Does Mode of Digital Contact Tracing Affect User Willingness to Share Information A Quantitative Study
2025-05-03
0
0
2.99MB
18 页
10玖币
侵权投诉
Does Mode of Digital Contact Tracing Aect User Willingness to
Share Information? A antitative Study
Camellia Zakaria
College of Information & Computer
Sciences, University of Massachusetts
Amherst
USA
Pin Sym Foong
Saw Swee Hock School of Public
Health, National University of
Singapore
Singapore
Chang Siang Lim
Future Health Technologies,
Singapore ETH Center
Singapore
Pavithren V. S. Pakianathan∗
Saw Swee Hock School of Public
Health, National University of
Singapore
Singapore
Gerald Koh Choon Huat
Saw Swee Hock School of Public
Health, National University of
Singapore
Singapore
Simon Tangi Perrault
Singapore University of Technology
and Design
Singapore
ABSTRACT
Digital contact tracing can limit the spread of infectious diseases.
Nevertheless, barriers remain to attain sucient adoption. In this
study, we investigate how willingness to participate in contact trac-
ing is aected by two critical factors: the modes of data collection
and the type of data collected. We conducted a scenario-based sur-
vey study among 220 respondents in the United States (U.S.) to
understand their perceptions about contact tracing associated with
automated and manual contact tracing methods. The ndings indi-
cate a promising use of smartphones and a combination of public
health ocials and medical health records as information sources.
Through a quantitative analysis, we describe how dierent modali-
ties and individual demographic factors may aect user compliance
when participants are asked to provide four key information pieces
for contact tracing.
CCS CONCEPTS
•Social and professional topics →User characteristics
;
•Human-
centered computing →Human computer interaction (HCI).
KEYWORDS
contact tracing, pandemic, willingness, trust, public health
ACM Reference Format:
Camellia Zakaria, Pin Sym Foong, Chang Siang Lim, Pavithren V. S. Pakianathan,
Gerald Koh Choon Huat, and Simon Tangi Perrault. 2022. Does Mode of
Digital Contact Tracing Aect User Willingness to Share Information? A
Quantitative Study. In CHI Conference on Human Factors in Computing Sys-
tems (CHI ’22), April 29-May 5, 2022, New Orleans, LA, USA. ACM, New York,
NY, USA, 18 pages. https://doi.org/10.1145/3491102.3517595
∗Also with Singapore University of Technology and Design.
Permission to make digital or hard copies of all or part of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for prot or commercial advantage and that copies bear this notice and the full citation
on the rst page. Copyrights for components of this work owned by others than the
author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior specic permission
and/or a fee. Request permissions from permissions@acm.org.
CHI ’22, April 29-May 5, 2022, New Orleans, LA, USA
©2022 Copyright held by the owner/author(s). Publication rights licensed to ACM.
ACM ISBN 978-1-4503-9157-3/22/04. . . $15.00
https://doi.org/10.1145/3491102.3517595
1 INTRODUCTION
Contact tracing is a public health strategy that plays a crucial role in
curbing the spread of communicable diseases [
10
]. Manual contact
tracing involves a series of steps, starting with contacting infected
persons and then interviewing the patient to gather a log of loca-
tions and persons with personal contact within a specied period.
These close contacts are identied and notied regarding potential
exposure and informed of further containment measures such as
testing or isolation. However, an incomplete or incorrect recall of
events, locations, and contact persons in the period of interest can
deter this process [
14
]. Often, contact tracing is resource-intensive
and time consuming [23].
In the U.S., the COVID-19 pandemic has particularly challenged
the limits of manual contact tracing. The novel clinical features
of COVID-19 infection, including long incubation period, asymp-
tomatic transmission, and high transmission rate [
65
] led to a bur-
geoning number of patients that easily outpaced manual contact
tracing eorts. These issues are exacerbated by less than 50% of
people being willing to participate in contact tracing eorts [32].
