Integrating Person-to-Person Social Support in Smartphone Apps for

Promoting Physical Activity

Bojan Simoski, Michel Klein, Aart T. van Halteren and Henri Bal

Dept. of Computer Science, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands

Keywords:

Mobile Health, Physical Activity, Social Support.

Abstract:

The epidemic of physical inactivity is a major health hazard in the modern society, therefore creating effec-

tive and innovative health programs and interventions is important. This paper presents a novel approach in

which person-to-person social support is incorporated in mHealth interventions for increasing physical activ-

ity. Social support is already used as a behaviour change technique in mHealth apps for inﬂuencing physical

activity, but mostly offered virtually. While virtually-offered social support is efﬁcient, we believe that person-

to-person communication, based on personal coaching, might open a new way of inﬂuencing the inactive users

and their motivation. Responding to this, we developed an Android application that facilitates physically inac-

tive users to connect with a real life coach to receive person-to-person social support. This paper explains the

motivation behind the system’s design decisions and discusses the potential of social support in mHeatlh apps.

In addition, we present the design of the evaluation study in which the hypotheses and research questions will

be evaluated.

1 INTRODUCTION

Physical inactivity is identiﬁed as one of the major

hazards related to many health problems including

cancer, diabetes and heart diseases. Consequently,

physical inactivity is the 4th global risk for mortal-

ity in the world, being responsible for over 3 mil-

lion deaths annually, as reported by the World Health

Organization [WHO] (2009). WHO recommends at

least 150 minutes of moderate-intensity aerobic phys-

ical activity throughout the week for adults between

16-64 years. Alternatively, this age group should do

at least 75 minutes of weekly vigorous-intensity ex-

ercises, or a combination between the two (WHO,

2017). Unfortunately, these guidances are not met

by a signiﬁcant proportion of the population (Hallal

et al., 2012), resulting in a common interest of de-

signing efﬁcient health programs and interventions.

Meanwhile by 2018, over a one third of the world

population is projected to own a smartphone, bring-

ing the estimated total of users to almost 2.53 bil-

lion (Statista, 2017). As of June 2017, there were

more than 3 million available smartphone applica-

tions (apps) at Google Play Store, almost 100000 of

them were categorized as health & ﬁtness apps (App-

Brain, 2017). Apps are undoubtedly popular, easy to

reach and offer cost-effective interventions, therefore

they could be considered as promising tool for inﬂu-

encing human health.

van den Dool et al. (2017) have shown that 31% of

the Dutch population uses electronic tools for sports

and moving activities, out of which the apps were the

most popular tool with 61% of total electronic tools

usage. However, there is a remarkable contrast in

tools usage between the sufﬁciently physically active

users and the insufﬁciently physically active users,

that we will refer to as inactives in the remaining part

of the paper. The usage of tools among active people

varies from 32% to 55%, depending on the type of

active person, compared with just 12% of inactives,

that have used any electronic tool for their sport ac-

tivities. In order to indulge inactives to use health &

ﬁtness apps, there is a necessity to better understand

their needs. van den Dool (2015) has investigated

the behaviour of this population group, and proposed

several guidances that inactives might ﬁnd useful re-

lated to goal setting, skills, social support and per-

sonal coaching, as detailedly explained in Section 3.

Following these guidances, we have created an

Android app that targets inactives and promotes

physical activity intervention predominantly via real-

life social support and having a personal motivator

(coach). Social support is proven as a powerful tool

for inﬂuencing people’s behavior change and is al-

Simoski, B., Klein, M., Halteren, A. and Bal, H.

Integrating Person-to-Person Social Support in Smartphone Apps for Promoting Physical Activity.

DOI: 10.5220/0006644504970504

In Proceedings of the 11th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2018) - Volume 5: HEALTHINF, pages 497-504

ISBN: 978-989-758-281-3

497

ready applied in some health & ﬁtness apps, but

mostly offered virtually: via forums, online social

networks or chat messages. While virtually-offered

social support is efﬁcient and was already linked with

increasing physical activity among users, we believe

that person-to-person communication, based on social

support and personal coaching, could bring additional

beneﬁt for the inactives and their motivation. There-

fore, in this work the virtually-offered social support

was joined by an additional layer, one of real-life so-

cial support - our app promotes group exercises and

support where dyad, a group of two people, can do

joint exercises, indulge in real life communication and

motivation. In the dyad, one individual being sufﬁ-

ciently physically active, is a personal motivator for

the other individual, namely the insufﬁciently physi-

cally active person. The dyad members should have

strong mutual social ties, being either friends, col-

leagues, family members or partners.

