial interest in using mobile platforms to fos-ter improved health. Although

most of the published research in this area has focused on a system’s efficacy in promoting health (see the sidebar), the kinds of experiences afforded by these systems are also important,1 particularly with regard to system adoption and

continued use. This question of how to design for sustained engagement in the context of supporting or encouraging health and well-being has been relatively underexplored.

To address this, we per-formed two studies using Virtual Environ-ments for Raised Awareness (VERA), a mobile social-health application we developed and then deployed for two years. VERA lets users share health-related activities and decisions with one another in an open-ended fashion—that is, it doesn’t focus specifically on one area, such as physical exercise or nutrition, and it doesn’t prescribe specific actions for users to take with respect to their health decisions. Rather, it lets the users determine what’s healthy, providing a

contrasting approach to existing systems that sense and analyze only a specific set of human behaviors.2

Our analysis of usage data from the two studies, along with user interviews, revealed lessons learned that serve as informal design suggestions for facilitating user engagement in the context of mobile-health applications. These lessons address common technology con-cerns—such as the application’s visible interface and the infrastructure behind it. However, they also suggest new strategies for deploying mo-bile social-health applications within real-world cohorts and for recognizing relevant external social contexts that govern social interactions within the mobile platform.

virtual environments for raised awarenessWe followed a sociotechnical systems approach in designing VERA—that is, we assumed the system design included not only the applica-tion’s interface and architecture but also the various social contexts in which the application is embedded. Thus, the two studies presented here, while involving few changes to the tech-nology itself, explore different social configura-tions among the system and user groups. This

How do you design for sustained engagement in the context of supporting or encouraging health and well-being? Two studies of a mobile social-health application reveal how different aspects of the sociotechnical design affect user engagement with the system.

Eric P.S. Baumer, Vera Khovanskaya, Phil Adams, John P. Pollak, Stephen Voida, and Geri GayCornell University

Page 2: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

july–sEptEmbEr 2013 PERVASIVE computing 33

I n the relatively short history of mobile health, much of the

work has aimed to be persuasive—encouraging users to im-

prove health or health-related behavior. such work often draws

on the work of b.j. Fogg, who contributed a set of guidelines,

derived from cognitive, experimental, and social psychology, for

creating persuasive technologies for changing user behavior.1

the persuasive technology paradigm has certainly been useful

for work in mobile health,2,3 and others have argued that mobile

devices are particularly useful in this context, because they

allow intervention at the moment of decision making.4 However,

although many of the guidelines put forth by Fogg can lead to

sustained user engagement (for example, the use of intrinsic and

extrinsic motivators, such as challenge or reward5), the intent of

persuasion is to drive behavior change, and the two don’t neces-

sarily go hand-in-hand.

A prime example of a persuasive mobile health system de-

signed with an eye toward sustaining user engagement is the

ubiFit Garden.6 the system epitomizes persuasive technology.

users’ activities are passively recorded using simple sensors

(wearable accelerometers and geolocation via cell towers), and

the application provides fun, real-time feedback and rewards.

What makes ubiFit Garden special from the perspective of user

engagement is the manner in which the feedback is delivered.

the application responds to physical activity on the part of the

user by growing a virtual garden on the phone’s home screen,

with different life-forms in the garden growing in response to

different types of activity. Given the frequency with which indi-

viduals access mobile phones, this “glanceable” display not only

provides a subtle, constant reminder of the user’s progress but

also encourages—or even forces—the user to engage with the

system many times each day.

Other successful mobile health systems have found different

ways to engage users. For some, the value and quality of the

content provided is enough to sustain engagement. text4baby

has been a tremendously successful mobile health campaign,

yet the “system” consists of nothing more than daily sms mes-

sages about the user’s pregnancy and forthcoming baby.7 many

diet-tracking systems, particularly those intended for individu-

als with acute dietary needs (for example, diabetics), fall into this

category as well.8

still other systems rely on sociality to involve and engage us-

ers. EatWell, for example, is a voicemail-based system in which

users call in to record and to listen to open-ended messages

about challenges and solutions to healthy eating in low-income,

urban communities.9 users’ attachment to their community,

their shared ground with other users, and their desire to over-

come significant challenges all play roles in making EatWell

engaging. Although not examining mobile health applications

per se, numerous studies have focused on online social support

systems to examine the features of these systems that lead to

growth and user engagement.10

Our research differs from most of this previous work in its

emphasis on user-defined health awareness. We suggest that a

minimally prescriptive system might be more engaging, not only

because it’s broadly applicable to a range of health behaviors (as

opposed to just food or exercise), but also because it motivates

users to continually self-define and explore what it means to be

healthy. Furthermore, the complex and highly situated nature

of health makes it particularly amenable not only to this open-

ended system design but also to the sociotechnical systems

design approach we employ. thus, this article explores how

not only the technology design but also how technology use is

embedded in and shaped by social practices might influence

engagement with these health systems.

