Digital Health Intervention for Asthma: Patient-Reported Value and Usability

Introduction

Cost of Asthma

The health impact and economic costs of asthma are significant, with the annual direct costs approaching US $50.1 billion and indirect costs such as lost productivity contributing an additional US $5.9 billion [1-3]. Despite asthma’s negative impact on health, patient self-management remains a challenge, with controller medication adherence rates of approximately 30% [4-7]. Asthma self-management requires daily work by the patient, including adherence to complex medication regimens and addressing multiple triggers of symptoms.

Digital Health Interventions for Asthma

Digital health interventions have been used increasingly in self-management interventions for asthma. For example, the development of mobile apps for asthma management doubled between 2011 and 2013 [8]. These interventions, which can include tools such as short message service, mobile apps, Web-based portals and websites, and electronic inhaler sensors, offer new ways for patients to monitor and manage their asthma. A limited but growing body of literature demonstrates the effectiveness of digital interventions in improving clinical outcomes, including asthma control, adherence, and symptom-free days [9-24].

Better understanding of how patients perceive and use digital health interventions to achieve improved outcomes is of utmost importance; however, there have only been a few studies published that focused on this topic. A recent review of digital health interventions across multiple diseases found that usability was assessed in only 33% of studies [25]. Within the asthma literature, a systematic review concluded that “patient perspectives have been largely ignored” [26]. A small number of studies have evaluated patient perceptions of asthma apps [19,24] or electronic inhaler sensors [27,28], but these have been limited to small samples over short periods of time (1-4 months). Little is known about how patients perceive the utility of combining electronic inhaler sensors with mobile apps and other digital tools. Furthermore, no studies have explored patient perceptions in a real-world setting over a sustained duration.

The study’s primary objective was to evaluate participant satisfaction, perception of usability and value, and overall experience after sustained use of a digital health intervention combining sensors, a mobile app, a Web dashboard, and email communication, to support asthma management in a real-world setting. The study also aimed to explore whether these perceptions would be influenced by specific participants’ characteristics such as age, gender, device ownership, insurance type, asthma control, and engagement. Improving the understanding of how diverse patients experience digital tools for asthma self-management will contribute to the success of these tools and the durability of their effects.

Methods

Digital Health Platform

Participants were enrolled in a randomized controlled study to evaluate the impact of a Food and Drug Administration–cleared digital health platform on asthma outcomes including short-acting beta-agonist (SABA) use and asthma control. The digital health platform includes electronic inhaler sensors (Propeller Health, Madison, Wisconsin) that attach to inhaled asthma medications (Figure 1). The sensor monitors the date, time, and frequency of medication use and transmits these data back to secure servers through a smartphone app or hub base station. Location data are collected on medication use among patients who have a smartphone. The sensors regularly transmit data back to the server, or sync, through the smartphone or hub. The version of the sensor that participants used in this study required them to charge their sensors every 2 weeks.

The data collected by the sensors are presented back to patients and health care providers through the digital health platform. The platform aims to promote disease awareness and self-management by enabling access to a patient’s own medication use data, including daily assessments of asthma control, adherence, identified triggers, and education based on the national guidelines (NHLBI 2007). The information is shared via a number of communication channels for all users, including weekly email reports, a Web-based dashboard, and a mobile app for smartphone users. Additional digital health platform details have been described elsewhere [13,18].

Participant Enrollment

The clinical study enrolled adult and pediatric asthma patients (N=495) in parallel arms from specialty and primary care clinics at 2 sites in the Dignity Health system. Clinic staff enrolled eligible participants with the following inclusion criteria: over the age of 5 years, a provider diagnosis of asthma, presence of a SABA prescription at study inception, Spanish or English fluency, and absence of significant comorbidity such as chronic obstructive pulmonary disease. Further study detail is described elsewhere [13].

Study Design

Participants were randomly assigned to either the intervention or control group and matched on sex, age, insurance type (public vs private), and baseline asthma control status (as defined by an Asthma Control Test [ACT] score ≤19 indicating a lack of asthma control) [29]. All participants received at least one sensor to monitor their medication over a 12-month period. If a participant had a smartphone, the study coordinator assisted them in downloading the app and conducting a first sync with the sensor. If a participant did not have a smartphone, they were provided a hub base station. This study used the earliest version of the sensor, which had a 15-day battery life, and participants were instructed to charge the sensor at regular intervals. Intervention group participants (n=250) received full access to the digital health platform described previously for 12 months. Physicians from the clinics could monitor the status of these intervention patients in real time and receive notifications about SABA overuse through Web-based dashboards. The Dignity Health Institutional Review Board reviewed and approved this study and survey (Trial Registration: ClinicalTrials.gov NCT01509183).