A modeling study by Feretti et al. [
14
] shows that a hypothetical,
fully automated digital contact tracing solution can slow or stop the
transmission of COVID-19. However, the extent of data gathering
that digital contact tracing requires leads to concerns with privacy,
user surveillance, and data leaks [
25
,
45
,
50
], worsened by the lack
of trust in government bodies and technology companies to handle
sensitive information required in digital contact tracing [
10
,
57
].
These concerns hamper the adoption of digital contact tracing and
place more lives at risk. Unfortunately, a recent scoping review
concludes that “there is a dearth of evidence regarding the barriers
and facilitators to uptake and engagement with COVID-19 digital
contact tracing applications” [
61
]. Thus it is essential to determine
the factors that increase acceptance of digital contact tracing.
Existing studies prospectively probed the desired feature set to
inuence users’ willingness to participate in digital contact tracing.
Researchers have begun to identify a few key facilitators. These are:
higher perceived public health benet, perceived individual benet,
and lower degree of privacy risk [
22
,
34
]. Missing in this body of
knowledge about the users’ perception of digital contact tracing
features is whether the stated mode of digital contact tracing may
arXiv:2210.13399v1 [cs.HC] 24 Oct 2022
CHI ’22, April 29-May 5, 2022, New Orleans, LA, USA Zakaria, et al.
aect willingness to participate. The perception that a particular
mode of delivery is more or less threatening to personal privacy
may derail digital contact tracing eorts even before actual usage.
While the particular features of an application aect a person’s
willingness to continue using an application, resistance to the mode
of delivery can be a barrier to the rst usage. To our knowledge,
this question of the users’ lens of digital contact tracing modes has
not yet been studied.
In this paper, we sought to
understand how the mode of dig-
ital contact tracing aects users’ willingness to share dier-
ent types of data
. We conducted a scenario-based survey of U.S.
respondents to understand users’ relative willingness to share pri-
vate information to help contact tracing and examine the modality
of data sharing that would be most acceptable by users. We studied
three types of information best collected by digital devices to sup-
port contact tracing. They are (1) the subject’s identity (name and
social security number), (2) the subject’s contact details (email and
contact number), and (3) details of exposure (location or person
of interest). We also oered four types of modalities for partici-
pants to indicate their willingness to share this information. They
are (1) in-person communication with public health ocials, (2)
providing access to their existing health records, (3) sharing their
information collected in their smartphone device (4) providing ac-
cess to analyze their internet activity. We also examined how such
willingness was inuenced by demographic variables such as age,
parenthood, income, and trust in public health ocials. As the
study was conducted in the U.S. between November and December
2020, we wanted to ensure that users’ responses were not solely
based on their experience of the COVID-19 pandemic. Hence, we
randomly assigned participants to a dierent potential epidemic
scenario involving one of six infectious agents.
Through a quantitative analysis, we present an overview of users’
willingness to share dierent types of private information to sup-
port contact tracing and factors that inuence such willingness,
supplemented with qualitative responses to help unpack our ob-
servations. We found no evidence of the eects of disease types
on users’ willingness to share their private information for digital
contact tracing. Notably, our ndings show that participants are
most willing to share their private information via smartphones
(
𝑝<.
01) and grant public health ocials access to individuals’
health records (
𝑝<.
01). However, sharing location information is
signicantly impacted by the level of trust in public health ocials
(
𝑝<.
01). These results add new dimensions to the well-established
importance of trust in public health ocials in the data collection
pipeline for digital contact tracing.
Our work contributes 1) an empirical study of user willingness
to adopt contact tracing strategies across various modes of collec-
tion, information types, and disease types, and 2) based on these
ndings, recommendations toward developing a national digital
contact tracing strategy in the U.S.
2 RELATED WORK
Our focus here was to summarize existing literature on barriers
and facilitators to digital contact tracing. Specically, we highlight
prior work that examined barriers and facilitators of digital contact
tracing and the importance of data collection modality.