The goal in the ﬁrst phase of this long term

project, is to investigate the effectiveness and accep-

tance of applying (real-life) social support in mo-

bile health (mHealth) interventions for increasing the

physical activity among inactives. In addition, we

will obtain the physical activity level (PAL) trend-

lines, and determine if there is an increase of phys-

ical activity for our participants over the experimental

period. Finally, we would like to investigate the corre-

lations between the PAL trendlines and the perceived

social support via our app.

The remainder of the paper is organized as fol-

lows: Section 2 discusses the potential of using so-

cial support for physical activity interventions, espe-

cially in physical activity apps. In Section 3 we ex-

plain the motivation behind the Social Coaching app

design choices, and give an extensive description of

the app’s modules. The evaluation plan is presented

in Section 4, where we describe the planned experi-

mental setup, the data nature and collection, followed

by the analysis plan where we deﬁne our hypotheses.

Finally, Section 5 ends the paper with a discussion.

2 BACKGROUND

This section starts by giving overview of previous re-

search studies that explored the implementation of so-

cial support in physical activity apps. As real-life

social support is often omitted when designing the

apps features, we continue by explaining the inﬂuence

that the social environment and real life contact might

have on people’s physical activity behavior. Finally,

we present one way in which person-to-person com-

munication could be integrated in apps, by explaining

the concept of social accountability.

Social support is a complex term and has be con-

ceptualized and deﬁned from multiple perspectives.

From the mHealth perspective, a valid social sup-

port interpretation is speciﬁed in the taxonomy con-

structed by Abraham and Michie (2008), that pro-

posed a set of standardized deﬁnitions of the tech-

niques most commonly used in behaviour change in-

terventions. Social support, one of the 26 techniques

described, is deﬁned as ”prompting consideration of

how others could change their behaviour to offer the

person help or (instrumental) social support, includ-

ing buddy systems and/or providing social support”

(Abraham and Michie, 2008, p.4). The presence and

effectiveness of social support as a behaviour change

technique in physical activity apps has already been

investigated. Matthews et al. (2016) investigated the

pervasive technologies used in physical activity apps,

concluding that social support is moderately repre-

sented, mostly via social comparison, social learn-

ing and competition. These techniques refer to com-

paring behavior and results between users, observing

and learning from behavior of other users, and ﬁnally

competing with other users. Another study (Bort-

Roig et al., 2014) has identiﬁed social support net-

working as one of the most effective behavior change

strategies for encouraging physical activity. King

et al. (2013) tested the effectiveness of different mo-

tivational apps (analytic, social and affective) for in-

creasing physical activity at users, and found the so-

cial app, being based on social comparison and social

normative feedback, as most effective.

A common approach for the above examined apps

is that they offer social support virtually, mostly via

online social networks like Twitter, Facebook, or by

creating online support communities. While applying

virtual social support has already been shown to be

effective, the potential of adapting the real-life social

support via the social environment in mHealth inter-

ventions is usually left out. The social environment

could offer social interactions that are missed in vir-

tual social support, for example, observational learn-

ing. Observational learning for physical activity can

be achieved by exercising with others and observing

their behavior, that was shown useful to build positive

social norms for physical activity (St

ahl et al., 2001).

People having low social support from the per-

sonal environment are more than twice as likely to

exhibit sedentary behavior than those whose personal

environment was highly supportive and motivational

(St

ahl et al., 2001). Connecting with the nearest sur-

roundings matter: having supportive partner, family

and/or friends all contribute to increased physical ac-

tivity (Eyler et al., 1999; Sternfeld et al., 1999). Social

HEALTHINF 2018 - 11th International Conference on Health Informatics

498

support interventions based on ”buddy” systems, hav-

ing a ”contract” with others to achieve a certain level

of physical activity, initiating walking or other types

of sport groups, showing conﬁdence in one’s ability,

were all associated with increasing levels of physical

activity (Kahn et al., 2002; Sternfeld et al., 1999).

