Participants were recruited from the Internal Medicine clinic and the Sleep Disorders Center at Walter Reed National Military Medical Center. Inclusion criteria included age between 18 and 75 years, being a military servicemember or Defense Enrollment Eligibility Enrollment System beneficiary, owning a smartphone, and having provider or self-referral for sleep problems. Exclusion criteria were pregnancy and untreated or uncontrolled (or both) medical or psychiatric illness.

Health care PCMs (including physicians, nurse practitioners, and physician assistants) were recruited via word of mouth from the Internal Medicine clinic at Walter Reed National Military Medical Center.

Following the onset of the COVID-19 pandemic, all research procedures were conducted remotely [13]. Potential participants were identified by health care providers or self-referred for having sleep problems and were contacted via telephone for eligibility screening. Interested volunteers were referred to a trained study coordinator who ensured informed consent, provided study instructions, and onboarded participants to the study app (downloaded from the Apple App store or Google Play store) using a secure code. Participants were provided Fitbit devices via priority courier. Study staff remained available to provide support or answer technical questions.

The in-app sleep assessment included a sleep history questionnaire; several validated research questionnaires that are widely used within the Department of Defense to assess symptoms of obstructive sleep apnea (Berlin Questionnaire [14]), insomnia (Insomnia Severity Index [15]), depression (Patient Health Questionnaire-9 [16]), anxiety (Generalized Anxiety Disorder-7 [17]), posttraumatic stress disorder (Primary Care PTSD [posttraumatic stress disorder] Screen for DSM-5 [Diagnostic and Statistical Manual of Mental Disorders] [18]), chronic pain (Defense and Veterans Pain Rating Scale [19]), and history of traumatic brain injury (Brief Traumatic Brain Injury Screen [20]); standardized daily sleep diaries; and brief symptom surveys administered twice a day (morning and evening) that assessed mood, cognition, and energy level. Symptom surveys followed the format “I feel ___” followed by descriptors such as happy, sad, clear-headed, and tired. Responses ranged from “not at all” to “very” and were scored from 0 to 4, respectively. Participants also wore a Fitbit Inspire 2 device throughout the study (Figure 1). Sleep data were obtained in 1-minute epochs daily from the Fitbit device and integrated directly into the WellTap platform engine via an application programming interface. Total sleep time, number of awakenings, wake after sleep onset, sleep efficiency, and sleep timing were reported to patients in the app and to providers in provider reports.

Following completion of the remote assessment, the platform generated separate sleep assessment reports for patients and PCMs. For patients, sleep assessment results were presented in the app (Table 1). For providers, sleep assessment results were provided (via store and forward) in an editable electronic document format (eg, PDF or DOCX [or both]; Table 1). The content for both reports were derived from the 10-day continuous sleep monitoring assessment and tailored for patients or providers on the basis of feedback gained in an earlier qualitative phase of this study [11,12].

Patients with sleep problems received their personalized sleep assessment results in the app, including integrated results from both patient-reported and COTS wearable data. Provider reports were provided in an editable electronic document format.

Patient engagement was defined a priori based on behavioral adherence metrics including (1) completion of the baseline assessment, (2) wearing the COTS sleep tracker, (3) completion of daily sleep diaries, (4) completion of daily symptom surveys (ie, EMA assessments), and (5) review of the personalized sleep report. In addition, following the 10-day remote monitoring assessment, patients completed a brief survey to assess perceived ease of completion. Individual items asked, “How easy was it for you to complete the assessment questions on your device?” and “How easy was it to navigate the assessment on your device?” Five possible responses ranged from “Not at all” to “Very.” A third question was asked about the length of assessment questions; responses ranged from “too short” to “good” to “too long” on a 5-point bipolar response scale. Finally, participants were asked (yes/no) if they needed human support to complete the remote monitoring assessment, as well as willingness to consider future sleep telehealth approaches (yes/no).

Effectiveness was defined a priori as usability, acceptability, perceived improvement, perceived credibility, and overall satisfaction. After reviewing sample assessment reports, providers completed an anonymous 16-item questionnaire developed for this study. All responses were presented on a 5-point scale ranging from “Not at all” to “Very” and were scored from 0 to 4, respectively. In addition, providers were asked, “How likely are you to recommend patients into this study?” Responses again ranged from “Not at all” to “Very.”