A recent survey of Americans indicates that just 42% are willing
to download and use a contact-tracing app [
71
]. For automated and
partly automated contact tracing to be eective, at least 56% uptake
is needed [
6
] within a population. What would the barriers and
facilitators of digital contact tracing be for sucient mass adoption
in the U.S.?
2.1 Background on Contact Tracing
Contact tracing is an essential strategy in public health manage-
ment, proving itself eective in limiting the spread of infectious
disease [
58
]. This process, however, must be performed expedi-
tiously. As dened by CDC, contact tracing involves trained case
investigators by the public health organization to evaluate a pa-
tient’s close contacts (i.e., person-of-interest, POI) [
12
]. A POI may
expect a case investigator to seek information on their health symp-
toms and possible exposures (i.e., places, partners) to determine
compromised entities during the interview. POI may also receive
instructions for isolation and follow-up sessions on their health.
Table 1 summarises the types of critical, but sensitive information
that serves as a general guideline by CDC in conducting such in-
terviews [
9
]. This laborious process can be slow [
24
] and often
relies on the POI’s willingness and ability to recall the relevant
information [
43
]. The race against time is critically challenged by
the shortage of human resources during a health crisis. Hence, more
recently, digital tools were explored to play a role in enhancing
contact tracing by either automating or semi-automating some of
these data collection tasks.
Table 1: A list of open-ended questions for contact tracing
by the CDC, and grouped according to information type.
Questions Information Type
1 What is their name?
Identity
2 What name do they go by?
3 What is their gender?
4 What is their race/ethnicity
5 What is their primary language?
6
What is the best way to reach them?
Contact
7 What is their cell number, email,
APP and username?
8 Where do they live?
Exposure Location
9 Who do they live with?
10 Where do they work?
11 Where is that located?
12 When did you see them last? Exposure Location
/ Intimate Partner
13 For how long have you spent
time with them?
14 What symptoms might have had?
Health Status15 What underlying medical
conditions might they have?
16 What do they know about your
infectious status?
Based on this established process, our study examines how an-
swers to these questions can be collected through dierent data
collection modalities and not restricted to human contact tracers
alone [9].
Does Mode of Digital Contact Tracing Aect User Willingness to Share Information? CHI ’22, April 29-May 5, 2022, New Orleans, LA, USA
2.2 Barriers and Facilitators of Digital Contact
Tracing
Various barriers and facilitators to digital contact tracing have been
identied. A discrete choice experiment prospective study in the
Netherlands suggested less modiable individual factors specic
to COVID-19 [
20
]. In this study, the factors that correlated with
predicted adoption rates were educational attainment, underlying
health conditions, and a perceived threat from COVID-19.
More promising, modiable facilitators can be gleaned with a
broader literature review that includes other diseases. Megnin et al.
[
34
] conducted a rapid review of qualitative studies on the factors
inuencing user uptake and engagement with any contact tracing
system across various infectious diseases. They identied four modi-
able factors that arise from the users’ perception of the application.
They are a perceived sense of collective responsibility, perceived per-
sonal benet, the presence of community co-production of contact
tracing systems, and the perceived capability of reaching contact
persons eciently and eectively. The authors also identied pri-
vacy concerns as a key barrier. These concerns were: mistrust with
the requester, unmet needs for information and support, fear of
stigmatization (due to an identied infection), and what they called
“mode-specic challenges”. These mode-specic challenges were
not communication channel issues per se. They were more akin to
accessibility issues (e.g., no smartphone) or usability issues (e.g.,
diculty downloading, using the application, or lack of technical
prociency).
These studies suggest two broader research areas that can be
brought to bear upon our knowledge about facilitators and barriers
of digital contact tracing. The rst area covers the perception of
trust in public bodies running the digital contact tracing operations.
The second area covers studies on multiple alternative channels of
user participation and data collection modes.
