Key Points
Question
Does a digital asthma self-management (DASM) program improve symptom control in adults with asthma?
Findings
In a randomized clinical trial enrolling 901 adults with asthma, those receiving a DASM program showed statistically significant improvements in symptom control after 12 months, compared with those receiving usual care. Differences in effectiveness were observed across some demographic groups.
Meaning
A DASM program may improve symptom control in adults with asthma, and further program development may be appropriate.
This randomized clinical trial investigates the impact of a digital asthma self-management program on symptom control and use of acute health care services in adults with asthma.
Abstract
Importance
Digital health technologies may improve asthma self-management, but evidence is limited in this area.
Objective
To investigate the effect of a digital asthma self-management (DASM) program on asthma symptoms in adults.
Design, Setting, and Participants
Patient-reported outcome results were reported from a randomized, pragmatic, parallel-arm, open-label, decentralized clinical trial. Adults with asthma were recruited via email, enrolled from October 29, 2020, through November 4, 2021, and were randomized to DASM or usual care (control). Participants completed study activities outside a clinical setting. Data were analyzed between October 13, 2023, and November 29, 2024.
Intervention
The app-based DASM program provided tailored notifications, symptom logging, wearable device integration, and other tools.
Main Outcomes and Measures
Change in the Asthma Control Test (ACT) was a primary outcome. The ACT is a validated measure of asthma control. Secondary outcomes included engagement and self-reported medication adherence.
Results
Nine hundred and one participants were enrolled, with data available for 899 (639 [71.1%] female; mean [SD] age, 36.6 [10.5] years). For subgroup analyses, 195 participants (21.7%) were African American; 125 (13.9%), Hispanic or Latino; 680 (75.6%), commercially insured; and 219 (24.4%), Medicaid insured. Prespecified analyses of participants with uncontrolled asthma at baseline (n = 550) showed improvements after 12 months by 4.6 (95% CI, 4.1-5.2) ACT points among DASM participants (P < .001) and 1.8 (95% CI, 1.3-2.4) ACT points among controls (P < .001) (adjusted difference, 2.8 [95% CI, 2.0-3.6] points; P < .001). Race moderated this effect. At 12 months, the difference between arms in ACT change favored DASM over control by 1.0 (95% CI, −0.7 to 2.7) points (P = .26) for African American participants and 3.3 (95% CI, 2.4-4.2) points (P < .001) for participants not endorsing African American race (adjusted difference, −2.3 [95% CI, −4.2 to −0.4] points; P = .02 for interaction). Moderation was not observed by insurance (Medicaid vs commercial; adjusted difference, 1.0 [95% CI, −0.8 to 2.8] points; P = .18 for interaction) or ethnicity (Hispanic or Latino vs non-Hispanic; adjusted difference, 1.0 [95% CI, −1.3 to 3.3] points; P = .70 for interaction).
Conclusions and Relevance
In this randomized clinical trial of DASM, improved asthma control was observed relative to usual care. Program adaptations may be appropriate to confer benefit throughout diverse populations.
Trial Registration
ClinicalTrials.gov Identifier: NCT04609644
Introduction
Asthma occurs in 8% of US adults1 and generates $80 billion annually in total costs.2 It is uncontrolled in more than 60% of adult cases,3 although fewer than 10% of cases are refractory to appropriate management.4,5 Scalable interventions are needed to address this gap.
Digital health technologies (DHTs) offer a scalable strategy to improve asthma outcomes, but studies of DHTs have been underpowered6 and limited in duration,7,8,9 with inconsistent results observed.10,11,12,13,14 Recent Global Initiative for Asthma guidelines note that high-quality studies are needed in this area.15
DHTs targeting asthma should provide benefit throughout diverse populations, including for groups that are disproportionately impacted by asthma.16 Such groups may include adults with low socioeconomic status (SES),17,18,19 Hispanic or Latino adults,20,21 and Black or African American21,22,23 adults. Low income is associated with frequent exacerbations and emergency visits.17,24 The incidence of asthma-associated emergency visits for Hispanic or Latino adults is nearly twice that of non-Hispanic adults,25,26 while incidence for African American adults is 5 times greater than that of White counterparts.27
Aspirations to reduce asthma disparities through DHTs began years ago,28 and some DHTs may support equitable care,29 but evidence in this area is also limited.30,31 Smartphone ownership exceeds 80% in underserved populations,32,33 which may provide an opportunity to deliver asthma self-management support.
The Asthma Digital Study was launched to build on this opportunity. A digital asthma self-management (DASM) program was developed, incorporating wearable devices for passive biometric monitoring and a smartphone app with multiple tools for asthma self-management. The objective of this study was to investigate the impact of DASM on symptom control and use of acute health care services among adults with asthma.
Herein we report patient-reported outcomes (PROs) and engagement metrics, as well as exploratory subgroup analyses. Preliminary outcomes for the use of health care services were published elsewhere.34 Full results for utilization and health care costs will be published separately, owing to space constraints.
Methods
Overview and Study Design
As detailed elsewhere,35 the Asthma Digital Study was a parallel-arm, open-label, randomized clinical trial. It was decentralized,36 and thus participants were not required to go to a clinical site. Recruitment was via email, and participants completed trial activities outside a clinical setting. Participants enrolled through a custom smartphone app that was designed to be compliant with applicable privacy laws. Enrollment took place from October 29, 2020, through November 4, 2021. PROs were measured during the 12 months following enrollment. We used a pragmatic trial approach to inform clinical decisions.37,38 The study was approved by the Institutional Review Board of the University of California, Irvine, and all participants provided written informed consent electronically. The trial protocol is found in Supplement 1. Results are reported per the Consolidated Standards of Reporting Trials (CONSORT) guideline.
Adult participants were recruited through a large US payer (Elevance Health, Inc). Plan members with commercial or Medicaid coverage were enrolled to increase generalizability. Coverage type served as a proxy for SES.