Social accountability refers to ”a person’s aware-

ness of another person’s goal and rendering him-

self/herself responsible to the goal’s successful ful-

ﬁllment” (Chen et al., 2014, p.1). Personal coach-

ing, where the coach observes and supervises the goal

progress set by another person (coachee), could be

one way of applying social accountability in mHealth

interventions. This type of coach-coachee relation-

ship could be used to integrate person-to-person com-

munication in physical activity apps. Commercial ap-

plications are already using social accountability to

help users achieve goals. CommitTo3 (2015) is an

app in which users build social accountability groups,

and try to motivate themselves in fulﬁlling three daily

goals, by sharing their own progress and looking at

other team members progress. In the GoalSponsor

(2012) app, users appoint their own ”accountability

buddy” to monitor and share their progress with. The

HealthyTogether project explores mutual accountabil-

ity in a gamiﬁcation mobile app, their results show-

ing that users improved their physical activity by 15%

when using the app compared with when they were

exercising alone (Chen and Pu, 2014).

3 THE SOCIAL COACHING APP

We present a ’Social Coaching’ app that offers a per-

sonalized social support experience for inactives. As

mentioned in Section 1, our app design is inspired by

the set of requirements introduced by van den Dool

(2015), that suggested several approaches that might

attract inactives to be more physically active, the most

relevant for our app being paraphrased as follows:

1. Include skills that the inactives already possess or

introduce them with less complicated new skills.

2. Deﬁne goals that are individual and where any

progress is good enough.

3. Apply social support in interventions as this tech-

nique can keep users motivated for longer time

and comfortable during the exercise.

4. Have a personal coach as this can be helpful to

overcome the initial obstacles.

The following subsections discusses the app func-

tionalities, and the technology behind the app.

3.1 Social Aspects between Dyad

Members

The dyad is composed by two individuals having

strong mutual social ties: family members, partners,

friends or colleagues. The dyad members are:

• Coachee - user that does not meet the minimum

150 minutes/week WHO requirement. Supported

by the motivator and using the app, the coachee

wants to improve its physical activity.

• Motivator - user that does meet the minimum 150

minutes/week WHO requirement. The motivator

should help the coachee in the process of becom-

ing more physically active.

As shown in Section 2, the social environment and

having supportive partners, family or friends were all

associated with increasing physical activity, therefore

socially tied dyads are preferred. The dyad is teamed

before participants are joining the experiment, it is a

responsibility of the coachee to decide whom to team

up with - preferably the coachee will choose a sup-

portive person as the motivator.

The relation between our dyad members is implic-

itly based on the social accountability factor intro-

duced in the previous section - even though the ex-

ercise results are team based, there is one dyad mem-

ber, the coachee, that wants to accomplish a goal, and

another dyad member, the motivator, that is aware of

this goal and tracks the goal progress.

The motivator should be a person who is already

physically active, therefore has awareness of the ben-

eﬁts of being ﬁt. Driven by the positive effects of ob-

servational learning for physical activity as discussed

in Section 2, we assume that having a ”role model”

motivator type can be helpful for the inactives.

3.2 Exercise Types

Following the ﬁrst requirement (related to skills), the

app promotes and records dyad’s walking and / or run-

ning exercise sessions. Walking and running are cho-

sen as physical activities because they do not require

any speciﬁc equipment, are budget friendly and have

signiﬁcant health effects if performed regularly.

The users could obviously do other types of phys-

ical activity during their day. In order to have this

information, the app supports manual logging of any

other physical activity types, as we want to take in

consideration all the activities that might inﬂuence

user’s PAL.