2.3 Issues of Trust and Data Collection
One commonly cited factor that predicts the adoption of contact
tracing is the trust in the entity conducting digital contact tracing
(i.e., institutional trust). Mayer’s body of work on institutional trust
denes it as the following: ‘the willingness of a party to be vul-
nerable to the actions of another party’ [
31
]. In a pandemic, the
willingness of users to be vulnerable must be particularly directed
towards public health ocials, who are providing practical guide-
lines and general awareness of the disease outbreak. Instead, trust
is often referred to as being vulnerable to government bodies and
private rms handling the sensitive information required in digital
contact tracing.
To overcome the mountain of skepticism towards utilizing con-
tact tracing apps, several public health organizations and researchers
[
11
,
21
,
25
,
67
] have proposed that these systems be transparent
and open to public scrutiny. Scientists and researchers across the
globe have also recommended that a privacy-by-design approach
be adopted whereby only necessary data is collected and stored
using secure encryption techniques to preserve the security and pri-
vacy of the data [
25
]. Additionally, the World Health Organization
(WHO) and American Civil Liberties Union (ACLU) collectively
agree with the CDC’s suggestion to make contact tracing voluntary,
with full user control over data management [11, 15, 67].
In response to these recommendations, Google and Apple pushed
for anonymous contact tracing [
16
]. The exposure notication
framework frequently exchanges anonymous identier beacons
through Bluetooth between and among smartphones whose users
are in close proximity; either by installing an ocial app from
their region’s government or directly through a verication proto-
col from public health authorities. NOVID [
40
], ICheckedIn [
59
]
and SaferMe [
53
] are similar eorts that followed in pursuit of
anonymity with an encrypted framework. For example, users’ names
and numbers are hashed in a "pin," maintaining relative obscurity
in businesses and places of visit [
59
]. While this information is
set to expire in limited duration (e.g., discarded after 30 days), per-
sonal information remains accessible to relevant government bodies
or pre-approved systems. As might be expected, these eorts did
not alleviate the tension associated with the lack of condence in
government and private companies [1, 10, 66, 67].
In a study specic to the U.S., only 37% of respondents found it
acceptable to share data with state and local ocials, compared to
75% of the sample who preferred sharing data with infectious dis-
eases researchers [
32
]. Yet, the National Academy for State Health
Policy found that contact tracing eorts are either led by state or
county (e.g., eight states, including California, are county-led) [
38
].
The suggestion to empower user autonomy in digital contact
tracing places the onus on the users to be suciently convinced
that the benets of digital contact tracing outweigh the concerns.
In a qualitative study on privacy concerns of a tool for online data
collection, Phelan et al. [
47
] drew on dual process theories to de-
scribe two kinds of user concern about privacy: intuitive concern
(when following a gut feeling) and considered concern (when weigh-
ing pros and cons). In considered concern, users regularly recognize
the benets of accepting a privacy intrusion, particularly when
there is trust in the requester or low assessed risk. However, in-
tuitive concern can override the more cognitive path of considered
concern. This can happen, for example, when the social presence
of the requester provides privacy assurance. Additionally, a study
on user responses to surveillance suggests that when these types
of concerns are not addressed, individuals may adopt protective
coping strategies [
54
]. These avoidance mechanisms would make
it harder to involve them in contact tracing eorts.
Overall the literature on trust indicates issues of trust seem to
be centered on perceptions of intrusiveness of the tool. Although
research in anonymous contact tracing actively pushes for compu-
tational and technological protocols that strive to guarantee user
anonymity and tool credibility, it remains unclear how willing users
are to share critical information that must fundamentally be dis-
closed to public health authorities upon identifying themselves
as infected. Thus, it led us to query if the users’ perception of
the data collection modality may also aect this intuitive concern.
Such knowledge could help pave the way to reducing resistance to
privacy intrusions of digital contact tracing for public health.