Participants were randomized (1:1) to DASM or usual care (control condition) using a permuted block randomization sequence, with block size of 4, to balance potential confounders related to recruitment wave or time of enrollment. The randomization sequence was generated before enrollment started, using the ralloc command in Stata/MP, version 16.1 (StataCorp LLC). Random assignment per prespecified sequence was implemented automatically, on enrollment, through a decentralized trial platform (MyDataHelps, CareEvolution LLC).
Recruitment, Screening, and Consent
Insurance claims were screened under an institutional review board–approved Health Insurance Portability and Accountability Act of 1996 (HIPAA) authorization waiver, to identify health plan members who had a diagnosis code indicating asthma and who met other criteria (eTable 1 in Supplement 2). Identified members were emailed a study description and app download link. Those who downloaded the app were provided with further study information. The app was used to administer eligibility questions, an informed consent form, and HIPAA authorization forms.
Participants and Setting
Participants were nonpregnant adults, aged 18 to 64 years, who confirmed a history of asthma by self-report. eTable 1 in Supplement 2 provides full eligibility criteria. Participants were recruited from throughout the contiguous US.
DASM Program
The DASM program uses wearable devices and a novel smartphone app. Features of the app include tailored notifications, symptom logging, and evidence-based education. eAppendix in Supplement 2 provides a detailed program description. Participants were shipped a smartwatch and sleep monitor (Apple Inc) at no cost. The sleep monitor uses a thin, piezoelectric strip, placed over a mattress. The devices and app (a custom version of MyDataHelps) (Figure 1A-C) were used in concert to track metrics including heart rate, respiratory rate, and sleep duration.
Figure 1. Selected Application Features and Components of the Digital Asthma Self-Management Program.
eAppendix in Supplement 2 provides a detailed program description.
During a 14-day baseline period, data from the wearable devices were collected to establish participant-specific norms. Subsequently, participants in the DASM group received tailored notifications (“smart nudges”) that were triggered by predefined deviations from these norms (eAppendix in Supplement 2). Each notification encouraged the participant to complete a daily entry, which presented questions about symptoms, triggers, and medications. Entries were designed to build awareness of asthma symptom control in relation to self-management behaviors.
Symptom logs from entries were used to provide assessments of asthma control per established guidelines.39 In each symptom log, the participant is presented with in-app questions around recent asthma symptoms (eAppendix in Supplement 2 provides details). Asthma was classified as not well controlled if entries from the past week showed any asthma-associated nighttime awakenings, reduced physical activity, or at least 3 days with asthma symptoms or rescue inhaler use. A calendar view (Figure 1C) visualized asthma control over time. While asthma control status at baseline was determined by the Asthma Control Test (ACT) score on enrollment, control status presented to participants in DASM via the app was based on daily entries.
Symptom logging paired with visualization of personal health data may improve awareness of symptoms and associated behaviors.40 This may highlight opportunities for proactive self-management. The approach used is consistent with the semi-automated tracking theoretical framework.41
Participants in DASM also received app-based tools to support asthma self-management (Figure 1D and eAppendix in Supplement 2). The program was designed to promote health behavior change, with a focus on symptom awareness, medication adherence, and trigger avoidance. Evidence-based education was provided on key topics.
The control arm received a modified app that delivered the ACT and other PROs but did not provide DASM features. Like participants in the DASM arm, participants in the control arm received and were asked to use both study devices. Control participants were not asked to complete daily entries. Both arms received incentives (eMethods in Supplement 2) to use the devices and complete PROs.
Outcomes
Coprimary outcomes were (1) 12-month change from baseline in ACT score42 among those with uncontrolled asthma at baseline and (2) asthma-associated allowed costs for unplanned asthma care. Baseline asthma control status was determined by the ACT score, with uncontrolled asthma defined as an ACT score of 19 or less.43,44
The ACT is a psychometrically validated PRO consisting of 5 items that address asthma control for a 4-week recall window.43,44,45 Responses are given on 5-point Likert scales (eg, not controlled at all [1] to completely controlled [5]). The score range is 5 to 25, with greater scores denoting better asthma control. The minimal clinically important difference is 3 points.46
Secondary outcomes included 2 engagement metrics: counts of symptom logs and counts of app use events (termed app opens). Each count was summed by participant-month, and the mean was calculated within participants. Other secondary outcomes included PRO measures of medication adherence, readiness to change self-management behaviors, confidence in one’s ability to self-manage health, and health-related productivity impairment. These were measured using validated self-report instruments, namely the Adherence to Refills and Medications Scale, 7-item Adaptation,47,48 Readiness Ruler,49 Health Confidence Score,50 and asthma-specific version of the Work Productivity and Activity Impairment Questionnaire (WPAI-AS), version 2.0.51 The WPAI-AS measures missed work time due to asthma and productivity impairment while working.
Further detail on each PRO is provided in eTable 2 in Supplement 2. All PROs were administered through the app. Demographic information (including race, ethnicity, and gender) was self-reported in the app as well, allowing for subgroup analyses within demographic groups that have shown poor outcomes in prior evidence.
Sample Size Determination
The sample size requirement was based on the coprimary outcome of asthma-related health care costs. Simulations were used to identify the enrollment target. A sample of 900 participants (450 per arm) yielded 80% power to detect a 20% relative cost reduction between study arms (pooled across insurance types). For the ACT end point, this sample size yielded power exceeding 90% to detect a between-arms difference of 3 points (the minimal clinically important difference46) in ACT change after 12 months. This assumed 35% prevalence of uncontrolled asthma by ACT,52 an ACT SD of 5,52 and 30% attrition. The 2-tailed α value was set at .05.