REfEREnCES

1. b.j. Fogg and D. Eckles, eds., Mobile Persuasion: 20 Perspectives on the Future of Behavior Change, stanford Captology media, 2007.

2. j.j. lin, “Fish’n’steps: Encouraging physical Activity with an Interac-tive Computer Game,” Proc. 8th Int’l Conf. Ubiquitous Computing (ubiComp 06), springer, 2006, pp. 261–278.

3. G. Gay et al., “pilot study of Aurora, A social, mobile-phone-based Emotion sharing and recording system,” J. Diabetes Science and Technology, vol. 5, no. 2, 2011, pp. 325–332.

4. s.s. Intille, “A New research Challenge: persuasive technology to motivate Healthy Aging,” IEEE Trans. Information Technology in Biomedicine, vol. 8, no. 3, 2004, pp. 235–257.

5. t.W. malone and m.r. lepper, “making learning Fun: A taxonomy of Intrinsic motivations for learning,” Aptitude Learning and Instruc-tion, Volume 3: Conative and Affective Process Analyses, lawrence Erlbaum Associates, 1987, pp. 223–253.

6. s. Consolvo et al., “Activity sensing in the Wild: A Field trial of ubiFit Garden,” Proc. SIGCHI Conf. Human Factors in Computing Systems (CHI 08), ACm, 2008, pp. 1797–1806.

7. E.t. jordan et al., “text4baby: using text messaging to Improve maternal and Newborn Health,” Nursing for Women’s Health, vol. 15, no. 3, 2011, pp. 206–212.

8. F. Andry et al., “Highly-Interactive and user-Friendly Application for people with Diabetes,” Proc. 10th IEEE Conf. e-Health Networking, Applications and Service (Healthcom 08), IEEE, 2008, pp. 118–120.

9. A. Grimes et al., “EatWell: sharing Nutrition-related memories in a low-Income Community,” Proc. ACM Conf. Computer Supported Cooperative Work (CsCW 08), ACm, 2008, pp. 87–96.

10. j. Galegher, l. sproull, and s. Kiesler, “legitimacy, Authority, and Community in Electronic support Groups,” Written Comm., vol. 15, no. 4, 1998, pp. 493–530.

related Work in Mobile health applications

Page 3: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

34 PERVASIVE computing www.computer.org/pervasive

Understanding and Changing Behavior

approach let us examine how different interpretations of the general open-ended design principle affected users’ engagement and experiences with the system.

VERA began as a tool for sharing images. It emerged from research in socially supportive emotion-sharing applications3 and resulted from our observation that users wanted to take pictures to help track their emotions. We initially piloted a prototype of the application as a food-sharing plat-form. Users photographed their food and tracked their emotions, which corresponded with their eating behav-ior. As we analyzed early data from the content that our beta testers were posting, we noticed that people tended to share pictures of general health be-haviors and decided to reframe VERA as a more open-ended mobile health- sharing platform.

In VERA, users’ photographs, com-ments, and self-assessments repre-sent an open-ended model of health

behavior, and the system serves as a platform for recording, sharing, and reflecting on these facets of day-to-day life, with the specific role of the system evolving and expanding through each of our studies.

study 1: social vs. individual UseWe initially developed VERA as an An-droid application, deploying it as part of a controlled experimental study. During the study, we instructed par-ticipants to document their health be-haviors, and we provided examples of such behaviors (such as “eating” or “exercising”).

To post an activity, participants opened the application (either on a per-sonal or a lab-issued mobile device) and were immediately prompted to take a picture using the built-in camera. This picture was intended to represent the health activity they were about to per-form (or had just performed), such as eating, exercising, using a technological device, or engaging in a social activity.

Users could also take a photograph to represent a health activity that they didn’t perform, such as choosing not to eat a piece of cake or deciding not to go to the gym. It was left entirely to the us-ers to decide what constituted a health activity and to report it as such.