2.4 Issues of Technology Access
Much research on developing the functions of digital contact trac-
ing is centered on the types of sensing modalities available on a
smartphone device. Specically, investigations have looked at Blue-
tooth and Global Positioning System (GPS) as two standard sensing
CHI ’22, April 29-May 5, 2022, New Orleans, LA, USA Zakaria, et al.
techniques utilized by most contact tracing mobile applications.
Other investigations for collecting location-based data include us-
ing a magnetometer, QR code, phone logs, and the most recent WiFi
logs of smartphones. [63, 66, 70].
However, the primary dependence on a smartphone device for
data collection may have the disadvantage of disproportionately
disenfranchising several demographic groups. In the U.S., the Pew
Research Center reported 66% smartphone ownership among users
aged 65 and above, and 71% ownership among lower-income earn-
ers (less than $30,000) [
46
]. In response to a similar situation of
low smartphone ownership and complaints about data and power
usage in some communities, Singapore’s digital contact tracing
eorts were supplemented by a dedicated hardware-based Blue-
tooth token. This self-contained token was distributed in pendant
form to the entire population, mainly beneting people who have
reservations about using their smartphone for contact tracing or do
not own one. The rollout was complex as it additionally required
the distribution of scanning hardware [
17
] at an estimated cost of
about USD 4.6 million [
69
]. Hence dierent digital contact tracing
modalities present challenges to balance accessibility and privacy
[11, 15, 50, 66].
Given the cost and complexity of digital contact tracing initia-
tives, it is important to understand how users perceive the modality
being used. By selecting modalities that improve the likelihood of
uptake, we may increase the chances of success for a digital contact
tracing program.
2.5 Summary
Researchers have identied a series of individual user factors that
promote the uptake and engagement with a digital contact tracing
program. Additionally, the lack of trust in the entities running a
digital contact tracing program can negatively inuence the accept-
ability. Finally, the choice of mode should take into account practical
concerns of cost and utility for the user and the supplier. Given
these considerations, we propose to complement previous work
by quantitatively studying how users’ perceptions of the mode of
digital contact tracing inuences users’ willingness to share digital
contact tracing data. We regard user identication as critical in
contact tracing for enforced containment and not just exposure
alert (in anonymous tracing); thus, we explicitly consider disclosing
personal data following CDC guidelines despite privacy concerns.
We also propose to understand the relative acceptability of the dif-
ferent modalities within a spectrum of manual to automated data
collection modes.
3 DATA & METHODOLOGY
We conducted an online scenario-based survey study utilizing Ama-
zon Mechanical Turk (MTurk) to understand how previously iden-
tied individual characteristics and modes might aect users’ will-
ingness to share contact tracing information in dierent disease
outbreak scenarios. Note: MTurk is a crowdsourcing marketplace
hosted by Amazon, allowing us to recruit participants (also known
as ‘workers’) for our survey research. Prior studies in HCI research
have leveraged MTurk to provide researchers with a diverse sample
of participants from tens [
55
] to hundreds [
5
], and thousands [
64
],
based on tasks complexity. Later in Section 3.4, we provide results
from conducting a power analysis to test the probability of our
design succeeding with the number of samples acquired from our
study.
By not limiting the study to users’ responses to one specic
disease, our goal was to obtain generalizable knowledge regarding
contact tracing for infectious diseases and how they are aected by
individual characteristics and data collection modalities. Our study
was approved by the university’s Institutional Review Board (IRB)
and took place between November and December of 2020.
3.1 Choice of Scenarios
We selected scenarios from the World Health Organization’s (WHO)
list of top ten threats to global health, where these communicable
diseases account for almost one-third of deaths worldwide [
37
].
The selection of infectious agents covers a broad range of trans-
mission features and participants’ level of familiarity given the
outbreak histories in the U.S. [
37
]. Diseases were categorized by
their transmission methods (i.e., fomite/surfaces, animal vector,
air/droplets, sexual transmission). We analyzed them by infectivity
and curated a mix of novelty diseases. Based on these criteria, we
narrowed the scenario list to six diseases – Human Immunode-
ciency Virus (HIV), Novel Coronavirus (nCov), Zika Virus, Ebola
Virus, Methicillin-resistant Staphylococcus Aureus (MRSA), and
Hepatitis causing virus (Hep). Table 2 shows the feature spread for
this selection of diseases.
