Asthma and chronic obstructive pulmonary disease (COPD) each affect approximately 330 million people worldwide, representing major public health challenges with substantial health and economic burdens [1,2]. Self-management interventions are key to mitigate the risk of symptom exacerbations in patients with asthma and COPD from escalating into life-threatening emergencies. To better support self-management, wearable technology can enable remote monitoring and early exacerbation detection, allowing patients to act before symptoms worsen [3-5].

Effective respiratory monitoring requires specialized sensing technologies that address unique challenges. While surface body sensors accurately capture vital signs including heart rate, skin temperature, and breathing patterns, continuous acoustic monitoring raises privacy concerns due to incidental speech recording [6-8]. Neck surface accelerometers (NSAs) provide an alternative approach, capturing purely mechano-acoustic signals without recording speech, making them suitable for long-term chronic disease monitoring [9-14]. Recent studies demonstrate that NSA-based systems can detect voice physiology, swallowing events, and COPD severity indicators with 93% to 96% accuracy using machine learning approaches [15-17].

The success of wearable systems depends on reliable mobile interfaces. In ubiquitous health solutions, raw sensor data (eg, from NSA) are stored onboard before transmission via Bluetooth Low Energy (BLE) to mobile devices for local processing or cloud-based analysis. Results are delivered through mobile health (mHealth) apps that often include interactive features such as self-assessment questionnaires and chatbots to promote healthier behaviors [18-21]. However, despite the increasing use of Android platforms for mHealth development, many studies lack comprehensive documentation of development workflows, design decisions, and validation procedures [22]. This transparency gap hinders clinical adoption, as mHealth apps serve as the primary interface between patients and wearable devices.

Clinical research shows that wearable tools can increase medication adherence by 50% and improve disease-related knowledge by 52% [23]. However, Cochrane systematic reviews report mixed benefits for mHealth interventions targeting patients with asthma and COPD specifically.

Belisario et al [24] evaluated smartphone and tablet apps versus paper-based asthma action plans in 2 randomized controlled trials (n=408 participants) over 6 months. While emergency visits decreased significantly, symptom control showed no improvement on the patient-reported Asthma Control Questionnaire (ACQ). Likewise, mixed results are also noted in COPD research. Janjua et al [25] reviewed 14 randomized controlled trials (n=1518 participants) comparing mobile-based self-management to usual care. Evidence to date suggested that digital technology might provide short-term but not long-term improvements in patients’ quality of life. Nevertheless, these 2 Cochrane reports noted key methodological limitations that undermined the evidence quality in asthma and COPD literature, such as heavy reliance on self-reported outcomes, unregistered protocols, and underpowered statistics from small sample sizes [24,25].

Despite that many mHealth solutions are available, most of them remain inadequate to fully support self-management interventions in asthma and COPD [26]. General-purpose wearables such as the Apple Watch [27] and Fitbit [28] provide basic physiological metrics but lack the disease-specific algorithms and personalized interventions essential for respiratory disease management. Specialized devices such as Afflo [29] and AcuPebble [30] capture respiratory signals but do not integrate data into self-management plans or provide timely exacerbation alerts. Stand-alone apps like Breathe [31], Sonde Health [32], and CoughPro [33] rely on smartphone microphones and questionnaires for symptom detection. These methods capture only intermittent snapshots of respiratory status, rather than providing continuous monitoring capabilities.

More importantly, most mHealth solutions did not disclose how well their apps adhere to regulatory frameworks. Multiple standards govern mHealth apps, including the European Commission’s Green Paper on Mobile Health [34] and the US Food and Drug Administration’s Mobile Medical Applications Guidance [35], which establish privacy, encryption, and data-transfer requirements. Beyond regulatory compliance, developers shall follow established design principles found in industrial guidelines such as Apple’s Human Interface Guidelines [36] and Android’s Material Design Guidelines [37]. Besides, apps handling protected health information must comply with the Health Insurance Portability and Accountability Act (HIPAA) or equivalent regional privacy regulations [38]. To address these multiple requirements, Llorens-Vernet and Miró [39] developed a comprehensive framework that consolidates these diverse requirements into 8 unified development criteria: usability, privacy, security, appropriateness and suitability, transparency and content, safety, technical support and updates, and technology.