Survey Data Collection

Surveys were administered electronically to participants in the intervention group who were aged 12 years or older at study completion at 12 months (N=207). The exit survey assessed asthma control with the ACT [29]. The study coordinator sent participants an email invitation to complete the survey electronically, and the coordinator made a single follow-up attempt with participants who did not respond within 1 week. Participants were given the option to take a paper version of the survey and return it through the mail. The participants did not receive any incentive to complete the exit survey.

The survey consisted of open- and closed-ended questions, which evaluated satisfaction with the reports and information, satisfaction with the sensor device, learnings as a result of platform use, identified triggers, quality of communications with health care providers, suggestions for improving the reports and sensor, and interest in continued use of the platform. Detailed survey questions are included in Multimedia Appendix 1.

Data Analysis

Closed-ended questions were analyzed by determining the percentage of respondents who selected each of the possible responses. Thematic analysis of the open-ended responses in the survey followed the structured approach described by Braun and Clarke [30], where primary themes and subthemes were identified and coded according to specific usability and value topic areas. Two reviewers (KH and MB) independently assessed the open-ended questions, met to address discrepancies, and agreed on primary themes and subthemes. We report the number and percentage of participants who reported the primary themes and subthemes.

We assessed participant engagement with the platform by evaluating the mean email open rate, number of push notifications received, and number of participant sign-ins to the dashboard or app. We used a logistic regression model to evaluate if there were differences between those participants who took the exit survey compared with those who did not across the following variables: age (12-17 years vs 18 years and over), sex, insurance type (public vs private), device type (smartphone vs hub), asthma control at intake (controlled vs uncontrolled), syncing duration (time between the first and last sync), and syncing frequency (number of syncs between the first and last sync). Using a logistic regression modeling approach, we also evaluated whether these variables were associated with participants’ responses to the closed-ended questions on device and report satisfaction, usefulness, and interest in continued use of the platform. Responses were grouped based on binary categories, for example, satisfied versus not satisfied, useful versus not useful.

‎View this figure

Results

Participants

Of the intervention group participants (N=250), 207 met the requirement of being aged 12 years or older at study completion to receive the exit survey. A total of 89 participants completed the exit survey (Table 1), which represented a 42.9% response rate (89/207). Among the participants who completed the survey (N=89), the average and median ages were 42.7 and 45 years, respectively; 85% (76/89) of participants were aged older than 18 years, 15% (13/89) were aged between 12 and 17 years, and 64% (57/89) of participants were female. Participants used a variety of device types to access the digital platform, including smartphones (49%, 44/89; 27%, 24/89 iOS and 23%, 20/89Android) and hub base stations for those without smartphones (51%, 45/89). In addition, 74% (66/89) of participants had private insurance and 26% (23/89) had public insurance. At the start of the study, 30% (27/89) of participants were well controlled, 61% (54/89) were not well controlled, and 9% (8/89) had an unknown control status.

Among the 207 intervention group participants, the logistic regression model identified that participants who used smartphones (P=.02), had a longer syncing duration (P<.001), and higher syncing frequency (P<.001) were more likely to take the exit survey (Table 2). There were no other significant differences across age, sex, asthma control, or insurance type between those participants who took the survey and those who did not.

Among the participants who completed the survey (N=89), on average, they synced their sensor for 361 days over the course of the study period (minimum: 264 days and maximum: 365 days). Participants synced their sensor 2149 times on average during the study or an average of 6 times per day. Participants also had a mean email open rate of 78%, received 35 push notifications, and had a mean 60 sign-ins to the patient dashboard, which represents more than 1 sign-in per week. The following results pertain to the 89 participants who completed the survey.

Satisfaction With Inhaler Sensor Device

Participants reported their degree of satisfaction with the inhaler sensor device, with 79% (70/89) of respondents reporting “very satisfied,” 20% (18/89) reporting “somewhat satisfied,” and 1% (1/89) reporting “not at all satisfied” (Figure 2). There were no significant differences in responses to this question across participants’ device type, age, sex, insurance, asthma control at intake, or syncing history (see Multimedia Appendix 2). A total of 37 participants (40%, 37/89) stated the inhaler device was “easy,” including 3 subthemes of “easy to use” (32%, 29/89), “easy to maintain” (6%, 6/89), and “easy to understand” (2%, 2/89). Moreover, 23 participants (26%, 23/89) described the size of the device as “small” and “compact” (Table 3). Of the subthemes relating to the size of the device, 10 participants (11%, 10/89) specified that it did not “obstruct” or “interfere” with their use of their inhaler and “was not bulky.” Furthermore, 17 participants (19%, 17/89) described the sensor’s convenience, with the 2 subthemes of the device being “portable” and functioning well.

The majority of the participants (56%, 50/89) made no recommendations for sensor improvements. Of the participants who reported specific ideas for improvements (44%, 39/89), the most frequent response was for a longer battery life (20%, 18/89). Participants also reported wanting a smaller size device (13%, 12/89). The remaining participants reported wanting a more secure fit (4%, 4/89) or recommended additional features such as adding a dose counter (6%, 5/89).