Statistical Analysis
The prespecified35 primary analysis used linear mixed-effects modeling of ACT scores for all participants with uncontrolled asthma at baseline (those with controlled asthma were included in secondary analyses). Fixed effects were included for time (baseline vs 12 months), arm, coverage type (Medicaid vs commercial insurance), ethnicity (Hispanic or Latino vs non-Hispanic), and race (dichotomized here as those who reported African American race vs those who did not report African American race). Analyses by race focused on African American individuals, given aforementioned disparities in this racial group. A time × arm interaction was included to investigate the difference between arms in ACT change after 12 months. This difference was the treatment effect of primary interest. Model-derived marginal estimates were used to calculate the treatment effect. Differences in the treatment effect were examined across demographic subgroups and coverage types, as prespecified.35 Subgroup analyses applied best practices summarized elsewhere.53 An interaction of time × arm × coverage type was included in the primary analysis model to investigate effect moderation by coverage type. Similar 3-way interactions were included to investigate moderation by ethnicity and race. The model controlled for age, gender, and smoking status at baseline. Random intercepts and slopes for time were included. Missing data were handled directly with maximum-likelihood estimation in mixed-effects models.54,55 This method generates unbiased estimates under the missing at random assumption. All available data were used in analyses.
Sensitivity analyses were performed with (1) longitudinal analysis of covariance56 (month 12 score modeled with adjustment for baseline), (2) filtering to the per-protocol sample (prespecified35 exclusion of those completing <70% of reporting requirements), and (3) multiple imputation of missing ACT data. Multiple imputation analyses used iterative Markov chain Monte Carlo methods57 to impute 100 datasets, which were combined for analyses per Rubin rules.58 Coverage type, ethnicity, race, age, gender, and baseline smoking status were auxiliary variables for imputation. The imputation and subsequent treatment effect estimation were performed with the Stata mi suite57 and mimrgns59 command. Additional methodological information and sensitivity analyses are reported in eMethods in Supplement 2.
In secondary analyses, the 2 engagement metrics were analyzed by quantile regression (more specifically, median regression), a nonparametric method. The 3 subgroup variables (coverage type, ethnicity, and race) were analyzed simultaneously, with the effect of each adjusted for the other 2. Exploratory analyses examined associations between engagement and change in ACT (eMethods in Supplement 2).
All analyses used original arm assignments. Analyses were performed in Stata/MP, version 16.1, and R, version 4.4.2 (R Foundation for Statistical Computing), with 2-tailed α = .05 indicating statistical significance. Secondary and exploratory analyses are not corrected for multiple comparisons and should be interpreted accordingly (eMethods in Supplement 2). Data were analyzed between October 13, 2023, and November 29, 2024.
Results
Nine hundred and one participants enrolled and were randomized, with data lost for 2 due to a technical issue, leaving 899 participants (450 in the DASM arm and 449 controls [Figure 2]) included in the analysis (mean [SD] age, 36.6 [10.5] years). Of these, 639 (71.1%) were female; 259 (28.8%), male; and 1 (0.1%), other or declined to answer. Twenty-nine participants (3.2%) were American Indian or Alaska Native; 27 (3.0%), Asian; 195 (21.7%), Black or African American; 125 (13.9%), Hispanic, Latino, or Spanish; 8 (0.9%), Middle Eastern or North African; 2 (0.2%), Native Hawaiian or Other Pacific Islander; 585 (65.1%), White; and 1 (0.1%), other or declined to answer. Among the 897 participants with ACT scores available (99.8%), 550 (61.3%) had uncontrolled asthma at baseline, while 347 (38.7%) had controlled asthma. ACT scores were available for 704 participants (78.3%) at 12 months. Baseline participant characteristics by arm are reported in the Table (eTables 3-7 in Supplement 2 present baseline characteristics by arm and subgroup). Participants resided in 41 US states (eTable 8 and eFigure 1 in Supplement 2).
Figure 2. Participant Flow Diagram.
DASM indicates digital asthma self-management; PRO, patient-reported outcome; and UC, usual care.
Table. Baseline Characteristics of Study Participants.
| Characteristic | Study group, No. (%) | ||
|---|---|---|---|
| All (N = 899) | Control (n = 449) | DASM (n = 450) | |
| Age, mean (SD), y | 36.6 (10.5) | 36.5 (10.6) | 36.7 (10.4) |
| Self-reported gender | |||
| Female | 639 (71.1) | 329 (73.3) | 310 (68.9) |
| Male | 259 (28.8) | 119 (26.5) | 140 (31.1) |
| Other or declined to answer | 1 (0.1) | 1 (0.2) | 0 |
| Self-reported race and ethnicitya | |||
| American Indian or Alaska Native | 29 (3.2) | 13 (2.9) | 16 (3.6) |
| Asian | 27 (3.0) | 13 (2.9) | 14 (3.1) |
| Black or African American | 195 (21.7) | 96 (21.4) | 99 (22.0) |
| Hispanic, Latino, or Spanish | 125 (13.9) | 63 (14.0) | 62 (13.8) |
| Middle Eastern or North African | 8 (0.9) | 2 (0.4) | 6 (1.3) |
| Native Hawaiian or Other Pacific Islander | 2 (0.2) | 2 (0.4) | 0 |
| White | 585 (65.1) | 291 (64.8) | 294 (65.3) |
| None of these or prefer not to answer | 26 (2.9) | 14 (3.1) | 12 (2.7) |
| Insurance type | |||
| Commercial | 680 (75.6) | 341 (75.9) | 339 (75.3) |
| Medicaid | 219 (24.4) | 108 (24.1) | 111 (24.7) |
| College graduate or higher by self-report | 425 (47.3) | 214 (47.7) | 211 (46.9) |
| Self-reported smoking statusb | |||
| Current nonsmoker | 781 (86.9) | 388 (86.4) | 393 (87.3) |
| Current smoker | 100 (11.1) | 48 (10.7) | 52 (11.6) |
| Unknown or declined to answer | 18 (2.0) | 13 (2.9) | 5 (1.1) |
| Asthma control statusc | |||
| Controlled asthma | 347 (38.7) | 173 (38.7) | 174 (38.7) |
| Uncontrolled asthma | 550 (61.3) | 274 (61.3) | 276 (61.3) |
Abbreviations: ACT, Asthma Control Test; DASM, digital asthma self-management.
Racial and ethnic categories are not mutually exclusive, so totals can exceed 100%.
Self-report items were adapted from validated patient-reported outcome measures.
