Systematic Selection and Prioritization of Communication Channels in

the Healthcare Sector

Francisco Casaca and Andr

e Vasconcelos

INESC-ID, Instituto Superior T

ecnico, Avenida Rovisco Pais 1, Lisbon, Portugal

Keywords:

Communication Channels, Multi-channel, Omni-channel, Chatbot, Seamless Experience, Healthcare.

Abstract:

Many industries are using multi-channel approaches to bring users and organizations closer. The services

provided through these channels can be leveraged by using chatbots, allowing users to have simpler and more

natural interactions. However, this entails designing architectures that make services available through mul-

tiple channels, while making the users’ experiences coherent, as they switch between them. Media Richness

Theory introduces a way to classify the richness of communication channels, resorting to objective factors,

however it does not provide a systematic channel selection and prioritization process. To address these needs,

this work proposes a systematic approach to select and prioritize communication channels based on six factors:

feedback, multiple cues, personal focus, language variety, accessibility and cost. To validate this approach,

the systematic process is applied to three use cases in the healthcare domain.

1 INTRODUCTION

When interacting with organizations, users resort to

a variety of communication channels. These include

traditional channels such as face-to-face communica-

tion and phone calls, as well as digital channels such

as e-mail, web portals, mobile applications and video-

conferences (Androutsopoulou et al., 2019). These

channels are distinguished by their richness (the ca-

pability of providing information with high levels of

understanding), which depends on a variety of fac-

tors: some of them are objective, such as rapidness of

feedback, available modalities, accessibility and cost;

whereas others are subjective to the experience of the

user, such as familiarity with a particular channel or

topic.

In recent years, multi-channel and omni-channel

approaches have been implemented in many indus-

tries (Caroll and Guzm

an, 2015), allowing consumers

to choose from a variety of channels (according to

their preferences) and providing a seamless experi-

ence when switching between them. Furthermore,

some of the communication tasks performed through

the aforementioned digital channels have been auto-

mated by using task-oriented dialog agents (or chat-

bots), which can help users complete tasks using natu-

ral language (Jurafsky and Martin, 2000). These sys-

tems are simpler to use when compared to GUI ap-

plications because they mimic human-human interac-

tion.

Chatbots have thus the ability to leverage organi-

zations’ services by being available through multiple

digital communication channels (Androutsopoulou

et al., 2019). As a result, organizations can target a

higher percentage of the population, by offering their

services through channels that are more convenient,

natural and easy to use.

Healthcare is a domain that can beneﬁt from

multi-channel approaches (Laranjo et al., 2018).

New channels can be easily integrated with the pre-

established workﬂow, providing services to patients

who would not, otherwise be able to access them

(Morris et al., 2018). They can also reduce patient

travel and allow more frequent follow-ups.

1.1 Context

When offering more channels to users, there is one

fundamental issue that arises. Since each channel has

its own characteristics, not all of them can be applied

in the same scenarios. For instance, for use cases in

which communication depends on the vocal modality,

a channel through which it is only possible to com-

municate via text is not suitable. Therefore, it is im-

portant to evaluate the suitability of channels regard-

ing the desired communication tasks. Additionally,

there is the preoccupation of evaluating which chan-

nels should be prioritized.

1.2 Objectives

The objectives of this work are thus the following:

668

Casaca, F. and Vasconcelos, A.

Systematic Selection and Prioritization of Communication Channels in the Healthcare Sector.

DOI: 10.5220/0010441706680676

In Proceedings of the 23rd International Conference on Enterprise Information Systems (ICEIS 2021) - Volume 1, pages 668-676

ISBN: 978-989-758-509-8

• To assess the factors of communication channels

that inﬂuence their suitability.

• To create a systematic process for organizations to

select and prioritize communication channels.

1.3 Article Organization

This article starts by discussing work related with

multi-channel architecture designs. This includes the

introduction of chatbots and understanding what use

cases are viable for a bot implementation. After-

wards, it focuses on multi-channel and omni-channel

strategies, exploring how channels are used in health-

care and what are the factors that make each channel

unique.