Next, users indicated whether they performed the health activity photo-graphed and were asked to briefly de-scribe the image or activity. Then, they assessed the healthiness of the activity (see Figure 1a). Users were instructed to assess their choice and not the item photographed—for example, if the im-age was of a food that they chose not to eat, they assessed the healthiness of not eating the photographed food.

After submitting the image of the health activity, users were prompted to reflect on their emotional state and choose, from an established array of thumbnails, an image that closely re-flected their emotional state at the time of performing (or not performing) the selected activity.4 After completing

Figure 1. The interface for the first iteration of Virtual Environments for Raised Awareness (VERA): (a) submitting post details, (b) the group view, and (c) the detailed view.

(a) (b) (c)

Page 4: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

july–sEptEmbEr 2013 PERVASIVE computing 35

their post, users were shown images of other health activities that had recently been posted (see Figure 1b).

We randomly and evenly split the 34 study participants into two groups—individual or social. Individual-condi-tion users could only access their own previous posts (Figure 1b) and, in es-sence, this group experienced VERA as a personal mobile health diary.

The participants in the social con-dition could see previous posts of all participants in this group. Because we recruited study participants through fli-ers and general email solicitation, the participants in the social condition, for the most part, didn’t know each other in advance. Users were encouraged to use the application daily for two weeks. At the end of the two-week trial, in ad-dition to completing a quantitative survey, we interviewed participants to learn about their views on healthiness and VERA’s role in their daily health decisions.

In the social condition, VERA pro-vided a variety of tools, letting us-ers browse and interact with the data shared by others in the study. Partici-pants could click a thumbnail to view a larger image, learn whether the other user performed the activity, review the photo’s description, and post comments (Figure 1c). They could also filter posts to see only the content that a specific user posted. The design of the posting interface and experience— using an arbitrary photograph, requesting mini-mally constrained details, and allowing for submission of a free-text caption—were intended to afford open-ended, user-driven definitions of health and healthiness.2

We were also interested in learning whether placing participants into social conditions would lead to differences in engagement with the system. To encour-age honest disclosure, users’ emotional and health self-assessments weren’t vis-ible to others. Participants were asked to post a certain number of health ac-tivities each day and encouraged to post comments on other peoples’ posts.

We compared the two conditions—individual versus social—with respect to changes in engagement from the first to second week of the study. Specifi-cally, we calculated the average number of posts made in the system during each day of the study.

In the social condition, although the number of daily posts decreased slightly from the first to the second week (from 44.4 to 36.3 posts), the change wasn’t statistically significant in a t-test (t(12) = 1.52, significance-level p = .15). In the individual condition, the number of posts per day dropped significantly from the first to second week (from 49.3 to 33.3, t(12) = 2.79, p = .016)— almost twice as much as in the social condi-tion (see Figure 2). These differences suggest that the social condition might have proven more engaging, a conclu-sion for which we found some support in our open-ended interviews.

During the interviews, participants in the social condition reflected on how other peoples’ posts affected their expe-rience with the application: they were curious about the other participants and noticed patterns in their posting behaviors and the locations of their posted images. This increased interest in others’ posts likely helped draw us-ers back to the system and facilitated sustained engagement.

However, this social inquisitiveness was somewhat stymied by the VERA interface. Participants had to click on the image thumbnail to see who posted it, which hindered the association be-tween the content and users. Also, users reported difficulty in filtering the posts to find content posted by a particular user. Although this functionality was supported, it wasn’t intended as a pri-mary interaction of the application.

Our initial deployment also prompted us to reconsider our recruiting process. In interviews, participants expressed concern about being asked to share photos and comment on strangers’ con-tent in the application and compared the relative lack of privacy settings on VERA with other social platforms.

These recruitment processes, and more generally, the social contexts in which the system was deployed, became a pri-mary focus in the subsequent system iterations.

study 2: shared goalsSeveral months later, we ran a second study using the VERA application. Given participants’ interest in running VERA on their own devices (in mid-2010, participants with smartphones most often used iOS), we developed an iPhone version of the application. Func-tionally, this version was identical to the Android app, albeit with a few aes-thetic differences resulting from varia-tion in the affordances provided by the operating systems’ interface toolkits.

Based on the interviews from our ini-tial study, we hypothesized that social engagement among users in this type of health application might be most productive in groups sharing similar health goals. That is, our decisions to support both sociality and open-ended

Figure 2. In the individual condition, the number of statuses submitted per day decreased significantly from the first to second week of the study, but it didn’t decrease significantly in the social condition.