Baseline ACT scores were available for 897 of 899 participants (99.8%). Controlled asthma indicates an ACT score of 20 or greater; uncontrolled, ACT score of 19 or less (ACT score range, 5-25).43
Primary Outcome
Prespecified analyses35 of participants with uncontrolled asthma at baseline included 978 ACT scores from 550 participants (274 controls and 276 DASM) (eTable 9 in Supplement 2). After 12 months, mean ACT score increased by 4.6 (95% CI, 4.1-5.2) points (P < .001) for DASM and 1.8 (95% CI, 1.3-2.4) points (P < .001) for controls (adjusted difference, 2.8 [95% CI, 2.0-3.6] points; P < .001 for time × arm interaction) (Figure 3A and eTable 10 in Supplement 2). This difference between arms in ACT change (treatment effect) was not moderated by coverage type (treatment effect difference by subgroup, 1.0 [95% CI, −0.8 to 2.8] points; P = .18 for time × arm × coverage type interaction) (Figure 4A) or ethnicity (treatment effect difference, 1.0 [95% CI, −1.3 to 3.3] points; P = .70 for time × arm × ethnicity interaction) (Figure 4A).
Figure 3. Change in Asthma Control Test (ACT) Score by Arm and Baseline Asthma Control Status.
Differences between arms in ACT change (treatment effects) are shown. Point estimates are marginal means derived from linear mixed-effects models. Both models control for insurance type, ethnicity, race, age, gender, and self-reported smoking status at baseline. A, Primary analysis of participants with uncontrolled asthma (ACT score, ≤19) at baseline (prespecified35). Mean change after 12 months was 4.6 (95% CI, 4.1-5.2) points (P < .001) for digital asthma self-management (DASM) arm and 1.8 (95% CI, 1.3-2.4) points (P < .001) for the control arm. B, Secondary analysis of those with controlled asthma (ACT score, ≥20) at baseline. Mean change was 0.7 (95% CI, 0.2-1.2) points (P = .005) for the DASM arm and −1.0 (95% CI, −1.4 to −0.5) points (P < .001) for the control arm. (the between-arm difference in change is slightly less than expected due to rounding). Error bars represent 95% CIs.
aP < .001 for time × arm interaction.
Figure 4. Treatment Effects and Symptom Log Counts by Subgroup.
Results are filtered to participants with uncontrolled asthma at baseline (Asthma Control Test [ACT], ≤1943), as prespecified.35 A, Results were estimated by linear mixed-effects modeling (Methods section and eTable 10 in Supplement 2). B, Counts by coverage type are adjusted for ethnicity and race to account for subgroup overlap. Counts by ethnicity and race are similarly adjusted (eTable 13 in Supplement 2 shows unadjusted results). App opens show the same pattern (eFigure 2 in Supplement 2). Boxes represent IQRs; vertical lines, medians; and whiskers, the most extreme values within 1.5 × IQR. DASM indicates digital asthma self-management.
aCalculated for treatment effect within subgroups.
bCalculated for differences in effect between subgroups.
cControl participants are excluded, as they did not have access to symptom logging or other DASM features.
In contrast, the treatment effect showed moderation by race. The treatment effect was 1.0 (95% CI, −0.7 to 2.7) points (P = .26) for African American participants and 3.3 (95% CI, 2.4-4.2) points (P < .001) for participants not endorsing African American race (treatment effect difference, −2.3 [95% CI, −4.2 to −0.4] points; P = .02 for time × arm × race interaction) (Figure 4A). This pattern of findings (significant treatment effect with moderation by race but not other subgroup variables) persisted in a longitudinal analysis of covariance (eTable 11 in Supplement 2), per-protocol analysis (eTable 12 in Supplement 2), multiple imputation analysis (eTable 12 in Supplement 2), and other sensitivity analyses (eResults in Supplement 2).
Engagement Metrics
Continuing with participants for whom asthma was uncontrolled at baseline (eTables 13 and 14 in Supplement 2 report on both groups), the median count of symptom logs per participant-month was 7.6 (IQR, 2.0-17.1) among the 276 DASM participants (control participants did not have access to symptom logging or other DASM features). No significant differences in median symptom log counts were observed between Medicaid and commercially insured subgroups (difference, −2.0 [95% CI, −6.1 to 2.1] logs; P = .33), nor between Hispanic or Latino and non-Hispanic subgroups (difference, −1.6 [95% CI, −7.0 to 3.8] logs; P = .56) (Figure 4B). African American participants logged symptoms less often than participants not endorsing African American race (difference, −4.9 [95% CI, −9.1 to −0.7] logs; P = .02) (Figure 4B). App open counts showed the same pattern; app open counts varied by race but not by other subgroup variables (eFigure 2 and eTable 14 in Supplement 2).
For both engagement metrics, a positive association was observed between engagement percentile and ACT change (eFigure 3 in Supplement 2). On average, the 12-month increase in ACT associated with a 75th percentile score on symptom logging exceeded that of a 25th percentile score by 1.9 (95% CI, 0.9-3.0) points (P < .001). The corresponding difference for app opens was 2.2 (95% CI, 1.3-3.1) points (P < .001).
Secondary PRO Outcomes
For the 347 participants with controlled asthma at baseline (173 control participants and 174 DASM participants) (eTable 15 in Supplement 2), mean ACT changes after 12 months were 0.7 (95% CI, 0.2-1.2) points for DASM participants (P = .005) and −1.0 (95% CI, −1.4 to −0.5) points for control participants (P < .001) (adjusted difference, 1.6 [95% CI, 1.0-2.3] points; P < .001 for time × arm interaction) (Figure 3B and eTable 16 in Supplement 2).
In the full sample, secondary PRO results at 12 months (eFigure 4 in Supplement 2) modestly favored the DASM over the control group for self-reported medication adherence (standardized means difference [SMD], 0.28; 95% CI, 0.13-0.43; P < .001), readiness to change (SMD, 0.27; 95% CI, 0.13-0.41; P < .001), confidence in the ability to manage one’s health (SMD, 0.33; 95% CI, 0.18-0.48; P < .001), and asthma-related work productivity impairment (SMD, 0.29; 95% CI, 0.12-0.45; P = .001). Odds of self-reported smoking at 12 months did not differ by arm (odds ratio, 1.20; 95% CI, 0.75-1.98; P = .48).
