Cystic fibrosis (CF) is the second most common inherited disorder in the United States. The disease affects multiple systems, including the respiratory, digestive, endocrine, and reproductive systems, and involves a complex, time-consuming daily care routine with multiple oral and inhaled medications, airway clearance therapy, diet, and exercise [1,2]. In the United States, clinical care guidelines recommend quarterly multidisciplinary CF clinic visits. During such visits, patients and families are asked to remember and communicate to the clinical team the most relevant aspects of their disease experience from the previous 3 months. From that snapshot, clinicians are expected to gather enough detail to make optimal treatment recommendations [3]. Although the majority of care takes place at home, exchange of information between the patient and the clinical team outside of the clinical setting is currently limited and occurs mostly via phone calls. This traditional, episodic model of care delivery is not well suited for a chronic condition such as CF and does not support optimal disease self-management, patient engagement, and shared decision making [3-6]. For optimal CF care, a bidirectional patient-clinician communication that takes place between visits is necessary. Such communication is particularly important for children and adolescents who are learning to transition from clinician-regulated care to autonomous self-management of their disease [7,8]. Ideally, the bidirectional communication would occur with electronic tracking of patient-reported symptoms, real-time sharing of this patient-generated health data controlled by the patient but matching the requirements of clinical workflows, and timely clinical feedback to such data [3,9].

Research in other complex chronic conditions has shown that patients benefit from routine collection of patient-generated health data. Studies have reported associations between the use of patient-generated health data and improved symptom control and quality of life [10-14], patient-clinician communication and satisfaction [15,16], reduced health care utilization, and increased survival [17-20]. Such an approach is not yet used in CF clinical practice, despite a strongly stated need by the CF patient community [21-23]. A previous assessment of preferences for remote collection and sharing of patient-generated health data among patients with CF or their caregivers reported that the majority are willing to share such data [24].

The opportunities for collaborative, continuous, and data-driven chronic care have increased exponentially with the application of mobile health (mHealth) technologies. Mobile apps allow patients to collect, monitor, and report symptoms and events outside of an office visit and to make these data available to providers for shared decision making. mHealth interventions have been widely used in the management of diabetes, hypertension, and chronic pain, among others [13,14,20]. Although technology has been used to facilitate earlier detection of pulmonary exacerbations and adherence monitoring in CF care [25-27], mHealth solutions for CF self-management have not been adopted [28,29]. CF care is uniquely suitable for mHealth solutions because of well-established clinical guidelines, multidisciplinary care teams, national patient registry, and an expansive network of accredited CF centers; however, barriers related to the integration of patient-generated health data with information technology (IT) systems and electronic medical records have been documented [21]. In addition, the few existing CF apps only partially satisfy the user needs reported in previous research [29,30]. mHealth solutions are even more important in the context of the COVID-19 pandemic, which has imposed limitations on in-person visits for chronic care. As such, mHealth approaches may be used to supplement telehealth as a risk-reduction strategy [31].

Genia is a mHealth platform designed to facilitate the coproduction of CF care (Figure 1). Developed by Upstream Dream in collaboration with the Swedish CF community at Karolinska Institute in 2014, it was first introduced to pediatric CF centers in Sweden and has been well received by patients, with 65% to 87% of families using the app at the most active clinics to track symptoms, activities, and aspects of daily care and share them with the clinical team [32].

Genia supports the upload of images (eg, a patient may take a photo of their sputum and upload it via the app) and can connect to other apps or devices, such as electronic spirometers and smart medication dispensers. A proof-of-concept test of the Genia platform in juvenile idiopathic arthritis reported that it improves patient engagement, patient-centered care, and practice-based learning and recommended the uptake of the platform in other chronic conditions [33,34].

This study adapted Genia to the needs and preferences of the CF community in a pediatric CF center in the United States. A proof-of-concept testing was conducted to optimize the platform, remove glitches, and ensure the security and privacy of patient-generated health data.

The project is informed by the model of health care service coproduction [5] whose central tenet is that health care services are coproduced by patients and providers in systems that support and constrain effective partnerships.