Integrating Person-to-Person Social Support in Smartphone Apps for Promoting Physical Activity

499

3.3 Goal Setting

The dyad sets weekly goals expressed in number of

meters of weekly walk and/or run. Before conﬁrming

the goal, the app gives a pop-up message showing the

approximate total time in minutes for achieving that

weekly goal. Considering the second requirement (re-

lated to goals), explicit goal recommendation is inap-

propriate. The goals are set by the coachee, according

to the ambition for the upcoming week, and possi-

bly, upon consulting the motivator. The app implic-

itly inﬂuences the coachee’s goal setting strategy by

showing the WHO recommendation of 150 minutes

of moderate-intensity physical activity per week as a

reference point for the user. We have used this partic-

ular reference point, as the app promotes walking and

running, that are both considered moderate-intensity

exercise types, making them a suitable choice for the

inactives. Users can track their goals on the app’s

home screen where we visualize pie charts of the cur-

rent weekly progress, presenting the percentage of the

remaining/accomplished goal. Moreover, in order to

support a timeline of the physical activity progress,

the app has a history screen that provides the dyad

members with a retrospective view of their weekly

walking / running goals accomplishments and their

individual exercise session records over the weeks.

3.4 Contract Agreement

Before they are allowed to exercise, the dyad mem-

bers needs to ’sign’ an exercise contract. This vir-

tual contract is imagined as deﬁning the available free

slots by both team members, during the week, that

would ideally be dedicated for performing team exer-

cise. The time slots should bring more responsibility

at both the motivator and the coachee ”to stick to the

plan”, as having a behavioural contract for a certain

amount of exercise was already related to increase of

physical activity (Kahn et al., 2002). The dyad mem-

bers can initiate exercise at any time, regardless the

time slots that are deﬁned in the contract agreement.

3.5 The Exercise Session

The exercise session is imagined as team activity,

where the motivator and coachee walk or run together.

The exercises can be initiated by both the motivator

and the coachee, upon mutual agreement. These ex-

ercises are an opportunity for the motivator to support

and teach the coachee. During the exercise session,

the app shows the progress made by the coachee in

time and meters passed, to both members of the dyad.

Therefore, the motivator is able to real-time track the

progress that the coachee is making and give instant

feedback, while the latter can simply ”forget” about

the phone and relax on actually performing the ex-

ercise. After each exercise, the app displays a short

questionnaire to the coachee, related to satisfaction

with the exercise and the team communication dur-

ing the exercise. When exercising together is not pos-

sible, the coachee can individually perform the exer-

cise session, and the motivator will be notiﬁed about

the end result.

Figure 1: Social Coaching app screenshots. Left screenshot

shows the different menu options, related to weekly goal

setting, initiating new exercises or logging extra exercises.

The right screenshot displays the home screen, where the

user can reﬂect on his weekly goal progress and see infor-

mation about upcoming exercises.

3.6 Chat Feature

The dyad can communicate to each other using the

chat feature. Having the messaging feature can be

beneﬁcial for the dyad to ease and enrich their com-

munication. The frequency of messages exchanged

was already linked to increase of physical activity

(Chen and Pu, 2014).

3.7 App Communication with

End-users

The app has a rather passive communication with the

users. This is a conscious design decision, since we

aimed at keeping the app interaction with the users at

minimum, as our goal is to investigate how the real-

life social support works for the coachee. However,

the app does communicate using notiﬁcations, and

there are several of them, mostly directed toward the

motivator. Sending these notiﬁcations, the app can be

HEALTHINF 2018 - 11th International Conference on Health Informatics

500

Figure 2: Architecture of the Social Coaching app. The app is supported by SWAN, an open source framework for building

context aware Android apps. SWAN enables ofﬂine real-time sensor data sharing between smartphones via Bluetooth, a

feature used for the exercise sessions of the app. In addition, the cloud layer is used to store and retrieve the app’s data.

considered as a rudimental virtual coach for the mo-

tivator: it reminds the motivator to send motivational

messages to the coachee in case this was not done yet

during the day, it reminds to initiate new exercise in

case when the dyad has a free slot assigned for that

particular day, and ﬁnally it notiﬁes if the dyad did

not do the planned exercise.

3.8 Technologies and Tools

The Social Coaching app is an Android based app.

We chose the Android platform due to its big market

share and wide acceptance among smartphone users.