Table 2: Disease feature spread selected for the scenarios.
DiseaseName Trans
mission
Infect
iousness
Detrimental
Eects
Comment(s)
HIV Sexual contact Varies by health
care context
High lifetime cost Need sexual
partner data for
contact tracing
nCov Air/Droplets High Age and health
prole dependent
Current pandemic
Zika Vector
(Mosquitoes)
Moderate Specically to
pregnant mothers
and fathers-to-be
-
Ebola Direct Contact,
bodily liquids
High Highly fatal Small U.S. out
-break in 2014
MRSA Surfaces High High, Currently
no Treatment
May occur in
hospitals
Hep Sexual contact,
bodily liquids
Moderate High Novel virus but
older vaccine hints
to possible new
variant outbreak
3.2 Scenario-based Survey
We chose to identify each disease in the given scenarios. We consid-
ered white-labeling (not naming) each disease scenario to reduce
users’ existing perceptions of the particular disease. However, after
discussion, we realized that presenting the specic characteristics
of the disease would still cause participants to recall their under-
standing or experiences of that characteristic. In contrast, naming
the actual disease would improve the likelihood that the responses
are real-world perceptions of the disease and the scenarios are NOT
associated with COVID-19, which would have had high saliency
during the study period.
A participant is randomly assigned to a scenario describing one
particular type of infectious disease outbreak. As shown in Figure
1, each scenario has the following structure: First, it describes the
摘要：
展开>>
收起<<
DoesModeofDigitalContactTracingAffectUserWillingnesstoShareInformation?AQuantitativeStudyCamelliaZakariaCollegeofInformation&ComputerSciences,UniversityofMassachusettsAmherstUSAPinSymFoongSawSweeHockSchoolofPublicHealth,NationalUniversityofSingaporeSingaporeChangSiangLimFutureHealthTechnologies,Sing...
声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！
相关推荐
-
【词汇变形总汇】2025高考词汇变形总汇 - 教师版VIP免费
2024-12-06 3 -
【超简37页】新课标高考英语考纲3500词汇VIP免费
2024-12-06 4 -
《高考英语3500词详解》（WORD版）VIP免费
2024-12-06 13 -
《高考英语3500词详解》VIP免费
2024-12-06 11 -
高中英语-[教师版]80天通关高考3500词汇VIP免费
2024-12-06 12 -
高中人教选修7课文逐句翻译VIP免费
2024-12-06 7 -
高中人教选修7课文原文及翻译VIP免费
2024-12-06 13 -
高中人教必修4课文逐句翻译VIP免费
2024-12-06 7 -
高中人教必修4课文原文及翻译VIP免费
2024-12-06 13 -
高考英语核心高频688词汇VIP免费
2024-12-06 10
分类：图书资源
价格：10玖币
属性：18 页
大小：2.99MB
格式：PDF
时间：2025-05-03
作者详情
相关内容
-
2014年高考物理试卷（安徽）（解析卷）
分类：中学教育
时间：2025-05-05
标签：无
格式：DOCX
价格：10 玖币
-
2014年高考物理试卷（北京）（空白卷）
分类：中学教育
时间：2025-05-05
标签：无
格式：DOC
价格：10 玖币
-
2014年高考物理试卷（北京）（解析卷）
分类：中学教育
时间：2025-05-05
标签：无
格式：DOC
价格：10 玖币
-
2014年高考物理试卷（大纲版）（空白卷）
分类：中学教育
时间：2025-05-05
标签：无
格式：DOCX
价格：10 玖币
-
2014年高考物理试卷（大纲版）（解析卷）
分类：中学教育
时间：2025-05-05
标签：无
格式：DOC
价格：10 玖币
渝公网安备50010702506394