50

40

30

Aver

age

stat

uses

per

day

20

10

0

Individual Social

Days 1–7Days 8–14

Page 5: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

36 PERVASIVE computing www.computer.org/pervasive

Understanding and Changing Behavior

definitions of “health” made sense, but users with varying or disparate moti-vations for using the application might be too motivationally disconnected for effective social engagement.

In this second study, we changed our participant recruitment strategy. Work-ing with a university Health and Well-ness Program offering classes and social support (email lists, walking groups, and so on) for staff, faculty, and their families, we recruited from two of their mailing lists: a general health and well-being list for all members of the pro-gram, and a mailing list for individuals trying to lose weight. Based on volun-teer responses, we assembled two exper-imental groups: a control group of 16 participants from the general list with no particular shared health goal, and a group of eight participants from an-other list, who all shared a goal of losing weight. Some dyads or triads of users previously knew each other, but most groups were comprised of strangers.

Participants used VERA for one month and were interviewed approxi-mately three weeks into the study. As described earlier, we didn’t implement changes to the system itself but rather sought to examine the effects of the so-ciotechnical design—the interaction of the application’s design and the social context in which it was adopted—on user engagement.

As with the first study, to assess sustained engagement, we examined changes in the number of posts between the first and last weeks of the study. In the group of general wellness pro-gram members (the “general” group), the number of posts per day dropped significantly from the first week to the last week (from 34.1 to 19.6, t(11.7) = 2.36, p = .036). However, in the weight-loss group, the change from the first to last week wasn’t statistically significant (from 10.9 to 7.6 posts, t(8.4) = 1.35, p = .21). In the general group, the decrease was several times larger (see Figure 3a).

Unlike in the first study, both of these groups were in social conditions, so we can also examine changes in the number of messages sent using VERA. Here, the story is somewhat different. The general group showed a decrease in the number of messages sent per day between the first and last week (from 7.6 to 1.8 messages), but the change was not statistically significant (t(4.4) = 2.29, p = .077). However, as Figure 3b shows, the weight-loss group demon-strated a significant decrease in the number of messages per day (from 17.9 to 6.2, t(9.5) = 3.26, p = .009). Thus, sharing a common health goal helped reduce the drop-off in health behavior posting but not the decrease in com-ments on others’ posts.

Again, we turn to the qualitative inter-view data to help interpret these findings and ensure that the observed differences weren’t solely due to the difference in the sizes of our two participant groups. Par-ticipants reported being aware of others’ posts inspiring them to be healthier—for example, seeing others going to the gym in the morning. Sometimes, this went further; participants spoke of feeling like they were “in competition” with ambiguous others (for example, seeing five sets of running shoes in one day) and accountable to these others, even within a space where they were identifiable only by their username (at most, first name and last initial, though often these were also obfuscated).

Several members of the weight-loss group indicated the value of belonging to a group of individuals with similar (health) motivations. Sometimes, this value related simply to a desire to be in a group of people “more like me,” with similar struggles—such as try-ing to maintain a lower weight after weight loss.

An activity-focused community often has its own set of language use and interaction mores. For example, participants reported appreciating that others in their group understood weight-loss jokes and jargon (for exam-ple, “reefy” is a colloquial abbreviation

Figure 3. Changes in the number of posts between the first and last weeks of the study. (a) The number of statuses posted each day decreased significantly from the first to the last week of the study, but only for participants without a shared goal. (b) However, those with a shared goal sent significantly fewer messages to one another from the first to the last week.

40

30

Aver

age

stat

uses

per

day

20

10

0

Shared goal No shared goal

20

15

Aver

age

mes

sage

s pe

r day

10

5

0

Shared goal No shared goal

Days 1–7Days 24–30

Days 1–7Days 24–30

(a) (b)

Page 6: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

july–sEptEmbEr 2013 PERVASIVE computing 37

of “re- feeding,” as a reference to over-eating) and talked about the carefully constructed positive and negative feed-back styles used in weight-loss groups to provide social support.

The extra attention needed to ensure that the comments were supportive and appropriate might have been one fac-tor contributing to the drop in messages among members of the weight-loss group; weight loss, as with many other health-focused areas, is one in which social support is especially valuable, but the potential for offense is high.