Discussion
In this randomized clinical trial of a DASM program, adults with asthma who received DASM showed statistically significant improvement in asthma symptoms, relative to usual care. The treatment effect was weaker for African American participants. Data do not suggest differences in treatment effect by insurance type or ethnicity.
To our knowledge, this is the largest randomized clinical trial to date investigating an app with consumer-grade wearables for asthma. A recent review16 cataloged more than 500 apps for asthma self-management. The review notes a dearth of evidence for these apps and a need to test them among racial and ethnic minority groups as well as patients with low SES. The present findings may be important given limited prior evidence on DHTs targeting asthma and on the effectiveness of these DHTs within underserved groups.
Inadequate sample size is a candidate explanation for nonsignificance of the treatment effect within the African American subgroup (125 with uncontrolled asthma at baseline). However, observed point estimates and 95% CIs (Figure 4A) suggest that reduced effect magnitude may be the primary reason for nonsignificance in this group (eDiscussion in Supplement 2). The treatment effect was statistically significant within the Hispanic-Latino subgroup despite its smaller size (72 with uncontrolled asthma at baseline).
Lower engagement may explain differing outcomes by race. We speculate that lower engagement among African American participants may have been observed for 3 reasons. First, while DASM was designed for a diverse population, some content may have been inconsistent with cultural norms and preferences of African American participants. Second, the documented trust deficit among African American individuals in clinical research60,61,62 may have diminished enthusiasm for DASM. Finally, DASM did not offer opportunities to partner with familiar peers or organizations. Programs that offer such opportunities may better establish trust and cultivate engagement.63,64,65
Findings reveal an opportunity to adapt the DASM program for more consistent effects throughout diverse populations. Future research is needed—quantitative and qualitative—to elucidate causes of differences in outcomes. Community-based participatory research66 may be useful to deepen partnerships in relevant communities and identify promising strategies for program adaptation.67 Future directions also include expanded reporting on the use of health care services and its association with asthma symptom control.
Limitations
This study has some limitations. The iPhone ownership requirement and exclusion of uninsured patients may limit generalizability, although representation from 41 states and 2 coverage types may improve generalizability. Lack of blinding is an additional limitation. Biased missingness remains possible despite robustness of findings in multiple imputation sensitivity analyses.
Conclusions
In this randomized clinical trial, a DASM program improved asthma symptom control on average, with a weaker effect observed for African American participants. Differences in treatment effects may relate to varying engagement. Future work is needed to adapt the program with the goal to confer benefit throughout diverse populations.
Trial Protocol
eTable 1. Eligibility Criteria
eTable 2. Patient-Reported Outcome Measures and Engagement Metrics
eTable 3. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline
eTable 4. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline, by Race
eTable 5. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline, by Ethnicity
eTable 6. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline, by Coverage Type
eTable 7. Baseline Characteristics of Participants With Controlled Asthma at Baseline
eTable 8. Participant Counts by US State
eTable 9. Counts of Asthma Control Test (ACT) Completions by Time, Trial Arm, and Subgroup, for Participants With Uncontrolled Asthma at Baseline
eTable 10. Adjusted Asthma Control Test (ACT) Outcomes by Subgroup, for Participants With Uncontrolled Asthma at Baseline (Primary Analysis)
eTable 11. Treatment Effects by Subgroup in the Primary (Linear Mixed-Effects) Analysis vs Longitudinal Analysis of Covariance
eTable 12. Treatment Effects by Subgroup in Primary, Per-Protocol, and Multiple Imputation Analyses
eTable 13. Symptom Logs per Participant-Month, With and Without Adjustment for All Subgroup Variables
eTable 14. App Opens per Participant-Month, With and Without Adjustment for All Subgroup Variables
eTable 15. Counts of Asthma Control Test (ACT) Completions by Time and Arm, for Participants With Controlled Asthma at Baseline
eTable 16. Adjusted Asthma Control Test (ACT) Outcomes for Participants With Controlled Asthma at Baseline
eFigure 1. Geographic Distribution of Participants
eFigure 2. App Open Counts by Subgroup
eFigure 3. ACT Change by Engagement Level
eFigure 4. Secondary PRO Results at 12 Months
eMethods. Participants, Study Design, and Analysis
eResults. Additional Findings
eDiscussion. Subgroup Analyses
eAppendix. Description of the Digital Asthma Self-Management Program
eReferences.