Following, the Communication Channel Selection

Process, which provides a way to select and prioritize

channels, is proposed.

Finally, the channel selection process is applied

to the healthcare domain and the conclusions are pre-

sented.

2 BACKGROUND AND RELATED

WORK

2.1 Chatbots

Dialog Systems: communicate with users through

different modalities (such as text or speech) and can

be classiﬁed in two categories (Jurafsky and Mar-

tin, 2000): (1) Task-oriented dialog agents help users

complete tasks using natural language, such as giving

directions, ﬁnding restaurants or making calls; and (2)

Chatbots, which have the purpose of having extended

conversations, being able to mimic human-human in-

teraction, thus providing a more natural communica-

tion. For simplicity, throughout this article, the term

chatbot is used when referring to a dialog system.

2.2 Chatbot Use Case Evaluator

The Chatbot Use Case Evaluator (Ferreira and Vas-

concelos, 2019) allows the evaluation of a set of use

cases, enabling the identiﬁcation of the more suitable

ones for a bot implementation. It takes into consid-

eration three factors that must apply for a use case to

be viable: (1) General factors, that divide into well-

deﬁned Business rules and integration with existing

systems; (2) Factors over GUI applications, which

encompass multiple steps or input parameters, notiﬁ-

cations and authentication; and, (3) Factors over Hu-

mans, which consider the repetitiveness, consistency

and scalability of the use cases. If these three factors

verify for a certain use case, then its implementation

in a chatbot is viable.

Additionally, use cases that meet a great num-

ber of factors, or that meet factors that are of greater

importance to the organization should be prioritized.

Organizations can thus deﬁne weights for each fac-

tor and give a higher priority to the ones that have

a greater sum of weighted factors. The described

methodology was applied to three use cases: (1)

Scheduling an appointment, (2) Paying for an ap-

pointment, and (3) Medical diagnosis. After applying

the Evaluation process, (1) and (2) were the only use

cases evaluated as suitable to be implemented. In fact,

(3) was considered to not have well-deﬁned business

rules, therefore, it did not comply with the General

factors assessment.

2.3 Multi-channel and Omni-channel

Strategies

Communication Channels are traditional ofﬂine or

digital online intermediaries used by consumers to

interact with services, facilitating the transmission

of information and content (Androutsopoulou et al.,

2019).

Multi-channel strategy is a trend in many indus-

tries toward the use of multiple channels when engag-

ing with customers. These approaches offer a diverse

experience across channels consumers can choose

from, according to their preferences. Accenture (Car-

oll and Guzm

an, 2015) explains the success of multi-

channel strategies with customers having more ac-

cess to the Internet and social networks, prioritizing

convenience and being constantly on the move, valu-

ing the availability of services anywhere and at any

time. However, multi-channel strategies have lim-

ited or no integration between channels, providing

a non-seamless experience as consumers switch be-

tween them. Thus, users may be forced to repeat cer-

tain procedures or receive ambiguous messages (De-

loitte, 2015).

Omni-channel strategy consists on a synchronized

operating model in which all of the organization’s

channels are aligned (Deloitte, 2015). Consequently,

it delivers a seamless, consistent and personalized

customer experience. This is achieved by integrating

different channels that can be used whenever, wher-

ever and however. Using this approach, organizations

are able to respond consistently to the consumers’

evolving needs, while enabling them with the oppor-

tunity of using the channel of their choice to access

the services they desire.

Nowadays there are several areas of activity that

Systematic Selection and Prioritization of Communication Channels in the Healthcare Sector

669

use multiple communication channels including re-

tail and commerce (Accenture, 2014), banking and

ﬁnance (Komulainen and Makkonen, 2018), educa-

tion (binti Mistar, 2016) and government organiza-

tions (Androutsopoulou et al., 2019).

2.4 Communication Channels in

Healthcare

In this section we will focus on channels patients use

when getting healthcare.