More generally, participants in the weight-loss group showed a particu-lar interest in getting to know others in their group. However, rather than seeking a medium for conversation, they cited a desire for profile pages to communicate a short slogan or other information that could be used to as-sess group-mates’ personalities. This request indicates interest in similar oth-ers and in self-presentation (weight loss members wanted to broadcast meta-information, such as a personal weight loss goal), but a profile page was also referenced when discussing a desire for control over membership in the group, or at least control over which users could see the posts shared in the system.

Some of the general group partici-pants reported converse desires. Spe-cifically, according to one participant, knowing “what other people are do-ing is irrelevant to me … [but it might be] different if I were in a weight loss program.” Several participants in this group said that VERA would be more useful (as a training tool, for example) if the user or group shared a more spe-cific health goal.

On the other hand, curiosity about others’ activities might explain the lack of a significant decrease in the average number of messages within the general group. We’re currently involved in qual-itatively coding the content of VERA messages, and early findings suggest that many messages in the general group were information requests and responses, commonly with the purpose

of clarifying, interrogating, and shar-ing more about the diverse health de-cisions and behaviors presented. Our interview data anecdotally supported this interpretation, with general well-ness participants reporting that they used messages to try and make sense of who did particular actions or cared about specific posts (for example, “oh, he’s the runner”).

Our second VERA deployment was thus somewhat more successful than the first. Specifically, deploying the sys-tem to groups of users who had similar motivations and health goals helped improve sustained engagement. How-ever, this approach didn’t necessarily help group members bond with one another. As a result, we argue that the social context into which a system is introduced should be viewed as part of the system design from a sociotechnical perspective.

Lessons LearnedAlthough many commercial applica-tions excel at encouraging a narrow range of behaviors (such as dieting or exercising), few offer a holistic, open-ended approach to documentation that lets users classify and define health activities on their own terms, using their own creative resources. Our stud-ies show how different sociotechnical

approaches to this overarching design principle affected engagement with the system. Here, we step back to consider lessons learned that could be of value in informing other work in the space of pervasive healthcare applications.

Make it socialA sense of social presence is impor-tant when engaging users in a mobile health application. In the first study,

participants in the social condition not only continued using the system but also reported feeling more motivated to post their own activities when there were other people posting. Making the VERA platform social decreased attri-tion and provided social support for users’ health behavior.

allow degrees of social sharingAffording a variety of social interactions helped meet different user needs. For prolonged browsing on a mobile device, user behavior is similar to Web brows-ing—aside from the different scale and ergonomics of phones—but when de-signing for short-term browsing, the designer should ensure that informa-tion isn’t obfuscated in the interface. In the first study, VERA users complained about having to click on the images to learn who uploaded them. If interaction is too cumbersome, it’s not reasonable to expect mobile users to participate with the same degree of engagement.

The application should also allow a variety of social interactions, such as commenting on a post or supporting (“liking”) it without having to type a comment. This is as important for dif-ferent browsing contexts (participating on the go) as it is for promoting social support. As noted in the second study, the design of the interactions must be

carefully curated so that there are easy ways to afford positive feedback about health decisions without unintended judgments on sensitive topics.

Another request that emerged from interviews about social sharing of health information was to allow varied degrees of social sharing, or privacy set-tings within the application, to provide more nuanced negotiations of what in-formation users disclose to whom.

Deploying the system to groups of users who

had similar motivations and health goals helped

improve sustained engagement.

Page 7: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

38 PERVASIVE computing www.computer.org/pervasive

Understanding and Changing Behavior

Leverage existing social structuresThe different contexts in which we de-ployed VERA had implications for how users adopt and appropriate it. Social users reported being more engaged when they felt connected, either by so-cial ties, common goals, demographics,

or location, with the other people posting content on the platform. Par-ticipants reported trusting other us-ers on the application if they felt like they had things in common with them. Participants who felt like they were working toward a common health goal or had similar lifestyles as the other

participants reported being more mo-tivated to participate in the application (and the surrounding community).

This suggests that the representation of social groups in pervasive health applications shouldn’t be random; in-stead, it’s helpful to leverage existing social structures, such as the weight-loss group in the second study.

afford individual- to-group relationsIn the second study, we observed the im-portance of letting participants explore other group members’ interests, goals, and contributions. In a social environ-ment, a mobile sharing platform for health should let the user relate to the general group and build relationships with specific members. Design choices have implications for users’ ability to build individual relationships on a so-cial media platform. These relationships are built both on visible or productive social interactions (such as commenting on another user’s posts) and on silent social interactions (such as simply view-ing or reading another user’s posts). The VERA interface didn’t explicitly sup-port these types of person-to-person or person-to-group interactions, and it was one of the most requested capabili-ties for future versions.