Data Sharing Statement
References
- 1.Summary Health Statistics: National Health Interview Survey, 2018. Accessed June 21, 2021. https://ftp.cdc.gov/pub/Health_Statistics/NCHS/NHIS/SHS/2018_SHS_Table_A-2.pdf
- 2.Nurmagambetov T, Kuwahara R, Garbe P. The economic burden of asthma in the United States, 2008-2013. Ann Am Thorac Soc. 2018;15(3):348-356. doi: 10.1513/AnnalsATS.201703-259OC [DOI] [PubMed] [Google Scholar]
- 3.Center for Disease Control and Prevention . AsthmaStats: uncontrolled asthma among adults, 2019. August 12, 2022. Accessed June 7, 2025. https://archive.cdc.gov/www_cdc_gov/asthma/asthma_stats/uncontrolled-asthma-adults-2019.htm
- 4.Trevor JL, Deshane JS. Refractory asthma: mechanisms, targets, and therapy. Allergy. 2014;69(7):817-827. doi: 10.1111/all.12412 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Global Initiative for Asthma (GINA) . Difficult-to-treat & severe asthma in adolescent and adult patients. April 2019. Accessed June 20, 2021. https://ginasthma.org/wp-content/uploads/2019/04/GINA-Severe-asthma-Pocket-Guide-v2.0-wms-1.pdf
- 6.Ainsworth B, Greenwell K, Stuart B, et al. Feasibility trial of a digital self-management intervention “My Breathing Matters” to improve asthma-related quality of life for UK primary care patients with asthma. BMJ Open. 2019;9(11):e032465. doi: 10.1136/bmjopen-2019-032465 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Newhouse N, Martin A, Jawad S, et al. Randomised feasibility study of a novel experience-based internet intervention to support self-management in chronic asthma. BMJ Open. 2016;6(12):e013401. doi: 10.1136/bmjopen-2016-013401 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ljungberg H, Carleborg A, Gerber H, et al. Clinical effect on uncontrolled asthma using a novel digital automated self-management solution: a physician-blinded randomised controlled crossover trial. Eur Respir J. 2019;54(5):1900983. doi: 10.1183/13993003.00983-2019 [DOI] [PubMed] [Google Scholar]
- 9.Kandola A, Edwards K, Straatman J, Dührkoop B, Hein B, Hayes JF. Randomized attention-placebo controlled trial of a digital self-management platform for adult asthma. meRxiv. Preprint published online July 23, 2023. [DOI] [PMC free article] [PubMed]
- 10.Ramsey RR, Plevinsky JM, Kollin SR, Gibler RC, Guilbert TW, Hommel KA. Systematic review of digital interventions for pediatric asthma management. J Allergy Clin Immunol Pract. 2020;8(4):1284-1293. doi: 10.1016/j.jaip.2019.12.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hui CY, Walton R, McKinstry B, Jackson T, Parker R, Pinnock H. The use of mobile applications to support self-management for people with asthma: a systematic review of controlled studies to identify features associated with clinical effectiveness and adherence. J Am Med Inform Assoc. 2017;24(3):619-632. doi: 10.1093/jamia/ocw143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Schulte MHJ, Aardoom JJ, Loheide-Niesmann L, Verstraete LLL, Ossebaard HC, Riper H. Effectiveness of eHealth interventions in improving medication adherence for patients with chronic obstructive pulmonary disease or asthma: systematic review. J Med Internet Res. 2021;23(7):e29475. doi: 10.2196/29475 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tong X, Zhang X, Wang M, et al. Non-pharmacological interventions for asthma prevention and management across the life course: umbrella review. Clin Transl Allergy. 2024;14(3):e12344. doi: 10.1002/clt2.12344 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Papadopoulou K, Gregory I. Digital interventions to improve adherence to maintenance medication in asthma. Clin Exp Allergy. 2023;53(6):605-609. doi: 10.1111/cea.14351 [DOI] [PubMed] [Google Scholar]
- 15.Global Initiative for Asthma . 2023 GINA main report. Updated July 10, 2023. Accessed November 3, 2024. https://ginasthma.org/2023-gina-main-report/
- 16.Himes BE, Leszinsky L, Walsh R, Hepner H, Wu AC. Mobile health and inhaler-based monitoring devices for asthma management. J Allergy Clin Immunol Pract. 2019;7(8):2535-2543. doi: 10.1016/j.jaip.2019.08.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cardet JC, Louisias M, King TS, et al. ; Vitamin D Add-On Therapy Enhances Corticosteroid Disparities Working Group members on behalf of the AsthmaNet investigators . Income is an independent risk factor for worse asthma outcomes. J Allergy Clin Immunol. 2018;141(2):754-760.e3. doi: 10.1016/j.jaci.2017.04.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Sahni S, Talwar A, Khanijo S, Talwar A. Socioeconomic status and its relationship to chronic respiratory disease. Adv Respir Med. 2017;85(2):97-108. doi: 10.5603/ARM.2017.0016 [DOI] [PubMed] [Google Scholar]
- 19.Kozyrskyj AL, Kendall GE, Jacoby P, Sly PD, Zubrick SR. Association between socioeconomic status and the development of asthma: analyses of income trajectories. Am J Public Health. 2010;100(3):540-546. doi: 10.2105/AJPH.2008.150771 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.US Department of Health and Human Services, Office of Minority Health . Asthma and Hispanic Americans. Updated February 13, 2025. Accessed June 07, 2025. https://minorityhealth.hhs.gov/asthma-and-hispanic-americans
- 21.Grant T, Croce E, Matsui EC. Asthma and the social determinants of health. Ann Allergy Asthma Immunol. 2022;128(1):5-11. doi: 10.1016/j.anai.2021.10.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Asthma and Allergy Foundation of America . Asthma disparities in America: a roadmap to reducing burden on racial and ethnic minorities. 2020. Accessed June 8, 2023. https://aafa.org/wp-content/uploads/2022/08/asthma-disparities-in-america-burden-by-race-ethnicity-executive-summary.pdf
- 23.Nazario S, González-Sepúlveda L, Telón-Sosa B, Suárez-Pérez EL. Inequalities in asthma mortality by ethnicity and race in the United States and Puerto Rico. J Allergy Clin Immunol Pract. 2022;10(8):2178-2180. doi: 10.1016/j.jaip.2022.04.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Redmond C, Akinoso-Imran AQ, Heaney LG, Sheikh A, Kee F, Busby J. Socioeconomic disparities in asthma health care utilization, exacerbations, and mortality: a systematic review and meta-analysis. J Allergy Clin Immunol. 2022;149(5):1617-1627. doi: 10.1016/j.jaci.2021.10.007 [DOI] [PubMed] [Google Scholar]