The aforesaid research gaps, namely, suboptimal clinical evidence, technical constraints, and inadequate development documentation, underscored the need for more rigorously developed mHealth solutions in respiratory disease monitoring. Our team previously developed privacy-preserving NSA-enabled wearable sensors for airway symptom detection [12,13,25,40]. However, these sensors require a companion mobile app to enable data visualization, cloud storage, user interaction, and so on. This study aimed to address these technological gaps by developing and evaluating a mobile app using transparent, framework-guided development processes.

This study presented the development and evaluation of AIrway, an mHealth app for asthma and COPD management. Research objectives were (1) to develop the app that would collect patient-reported outcomes, environmental data, and peripheral accelerometer sensor data; (2) to evaluate the functionality of the app including wireless connectivity, cloud storage, and literacy analysis; and (3) to perform a pilot usability study with technical raters, namely, app developers.

This study consisted of 2 phases, namely, frontend and backend development (phase 1) and a pilot usability test (phase 2). The AIrway development followed rigorous academic and industry best practices to ensure compliance with the mHealth development standard. Comprehensive documentation of app requirements, design, implementation and validation, and benchmark were reported against the framework of Llorens-Vernet and Miró [39].

The backend development was programmed using Java (version 11.0.12) [76] in Android Studio (Chipmunk 2021.2.1 version) [42] to implement (1) local weather data retrieval, (2) asthma and COPD control zone calculations, (3) cloud database storage and retrieval, and (4) BLE data transmission.

A pilot usability study with technical raters (app developers) was performed to evaluate the app’s pilot functionality.

The human study was approved by the institutional review board of the Faculty of Medicine and Health Sciences at McGill University (protocol A12-E39-22B). Informed consent was obtained from all participants. Participant data were deidentified and anonymized for privacy protection. Each received a US $35 Amazon gift card as compensation.

AIrway met 7 of the 8 best practices for mHealth app development (Table 2). Usability compliance was confirmed through a technical-expert evaluation and strict adherence to Android design guidelines. As the app targeted English-speaking adults with respiratory conditions, multilingual support was not included.

Privacy is protected by a comprehensive policy that covers data collection, terms of use, and secure cloud storage. Security relies on password-protected Google Firestore access, although broader cyberthreat defenses remain to be implemented. The app’s appropriateness for English-speaking adults with respiratory conditions was demonstrated via targeted performance testing.

Transparency requirements were satisfied through dedicated Help and FAQ sections providing author credentials, mission statements, funding information, and scientific evidence supporting the asthma and COPD control zone calculations. Safety protocols included explicit disclaimers regarding clinical diagnosis limitations and emergency response guidance for users experiencing severe symptoms.

Technical support is provided via monitored email channels and app updates distributed through Firebase App Distribution. The only area needing enhancement is technology—specifically, adding cross-platform compatibility and offline functionality, since AIrway is currently Android-only and requires a constant internet connection.

Based on the readability calculator [69], the Flesch Reading Ease score [103] for the privacy policy, FAQs, the asthma action plan, and the COPD action plan were 58.63, 56.69, 57.72, and 59.91, respectively. With an average score near 60 (1.4), these materials are accessible to users with a formal education level of grades 9-10. The Gunning Fog Index [104] indicated an 8th- to 10th-grade reading level, while the Coleman Liau Index [105] suggested suitability for a 7th- to 10th-grade audience. The Flesch-Kincaid Grade Level [106] indicated a 6th- to 8th-grade range, and the Automated Readability Index [72] estimated a 5th- to 7th-grade level. The Simple Measure of Gobbledygook Index measured the texts based on syllables [107] and concluded that the materials were suitable for a 9th- to 10th-grade audience (Table 3).