T he specific findings high-light important relation-ships between participants’ engagement with the VERA

system and the social contexts in which the system was used. The quantitative and qualitative data that we collected helped underscore the importance of these sociotechnical facets of the VERA system’s design and the context in which it was deployed. Although our focus has been on a system for promot-ing and supporting health, the design recommendations presented here might be applicable in other contexts as well.

Previous work on mobile applications for supporting health has documented problems with user drop-off over time.1

the AuTHoRsEric P.S. Baumer is a postdoctoral associate in communication and informa-tion science at Cornell university. His research involves designing, implement-ing, and evaluating information technologies, especially interactive visualiza-tions based on computational text analysis, to promote critical thinking and reflection. baumer received his phD in information and computer sciences from the university of California, Irvine. Contact him at [email protected].

Vera Khovanskaya is a researcher in in the Interaction Design lab at Cornell university. Her work combines critical reflection on the collection and process-ing of data with the design of personal informatics technologies. Khovanskaya received her bs in information science from Cornell university. Contact her at [email protected].

Phil Adams is a graduate student working with Geri Gay’s Interaction Design lab in Information science at Cornell university. His research interests are in social computing and mobile health; specifically, he is working on ways to measure and manage stress. Adams received his bs in information science from Cornell university. Contact him at [email protected].

John P. Pollak is a postdoctoral associate in information science at Cornell uni-versity and at Weill Cornell medical College. His research interests include how mobile phones can be used to both encourage healthy behaviors and collect large quantities of contextually relevant health data. specifically, he’s develop-ing mobile phone apps to track emotion, stress, and daily health behaviors and examining those data in the context of health outcomes, social network prop-erties, and job performance. pollak received his phD in information science from Cornell university. Contact him at [email protected].

Stephen Voida is a postdoctoral associate in the Cornell university Depart-ment of Information science, where he studies personal information manage-ment, pervasive healthcare, and ubiquitous computing as a member of the Interaction Design and people-Aware Computing laboratories. He received his phD in computer science from the Georgia Institute of technology. Contact him at [email protected].

Geri Gay is the Kenneth j. bissett professor of Communication and holds a joint appointment in Information science at Cornell university. she also directs the Interaction Design lab. Gay received her phD in education from Cornell uni-versity. Her research is funded by the us National science Foundation, National Institutes of Health, Intel, Google, and the us Department of Agriculture. Con-tact her at [email protected].

Page 8: designing for engaging experiences in Mobile social- health ...homepage.ufp.pt/rmoreira/CM/PapersPerCom201415/Designing...july–sEptEmbEr 2013 PERVASIVE computing 33I n the relatively

The recommendations reflectively dis-tilled from our experiences—emphasiz-ing social experiences in an application that allows degrees of social sharing, leverages existing social structures, and affords individual to group rela-tions—might be useful in a variety of health systems, and even perhaps more broadly. Thus, this work represents an effort to design systems that not only help promote health but also provide for sustained, engaging user experiences.

In the future, we hope to gain a better understanding how different types and sources of feedback, monitoring, emo-tional states, and social networks can affect behavior. The sociotechnical de-sign insights provided here can be gen-eralized and used effectively in a variety of other health-related applications, in-cluding tobacco cessation, treatment of alcoholism, chronic pain management, and chronic disease management. With a better understanding of the relation-ships between system design and the mechanisms of persuasion and social influence, we can further develop com-prehensive and empirically validated design guidelines to effectively support maintaining a healthy lifestyle.

REFEREnCEs 1. P. Klasnja, S. Consolvo, and W. Pratt,

“How to Evaluate Technologies for Health Behavior Change in HCI Research,” Proc. SIGCHI Conf. Human Factors in Com-puting Systems (CHI 11), ACM, 2011, pp. 3063–3072.

2. E.P.S. Baumer et al. “Prescriptive Persua-sion and Open-Ended Social Awareness: Expanding the Design Space of Mobile Health,” Proc. ACM Conf. Computer-Supported Cooperative Work (CSCW 12), ACM, 2012, pp. 475–484.