‎View this figure

Perception of the Reports and Information

Participants reported on their satisfaction and perceptions of the reports’ usability and value in supporting their asthma management. A total of 60% (54/89) of the participants reported that they were “very satisfied” with the reports, 33% (29/89) were “somewhat satisfied,” and 7% (6/89) were “not at all satisfied” (Figure 2). There were no significant differences in responses to this question across participants’ device type, age, sex, insurance, asthma control at intake, or syncing history (see Multimedia Appendix 3).

Moreover, 33 (37%, 33/89) of participants reported liking the content of the reports, including informative (9%, 8/89), actionable (6%, 5/89), and easy to understand (8%, 7/89; Table 4). They also reported that they liked the report display (10%, 9/89) and the frequency (4%, 4/89). In addition, 23 participants (26%, 23/89) shared that the reports provided new information that increased their awareness of their asthma management. Some participants did not see significant value in the reports, responding that they did not view the reports (9%, 8/89), found the reports to be only “generally helpful” (4%, 4/89), did not receive the reports (3%, 3/89), or did not have any feedback to share (9%, 8/89).

Perceived Utility and Participant Learnings

Participants responded the information was “very useful” (45%, 40/89), “somewhat useful” (45%, 40/89), or “not at all useful” (10%, 9/89) in learning about their asthma (Figure 2). There were no significant differences in responses to this question across participants’ device type, age, sex, insurance, asthma control at intake, or syncing history (see Multimedia Appendix 4).

The participants reported several learnings from using the digital platform (Table 5). A primary theme focused on understanding, identifying, and managing asthma triggers (23%, 21/89). A total of 53 participants (60%, 53/89) reported identifying up to 7 new triggers. Weather, allergies, and exercise were the most commonly identified triggers, documented across 79% (42/53) of the participants who identified new triggers.

A second primary theme focused on improved self-awareness. A total of 19 participants (21%, 19/89) specifically reported learning from monitoring the timing, location, and frequency of their medication use. Moreover, 10 participants (11%, 10/89) described learning strategies for managing their asthma such as taking their controller medications and avoiding specific triggers. Furthermore, 6 participants (7%, 6/89) reported that they discovered they were not as in control of their asthma as they had believed or they confirmed that their treatment plan helped them stay controlled. In addition, 5 participants (6%, 5/89) learned that they are accountable for their self-management and have the capacity to control their asthma. Participants could opt to discuss their data with their providers; 46% (41/89) of the participants reported that they had conversations with their providers about their data.

Some participants reported not learning anything from the reports: 16 participants (18%, 16/89) reported “none,” and 3 participants (3%, 3/89) reported not learning anything because they were already well controlled.

Participants were asked to provide suggestions for improving the reports or platform. A total of 59 participants (66%, 59/89) reported no improvements were needed, and 11 participants (12%, 11/89) responded with “not sure” or “no comment.” The remaining responses (21%, 19/89) offered specific suggestions including more educational content on how to improve asthma, information about how to subsidize medication costs, improvements in the app login process, additional customization options for how to receive reports, changing the cadence of the reports, and extending the sensor battery life.

The final question of the survey asked participants whether they would like to continue using the digital intervention: 72% (64/89) of the participants reported that they were interested in continuing to use the sensor and platform beyond the study. There were no significant differences among those participants who were interested in continuing or those who were not based on a participant’s device type, age, gender, insurance type, syncing history, or asthma control at intake (see Multimedia Appendix 5).

Discussion

Summary of Findings

Participants were satisfied with the digital health intervention for managing their asthma and perceived value in using the technology to support their self-management. Participants reported improving their awareness, learning about trends in their medication use, identifying new triggers, and objectively monitoring how well their asthma is controlled, which they could discuss with their doctor. Participants shared helpful feedback regarding areas for improvement including interest in longer battery life and smaller size of the device.

We saw no significant differences in the degree of satisfaction and perceived usability across participants regardless of participant age, gender, insurance type, asthma control, technology adoption, and syncing duration. However, we did observe that those with a smartphone and a longer record of syncing were more likely to take the survey in the first place, which could have biased the sample toward more engaged, more technologically savvy participants. Despite this potential bias, these results add to the limited evidence exploring patient perspectives on digital health tools across diverse participants and shed light on the perceived benefits of the tools as well as opportunities for improvement.

Mechanisms for Supporting Self-Management

These findings align with and supplement the existing evidence [19,24,31], in particular adding to the literature by demonstrating participants’ perceptions among a larger cohort in a real-world setting over a prolonged period of time. We explore 4 learnings that emerged from the participants’ responses.