The accessibility analysis identified 3 primary areas for further enhancing the app’s UIs (Multimedia Appendix 7). First, it recommended modifying the width layout of test view components from a fixed value to an adjustable parameter, such as wrap_content. This modification enabled dynamic text expansion, which could prevent component overlap when screen sizes change. Second, it suggested increasing the text size of input items from the current range of 24-44 density-independent pixels to a minimum of 48 density-independent pixels for better readability. Third, it recommended enhancing the contrast colors of the “Today,” “Report,” “Profile,” and “Help” icons to achieve a minimum ratio of 3.00:1 as well as the contrast of unselected time-scrolling texts within the diary reminder interface to at least 4.50:1 with respect to the background.

The BLE experiment evaluated data transmission between AIrway (client) and the nRF52840 development board (peripheral; Tables 4 and 5). First, the app’s scan feature accurately detected the board by its device name, “NSA BLE,” along with the unique manufacturing MAC address (D5:D2:5D:62:AF:48). Signal strength, measured in decibel-milliwatts on a logarithmic scale (with values closer to 0 dBm indicating stronger signals [109]), showed that the board was advertising the highest signal value of –19 dBm among the neighboring BLE devices, which confirmed its closest proximity to the app.

Once connected, the data read characteristic feature was successfully displayed and verified visually. As described in the experiment setup, simulated battery level data were transmitted using random integers and set to decrement by 1 every 2 seconds. The app accurately retrieved the real-time data timestamp along with the corresponding hexadecimal data value (eg, 0x0D representing integer 13, followed by 0x0C representing integer 12 after 2 seconds). This result confirmed the successful integration of the scan, connect, and read characteristic features.

Furthermore, physical distance, connection time, and obstacles were found to impact BLE performance (Table 6). The best performance for stable and continuous data streaming occurred when the devices were adjacent without obstacles. However, after connections exceeded 10 minutes, some packet loss was observed. Additionally, occasional disconnections, particularly when they were farther apart, were likely due to interference from nearby BLE devices. In these cases, the app was able to manually reconnect to the board.

A total of 5 app developers participated in the evaluation of AIrway usability. Most technical raters were female, identified as visible minorities, and resided in Canada. They held at least a bachelor degree, worked part-time, and had an average age of 29.0 (SD 5.61) years. The majority were computer science PhD students or postdoctoral fellows with 1-2 years of Android mobile app development experience, and each had developed at least 1 app before (Table 7).

The overall mean uMARS score for AIrway was 3.6 of 5.0 (SD 0.2), indicating above-average quality (>3.0; Figure 5A). The information quality section received the highest mean score of 4.4 of 5.0 (SD 0.2), which implied that the app’s quality and visual information were highly appealing to users. Similarly, the functionality section also received the highest mean score of 4.2 of 5.0 (SD 0.4), which indicated excellent performance, ease of use, navigation, and gestural design for the app. The aesthetics section received a mean score of 3.5 of 5.0 (SD 0.6), while the engagement section received a mean score of 3.4 of 5.0 (SD 0.4). Both the aesthetics and engagement sections indicated appropriate layout elements and customization aspects. Among all the sections, the subjective quality section received a score of 2.6 of 5.0 (SD 0.3), which indicated a moderate willingness to use the app in the next 12 months and pay for its use.

The IQVIA functionality score ranged from 6 to 11, with a median of 8 (IQR 7-10; Figure 5B). All technical raters confirmed that the app included recording, data collection, intervention, display, and alert functions (5/5, 100%). Most agreed (4/5, 80%) that the user data could be evaluated by others, and over half (3/5, 60%) expressed that the app supported instructing, guiding, and sharing data functions. However, only 1 (20%) rater observed a communication feature that allowed users to interact with others via social networks and provided information in a variety of formats (eg, text, photo, and video).

Feedback was collected through 5 open-ended questions regarding the app’s engagement, functionality, aesthetics, information quality, and potential improvements. Overall, the technical raters expressed positive impressions of the app and a clear understanding of its purpose. The results are summarized in Multimedia Appendix 8. A key suggestion was to add an option to show and hide passwords in the “Enter a password” and “Retype Your Password” fields on the register interface, which would help users verify that their passwords match.