3. G. Gay et al., “Pilot Study of Aurora, a Social, Mobile-Phone-Based Emotion Sharing and Recording System,” J. Diabe-tes Science and Technology, vol. 5, no. 2, 2011, pp. 325–332.

4. J.P. Pollak, P. Adams, and G. Gay, “PAM: A Photographic Affect Meter for Frequent, In Situ Measurement of Affect,” Proc. SIGCHI Conf. Human Factors in Computing Sys-tems (CHI 11), ACM, 2011, pp. 725–734.

PURPOSE: The IEEE Computer Society is the world’s largest association of computing

professionals and is the leading provider of technical information in the field.

MEMBERSHIP: Members receive the monthly magazine Computer, discounts, and

opportunities to serve (all activities are led by volunteer members). Membership is open to

all IEEE members, affiliate society members, and others interested in the computer field.

COMPUTER SOCIETY WEBSITE: www.computer.org

Next Board Meeting: 13–14 June 2013, Seattle, WA, USA

EXECUTIVE COMMITTEEPresident: David Alan Grier

President-Elect: Dejan S. Milojicic; Past President: John W. Walz; VP, Standards Activities: Charlene (“Chuck”) J. Walrad; Secretary: David S. Ebert; Treasurer: Paul K.

Joannou; VP, Educational Activities: Jean-Luc Gaudiot; VP, Member & Geographic Activities: Elizabeth L. Burd (2nd VP); VP, Publications: Tom M. Conte (1st VP); VP, Professional Activities: Donald F. Shafer; VP, Technical & Conference Activities: Paul

R. Croll; 2013 IEEE Director & Delegate Division VIII: Roger U. Fujii; 2013 IEEE Director & Delegate Division V: James W. Moore; 2013 IEEE Director-Elect & Delegate Division V: Susan K. (Kathy) Land

BOARD OF GOVERNORSTerm Expiring 2013: Pierre Bourque, Dennis J. Frailey, Atsuhiro Goto, André Ivanov,

Dejan S. Milojicic, Paolo Montuschi, Jane Chu Prey, Charlene (“Chuck”) J. Walrad

Term Expiring 2014: Jose Ignacio Castillo Velazquez, David. S. Ebert, Hakan

Erdogmus, Gargi Keeni, Fabrizio Lombardi, Hironori Kasahara, Arnold N. Pears

Term Expiring 2015: Ann DeMarle, Cecilia Metra, Nita Patel, Diomidis Spinellis,

Phillip Laplante, Jean-Luc Gaudiot, Stefano Zanero

EXECUTIVE STAFFExecutive Director: Angela R. Burgess; Associate Executive Director & Director, Governance: Anne Marie Kelly; Director, Finance & Accounting: John Miller;

Director, Information Technology & Services: Ray Kahn; Director, Membership Development: Violet S. Doan; Director, Products & Services: Evan Butterfield;

Director, Sales & Marketing: Chris Jensen

COMPUTER SOCIETY OFFICESWashington, D.C.: 2001 L St., Ste. 700, Washington, D.C. 20036-4928

Phone: +1 202 371 0101 • Fax: +1 202 728 9614

Email: [email protected]

Los Alamitos: 10662 Los Vaqueros Circle, Los Alamitos, CA 90720 •

Phone: +1 714 821 8380 • Email: [email protected]

Membership & Publication OrdersPhone: +1 800 272 6657 • Fax: +1 714 821 4641 • Email: [email protected]

Asia/Pacific: Watanabe Building, 1-4-2 Minami-Aoyama, Minato-ku, Tokyo 107-

0062, Japan • Phone: +81 3 3408 3118 • Fax: +81 3 3408 3553 • Email: tokyo.ofc@

computer.org

IEEE BOARD OF DIRECTORSPresident: Peter W. Staecker; President-Elect: Roberto de Marca; Past President: Gordon W. Day; Secretary: Marko Delimar; Treasurer: John T. Barr; Director & President, IEEE-USA: Marc T. Apter; Director & President, Standards Association: Karen Bartleson; Director & VP, Educational Activities: Michael R. Lightner; Director & VP, Membership and Geographic Activities: Ralph M. Ford; Director & VP, Publication Services and Products: Gianluca Setti; Director & VP, Technical Activities: Robert E. Hebner; Director & Delegate Division V: James W. Moore; Director & Delegate Division VIII: Roger U. Fujii

revised 22 Jan. 2013