Presentation
Designing Exposure Notification Applications as Persuasive Technologies to Improve Uptake and Effectiveness
Event Type
Oral Presentations
TimeWednesday, April 143:30pm - 3:50pm EDT
LocationDigital Health
DescriptionINTRODUCTION
The COVID-19 pandemic has resulted in the deployment of several contact tracing apps (CTAs) in many countries worldwide to combat the spread of the coronavirus. However, most of the current CTAs are only aimed at: (1) notifying users who may have been exposed to the virus—thus the name “Exposure Notification App” (ENA); and (2) providing information on what to do next. The uptake of this ENAs has not been very impressive, thereby delaying the lifting of public health restrictions and the opening of the global economy. We argue that CTAs can be designed as persuasive technologies (PTs) to make them more visually appealing, motivating and effective in curbing the spread of the virus. In this paper, using the Government of Canada’s “COVID Alert” app as a case study, we: (1) propose PT design guidelines for ENAs; (2) implement the PT design guidelines in the COVID Alert app; and (3) describe a planned study aimed at evaluating the effectiveness of the persuasive design.
OBJECTIVES
Our paper sets out to achieve three main objectives. The first objective is to show how persuasive design can be useful to human factor design of public healthcare systems. The second objective is to propose PT design guidelines to improve ENA user interface (UI) design, which are currently experiencing low adoption rate. The third objective is to describe our future study aimed at evaluating the effectiveness of the persuasive design of ENAs using COVID Alert as a case study.
OVERVIEW OF COVID ALERT
The COVID Alert app is the official ENA endorsed by the Canadian government. It leverages the Google/Apple Exposure Notification application programming interfaces and uses strong privacy measures to protect the user data it collects. It does not track the user’s location or collect personally identifiable information such as name, contacts, address or health information. The COVID Alert app, just as most ENAs, comprises three main UIs: “No Exposure Status,” “Exposure Status,” and “Diagnosis Report.” These UIs are minimalist and contain no motivational affordances. They are briefly explained as follows:
No Exposure Status UI: This interface lets the user know that they have not been exposed to COVID-19 by being in close contact with someone having COVID-19.
Exposure Status UI: This interface notifies the user that they may have been exposed to COVID-19 and provide information on what to do next.
Diagnosis Report UI: This interface allows the user to enter their one-time key given to them by the public health authority after testing positive.
HUMAN FACTORS DESIGN
People use, maintain, abandon or dispose of products. To ensure enduring usage, the creation of products should take into consideration the human factors to improve the user experience. In order to fully benefit from the functional capability of a product, users have to interact with its interface and enjoy using it. Hence, a user-centered design is recommended in the design of human-computer-interaction systems to understand the needs and requirements of the users, which are both functional and motivational. However, in the design of the current ENAs on the market, little attention has been paid to the motivational needs of the users, partly due to the urgency to roll them out to curb the spread of the virus. Particularly, the application of persuasive design to human factor design has been overlooked. This may be partly responsible for the low uptake of the current ENAs on the market as users’ motivational needs are not taken into consideration in their design.
PERSUASIVE DESIGN OF ENAs
Due to the current poor uptake, we propose a number of ways by which ENAs can incorporate persuasive features. The features, which include Self-Monitoring, Praise, Feedback, Social Comparison, Social Learning and Normative Influence, are drawn from the persuasive system design framework proposed by Oinas-Kukkonen and Harjumaa. The following shows the PT design guidelines for the three key UIs of ENAs:
No Exposure Status UI: Apart from the status information, “You have not been near anyone who reported a COVID-19 diagnosis through this app,” which the current No Exposure Status UI contains, the UI can be made persuasive through the following PT design guidelines:
- Self-Monitoring: Track daily contacts, exposure time, and showcase historical behavior.
- Social Comparison: Compare the user’s daily exposure levels with others’ in the community.
- Feedback: Notify user when they reach or cross a certain (preset) exposure level.
Exposure Status UI: Apart from the status information, “Someone you’ve been near has reported a COVID-19 diagnosis through the app. You were close them for 15 minutes or more,” which the Exposure Status UI contains, the UI can be made persuasive through the following PT design guidelines:
- Self-Monitoring: Track total contacts and exposure time for the last 14 days within which the user must have been exposed.
- Social Comparison: Compare the user’s total exposure levels in the last 14 days with others’ in the community.