First, participants frequently emphasized ease of use for the inhaler sensor in terms of deployment, comprehension, maintenance, size, and convenience. Similarly, in a small study of adolescent asthma participants, patients cited the importance of the device not drawing any attention, being small in size, and being easily portable [28]. Furthermore, a review of previous digital health intervention studies suggested that ease of use and adaptation to an individual’s personal lifestyle are essential factors for engagement and persistent use [32]. These findings align with the concept of minimally disruptive medicine, which aims to find ways to reduce patient treatment burden [33]. To support this approach, digital tools enable passive data collection of medication use, eliminating the burden of manual tracking. By providing objective documentation of medication use, trends, and specific information about potential triggers, the self-monitoring process can become more efficient and actionable.

Second, participants in this study and others have reported that having objective data leads to increased self-awareness. This finding aligns with the self-regulation model for chronic disease and the more refined version for asthma management [24,34], in which a patient goes through a process of self-monitoring and assessing feedback to inform behavior. Participants reported gaining insights from monitoring medication use, especially the timing and location of medication use events, to identify patterns, times of worsening, and potential triggers. Almost one-quarter of patients valued learning about their asthma triggers including specific environmental conditions, such as temperature, humidity, and air quality, through the digital tool to help inform their management.

Improved confidence in one’s ability to self-manage, as was seen in some of the open-ended responses, is promising, but further investigation is needed to evaluate the intervention’s impact on self-efficacy. The survey design did not include specific questions to assess self-efficacy; therefore, we are unable to determine if the intervention had an impact on self-efficacy. Future studies evaluating digital interventions for asthma should thoroughly evaluate self-efficacy using validated self-efficacy measures and surveys, such as the mini-Asthma Quality of Life Questionnaire [35], Self-efficacy and Situational Barrier Survey Questionnaire (KASE-AQ) [36], and Mobile App Rating Scale [37].

Third, this study found that participants value having information they can share with their clinical providers. Almost half of all participants had spoken with their clinical provider about the data collected. A smaller subset reported using the data to request medication adjustments, therefore taking more responsibility for their management. These findings are supported in a study that found that 70% of patients wished to share their data with health teams [31].

Digital health tools can support the patient-provider relationship through shared decision making [38,39]. By enabling the sharing of data and the ability of patients to discuss specific issues as they arise, patients and providers can work collaboratively to adjust their management plan and achieve asthma control. A recent study that surveyed clinical providers about digital health tools, such as sensors, documented similar perceived benefits. Clinicians reported that objective data from sensors have the potential to enable patients to see trends and patterns in their medication use, increase accountability and confidence in their management, and support discussion with clinical providers [40].

Fourth, this study identified areas for improvement that can inform digital health tool development in the future. Participants reported wanting a longer battery life for the sensors. This feedback helped inform the development of the latest version of the sensors, which use Bluetooth low energy technology, do not require charging, and have a battery life of 12-18 months. In addition, some participants reported an interest in a smaller device, which resulted in the development of a sensor with 30% reduced volume. Participants also reported desiring more personalization and customization in the notifications and reminders and requested multiple methods of communication so that they can select the method that works best for them at the right time. Ongoing work is underway to add personalization, customization, and new communication options. There is growing evidence in this study and others [24,31] indicating how much patients value personalization and customization; however, more research is needed to determine best practices for identifying customizable features that will enhance patient satisfaction and engagement.

Limitations

First, although the participants’ cohort was fairly diverse in age, insurance type, technology use, gender, and asthma control status, additional studies are needed to evaluate preferences in larger samples of different types of patients across race, pediatric versus adult, geographic diversity, and health literacy levels. For example, this survey was only administered to participants aged 12 years and older and therefore did not capture the young pediatric patient or caregiver perspective. Second, the response rate was 42.9%, and selection bias may have influenced the results. We did find that participants who used smartphones and those who remained in the program longer (defined by their sync duration) were more likely to take the survey in the first place. They may have experienced more value from a longer interaction with the intervention and reported as such on the survey. Future research should target individuals with early study attrition to ensure a more complete understanding of their perceptions of usability and satisfaction to enable scalability to all populations. Third, the infrastructure to capture more robust individual-level participant engagement data was limited, which prevented a more complete understanding of participant engagement and how this engagement may have influenced overall satisfaction and perception of utility. Fourth, the survey questions were not derived from validated measures and did not use standard Likert scales, which limited the generalizability of the results.

Conclusions

This study offers new insights about patient satisfaction, preferences, and perceptions of a digital health intervention for respiratory disease among a larger, fairly diverse cohort in a real-world setting after a prolonged period of use. Participants reported satisfaction with the sensor device and reports and particularly valued the ability to integrate the device into everyday life, the provision of accessible and meaningful information to improve self-awareness, and the enhancement of communication with health care providers. As the use of digital tools in health care expands, it will be critical to learn from patient preferences and experiences to ensure that the tools fit well into their daily life and support their self-management.