Increasing long-term use of wearable sensors using targets and reminders at different frequencies

Last registered on March 09, 2021

View Trial History

Pre-Trial

Trial Information

General Information

Title

Increasing long-term use of wearable sensors using targets and reminders at different frequencies

RCT ID

AEARCTR-0007311

Initial registration date

March 08, 2021

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

March 09, 2021, 6:23 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country

United States of America

Region

California

Primary Investigator

Name

Janna Ter Meer

Affiliation

Scripps Research Translational Institute

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Matthew Thombs

PI Affiliation

Scripps Research Translational Institute

Contact Investigator

PI Name

Julia Moore Vogel

PI Affiliation

Scripps Research Translational Institute

Contact Investigator

Additional Trial Information

Status

In development

Start date

2021-03-31

End date

2022-05-31

Keywords

Behavior, Health

Additional Keywords

JEL code(s)

Secondary IDs

Abstract

A wide range of wearable sensor technologies exist that can capture a variety of physiological and behavioral measurements that may provide valuable individualized health information. Some examples include continuous glucose and blood pressure trackers, personal environmental sensors, and lung function monitors. Today, the best known and most widely used wearable sensor technologies are comprised of a wide array of wrist-worn activity, sleep, and heart rate monitors. Market research claims one-in-five Americans use a smartwatch or fitness tracker and continuous use of these devices drops to 70 percent after 6 months of purchase and 55 percent after 1 year of purchase. This attrition limits the usefulness of wearable sensors for health research purposes, which can longitudinal data from the same individual to predict illness, such as the onset of influenza and Covid-19. We run a Randomized Controlled Trial to reduce attrition of a Fitbit wearable sensor over a one year period with 10,000 people enrolled in a health research program. As part of the study, individuals are invited to receive a Fitbit free of charge that is theirs to keep. Individuals are randomized to two long-term engagement interventions: a wear target for the device (2 conditions); and a reminder frequency condition (3 conditions). For the wear target, individuals are encouraged to wear their device at least 4 consecutive days or 12 total days per month for a period of one year. For the reminder frequencies, individuals either receive no reminders, weekly reminders or can choose their preferred cadence (twice a week; once a week; twice a month or once a month) following a 3-6 week onboarding period where they receive messages once a week. Individuals in this latter group are explicitly prompted to declare a communication frequency preference after the end of the first and sixth month of the study.

Through this design, we evaluate the number of days a device is worn in a given month as well as over the duration of the one year study period. Our work contributes to literature on the use of targets on long-term behavior change. In addition, we add insights on whether giving individuals the ability to customize reminder message frequency improves participation in the study.

External Link(s)

Registration Citation

Citation

Moore Vogel, Julia, Janna Ter Meer and Matthew Thombs. 2021. "Increasing long-term use of wearable sensors using targets and reminders at different frequencies." AEA RCT Registry. March 09. https://doi.org/10.1257/rct.7311-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

We deploy three sets of interventions with the 10,000 individuals in the study.

1) Framing of the invitation email
* Neutral: "All of Us participants who join the WEAR Study receive a new Fitbit to wear at no cost. "
* Emphasizing value to the participant: This will frame the opportunity around the value to the participant. "All of Us participants who join the WEAR Study receive a new Fitbit to wear at no cost. The All of Us Research Program is only giving out Fitbit devices through this study. The Fitbit device is yours to keep."
* Emphasizing value to the AoURP: This will frame the opportunity to wear the device as a benefit to the goals of the research program. "All of Us participants who join the WEAR Study receive a new Fitbit to wear at no cost. Wearing your Fitbit will really help our research by providing data for thousands of research studies."

2) Wear target
* 4 consecutive days per month: Participants are encouraged to wear their device for a minimum of 4 consecutive days per month. This is communicated to all participants when they start using their device and during the first month of the onboarding period, during which everyone receives reminder messages once a week. Participants who are either in the once-a-week message condition or the message frequency choice condition, will continue to receive feedback about their wear days relative to the target.
* 12 total days per month: Participants are encouraged to wear their device for a minimum of 12 total days per month. This is communicated to all participants when they start using their device and during the first month of the onboarding period, during which everyone receives reminder messages once a week. Participants who are either in the once-a-week message condition or the message frequency choice condition, will continue to receive feedback about their wear days relative to the target.