When patients have health-related issues, they

seek healthcare in order to obtain a diagnosis (Fer-

reira and Vasconcelos, 2019). Diagnosis is typically

performed in-person, however, alternative channels

have been adopted in recent years, in order to pro-

vide services to patients who would not, otherwise,

be able to access them. Video-conference has been

regarded as a suitable alternative, since it has clear

similarities when compared with the face-to-face in-

teraction. These two channels have been compared

in post-traumatic stress disorder treatment (Germain

et al., 2010). The results show that therapeutic al-

liances developed similarly in both channels.

When a patient has already been diagnosed, it is

important to conduct follow-up consultations (post-

operative follow-ups, disease management or behav-

ior change (Morris et al., 2018)). Moreover, med-

ical follow-ups are typically used to guarantee that

there are no interruptions in medications, identify the

need for adjustments, and provide a better compliance

with the surveillance of various health parameters and

healthy lifestyle habits (Ferreira, 2020). Follow-ups

are typically performed in-person, however, research

has shown that using telemedicine has the advantage

of allowing a more frequent surveillance (Ferreira,

2020). Instant messaging (IM) is also being adopted

by clinicians, in particular, dermatologists (Morris

et al., 2018).

Before attending an in-person or video-conference

consultation (either for diagnosis or follow-up), pa-

tients should book their appointments. The schedul-

ing itself is typically performed with clerks (in-person

or through phone calls) or by using clinics’ portals

and applications.

Providing healthcare services via digital online

channels is a cost effective and ﬂexible alternative

that can be easily integrated with the pre-established

workﬂow. Furthermore, it has enabled clinical ser-

vices between rural areas and urban hospitals, reduc-

ing patient travel and the associated time and ﬁnancial

costs (Morris et al., 2018).

2.5 Communication Channels Factors

To provide channels with higher level of richness to

Greek citizens (Androutsopoulou et al., 2019), while

accounting for the equivocality and uncertainty of

communication tasks, the authors decided to use the

notions of Media Richness Theory (Daft and Lengel,

1983) and Channel Expansion Theory (Carlson and

Zmud, 1999) to provide channels with a higher level

of richness. These theories helped them analyze com-

munication channels both from an objective and from

a subjective standpoint respectively.

In recent years, other subjective factors regard-

ing communication channels have been studied (Ishii

et al., 2019) (apart from the ones introduced by Chan-

nel Expansion Theory). These include the channels’

accessibility, competency of use, time saving and cost

saving. This research concludes that evaluating the

richness of a channel according to objective features

is not enough. In fact, it is important to regard the ob-

jective characteristics of a medium, while keeping in

mind subjective experience and perceived richness.

3 SELECTION AND

PRIORITIZATION OF

COMMUNICATION

CHANNELS APPROACH

3.1 Communication Channel Selection

Process

This section presents a systematic way to select

and prioritize channels, depending on the use case.

Firstly, factors that inﬂuence channels suitability are

discussed. And following, the evaluation process is

presented.

3.1.1 Channel Suitability Factors

Before evaluating what channels should be used for

each use case, there is the need to identify what fac-

tors inﬂuence suitability (Tables 1 and 2. Based on

(Daft and Lengel, 1983) and (Ishii et al., 2019), it is

possible to identify the following factors:

Feedback. Enables individuals to check under-

standing and correct misunderstood interpretations.

Depending on the channel, it can be classiﬁed as im-

mediate, fast, medium, slow or very slow. In a face-

to-face interaction, for instance, feedback is immedi-

ate, whereas through a formal email it is very slow

(Daft and Lengel, 1983).

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

670

Table 1: Classiﬁcation of communication channels, according to objective factors.

Channel Feedback Cues Focus Language

Face-to-face Immediate Full visual/audio Personal Body, Natural

Video-conference Immediate Limited visual/audio Personal Body, Natural

Phone call Fast Limited audio Personal Natural

Instant messaging platform Slow Limited visual/audio Personal Body, Natural, GUI

Web portal without IM Medium Limited visual Impersonal GUI

Web portal with IM Slow Limited visual/audio Impersonal Natural, GUI

Mobile app without IM Medium Limited visual Impersonal GUI

Mobile app with IM Slow Limited visual/ audio Impersonal Natural, GUI

E-mail Very slow Limited visual Impersonal Natural

SMS Slow Text Personal Natural

Table 2: Classiﬁcation of communication channels, according to accessibility and cost.