Diagnosis Report UI: Apart from the key-entry information, “Enter the key you got when you were diagnosed. To prevent false notifications, you can only get a key if you test positive for COVID-19,” which the current Diagnosis Report UI contains, the UI can be made persuasive through the following PT design guidelines:
- Social Learning: Raise awareness about the number of people that have reported their COVID-19 diagnosis status.
- Normative Influence: Share how other people in the community have reduced their exposure.
- Praise: Praise or thank user for reporting their diagnosis, even before uploading their one-time key.
Moreover, as a proof of concept, we implemented the proposed PT design guidelines for each of the three key UIs of the COVID Alert app. However, in this summary, we could not show the images of the designs.
METHOD
We are currently running our planned study on the six UIs (three intervention and three control versions) among Canadian residents on Amazon Mechanical Turk: a crowd-sourcing commercial platform for recruiting non-convenience samples. The research design for each interface has two levels of persuasive design (intervention and control) and two levels of adoption status (adopters and non-adopters). We aim to use a 2X2 factorial design to evaluate the effectiveness of the persuasive design by comparing the perceived persuasiveness of the intervention versions and their effect on installing the app, self-isolating and reporting COVID-19 diagnosis intentions with those of the control (original) versions. Through this research design, we will be able to evaluate the main effect of persuasive design and adoption status (and their interaction) on the target behaviors: adoption, self-isolation and diagnosis reporting. We hope to obtain the preliminary results of the study by March and present them at the symposium in April 2021.
EXPECTED RESULTS
The expected results from the study will validate or invalidate the following hypotheses:
H1: The perceived persuasiveness of the intervention version of the No-Exposure-Status UI will be higher than that of the control version.
H2: The perceived persuasiveness of the intervention version of the Exposure-Status UI will be higher than that of the control version.
H3: The perceived persuasiveness of the intervention version of the Diagnosis-Report UI will be higher than that of the control version.
H4: The intervention group will be more likely to install the app using the No-Exposure-Status UI than the control group.
H5: The intervention group will be more likely to self-isolate using the Exposure-Status UI than the control group.
H6: The intervention group will be more likely to report their COVID-19 diagnosis using the Diagnosis-Report UI than the control group.
H7: Adopters will perceive the three UIs as more persuasive than the non-adopters.
The COVID-19 pandemic has resulted in the deployment of several contact tracing apps (CTAs) in many countries worldwide to combat the spread of the coronavirus. However, most of the current CTAs are only aimed at: (1) notifying users who may have been exposed to the virus—thus the name “Exposure Notification App” (ENA); and (2) providing information on what to do next. The uptake of this ENAs has not been very impressive, thereby delaying the lifting of public health restrictions and the opening of the global economy. We argue that CTAs can be designed as persuasive technologies (PTs) to make them more visually appealing, motivating and effective in curbing the spread of the virus. In this paper, using the Government of Canada’s “COVID Alert” app as a case study, we: (1) propose PT design guidelines for ENAs; (2) implement the PT design guidelines in the COVID Alert app; and (3) describe a planned study aimed at evaluating the effectiveness of the persuasive design.
OBJECTIVES
Our paper sets out to achieve three main objectives. The first objective is to show how persuasive design can be useful to human factor design of public healthcare systems. The second objective is to propose PT design guidelines to improve ENA user interface (UI) design, which are currently experiencing low adoption rate. The third objective is to describe our future study aimed at evaluating the effectiveness of the persuasive design of ENAs using COVID Alert as a case study.
OVERVIEW OF COVID ALERT
The COVID Alert app is the official ENA endorsed by the Canadian government. It leverages the Google/Apple Exposure Notification application programming interfaces and uses strong privacy measures to protect the user data it collects. It does not track the user’s location or collect personally identifiable information such as name, contacts, address or health information. The COVID Alert app, just as most ENAs, comprises three main UIs: “No Exposure Status,” “Exposure Status,” and “Diagnosis Report.” These UIs are minimalist and contain no motivational affordances. They are briefly explained as follows:
No Exposure Status UI: This interface lets the user know that they have not been exposed to COVID-19 by being in close contact with someone having COVID-19.
Exposure Status UI: This interface notifies the user that they may have been exposed to COVID-19 and provide information on what to do next.
Diagnosis Report UI: This interface allows the user to enter their one-time key given to them by the public health authority after testing positive.
