The objective of this study is to conduct a randomized controlled trial (RCT) to compare the adapted and refined ASTHMAXcel Voice platform to usual care (UC). It is hypothesized by the investigator team that ASTHMAXcel Voice will be associated with improved clinical and process outcomes, asthma quality of life (QOL), medication adherence, and self-efficacy as compared to UC.

Poor outcomes for minority patients with asthma have been linked to poverty and other social determinants of health (SDOH), environmental exposures, and poor self-management. In a previous Agency for Healthcare Research and Quality (AHRQ)-funded study, the researchers developed and pilot tested ASTHMAXcel PRO, a mobile app that promotes self-management of asthma (NCT03847142). The app was optimized for outpatient settings and promoted asthma self-management through the collection of patient-reported outcomes (PROs), animated videos, goal setting, personalized algorithms, and push notifications. The use of the app led to significant decreases in the need for steroids, visits for asthma to the emergency department, and hospitalizations for asthma.

In this current research, ASTHMAXcel Voice, an app developed and refined during enhancement of ASTHMAXcel PRO, will make use of voice biomarkers to detect worsening symptoms. This technology uses machine learning to assess respiratory dysfunction, including asthma, based on a 6-second voice sample. From the sample, a Respiratory Symptoms Risk Score (RSRS) is calculated that correlates with the speaker's risk of respiratory impairment. The updated platform will calculate the patient's RSRS; facilitate shared decision making, screen for SDoH, and referrals; improve the ability of patients to self-manage; and allow for remote care coordination.

This program draws upon the Common Sense Model (CSM) of Self-Regulation which describes a cognitive processing system that includes situational stimuli (asthma symptoms), objective representation of the health threat (illness representations) with its treatment decision (controller medication use), and appraisal of outcomes (asthma control) for the success/failure of those treatment decisions. The model contains a feedback loop with illness representations changing over time as patients gain experience with asthma management. Social Determinants of Health (SDoH) may also affect the representation of the health threat, treatment decisions, and appraisal of outcomes. As an example, a patient with depression, a poor social support network, insecure housing, and financial stress may view asthma as an acute disease that is uncontrollable, which in turn leads to negative beliefs about medications and low self-efficacy towards asthma management. ASTHMAXcel Voice strives to shift illness representations away from the belief that asthma only exists when there are active symptoms and change behavior towards daily controller medication use over the long term to prevent asthma symptoms. Realtime feedback based on voice samples that yield a RSRS (voice biomarker) will help the patient to accurately detect perceived threats and manage asthma exacerbations during earlier stages. ASTHMAXcel Voice is also based on the SEM that addresses causes of poor asthma control across four interconnected domains: community, medical system, interpersonal, and individual level factors. ASTHMAXcel Voice is a multilevel approach to address these barriers with intervention components that are directly applied at each level.

There is growing recognition that mobile health interventions can be applied across all these levels to facilitate health behavior change through the use of push notifications and interactive educational content. ASTHMAXcel Voice works on the individual and interpersonal levels by providing targeted asthma education and push notifications to assist with medication adherence and asthma management. Worse outcomes assessed by PROs (asthma control) and voice biomarkers may heighten the perceived threat level of asthma and prompt Just-in-Time Adaptive Interventions (JITAIs) to seek out the educational content more frequently to improve asthma control. On the organizational (medical system) level, ASTHMAXcel Voice will facilitate shared decision-making and ongoing communication between the patient, Community Health Worker (CHW) or Social Worker (SW), and Health Care Provider (HCP). For example, a monthly visual dashboard display will increase HCP awareness of deteriorating trends assessed from PROs and voice biomarkers. On the community level, the CHW or SW will provide patients with SDoH relevant community resources (e.g., pest remediation services, smoking cessation programs, support groups, food pantries) to address SDoH concerns reported in the mobile platform. Finally, to inform more effective design and implementation of ASTHMAXcel Voice, the study team will use the Unified Theory of Acceptance and use of Technology (UTAUT) health IT framework in determining a user's technology acceptance and adoption behavior.

Inclusion Criteria:

• English speaking
• Persistent asthma (diagnosed by a healthcare provider) on a daily controller medication
• Able to provide informed consent
• Smartphone access (iOS or Android) with data plan

Exclusion Criteria:

• Pregnancy
• Severe psychiatric or cognitive problems that would prohibit completion of protocol

Once primary study findings have been published, deidentified data publicly available along with its associated documentation. Public Use Data: All de-identified study data that are not designed as restricted use will be made publicly available as supplementary material of articles in PubMed Central. All datasets will require detailed metadata documentation (Word documents), including (but not limited to) the methodology and procedures used to collect the data, details about codes (codebooks), definitions of variables, variable field locations, frequencies, etc. All data sets that are shared will be accompanied by the relevant metadata. No other specialized tools will be needed to access or manipulate shared scientific data to support replication or reuse.

Participant-level data from this study will be fully de-identified and represented as an OMOP-CDM v5.3.1 dataset (mapped to OHDSI/Athena vocabulary system)

Results generated from this grant will be presented at national or international scientific meetings and conferences and will be published in a timely fashion. Data sets and associated documents will be provided to the NIH study Program Official within timelines described in the NIH Policy for Data Sharing from Observational Epidemiology Studies - no later than 3 years after the completion of each examination or follow-up cycle or 2 years after the baseline, follow-up, genetic, ancillary study, or other data set is finalized within the study for analysis for use in publication, whichever comes first. Study data will be made available to other users after publication of the study's main results have been published (estimated to be within 12 months of study closure). Once available, no time limit will be placed on accessibility of the study data.

Research assistants who administer questionnaires via the Electronic Data Capture (EDC) system will be blinded. Outcomes Assessors will not be blinded.