In recent years, mHealth apps targeting asthma monitoring have doubled from 2015 to 2019 [110], and COPD apps increased 6-fold between 2017 and 2023 [111,112]. The SARS-CoV-2 pandemic has notably increased clinicians’ and patients’ familiarity and positive attitudes toward self-monitoring and home-monitoring apps [113]. COVID-19 pandemic restrictions, such as social distancing mandates, shortages of protective equipment, and concerns about disease exposure, had increased the use and adoption of telehealth and remote monitoring technology from 11% to 46% among patients and health care providers [114]. In a study of 162 patients with acute COVID-19 infections, almost 79% of them were reported to engage in the digital platform of remote patient monitoring [114]. Further, to respond to this pandemic, the US Centers for Medicare and Medicaid Services approved over 80 new telehealth services and began reimbursement for telehealth visits since 2020 [114,115].

In this study, AIrway was designed to support NSA wearable devices for active asthma and COPD self-management purposes. AIrway did not provide diagnostic recommendations or treatment prescriptions to users and was not intended for medical use. Target user characteristics and needs were incorporated into the AIrway app UI design. For instance, the content of AIrway including the privacy policy, FAQ, and action plan was written at a 9th- to 10th-grade reading level that would enhance the comprehensibility and accessibility to general and geriatric users. Meanwhile, about 6.4% of patients with COPD or asthma using bronchodilator medications (eg, β₂-adrenergic agonist bronchodilators) experience hand tremors, which can impair their ability to navigate touchscreen interfaces effectively [116]. The AIrway app has implemented the recommended text sizes (20 scalable pixels for titles, 16 scalable pixels for subtitles, and 14 scalable pixels for buttons and body) and constraint layout to enhance accessibility and adaptability across various screen sizes, following Android Material Design Guidelines. However, these measures were limited, as the app did not include a dynamic font sizing feature to adapt to individual user needs.

BLE tests with the nRF52840 development board demonstrated reliable real-time data exchange and accurate decoding of hexadecimal values. These results support future integration with NSA wearables using nRF5 series microcontrollers. Meanwhile, performance tests revealed that distance, connection time, and obstacles affect BLE stability. Future work will prioritize optimizing critical parameters like the maximum transmission unit and transmission frequency while implementing auto-reconnection functionality to minimize packet loss.

In the usability evaluation, AIrway scored 3.6 of 5.0 on the uMARS and a median IQVIA functionality score of 8 of 11 (IQR 7-10). These results were compared favorably with commercial mHealth apps for respiratory, cardiac, and sleep disorders, which typically report scores ranging from 3.0 to 4.2 on the Mobile App Rating Scale and 6 to 10 on IQVIA [101,116-118]. However, the lower uMARS subjective quality score (2.6/5.0) indicated areas for enhancement. Technical raters valued AIrway’s utility for condition management but showed limited interest in long-term use or payment. Future work will prioritize clinical validation studies co-designed with people living with asthma or COPD and their clinicians, using appropriately powered trials to assess safety and demonstrate meaningful clinical benefits before broader deployment.

The usability study was also limited by a small sample size (n=5) compared to the 15-40 participants typically included in similar research [119-121]. This sampling limitation was attributed to recruitment challenges through university mailing lists and paid research groups on Facebook, which did not effectively reach the intended audience. In addition, the usability evaluation relied on subjective uMARS and IQVIA ratings, which are prone to perception variability and bias. A more comprehensive assessment could incorporate objective use metrics such as system interaction patterns (eg, login frequency and session duration), error occurrence rates, feature use statistics, and configuration preferences to provide a more accurate view of UXs and app performance.

At present, AIrway is Android-specific; thus, future development is needed for cross-platform frameworks like Flutter to improve accessibility and equity by extending availability to a broader user base. Our further accessibility review of the UIs recommended adjusting layout widths to accommodate dynamic text and increasing text sizes to a minimum of 48 density-independent pixels. Future studies should also include patients with asthma and COPD as well as their attending clinicians and caretakers to collect direct feedback from end users, ensuring that AIrway addresses the specific needs and preferences of its target populations.

This AIrway app followed mHealth best practices to ensure accessibility and seamless wireless integration with wearables. Future work will include cross-platform support and assess usability with patients to validate its clinical effectiveness.