The app is supported by a back-end API system, built

using the Play Framework and deployed as a Heroku

cloud application. The REST API calls are used

to store and retrieve information to/from the cloud

database (Heroku Postgres database), that contains all

the app related data, except the chat messages. To de-

ploy the chat we used Google Firebase and stored the

chat messages in the Firebase Realtime Database.

One of the biggest challenges we faced during the

app development was how to build an exercise track-

ing system, that can work in ofﬂine mode, has low

battery consumption and enables real-time sharing of

exercise progress between the coachee and the moti-

vator. We have used the open-source SWAN project

(SWAN, 2010), a framework for developing context-

aware Android apps, in order to manage the smart-

phone’s sensor data processing during the exercises.

SWAN runs as a background service and can be

simultaneously accessed by multiple Android apps

on the phone. SWAN creates powerful abstractions

for managing the otherwise complex (Android) sen-

sor APIs: it supports smartphone sensors, software

sensors (e.g. open-source weather API data) or exter-

nal sensors connected to the smartphone (e.g. wear-

ables data via Bluetooth). SWAN offers more than 20

predeﬁned sensors, moreover the developer can eas-

ily plug-in any new sensor to the framework accord-

ing to the application needs. As example, in order to

satisfy our application requirements, we have manu-

ally created the Distance Covered sensor that uses the

smartphone’s accelerometer sensor data obtained dur-

ing the exercise and translates it to number of steps,

using the Pedometer (2013) app’s open source step

counter algorithm. The number of steps are then rep-

resented in meters depending on the walk/run stride

length of the coachee.

Context-aware apps communicate with SWAN us-

ing the SWAN API to register and unregister sen-

sor expressions. These expressions are build using

the domain speciﬁc language of SWAN: the SWAN-

Song. In our app we use the following sensor expres-

sion for calculating the distance covered during the

exercise session, at the coachee phone:

Integrating Person-to-Person Social Support in Smartphone Apps for Promoting Physical Activity

501

self@distancecovered:meters

where the ﬁrst component indicates the source of

the sensor - self means that we get the data locally

from the smartphone. The second component deﬁnes

the sensor, which is the Distance Covered sensor in

our case. Finally, the third component speciﬁes the

value of interest within the chosen sensor, in our case

we are interested in the number of meters, obtained

from the Distance Covered sensor.

On the motivator side, during the joint exer-

cises, we want to show the real-time progress of the

coachee. SWAN supports Bluetooth-enabled data

sharing between phones, as simple as registering a

sensor expression in the following manner:

coachee_BT_ID@distancecovered:meters

here the source of the sensor is the coachee’s Blue-

tooth MAC address, meaning that we are interested

in getting the meters of the Distance Covered sensor

from the coachee’s smartphone. After the Bluetooth

connection is established, the motivator will be able

to see the coachee’s real-time progress on the screen.

Once the expressions are registered, they are eval-

uated on every new sensor event, using the Evaluation

Engine. The Evaluation Engine sends broadcast mes-

sages to the app, whenever it needs to notify sensor

data changes. The sensor data is read until an un-

ﬁnish an exercise session.

4 EVALUATION PLAN

4.1 Experimental Setup

The proposed app in this paper will be evaluated in

a user experiment, to test if promoting real-life social

support via mHealth apps can be considered as an use-

ful tool for physical activity interventions. There is no

control group (i.e. dyads without the app), as the aim

of the experiment is broader than comparing the ef-

fectiveness of using such a system with face-to-face

coaching.

The goal of our experiment is to investigate the

feasibility of real-life social support via an app. We

will do this in a small user study of approximately 50

participants, i.e. at least 25 dyads, in which we both

look at the user experiences and the change in physi-

cal activiity. At the time of writing this paper, we have

three dyads that have already started the experiment.

The ongoing recruitment is done via online tools like

websites, forums, social media, and by recruiting at

the university campus. The participants should be

between 16-64 years old, similar to the age group

speciﬁed by WHO in Section 1, be healthy enough

to perform physical activities and possess an Android

smartphone.

In order to get additional step-based physical ac-

tivity data, each participant will be given an activity

tracker, a Fitbit One device. This is a sufﬁciently

reliable activity tracker for continuous measurement

of physical activity (Takacs et al., 2014; Paul et al.,

2015), and will be used during the whole experiment

duration, as a support tool besides the apps.