- 25.Centers for Disease Control and Prevention . Asthma Surveillance—United States, 2006-2018. September 17, 2021. Accessed February 24, 2023. https://stacks.cdc.gov/view/cdc/109086
- 26.Greiner B, Cronin K, Salazar L, Hartwell M. Asthma-related disparities in emergency department use and clinical outcomes among Spanish-speaking Hispanic patients. Ann Allergy Asthma Immunol. 2023;130(2):254-255. doi: 10.1016/j.anai.2022.11.002 [DOI] [PubMed] [Google Scholar]
- 27.CDC . Healthcare Use Data 2020. Accessed May 29, 2023. https://www.cdc.gov/asthma/healthcare-use/2020/table_a.html
- 28.Chin MH, Clarke AR, Nocon RS, et al. A roadmap and best practices for organizations to reduce racial and ethnic disparities in health care. J Gen Intern Med. 2012;27(8):992-1000. doi: 10.1007/s11606-012-2082-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nanda A, Siles R, Park H, et al. Ensuring equitable access to guideline-based asthma care across the lifespan: tips and future directions to the successful implementation of the new NAEPP 2020 guidelines, a Work Group Report of the AAAAI Asthma, Cough, Diagnosis, and Treatment Committee. J Allergy Clin Immunol. 2023;151(4):869-880. doi: 10.1016/j.jaci.2023.01.017 [DOI] [PubMed] [Google Scholar]
- 30.Codispoti CD, Greenhawt M, Oppenheimer J. The role of access and cost-effectiveness in managing asthma: a systematic review. J Allergy Clin Immunol Pract. 2022;10(8):2109-2116. doi: 10.1016/j.jaip.2022.04.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Harper LJ, Kidambi P, Kirincich JM, Thornton JD, Khatri SB, Culver DA. Health disparities: interventions for pulmonary disease—a narrative review. Chest. 2023;164(1):179-189. doi: 10.1016/j.chest.2023.02.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Deloitte Insights . Medicaid and digital health. September 7, 2018. Accessed April 1, 2021. https://www2.deloitte.com/us/en/insights/industry/public-sector/mobile-health-care-app-features-for-patients.html
- 33.Pew Research Center . Mobile Fact Sheet. November 13, 2024. Accessed June 07, 2025. https://www.pewresearch.org/internet/factsheet/mobile/
- 34.Harris B, Silberman J, Sarlati S, et al. A digital asthma self-management tool reduced emergency visit rates in a Medicaid population. Ann Allergy Asthma Immunol. 2023;131(5):S230-S231. doi: 10.1016/j.anai.2023.10.021 [DOI] [Google Scholar]
- 35.Silberman J, Sarlati S, Harris B, et al. A digital approach to asthma self-management in adults: protocol for a pragmatic randomized controlled trial. Contemp Clin Trials. 2022;122:106902. doi: 10.1016/j.cct.2022.106902 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Apostolaros M, Babaian D, Corneli A, et al. Legal, regulatory, and practical issues to consider when adopting decentralized clinical trials: recommendations from the clinical trials transformation initiative. Ther Innov Regul Sci. 2020;54(4):779-787. doi: 10.1007/s43441-019-00006-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ford I, Norrie J. Pragmatic trials. N Engl J Med. 2016;375(5):454-463. doi: 10.1056/NEJMra1510059 [DOI] [PubMed] [Google Scholar]
- 38.Sox HC, Lewis RJ. Pragmatic trials: practical answers to “real world” questions. JAMA. 2016;316(11):1205-1206. doi: 10.1001/jama.2016.11409 [DOI] [PubMed] [Google Scholar]
- 39.Asthma care quick reference: diagnosing and managing asthma. Revised September 2012. Accessed October 21, 2023. https://www.nhlbi.nih.gov/files/docs/guidelines/asthma_qrg.pdf
- 40.Kim SH. A systematic review on visualizations for self-generated health data for daily activities. Int J Environ Res Public Health. 2022;19(18):11166. doi: 10.3390/ijerph191811166 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Choe EK, Abdullah S, Rabbi M, et al. Semi-automated tracking: a balanced approach for self-monitoring applications. IEEE Pervasive Comput. 2017;16(01):74-84. doi: 10.1109/MPRV.2017.18 [DOI] [Google Scholar]
- 42.Nathan RA, Sorkness CA, Kosinski M, et al. Development of the asthma control test: a survey for assessing asthma control. J Allergy Clin Immunol. 2004;113(1):59-65. doi: 10.1016/j.jaci.2003.09.008 [DOI] [PubMed] [Google Scholar]
- 43.Schatz M, Sorkness CA, Li JT, et al. Asthma Control Test: reliability, validity, and responsiveness in patients not previously followed by asthma specialists. J Allergy Clin Immunol. 2006;117(3):549-556. doi: 10.1016/j.jaci.2006.01.011 [DOI] [PubMed] [Google Scholar]
- 44.Cloutier MM, Schatz M, Castro M, et al. Asthma outcomes: composite scores of asthma control. J Allergy Clin Immunol. 2012;129(3)(suppl):S24-S33. doi: 10.1016/j.jaci.2011.12.980 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Juniper EF, O’Byrne PM, Guyatt GH, Ferrie PJ, King DR. Development and validation of a questionnaire to measure asthma control. Eur Respir J. 1999;14(4):902-907. doi: 10.1034/j.1399-3003.1999.14d29.x [DOI] [PubMed] [Google Scholar]
- 46.Schatz M, Kosinski M, Yarlas AS, Hanlon J, Watson ME, Jhingran P. The minimally important difference of the Asthma Control Test. J Allergy Clin Immunol. 2009;124(4):719-23.e1. doi: 10.1016/j.jaci.2009.06.053 [DOI] [PubMed] [Google Scholar]
- 47.Kripalani S, Risser J, Gatti ME, Jacobson TA. Development and evaluation of the Adherence to Refills and Medications Scale (ARMS) among low-literacy patients with chronic disease. Value Health. 2009;12(1):118-123. doi: 10.1111/j.1524-4733.2008.00400.x [DOI] [PubMed] [Google Scholar]
- 48.Mixon AS, Myers AP, Leak CL, et al. Characteristics associated with postdischarge medication errors. Mayo Clin Proc. 2014;89(8):1042-1051. doi: 10.1016/j.mayocp.2014.04.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Center for Evidence-Based Practices, Case Western Reserve University . Readiness ruler. 2010. Accessed February 1, 2020. https://case.edu/socialwork/centerforebp/resources/readiness-ruler