Channel Accessibility Cost

Face-to-face Very low Commute

Video-conference Low Internet

Phone call Very high Per call

Instant messaging platform High Internet

Web portal without IM Medium Internet

Web portal with IM Medium Internet

Mobile app without IM High Internet

Mobile app with IM High Internet

E-mail High Internet

SMS Very high Per message

Multiple Cues. Information can be conveyed be-

yond the spoken or written message, using modali-

ties such as facial expression or tone of voice. The

considered cues in the MRT (Daft and Lengel, 1983)

are visual, audio and limited visual (referring to writ-

ten messages). However, when considering modern

communication channels these cues are not enough.

Communicating through video, for instance, repre-

sents visual and audio cues, however, it is far less rich

than communicating in-person, since it is not possible

to perceive modalities such as facial expressions with

the same level of precision. Because of this, the fol-

lowing types of cues will be considered: full visual,

full audio, limited visual, limited audio, and text.

Personal Focus. This factor is concerned with the

information source. Personal sources are related with

informal communication such as face-to-face inter-

action, while impersonal sources encompass formal

communication such as emails.

Language Variety. Each communication channel

has different possibilities of language use. Face-

to-face interaction, for instance, allows to commu-

nicate using natural and body language. However,

some mediums only allow the use of natural language

(SMS, for example). Additionally, the introduction

of instant messaging apps also allows interaction to

be performed via UI elements (Klopfenstein et al.,

2017).

Accessibility. Measures how accessible channels

are to users. Instant messaging apps, for instance,

are more accessible than face-to-face communication

(since the latter requires the interaction actors to meet

in-person), but less accessible than SMS (since IM

apps require an Internet connection).

Cost Saving. Is an important aspect for users and

organizations alike. Each communication channel has

different costs associated: the cost of a face-to-face

interaction is expressed in terms of commute; the

cost of using instant messaging apps is related with

internet access; and communicating via SMS or

phone calls has costs per message and per phone call,

respectively.

When communicating in-person (face-to-face),

the actors of the interaction have access to immediate

feedback and full visual and audio cues. The focus of

the communication is personal and they can use both

body and natural language to communicate. However,

Systematic Selection and Prioritization of Communication Channels in the Healthcare Sector

671

Figure 1: Use cases performed by patients in the healthcare domain.

communicating in-person requires the actors to meet,

therefore it has very low accessibility and costs asso-

ciated with commute.

Video-conference provides a similar experience to

face-to-face, however, it only provides limited visual

and audio cues, since the interaction occurs through a

screen. It has low accessibility since, usually, a spe-

ciﬁc software is required and has costs associated with

internet access.

Phone calls only provide limited audio cues and

the feedback is not as immediate as the previous al-

ternatives, thus being classiﬁed as fast. The focus is

still personal, however, it only allows the use of nat-

ural language. The accessibility is very high, since

the actors only need their mobile phones and no addi-

tional software is required and the costs are associated

with the cost of each call.

As for instant messaging platforms, feedback is

slower than in the previous alternatives, since it re-

quires typing a message or recording an audio or

video. It provides limited visual and audio cues,

personal focus and a great variety of language types

(body - in case of a recorded video -, natural and GUI

- it is possible to use interface elements such as quick-

replies or carousels). Their accessibility is also high,

since most people use instant messaging platforms in

their every day life, however, not as high as phone

calls, since they require Internet access.

Web portals and mobile apps are similar in fac-

tor classiﬁcation. They both provide medium feed-

back, since the interaction is faster when compared

to instant messaging. Typically they rely on visual

elements, providing limited visual cues and allowing

users to communicate only via traditional GUI. Fur-

thermore, they have an impersonal focus and require

Internet access. Mobile apps are more accessible than

web portals, since they can be accessed through a mo-

bile phone.

