A Comparison of Cosmos Rx Medication Adherence Platform vs. Pharmacy Fill Adherence Measures

4.1 Background and Rationale Medication non-adherence is associated with greater morbidity and mortality in chronic disease, and has been estimated to increase healthcare costs by over $170 billion annually in the United States.1 Medication adherence is challenging for patients across the spectrum of medical disorders that are treated with long-term use of oral medications. Poor medication adherence can result in increased risk of complications, disease progression, increased healthcare utilization and poor therapeutic outcomes.2 Adherence is typically tracked using measures based on pharmacy fill data such as the medication possession ratio (MPR) and/or proportion of days covered (PDC).3 Although MPR and PDC are associated with a higher likelihood of using medication as prescribed at the population level, when compared to direct assessments or patient reports, they are not accurate enough for making individual management decisions and likely to underestimate non-adherence.4 However, real-time documentation methods using direct video observation are expensive and cumbersome for patients and providers.

Gastroesophageal reflux disease (GERD) is among the most prevalent gastrointestinal conditions in the United States, affecting an estimated 20-30% of the adult population.5 Proton pump inhibitors (PPIs) are the cornerstone of pharmacologic management for GERD forming the standard of care for erosive esophagitis, Barrett's esophagus, and symptomatic non-erosive reflux disease.6 Despite their established efficacy, suboptimal medication adherence remains a significant clinical barrier: estimates of PPI adherence in real-world populations range from 50-70% at one year7, with non-adherence associated with symptom recurrence, mucosal injury, and increased healthcare utilization.8 FORTISKAP™ (Cosmos Rx, Inc.) is a digital adherence monitoring device consisting of a sensor-embedded cap that attaches to standard prescription medication bottles. The device records the date and time of each bottle opening event and transmits this data wirelessly to a paired smartphone application, which generates adherence metrics and patient-facing reminders. Prescribers may access aggregated adherence data through a connected provider portal.

The goal of this study is to demonstrate that the Fortiskap™ prescription bottle-top, medication adherence device results in a higher detection rate of medication non-adherence compared to pharmacy fill-data assessment and/or compared to ARMS adherence self-assessment.

4.2 Study Objectives

Primary outcomes:

1. Comparison of medication adherence between Fortiskap™ tracking of prescription bottle opening and pharmacy fill data: The study measure is defined as the proportion of study days with at least one recorded medication removal event ("Fortiskap adherence rate") compared with the Proportion of Days Covered(PDC): the number of days during which pharmacy fill records demonstrate medication available divided by the number of study days. The primary comparison is a within subject paired analysis of continuous variables. 2. Comparison of medication adherence between Fortiskap™ tracking of prescription bottle opening and a validated adherence self-reporting tool: The study measure is defined as the proportion of study days with at least one recorded medication removal event ("Fortiskap adherence rate") compared with vs. Adherence to Refills and Medications Scale (ARMS) adherence self-reporting instrument.9 The primary comparison is a within subject paired analysis of the industry-standard dichotomous outcomes of 80% for daily medication consumption and \>=16 for the ARMS.

Secondary Outcomes:

* Comparison of the proportion of subjects who achieve 90% adherence based on Fortiskap™ tracking, MPR and PDC calculations * Day 7, Day 30 and Day 90 ARMS and Usability Instrument trends and associations with adherence outcomes, demographic, and SDOH variables * Safety: Number of dose deviation reminders provided to subjects via the Fortiskap™ companion app * Safety: Outcomes of alerts provided to clinical providers via the Fortiskap™ physician dashboard

Human Factor Outcomes:

1. 7 days after onboarding, proportion of subjects who can independently complete medication access, reminder acknowledgement and respond to study data collection requests 2. 14 days after provider onboarding, the proportion of providers who can independently access the physician dashboard. The physician dashboard is a standalone, password-protected secure web portal that displays subjects' device status and device-recorded events 4.3 Study Design This is a prospective, non-randomized, single-arm interventional study. The FORTISKAP™ device is the intervention. There is no separate control group; each subject serves as their own control: subjects continue their prescribed PPI regimen without modification. The study observes and compares three methods of measuring adherence to that regimen. The three methods, described in Section 4.2 and the Appendices, are: (1) FORTISKAP™ device-recorded bottle-access events (Section 4.1); (2) pharmacy-fill-based PDC from TidalHealth Community Pharmacy dispensing records (Appendix 5.4); and (3) the ARMS self-report instrument (Appendix 3). Comparisons are paired and within-subject (Section 4.5), with the accepted 80% PDC threshold as the reference standard for adherent/non-adherent classification.

4.4 Study Duration Individual subject participation 90 days from enrollment (Day 0) through the Day 90 closeout visit. Total study timeline: Approximately 270 days (\~9 months), accounting for a rolling enrollment period of up to 180 days followed by the 90-day follow-up window for the last enrolled subject. (for details of recruitment plans, see section 6).

4.5 Statistical Approach The primary statistical analysis will compare paired adherence measurements within the same subject across the three measurement modalities. The Wilcoxon signed-rank test (non-parametric, within-subject, two-tailed, α = 0.05) will be used as the primary test for the FORTISKAP™ vs. PDC comparison and the FORTISKAP™ vs. ARMS comparison. This non-parametric test is chosen given modest sample size and expected non-normal distribution of adherence scores.