- 50.Benson T, Potts HWW, Bark P, Bowman C. Development and initial testing of a Health Confidence Score (HCS). BMJ Open Qual. 2019;8(2):e000411. doi: 10.1136/bmjoq-2018-000411 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Bousquet J, VandenPlas O, Bewick M, et al. The Work Productivity and Activity Impairment Allergic Specific (WPAI-AS) questionnaire using mobile technology: the MASK study. J Investig Allergol Clin Immunol. 2018;28(1):42-44. doi: 10.18176/jiaci.0197 [DOI] [PubMed] [Google Scholar]
- 52.Kosinski M, Bayliss MS, Turner-Bowker DM, Fortin EW. Asthma Control Test: A User’s Guide. QualityMetric Inc; 2008. [Google Scholar]
- 53.Sherry AD, Hahn AW, McCaw ZR, et al. Differential treatment effects of subgroup analyses in phase 3 oncology trials from 2004 to 2020. JAMA Netw Open. 2024;7(3):e243379. doi: 10.1001/jamanetworkopen.2024.3379 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Newgard CD, Lewis RJ. Missing data: how to best account for what is not known. JAMA. 2015;314(9):940-941. doi: 10.1001/jama.2015.10516 [DOI] [PubMed] [Google Scholar]
- 55.Little R, Rubin R. Statistical Analysis With Missing Data. 2nd ed. Wiley; 2002. doi: 10.1002/9781119013563 [DOI] [Google Scholar]
- 56.Twisk J, Bosman L, Hoekstra T, Rijnhart J, Welten M, Heymans M. Different ways to estimate treatment effects in randomised controlled trials. Contemp Clin Trials Commun. 2018;10:80-85. doi: 10.1016/j.conctc.2018.03.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Stata multiple imputation . Release 16. StataCorp LLC; 2019. Accessed November 19, 2023. https://www.stata.com/manuals16/mi.pdf
- 58.Rubin DB. Multiple imputation for nonresponse in surveys. Wiley Series in Probability and Statistics. Wiley; 1987. Accessed November 9, 2024. https://onlinelibrary.wiley.com/doi/book/10.1002/9780470316696
- 59.Klein D. MIMRGNS: Stata module to run margins after mi estimate. Statistical Software Components. July 25, 2022. Accessed November 9, 2024. https://ideas.repec.org//c/boc/bocode/s457795.html
- 60.Melikam ES, Magwood GS, Ford M, et al. Community trust, attitudes and preferences related to participation in cancer research in South Carolina. J Community Health. 2024;49(1):100-107. doi: 10.1007/s10900-023-01251-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61.Stallings SC, Cunningham-Erves J, Frazier C, et al. Development and validation of the perceptions of research trustworthiness scale to measure trust among minoritized racial and ethnic groups in biomedical research in the US. JAMA Netw Open. 2022;5(12):e2248812. doi: 10.1001/jamanetworkopen.2022.48812 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.Scharff DP, Mathews KJ, Jackson P, Hoffsuemmer J, Martin E, Edwards D. More than Tuskegee: understanding mistrust about research participation. J Health Care Poor Underserved. 2010;21(3):879-897. doi: 10.1353/hpu.0.0323 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 63.Victor RG, Lynch K, Li N, et al. A cluster-randomized trial of blood-pressure reduction in Black barbershops. N Engl J Med. 2018;378(14):1291-1301. doi: 10.1056/NEJMoa1717250 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Wippold GM, Frary SG, Abshire D, Wilson DK. Peer-to-peer health promotion interventions among African American men: a scoping review protocol. Syst Rev. 2021;10(1):184. doi: 10.1186/s13643-021-01737-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Palmer KNB, Rivers PS, Melton FL, et al. Health promotion interventions for African Americans delivered in US barbershops and hair salons—a systematic review. BMC Public Health. 2021;21(1):1553. doi: 10.1186/s12889-021-11584-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 66.Chen E, Leos C, Kowitt SD, Moracco KE. Enhancing community-based participatory research through human-centered design strategies. Health Promot Pract. 2020;21(1):37-48. doi: 10.1177/1524839919850557 [DOI] [PubMed] [Google Scholar]
- 67.Sankaré IC, Bross R, Brown AF, et al. Strategies to build trust and recruit African American and Latino community residents for health research: a cohort study. Clin Transl Sci. 2015;8(5):412-420. doi: 10.1111/cts.12273 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol
eTable 1. Eligibility Criteria
eTable 2. Patient-Reported Outcome Measures and Engagement Metrics
eTable 3. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline
eTable 4. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline, by Race
eTable 5. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline, by Ethnicity
eTable 6. Baseline Characteristics of Participants With Uncontrolled Asthma at Baseline, by Coverage Type
eTable 7. Baseline Characteristics of Participants With Controlled Asthma at Baseline
eTable 8. Participant Counts by US State
eTable 9. Counts of Asthma Control Test (ACT) Completions by Time, Trial Arm, and Subgroup, for Participants With Uncontrolled Asthma at Baseline
eTable 10. Adjusted Asthma Control Test (ACT) Outcomes by Subgroup, for Participants With Uncontrolled Asthma at Baseline (Primary Analysis)
eTable 11. Treatment Effects by Subgroup in the Primary (Linear Mixed-Effects) Analysis vs Longitudinal Analysis of Covariance
eTable 12. Treatment Effects by Subgroup in Primary, Per-Protocol, and Multiple Imputation Analyses
eTable 13. Symptom Logs per Participant-Month, With and Without Adjustment for All Subgroup Variables
eTable 14. App Opens per Participant-Month, With and Without Adjustment for All Subgroup Variables
eTable 15. Counts of Asthma Control Test (ACT) Completions by Time and Arm, for Participants With Controlled Asthma at Baseline
eTable 16. Adjusted Asthma Control Test (ACT) Outcomes for Participants With Controlled Asthma at Baseline
eFigure 1. Geographic Distribution of Participants
eFigure 2. App Open Counts by Subgroup
eFigure 3. ACT Change by Engagement Level
eFigure 4. Secondary PRO Results at 12 Months
eMethods. Participants, Study Design, and Analysis
eResults. Additional Findings
eDiscussion. Subgroup Analyses
eAppendix. Description of the Digital Asthma Self-Management Program
eReferences.
Data Sharing Statement