There are also web portals and mobile apps that

support instant messaging. This is the case, for in-

stance, of applications who provide both the tradi-

tional GUI and a chat window, through which users

can communicate with a human operator or chatbot.

In this case, the factors are similar to the traditional

web portals and mobile apps, however, feedback be-

comes slower, since messages take longer to write

than GUI interaction. Moreover, they support limited

visual and audio cues and the possibility of commu-

nicating via natural language.

E-mail is the slowest of all mediums, since it usu-

ally relies on formal messages written using natural

language, which take longer to write. It is a channel

that provides limited visual cues and is impersonal.

It has high accessibility, since it is available on mo-

bile phones and its costs are associated with Internet

access.

SMS is a similar medium to instant messaging

platforms, however, its cues only include text, there-

fore it only supports natural language. Its accessibil-

ity is very high and similar to phone calls, since SMS

is available on all mobile phones and has a cost per

message sent.

It is worth mentioning that the digital channels

through which it is possible to communicate using

natural language (such as instant messaging applica-

tions or SMS) can be integrated with chatbots. For

certain use cases, the bot’s performance may present

advantages when compared with human operators,

thus providing more consistent and scalable services

(Ferreira and Vasconcelos, 2019).

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

672

3.1.2 Channel Selection Process

Using the factors presented in the previous section,

it is possible to select and, afterwards, prioritize the

channels that are suitable for the given use cases.

Firstly, it is necessary to ﬁlter out channels that

can not be used in a use case. For instance, if a

use case has to be performed in-person, requiring full

visual and audio cues, it is possible to ﬁlter out all

other alternatives. Afterwards, the remaining chan-

nels can be prioritized. The Analytic Hierarchic Pro-

cess (AHP) (Kosteln

ık et al., 2019) is used to tackle

this issue, not only because it is widely use in in-

formation technology related ﬁelds when making de-

cisions, but also because it allows to measure non-

quantiﬁable criteria. Additionally, it has the advan-

tage of being easily understood by general audiences

(Kosteln

ık et al., 2019).

The channel selection process is thus deﬁned as

follows:

1. Assessment of the channels. For each use case:

(a) Deﬁne the minimum requirements (if neces-

sary) for the ﬁrst 4 factors: feedback, multiple

cues, personal focus and language variety.

(b) Filter out the channels that do not comply with

the minimum requirements.

i. If only one channel remains, the channel se-

lection process stops and that channel is con-

sidered as the only suitable option.

ii. If more than one channel remains, continue to

step 2.

2. Prioritization of the channels (using the Analyti-

cal Hierarchy Process). For each use case:

(a) Perform a series of pairwise comparisons, by

comparing all factors to one another, using the

Saaty method.

(b) Compute the priority scores.

scores.

3.2 Healthcare Use Cases

This section presents use cases in healthcare (see Fig-

ure 1) that can beneﬁt from the application of the

Communication Channel Selection Process. These

use case deﬁnitions are based on the Related Work

presented in Section 2.4.

3.2.1 UC1: Medical Diagnosis

Use Case Deﬁnition. Diagnosis is a patient-centered

activity that involves information gathering and clini-

cal reasoning to determine a patient’s health problem

(Ferreira and Vasconcelos, 2019). This use case is

performed by patients who have a health issue and are

looking for a diagnosis. In order to understand and di-

agnose the patient, health professionals may resort to

several approaches: (1) perform clinical history and

interview, (2) conduct physical exams, (3) perform

diagnostic testing, and (4) consulting with other clini-

cians (Ferreira and Vasconcelos, 2019). Additionally,

this use case includes ”Book Medical Appointment”

(see Figure 1). It is typically performed in-person or

through a video-conference.

3.2.2 UC2: Medical Follow-up

Use Case Deﬁnition. Follow-up is a patient-centered

activity dedicated to patients who have already been

diagnosed and are currently being monitored. Patients

may schedule an appointment beforehand (”Book

Medical Appointment”) or they may contact the

health professional directly. Furthermore, this use

case is typically performed in-person or through a

video-conference, however, instant messaging appli-

cations, SMS and e-mail have also been used in the

past.