Adherence will be expressed as a continuous percentage (0-100%) for FORTISKAP™ and PDC comparisons. ARMS total score (range 12-48) will be treated as a continuous variable for the FORTISKAP™ vs. ARMS comparison; scores ≥ 16 will additionally be classified as indicating non-adherence for categorical analyses.

Agreement between modalities will be further assessed using Bland-Altman plots and intraclass correlation coefficients (ICC). Categorical concordance (adherent vs. non-adherent) across methods will be reported using Cohen's kappa.

Secondary outcomes will be summarized descriptively. PUASQ scores will be reported by timepoint (Day 7, 30, 90) with mean, standard deviation, and range.

4.6 Sample Size A formal sample size calculation was performed using the Wilcoxon signed-rank test under the following assumptions: two-tailed α = 0.05; power = 0.80; expected medium effect size (r = 0.3) representing a clinically meaningful difference between adherence measurement modalities, based on analogous studies comparing electronic monitoring to pharmacy fill data in chronic disease populations.

Under these assumptions, a minimum of 37 evaluable subjects is required. To account for an estimated 10-15% dropout and data incompleteness rate over the 90-day follow-up, a target enrollment of 45 subjects (evaluable target: 40) is established. A sensitivity analysis will be conducted including all subjects with at least 30 days of FORTISKAP™ data.

4.7 Stopping Rules

The study may be stopped early if:

* FORTISKAP™ device malfunction that interrupts expected medication access for 2 or more days exceeds 20% of enrolled subjects. Malfunction is defined as failure to record or transmit bottle-access events and/or physical failure of the cap mechanism rendering primary outcome data collection infeasible. * Any unanticipated adverse event or serious adverse event that is attributed to the FORTISKAP™ device. * Enrollment remains below 25 subjects after 180 days from the start of recruitment.

The Site PI makes the final determination if study termination is warranted. SECTION 5: SUBJECT POPULATION, INCLUSION/EXCLUSION CRITERIA, AND VULNERABLE POPULATIONS 5.1 Target Population Adult patients (≥ 18 years of age) presenting to TidalHealth Gastroenterology Clinic with a confirmed diagnosis of GERD who are currently prescribed a once-daily oral PPI.

5.2 Inclusion Criteria

1. Age ≥ 21 years at time of enrollment 2. Confirmed diagnosis of GERD and/or erosive esophagitis documented in the Epic medical record (ICD-10 code K21.0 or K21.9) 3. Currently prescribed a once-daily oral PPI (omeprazole, esomeprazole, pantoprazole, lansoprazole, rabeprazole, or dexlansoprazole) for a minimum of 30 days prior to enrollment 4. Willing and able to fill PPI prescription exclusively at TidalHealth Community Pharmacy for the 90-day study period 5. Owns a smartphone (iOS or Android) capable of running the FORTISKAP™ companion application 6. Able to participate in the study enrollment and ongoing requirements in English.

5.3 Exclusion Criteria

1. Current diagnosis of active esophageal malignancy, gastric malignancy, or other upper gastrointestinal malignancy requiring active systemic treatment 2. Cognitive impairment, dementia, or other neurological condition that, in the judgment of the Site PI, would prevent proper use of the FORTISKAP™ device or completion of self-report questionnaires 3. Institutionalized subjects (nursing home, assisted living, correctional facility) or subjects whose medications are managed by a caregiver who would need to operate the device on their behalf 4. Life expectancy \< 6 months in the judgment of the treating physician 5.4 Vulnerable Populations Pregnant Women: Pregnant women are not specifically excluded. PPI use in pregnancy is not contraindicated, and there is no additional risk from FORTISKAP™ device use in pregnancy. If a subject becomes pregnant during the study, she may continue participation at the discretion of her treating physician and the Site PI.

Prisoners / Incarcerated Individuals: Incarcerated individuals are excluded. Children / Minors: Minors (\< 18 years) are excluded. Cognitively Impaired Individuals: Individuals with documented cognitive impairment that would prevent meaningful study participation are excluded.

Economically or Educationally Disadvantaged Subjects: No subjects are targeted based on economic or educational status. The compensation offered ($175 maximum) is not considered coercive given the minimal burden of participation. See Section 11 for compensation details.

SECTION 6: RECRUITMENT METHODS 6.1 Recruitment Strategy

Subjects will be recruited from the existing patient panel of TidalHealth Gastroenterology Clinic by the following methods:

6.1.1 Identification via Epic The Site PI and/or study coordinator will use Epic reporting tools to generate a list of patients with active GERD diagnoses and current PPI prescriptions. This list will be used to identify potentially eligible subjects for outreach. Dr. Canakis; the sole physician enrolling subjects into this study.

6.1.2 Potential Subject Communication Eligible subjects identified through Epic reporting may receive a brief outreach letter or email describing the study and inviting them to contact the study coordinator. A template for this communication is included in Appendix 6. Outreach communications will be sent through TidalHealth's secure patient messaging infrastructure (MyChart) or by US mail. No unsolicited phone contact will be attempted.

6.1.3 Point-of-Care Recruitment The study coordinator will identify potentially eligible subjects at the time of a scheduled gastroenterology clinic appointment. The coordinator will introduce the study to inter