HUMAN FACTORS DESIGN
People use, maintain, abandon or dispose of products. To ensure enduring usage, the creation of products should take into consideration the human factors to improve the user experience. In order to fully benefit from the functional capability of a product, users have to interact with its interface and enjoy using it. Hence, a user-centered design is recommended in the design of human-computer-interaction systems to understand the needs and requirements of the users, which are both functional and motivational. However, in the design of the current ENAs on the market, little attention has been paid to the motivational needs of the users, partly due to the urgency to roll them out to curb the spread of the virus. Particularly, the application of persuasive design to human factor design has been overlooked. This may be partly responsible for the low uptake of the current ENAs on the market as users’ motivational needs are not taken into consideration in their design.
PERSUASIVE DESIGN OF ENAs
Due to the current poor uptake, we propose a number of ways by which ENAs can incorporate persuasive features. The features, which include Self-Monitoring, Praise, Feedback, Social Comparison, Social Learning and Normative Influence, are drawn from the persuasive system design framework proposed by Oinas-Kukkonen and Harjumaa. The following shows the PT design guidelines for the three key UIs of ENAs:
No Exposure Status UI: Apart from the status information, “You have not been near anyone who reported a COVID-19 diagnosis through this app,” which the current No Exposure Status UI contains, the UI can be made persuasive through the following PT design guidelines:
- Self-Monitoring: Track daily contacts, exposure time, and showcase historical behavior.
- Social Comparison: Compare the user’s daily exposure levels with others’ in the community.
- Feedback: Notify user when they reach or cross a certain (preset) exposure level.
Exposure Status UI: Apart from the status information, “Someone you’ve been near has reported a COVID-19 diagnosis through the app. You were close them for 15 minutes or more,” which the Exposure Status UI contains, the UI can be made persuasive through the following PT design guidelines:
- Self-Monitoring: Track total contacts and exposure time for the last 14 days within which the user must have been exposed.
- Social Comparison: Compare the user’s total exposure levels in the last 14 days with others’ in the community.
Diagnosis Report UI: Apart from the key-entry information, “Enter the key you got when you were diagnosed. To prevent false notifications, you can only get a key if you test positive for COVID-19,” which the current Diagnosis Report UI contains, the UI can be made persuasive through the following PT design guidelines:
- Social Learning: Raise awareness about the number of people that have reported their COVID-19 diagnosis status.
- Normative Influence: Share how other people in the community have reduced their exposure.
- Praise: Praise or thank user for reporting their diagnosis, even before uploading their one-time key.
Moreover, as a proof of concept, we implemented the proposed PT design guidelines for each of the three key UIs of the COVID Alert app. However, in this summary, we could not show the images of the designs.
METHOD
We are currently running our planned study on the six UIs (three intervention and three control versions) among Canadian residents on Amazon Mechanical Turk: a crowd-sourcing commercial platform for recruiting non-convenience samples. The research design for each interface has two levels of persuasive design (intervention and control) and two levels of adoption status (adopters and non-adopters). We aim to use a 2X2 factorial design to evaluate the effectiveness of the persuasive design by comparing the perceived persuasiveness of the intervention versions and their effect on installing the app, self-isolating and reporting COVID-19 diagnosis intentions with those of the control (original) versions. Through this research design, we will be able to evaluate the main effect of persuasive design and adoption status (and their interaction) on the target behaviors: adoption, self-isolation and diagnosis reporting. We hope to obtain the preliminary results of the study by March and present them at the symposium in April 2021.
EXPECTED RESULTS
The expected results from the study will validate or invalidate the following hypotheses:
H1: The perceived persuasiveness of the intervention version of the No-Exposure-Status UI will be higher than that of the control version.
H2: The perceived persuasiveness of the intervention version of the Exposure-Status UI will be higher than that of the control version.
H3: The perceived persuasiveness of the intervention version of the Diagnosis-Report UI will be higher than that of the control version.
H4: The intervention group will be more likely to install the app using the No-Exposure-Status UI than the control group.
H5: The intervention group will be more likely to self-isolate using the Exposure-Status UI than the control group.
H6: The intervention group will be more likely to report their COVID-19 diagnosis using the Diagnosis-Report UI than the control group.
H7: Adopters will perceive the three UIs as more persuasive than the non-adopters.
Authors