3.2.3 UC3: Book Medical Appointment

Use Case Deﬁnition. To schedule a medical appoint-

ment, patients have to specify multiple parameters

(include ”Gather Appointment Information”, see Fig-

ure 1): the desired specialty, the desired and available

doctor, the desired and available time slot and whether

they desire a video-consultation (see Figures 2 and 3).

Some days prior to the consultation, patients are no-

tiﬁed of the upcoming appointment (include ”Notify

Patient”) and, if it is a video-conference, how to ac-

cess the consultation.

This use case is typically achieved with clerks at

the clinic (in-person or through phone calls) or using

clinic’s systems (such as portals). As for the notiﬁca-

tions, phone calls, SMS or e-mail are usually used.

4 CHANNEL SELECTION

APPROACH IN THE

HEALTHCARE CONTEXT

The selection process is applied to the healthcare

use cases introduced previously: medical diagnosis

(UC1), medical follow-up (UC2) and appointment

scheduling (UC3).

The ﬁrst step is to ﬁlter out the channels that

do not comply with the minimum requirements for

each use case. UC1 and UC2 are more complex

Systematic Selection and Prioritization of Communication Channels in the Healthcare Sector

673

Figure 2: Scheduling an appointment business process (based on (Ferreira and Vasconcelos, 2019)).

Figure 3: Gather Appointment Information subprocess (based on (Ferreira and Vasconcelos, 2019)).

and patient-centered when compared to the other use

cases. In fact, UC1 ideally requires immediate feed-

back, personal focus, body and natural language and

at least, limited visual and audio cues. This is impor-

tant because during the phase of diagnosis it is im-

portant for patients to understand the information be-

ing conveyed to them as unequivocally as possible.

Therefore, only face-to-face and videoconferencing

are suitable for UC1. UC2 is more ﬂexible, however,

it still requires personal focus, may require body and

natural language, and at least, limited visual and au-

dio cues. As a result, face-to-face, video-conference,

phone call and instant messaging apps are suitable for

UC2. UC3 can, in theory, be performed using any of

the channels, so it requires no restrictions.

The second step is prioritizing the remaining

channels. One can start by comparing all factors in

a pairwise manner, using the Saaty method. The

computed weights, resulting from this method, can

be observed in Table 3. Next, the channels need

to be scored from 1-5 across the 6 considered fac-

tors. Based on Tables 1 and 2, it is possible to grant

with them scores shown in Table 4. Finally, the

weighted scores for each communication channel are

computed, as presented in Table 5.

From the obtained results, we can verify that UC1

and UC2 achieve similar results in terms of prioriti-

zation. Therefore, if possible, patients should opt to

have diagnosis and follow-up consultations in person

before opting for video-conferences. Additionally, for

follow-up, patients can also resort to phone calls or

instant messaging platforms.

As for UC3, all channels can, in theory, be

used. However, instant messaging applications and

web/mobile applications with instant messaging rank

higher than the other ones. This result can be at-

tributed to their higher scores in language variety, ac-

cessibility and cost, which are the most important fac-

tors when performing this use case. Therefore, when

implementing UC3 in a multi-channel architecture,

although all channels are considered suitable, these

should be prioritized.

5 CONCLUSIONS

Multi-channel and omni-channel strategies are trends

in many industries toward the use of multi-channel

approaches to bring users and organizations closer.

The services provided through these channels can

be leveraged by using chatbots, allowing the user

for a simpler and more natural experience. How-

ever, this entails designing architectures that make

services available through multiple channels, while

making the users’ experiences coherent as they switch

between them. This also means that organizations

should be ready to select suitable channels and pri-

oritize them.

As presented in the Related Work of this article,

there are a couple of works that aim to address these

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

674

Table 3: Weights of each factor per Use Case.