A total of 5 focus groups with 21 participants were conducted: 1 with pediatric clinical team members (n=5), 1 with adolescent patients with CF treated at COA (n=5), and 3 with familial caregivers of pediatric patients with CF (n=11). Clinical team members represented the specialties included in a multidisciplinary CF clinic (pulmonology, nursing, nutrition, respiratory therapy, and social work). Patients with CF, mean age 17.8 years (SD 1.7; range 15-20 years) and 80% (4/5) female, resided in 4 different geographic areas of the state. Familial caregivers, mean age 42 years (SD 6.9; range 33-59 years), 90% (10/11) female, and 90% (10/11) parents, represented 3 different geographic areas of residence.

The results of the focus groups highlighted both overlapping and differing preferences among clinical staff, caregivers, and patients. Overall, when making decisions about modifications, the research and development teams prioritized the perspectives of patients and caregivers—the app’s end users—over those of clinical team members, which resulted in a greater number of trackers. Although not every recommendation made by participants was feasible, most suggestions were successfully incorporated into the modified app, highlighting the necessity of identifying the needs and preferences of the end users before entering a test environment. A summary of the modifications made as a result of focus group recommendations is presented in Table 2. A full list of all changes is provided in Multimedia Appendix 1.

Testing of the HIPAA-compliant data flow from the patient-facing app to a secure US-based cloud server to the REDCap study database was necessary to ensure the protection and privacy of patient-generated health data. The research and development teams also tested that patient-generated reports in the mobile app were correctly uploaded to the REDCap database. The issue of data readability for the clinical team and the research team had to be addressed as well. The development team successfully converted the incoming patient-generated reports into PDF files with charts, graphs, and text boxes of tracked symptoms and raw data in .xls files. This ensured that the clinical team had an easy way to interpret the output of patient symptoms and that the research team had easily accessible data for data analysis.

Beta testing of the modified Genia platform for technical issues was conducted for 30 days with 6 participants (mean age 41 years, SD 2.4; range 43-51) from the pool of caregiver focus group participants, as they were already familiar with the platform. Specifically, we asked participants to test every tracker, submit notes with and without pictures, test if the treatment reminders feature is functional, and make note of anything that needs to be corrected. The research team informed testers that data entered into the app would not be acted upon by the clinical team during the beta testing period and that they should contact the clinical team through usual channels if health concerns arose.

Beta testers reported bugs and issues to a contact person on the research team who then funneled that information to the development team. The most common issues reported by beta testers included metric units for height and weight, unnecessary caps on information input, medication reminders not going off, and leftover Swedish language.

The developers removed the reported glitches and worked collaboratively with the research and clinical team members to make adaptations that ensure smooth user experience. For example, beta testers reported a medication missing from the menu of options. The CF nurse coordinator reviewed and updated the medication list, and the developers updated the app accordingly. In addition, beta testers reported an issue with attaching images to previsit reports, and the research team detected a glitch in the upload of previsit reports to the REDCap database. All these issues were fixed by app developers. A complete list of optimizations and removed glitches is provided in Multimedia Appendix 1.

The clinical and research teams met with the development team during a site visit in November 2019. This meeting was instrumental in laying the groundwork of clinical protocols. For example, clinical providers had concerns about how the patient-generated health data from the app would fit into clinical planning and flow. To address these concerns, a clinical management protocol was developed that summarizes patient reports from the app and incorporates them into the weekly previsit overview of patients with an upcoming clinical appointment that week. To ease strain on the providers, an appointed research team member organizes the patient reports from REDCap and provides the PDFs to the CF nurse coordinator on a weekly basis. The CF nurse coordinator then follows a patient support protocol that directs what needs to be done with specific types of information (eg, >3 increased symptoms for >2 days→report to treating pulmonologist and add antibiotic and increased depression reports→initiate contact with a social worker or mental health counselor).

The effect of the platform for disease self-management, patient-reported outcomes, and patient-centered care is being assessed in an ongoing clinical trial (NCT03910881).

In this study, we adapted an mHealth platform developed for the CF community in Sweden to the needs and preferences of the US CF community. We also conducted proof-of-concept testing to optimize the platform, remove glitches, and ensure the security and privacy of patient-generated health data. The project demonstrated the importance of identifying the needs and preferences of end users before using existing mHealth solutions and the need to consider all relevant perspectives, including those of clinicians, patients, caregivers, and app developers.