Before the experiment, the participants will an-

swer online questionnaires regarding team relation-

ship; personality traits; and current physical activity

status, motivation and goals. The overall participa-

tion will take 5 weeks, with one week of assessment

period and four weeks of intervention period.

For the assessment week, the participants will in-

stall the ActivityLogger app, a simple physical activ-

ity diary app, where they will log their individual ex-

ercises over the period of 7 days. We will use the

data obtained from this app, combined with the activ-

ity tracker data, to determine the pre-intervention PAL

for each participant. This approach will give us more

reliable and accurate baseline for determining the ini-

tial PALs of the participants, compared to using phys-

ical activity self-assessment questionnaires like IPAQ

(Lee et al., 2011). Using the initial PAL values of our

participants, we can test if the dyad indeed consists of

one sufﬁciently active person and one insufﬁciently

active person.

After the dyads are conﬁrmed, the participants

will enter the four weeks of intervention period, dur-

ing which they will use the Social Coaching app, sup-

ported by the activity tracker. At the end of the in-

tervention period, the participants will be asked to ﬁll

another set of questionnaires, regarding their satisfac-

tion of using the app as a tool for increasing their

physical activities, focusing on the social support con-

text of the app.

4.2 Data Collection

We are interested in gathering relevant data regarding

physical activity and perceived social support. The

physical activity data is collected via the Android

apps, and the activity tracker, and enables us to calcu-

late the PAL of each participant. The pre-intervention

PAL value will be calculated combining the data from

the ActivityLogger app and the activity tracker. Dur-

ing the intervention period, the PAL will be obtained

on weekly basis, by gathering the data from the Social

Coaching app and the activity tracker. Using these

PAL values we will calculate the PAL trendlines of

each participant, which will be then used in the anal-

ysis. The activity tracker enables data extraction in

HEALTHINF 2018 - 11th International Conference on Health Informatics

502

spreadsheets, where we will obtain daily information

about number of steps, distance in km, ﬂoors climbed;

and number of lightly, fairly and very active minutes

over the day. The ActivityLogger app will provide in-

formation about exercise type, duration and date. The

Social Coaching app gives data about exercise type,

time spent in exercise, number of meters, date for the

walking and running sessions. In addition, this app al-

lows the user to log extra exercises (besides walking

and running), in a similar way as in the ActivityLog-

ger app, therefore an additional data about exercise

type, duration and date, will be obtained for them.

The Social Coaching app, will further give us in-

sights about the perceived social support. The social

support factor will be calculated by quantifying the

communication patterns between dyad members, and

by using the app questionnaire responses. We have

identiﬁed several suitable data sources for calculat-

ing the perceived social support: the frequency of

joint exercises; frequency of chat messages and senti-

ment analysis of the messages; weekly questionnaires

about dyad satisfaction sent as notiﬁcation; after ex-

ercise satisfaction questionnaire sent as a notiﬁcation

to the coachee.

Additional data will be collected via online ques-

tionnaires both before and after the experiment, as ex-

plained in the previous subsection.

4.3 Analysis Plan

The multidimensionality of the collected data will

give an opportunity for data analysis from different

perspectives. Calculating the individual PAL trend-

lines, as explained in Section 4.2, will help us de-

termine if there was an increase of physical activity

at the participants, during the experimental period.

We will test the signiﬁcance of the PAL trends us-

ing Mann-Kendall (MK) trend tests, that can statis-

tically access trend presence in a variable over time.

Moreover, we will investigate the correlation between

the PAL trends and the perceived social support over

time. For this, we have deﬁned several hypotheses:

H1. Inactives that show satisfaction of the per-

ceived social support will show an increase in

PAL trends.

H2. Inactives that show dissatisfaction of the per-

ceived social support will not show an increase in

PAL trends.

H3. High frequency and positive sentiment of ex-

changed messages between dyad members, will

result with increase in PAL trends at inactives.

H4. Low frequency and negative/neutral senti-

ment of exchanged messages between dyad mem-

bers, will not result with increase in PAL trends at

inactives.