3) Support message frequency delivered via text:
* Standard Fitbit messaging: Participants receive the standard communication they receive as a Fitbit device user, from Fitbit, which includes an activity report at the end of week. This serves as the control condition. This condition does not include SMS messages, as Fitbit does not send SMS messages.
* Default AoU Message Frequency: Participants receive one SMS message per week. The content of the SMS message may depend on the number of days they have worn their device in the previous 7-28 days.
* Participant Choice Message Frequency: Participants will be prompted to choose how often they want to receive messages from the research program to support them in wearing their device. The options are twice a week, once a week, twice a month or once a month. The content of the SMS message may depend on the number of days they have worn their device in the previous 7-28 days. If participants who are offered the choice to choose message frequency do not make a selection, they will receive one SMS message per week. This participant group will receive the a similar prompt once more at the 6 month mark to either confirm or modify how often they want to receive messages from the study.

Note that in all conditions participants can unsubscribe from study support messages. However, only those in the message frequency choice condition are explicitly prompted to amend their notification preferences and have more options in terms of the cadence.

Intervention Start Date

2021-03-31

Intervention End Date

2022-05-31

Primary Outcomes

Primary Outcomes (end points)

The number of months (0-12) that contain at least one wear day. This represents the length of time the participant is engaged in the research study.

Primary Outcomes (explanation)

We record a participant as having worn their device on a given day if they have logged 1 hour or more of device data, or registered a non- zero step count, between 00:00:00 and 23:59:59 on that day. A participant is considered to have worn their device in a given week if we register 1 wear day or more between Monday 00:00:00 and Sunday 23:59:59 in that week.

Secondary Outcomes

Secondary Outcomes (end points)

The average number of days that the participant wears the device in the months in which at least one wear day is logged. This captures the level of engagement for a participant that is active during a given month in the study.

The average number of hours that the participant wears the device on a given day during a given period. This is another important measure of the level of engagement for a participant when they are active during a particular period in the study.

Whether communication preferences about support messages are changed throughout the study period.

How often communication preferences about support messages are changed throughout the study period.

Secondary Outcomes (explanation)

See explanation under primary outcomes for how we construct a wear day.

Experimental Design

Individuals in this study are enrolled in a large health research program called the All of Us Research Program (AoURP). Eligible individuals receive an invitation email for the WEAR sub-study. To enroll, they need to complete a separate consent module, agree that they are willing to receive text messages as part of study and confirm that they understand that they are supposed to wear their Fitbit device for at least one year. After completing these steps, individuals set up a Fitbit account (or log in to an existing one) and provide authorization to share their Fitbit data with AoURP. They are then taken to a dedicated Fitbit store when they can order a Fitbit free of charge. At study launch the two models are on offer were the latest versions of Fitbit's smartwatch (Versa 3) and the slim wristband (Charge 4), though we may upgrade to newer models as they are released by Fitbit. The Fitbit order is processed and delivered by Fitbit and the device will automatically start sharing data with the research program the moment that it is set up by the individual (following the usual set-up instructions provided by Fitbit).

Once individuals start using their device, they will start a 3-6 week onboarding period during which they receive support messages via text once a week (Sunday). The length of the onboarding period differs depending on when the individual starts using their device. We want to ensure that each participant receives a minimum of 3 support messages. If they start using their device when there are fewer than 3 Sundays remaining in that calendar month, their onboarding period will be extended to cover the next month as well. During the onboarding period, individuals receive information about their wear target, that wearing their device is important for health research, how they need to sync their device, information about the study support center and feedback on how they have performed relative to their target. This onboarding period is important to ensure that all participants are informed about the purpose of the study and what is expected of them. In addition, it gives participants in the message choice condition familiarity with the reminder messages so that they can make a more informed choice when they are prompted about their communication frequency choice.

At the end of the onboarding period, individuals receive a support message informing them that they have 11 more months remaining in the study. Depending on the support message frequency condition they have been assigned, they are informed that they will receive no more support messages for the remaining 11 months (standard Fitbit message condition), will continue to receive support messages once a week (default AoURP message frequency condition) or are prompted to declare their preferred cadence of support messages (participant choice message frequency condition). All participants are informed that they can change their study notification preferences via the research portal, but only the choice group can set their preferences by responding to the text message prompt. This group will receive a prompt to confirm or modify their chosen cadence at the 6-month mark.

Experimental Design Details

Randomization Method

Randomization is implemented by the technology partner that runs the research study platform. When individuals receive their invitation for the study, they are randomized to all three treatment sets: invitation email variant; wear target; message frequency condition.

Randomization Unit

Individual

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

10000 individuals

Sample size: planned number of observations

10000 individuals

Sample size (or number of clusters) by treatment arms

We are cross-randomizing individuals to the wear and message frequency conditions. As such, we expect the following numbers per treatment group:

Wear target (2 conditions: 4 consecutive days per month; 12 total days per month): ~5000 individuals per condition.
Message frequency (3 conditions: no messages; messages once a week; messages at cadence determined by the participant (2x a week; 1x a week; 2x a month; 1x a month): ~3,333 individuals per condition.

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Supporting Documents and Materials

IRB