Feedback Multiple cues Personal focus Language variety Accessibility Cost saving

UC1 0.234 0.234 0.234 0.234 0.042 0.021

UC2 0.228 0.228 0.228 0.228 0.062 0.024

UC3 0.061 0.025 0.028 0.289 0.298 0.298

Table 4: Classiﬁcation of communication channels, according to the factors that inﬂuence suitability.

Channel Feedback

Multiple

Cues

Personal

Focus

Language

Variety

Accessibility

Cost

Saving

Face-to-face 5 5 5 5 1 1

Videoconference 4 4 4 4 2 5

Phone call 4 3 3 3 5 2

OIS 2 4 4 4 4 5

Web app (without IM) 3 2 1 1 3 5

Web app (with IM) 2 4 1 4 3 5

Mobile app (without IM) 3 2 1 1 4 5

Mobile app (with IM) 2 4 1 4 4 5

E-mail 1 2 2 3 4 5

SMS 2 1 2 3 5 3

Table 5: Channel prioritization scores.

Face-

to-

Face

Video

Confe-

rence

Phone

call

Instant

messaging

platform

Web

portal

without

Web

portal

with

Mobile

app

without

Mobile

app

with

E-mail SMS

UC1 4.743 3.933 - - - - - - - -

UC2 4.646 3.892 3.322 3.356 - - - - - -

UC3 2.611 3.698 3.356 4.172 2.934 3.79 3.232 4.088 3.716 3.454

issues. The one that is the closest to reaching this

goal (Androutsopoulou et al., 2019) aims to expand

the channels through which Greek citizens interact

with governmental services. Even though the authors

present Media Richness Theory and Channel Expan-

sion Theory as a basis for their work, it is not clear

how they are able to select and prioritize channels.

However, both these theories present ways to classify

the richness of communication channels, resorting to

objective and subjective factors.

5.1 Major Contributions

In Section 2.2 we have explored the Chatbot Use Case

Evaluator, which allows the evaluation of a set of use

cases and enables the identiﬁcation of the more suit-

able ones for a bot implementation. This process, al-

lied with the Channel Selection Process, could allow

organizations to identify not only whether a use case

should be implemented in a bot, but also prioritize

channels through which a use case should be avail-

able.

To address these needs, this work proposes a sys-

tematic approach to select and prioritize communica-

tion channels based on six factors: feedback, multiple

cues, personal focus, language variety, accessibility

and cost.

Finally, the Channel Selection Process was ap-

plied to three use cases in the healthcare domain (di-

agnosis, follow-up and appointment booking). From

the obtained results, patients should opt to have di-

agnosis and follow-up consultations in-person be-

fore opting for video-conferences. Additionally, for

follow-up, patients can also resort to phone calls or

instant messaging platforms. As for the appointment

scheduling use case, all channels were concluded to

be suitable, however, instant messaging applications

and web/mobile applications with instant messaging

should be prioritized.

5.2 Limitations of the Current Study

The ﬁrst limitation of this study is that, although there

are many works addressing the classiﬁcation of com-

munication channels according to their richness, there

is a lack of research regarding processes for channel

Systematic Selection and Prioritization of Communication Channels in the Healthcare Sector

675

selection and prioritization. Moreover, there is also a

lack of experimental results. Therefore, it is difﬁcult

to evaluate the results of this process when compared

with other selection methodologies.

5.3 Future Work

Future work can be developed in order to further

improve and validate the proposed channel selection

process. In terms of evaluation, another use cases

should be assessed. Furthermore, another communi-

cation channel factors can be explored in the health-

care domain (for instance, regarding the privacy and

conﬁdentiality of the mediums).

Additionally, an architecture that is able to in-

tegrate multiple channels, while providing a seam-

less experience to users can be proposed and imple-

mented. This architecture could not only provide co-

herent information across channels, but also allow

users to ﬁnish uncompleted actions across them. Be-

sides, this implementation would allow the evaluation

of the selection process. User tests could then be con-

ducted, measuring the success and efﬁciency of task

completion across different channels and allowing a

performance comparison between them.