Although end user preferences guided the modification of app functionalities and presentation, clinician perspectives governed the development of clinical protocols that detailed how patient-generated health data are used in clinical settings. Successful coproduction of health care services for managing a chronic health condition [5] requires the perspectives of both parties. Engaging end users in the development or modification of mHealth apps can help prevent implementation challenges and increase the likelihood of app uptake by the target audience [40]. Engaging clinicians can ensure that mHealth solutions are practical in clinical settings and can be integrated into the clinical data flow within existing IT environments [21]. Other barriers and challenges also exist. For example, not all end users may be able or willing to self-monitor and report symptoms or activities, particularly those with fewer resources or lower literacy, whereas clinicians may find the use of patient-generated health data burdensome or having little measurable effect on health outcomes. Addressing the priorities and concerns of both parties will facilitate realistic and attainable goals in the context of coproduction facilitated by technology-enabled data sharing [5]. For successful mHealth interventions, patient-generated health data must be both meaningful to patients and useful to clinicians [22].

The preferences of end users of Genia mirrored the preferences of patients with CF and their caregivers reported in a previous study [30]. Specifically, end users preferred an app with multiple functions that facilitates access to information, automates disease management, integrates with other apps, facilitates communication with the health care team, and is highly customizable to meet individual goals and preferences. Involving end users in all stages of mHealth development and collaborating with clinicians and health care system experts may result in apps that maintain engagement, improve coproduction of services, and ultimately impact self-management and health outcomes [30].

The described modification process represents a scalable approach to adapting mHealth solutions to local needs and changing contexts. Feedback from patients, caregivers, health care providers, and experts gave input to developers on how to prepare the mobile platform for further adaptations to local needs and changing contexts. For example, a file-based approach was developed to support a multilanguage platform, and medication and treatment lists were prepared to allow for dynamic updates and changes in response to clinician requests and patient preferences. The collection of patient-generated health data was implemented according to a structure that could be used to create new trackers and new report questions. The result was a more scalable and dynamic platform, facilitating further adaptations. Such an approach can be replicated in further modification of this and other mHealth platforms to track home spirometry, home intravenous antibiotics, oral and nebulized medications, or airway clearance therapy via smart devices.

App repositories include hundreds of apps that claim to improve disease management, health outcomes, and health-related behaviors. However, few are evidence-based solutions developed with the involvement of health care professionals, patients, caregivers, and behavioral scientists [33,34]. Even fewer are tailored to the components of daily CF care and the symptoms typically experienced by people with CF. The few apps that collect CF-specific patient-generated health data are not integrated in the CF care plan or the clinical workflow. Without such integration, mHealth technologies can have only a minimal impact on chronic disease care and management [41]. At the same time, integration of patient-generated health data in clinical care via mHealth solutions needs to be done in a way that safeguards the privacy and integrity of patients in their domestic domain [42]. mHealth platforms and patient support systems, such as Genia, provide new opportunities for improved self-care and clinical management of CF. However, for the integration of mHealth solutions in clinical care, their feasibility, their acceptability to patients and providers, and their clinical effectiveness need to be tested in rigorous clinical trials.

Participants included English-speaking patients, caregivers, and health care providers recruited from a single pediatric CF center in the United States. The perspectives of this convenience sample may not be representative of all CF centers or geographic areas in the country. However, as CF disease management and care delivery are highly protocolized across the network of accredited CF centers, vast differences between CF stakeholders from other centers and regions are unlikely. In addition, focus group participants had an opportunity to use the app only during focus group discussions, which may have limited the scope and depth of their experience and feedback. Finally, the project did not address data mapping of patient-generated health data to interoperable standards to integrate Genia-collected patient-generated health data in the electronic health record, a step that will be addressed in a future study phase.

A currently ongoing pilot clinical trial (NCT03910881) assesses whether the use of Genia over 6 months enhances patient-centered care (shared decision making, patient activation, patient satisfaction, and patient self-efficacy) and improves patient-reported outcomes (symptom scales and quality-of-life domains of the Cystic Fibrosis Questionnaire-Revised [43]). A randomized clinical trial will evaluate the impact of Genia on respiratory and nutritional outcomes. Informed by the model of health care service coproduction [5], these study phases are predicated on the hypotheses that (1) use of patient-generated health data will affect providers’ actions, which will have downstream benefits, and (2) use of patient-generated health data will affect patients’ disease management behaviors and interaction with the health system.

mHealth offers new opportunities to support self-management of CF, facilitate patient-physician communication, and promote coproduction of care. Broad and successful uptake of mHealth solutions in CF care requires careful consideration of patient and family perspectives and active involvement of CF clinical teams.