These conditions will be tested by performing MK

non-parametric tests for statistical dependence.

Finally, we would like to investigate the ac-

ceptance of applying (real-life) social support in

mHealth interventions for increasing the physical ac-

tivity among inactives. In order to answer this, the

participants will be given questionnaires regarding

app usability and acceptance, with special emphasis

on the (real-life) social support context of the app.

5 DISCUSSION

This paper presents a novel approach in which person-

to-person social support is incorporated in mHealth

interventions for increasing physical activity. In or-

der to ﬁght the epidemic of physical inactivity, in-

novative and effective tools which are accepted by

end-users are required. Real-life social support via

personal coaching has the potential to motivate inac-

tives, but is hardly used in mHealth apps. In this work

we focused on explaining the design decisions behind

the system and we discussed the potential of social

support in mHeatlh apps. Furthermore, we have in-

troduced our evaluation plan in which the hypotheses

and research questions will be evaluated.

There are several considerations for future work,

inspired by the current limitations of our research, that

we would like to discuss. The social context play key

role in this research, and we can imagine incorporat-

ing different social interactions, in addition to the cur-

rent dyad relationship, for example:

• create groups of arbitrary number of members, for

example enable the coachee to have more than one

motivator.

• match with an anonymous person instead of a per-

son from the social circle.

• instead of having a motivator, match inactive per-

sons with each other.

With these combinations, we could test different

types of social interactions, which is important in or-

der to ﬁnd the ideal combinations of motivators and

coachees.

One important aspect that could inﬂuence the per-

ceived social support by people, is the effect of social

comparison. Not all people get motivated by having

a role model that performs better then they do. Some

individuals gets more motivated by comparing them-

selves with others performing worse (downward so-

cial comparison), and other individuals are motivated

Integrating Person-to-Person Social Support in Smartphone Apps for Promoting Physical Activity

503

by comparing themselves with someone who is better

then they are (upward comparison). We would like to

consider this theory when thinking about supporting

new social ties types.

Finally, the focus of this research is on the inac-

tives, but in order to have an effective intervention

that is based on social support, we should consider

the satisfaction of the motivators as well. We have

build the current app under the assumption that the

motivators will use the app in order to help their ac-

quaintances, but the question is what will keep them

motivated to use the app for a longer period? One pos-

sible approach is to extend the app by implementing

gamiﬁcation elements, i.e having rewards, trophies,

challenges, or the possibility to compare their perfor-

mance with the performance of other motivators.

REFERENCES

Abraham, C. and Michie, S. (2008). A taxonomy of behav-

ior change techniques used in interventions. Health

psychology, 27(3):379.

AppBrain (2017). Most popular google play cate-

gories. Retrieved on 01/10/2017 from website:

https://www.appbrain.com/stats/android-market-app-

categories.

Bort-Roig, J., Gilson, N. D., Puig-Ribera, A., Contreras,

R. S., and Trost, S. G. (2014). Measuring and inﬂu-

encing physical activity with smartphone technology:

a systematic review. Sports Medicine, 44(5):671–686.

Chen, Y. and Pu, P. (2014). Healthytogether: exploring

social incentives for mobile ﬁtness applications. In

Proceedings of the Second International Symposium

of Chinese CHI, pages 25–34. ACM.

Chen, Y., Zhang, J., and Pu, P. (2014). Exploring social

accountability for pervasive ﬁtness apps. In Proceed-

ings of the Eighth International Conference on Mobile

Ubiquitous Computing, Systems, Services and Tech-

nologies.

CommitTo3 (2015). Committo3. Retrieved on 20/08/2017

from: https://www.committo3.com/.

Eyler, A. A., Brownson, R. C., Donatelle, R. J., King, A. C.,

Brown, D., and Sallis, J. F. (1999). Physical activ-

ity social support and middle-and older-aged minority

women: results from a us survey. Social science &

medicine, 49(6):781–789.

GoalSponsor (2012). Goal sponsor. Retrieved on

20/08/2017 from: http://signup.goalsponsors.com/.