ACKNOWLEDGEMENTS

This work was supported by national funds through

Fundac¸

ao para a Ci

encia e a Tecnologia (FCT) with

reference UIDB/50021/2020 and by the European

Commission program H2020 under the grant agree-

ment 822404 (project QualiChain).

REFERENCES

Accenture (2014). Capturing the potential of omni-channel

commerce (white paper). https://www.accenture.

com/ acnmedia/accenture/conversion-assets/

dotcom/documents/global/pdf/technology 1/

accenture-capturing-potential-omni-channel-commerce.

pdf. Accessed: 2020-12-16.

Androutsopoulou, A., Karacapilidis, N., Loukis, E., and

Charalabidis, Y. (2019). Transforming the commu-

nication between citizens and government through ai-

guided chatbots. Gov. Inf. Q., 36:358–367.

binti Mistar, I. (2016). Students ’ perception on the use of

whatsapp as a learning tool in esl.

Carlson, J. and Zmud, R. (1999). Channel expansion theory

and the experiential nature of media richness percep-

tions. Academy of Management Journal, 42:153–170.

Caroll, D. and Guzm

an, I. (2015). The new

omni-channel approach to serving customers

(white paper). https://www.accenture.com/id-

en/∼/media/Accenture/Conversion-Assets/DotCom/

Documents/Global/PDF/Industries 2/accenture-new

-omni-channel-approach-serving-customers.pdf.

Accessed: 2020-12-16.

Daft, R. and Lengel, R. (1983). Information richness. a new

approach to managerial behavior and organization de-

sign. Research in Organizational Behavior.

Deloitte (2015). Omni-channel: Ex-

plore new heights (white paper).

https://www2.deloitte.com/content/dam/Deloitte/us/

Documents/life-sciences-health-care/us-lshc-

consumer-transformation-pentalogy-omni-

channel.pdf. Accessed: 2020-12-16.

Ferreira, D. (2020). The role of telehealth in the covid-19

era: Great ills require great remedies. Revista da So-

ciedade Portuguesa de Medicina Interna, Publicac¸

Especial COVID-19.

Ferreira, P. and Vasconcelos, A. (2019). Evaluating use

cases suitability for conversational user interfaces.

Proceedings of the 21st International Conference on

Enterprise Information Systems (ICEIS 2019), pages

431–437.

Germain, V., Marchand, A., Bouchard, S., Guay, S., and

Drouin, M.-S. (2010). Assessment of the therapeutic

alliance in face-to-face or videoconference treatment

for posttraumatic stress disorder. Cyberpsychology,

behavior and social networking, 13 1:29–35.

Ishii, K., Lyons, M., and Carr, S. A. (2019). Revisiting

media richness theory for today and future.

Jurafsky, D. and Martin, J. H. (2000). Speech and language

processing.

Klopfenstein, L. C., Delpriori, S., Malatini, S., and Bogli-

olo, A. (2017). The rise of bots: A survey of conver-

sational interfaces, patterns, and paradigms. Proceed-

ings of the 2017 Conference on Designing Interactive

Systems.

Komulainen, H. and Makkonen, H. (2018). Customer expe-

rience in omni-channel banking services. Journal of

Financial Services Marketing, 23:190–199.

Kosteln

ık, P., Pisa

rovic, I., Muro

n, M., Darena, F., and

Prochazka, D. (2019). Chatbots for enterprises: Out-

look. Acta Universitatis Agriculturae et Silviculturae

Mendelianae Brunensis, 67:1541–1550.

Laranjo, L., Dunn, A., Tong, H. L., Kocaballi, A., Chen,

J. A., Bashir, R., Surian, D., Gallego, B., Magrabi, F.,

Lau, A., and Coiera, E. (2018). Conversational agents

in healthcare: a systematic review. Journal of the

American Medical Informatics Association : JAMIA,

25:1248 – 1258.

Morris, C., Scott, R., and Mars, M. (2018). Instant mes-

saging in dermatology: A literature review. Studies in

health technology and informatics, 254:70–76.

ICEIS 2021 - 23rd International Conference on Enterprise Information Systems

676