Hallal, P. C., Andersen, L. B., Bull, F. C., Guthold, R.,

Haskell, W., Ekelund, U., Group, L. P. A. S. W.,

et al. (2012). Global physical activity levels: surveil-

lance progress, pitfalls, and prospects. The lancet,

380(9838):247–257.

Kahn, E. B., Ramsey, L. T., Brownson, R. C., Heath, G. W.,

Howze, E. H., Powell, K. E., Stone, E. J., Rajab,

M. W., and Corso, P. (2002). The effectiveness of

interventions to increase physical activity: A system-

atic review. American journal of preventive medicine,

22(4):73–107.

King, A. C., Hekler, E. B., Grieco, L. A., Winter, S. J.,

Sheats, J. L., Buman, M. P., Banerjee, B., Robinson,

T. N., and Cirimele, J. (2013). Harnessing differ-

ent motivational frames via mobile phones to promote

daily physical activity and reduce sedentary behavior

in aging adults. PloS one, 8(4):e62613.

Lee, P. H., Macfarlane, D. J., Lam, T., and Stewart, S. M.

(2011). Validity of the international physical activity

questionnaire short form (ipaq-sf): A systematic re-

view. International Journal of Behavioral Nutrition

and Physical Activity, 8(1):115.

Matthews, J., Win, K. T., Oinas-Kukkonen, H., and Free-

man, M. (2016). Persuasive technology in mobile ap-

plications promoting physical activity: a systematic

review. Journal of medical systems, 40(3):72.

Paul, S. S., Tiedemann, A., Hassett, L. M., Ramsay, E.,

Kirkham, C., Chagpar, S., and Sherrington, C. (2015).

Validity of the ﬁtbit activity tracker for measuring

steps in community-dwelling older adults. BMJ open

sport & exercise medicine, 1(1):e000013.

Pedometer (2013). Pedometer : Lightweight pe-

dometer app for android using the hardware

step sensor. Retrieved on 12/09/2017 from:

https://github.com/j4velin/pedometer.

ahl, T., R

utten, A., Nutbeam, D., Bauman, A., Kannas, L.,

Abel, T., L

uschen, G., Rodriquez, D. J., Vinck, J., and

van der Zee, J. (2001). The importance of the social

environment for physically active lifestyleresults from

an international study. Social science & medicine,

52(1):1–10.

Statista (2017). Number of smartphone users world-

wide from 2014 to 2020 (in billions). Retrieved on

04/10/2017 from website: https://www.statista.com/

statistics/330695/number-of-smartphone-users-

worldwide/.

Sternfeld, B., Ainsworth, B. E., and Quesenberry Jr, C.

(1999). Physical activity patterns in a diverse popula-

tion of women. Preventive medicine, 28(3):313–323.

SWAN (2010). The SWAN project. Retrieved on

20/08/2017 from: http://www.cs.vu.nl/swan/.

Takacs, J., Pollock, C. L., Guenther, J. R., Bahar, M.,

Napier, C., and Hunt, M. A. (2014). Validation of

the ﬁtbit one activity monitor device during treadmill

walking. Journal of Science and Medicine in Sport,

17(5):496–500.

van den Dool, R. (2015). Gedragsverandering bij moeilijke

groepen kansen om sport en bewegen te stim-

uleren? Retrieved on 25/08/2017 from: https://

www.kennisbanksportenbewegen.nl/?ﬁle=4720&

m=1437571688&action=ﬁle.download.

van den Dool, R., Hover, P., and Vos, S. (2017). Websheet

2017/1: Apps&devices. Retrieved on 20/08/2017

from: http://www.mulierinstituut.nl/publicaties/

websheets/websheet-apps-devices/.

WHO (2009). Global health risks: Mortality and

burden of disease attributable to selected ma-

jor risks. Retrieved on 29/08/2017 from web-

site: http://www.who.int/healthinfo/global burden

disease/globalhealthrisks report full.pdf.

WHO (2017). Physical activity and adults: Recommended

levels of physical activity for adults aged 18 - 64

years. retrieved on 29/08/2017 from website: http://

www.who.int/dietphysicalactivity/factsheet adults.

HEALTHINF 2018 - 11th International Conference on Health Informatics

504