A Comparison of a Medication Adherence Platform (FORTISKAP™) vs. Usual Care in Subjects on Oral Medications for the Treatment of Interstitial Lung Disease, Sarcoid and Pulmonary Hypertension

Overview

This study will evaluate whether participants with serious lung diseases such as idiopathic pulmonary fibrosis, sarcoidosis, and pulmonary hypertension who use the FORTISKAP™ smart medication cap - a bottle-top device that tracks prescription bottle openings and sends dose reminders to participants and their care team - take their medications more consistently and experience better health outcomes compared to similar participants receiving standard care without the device. Participation requires no changes to prescribed medications, testing or clinical visits beyond what is already part of routine care; participants use a modified medication bottle equipped with the FORTISKAP™ cap for nine months.

Interstitial Lung Disease (ILD) is a group of conditions that cause progressive inflammation and often irreversible scarring of lung tissue. This fibrotic process disrupts oxygen and carbon dioxide exchange, leading to symptoms including dyspnea, cough, wheezing, chest discomfort, and hemoptysis. Patients with ILD experience substantial functional impairment; hospitalizations for respiratory failure carry a one-year mortality exceeding 50%. Long-term management frequently involves glucocorticoids, immunosuppressants (e.g., azathioprine), and antifibrotic agents - all of which carry significant side effects and substantial costs. Pulmonary hypertension(PH) associated with lung disease is a serious and increasingly recognized complication of chronic parenchymal lung diseases, most notably interstitial lung disease (ILD) and chronic obstructive pulmonary disease. A recent systematic review and meta-analysis found a pooled prevalence of PH of 36% in ILD populations. Clinically, patients present with progressive exertional dyspnea and fatigue that is disproportionate to the severity of their underlying ILD, markedly reduced exercise capacity, worsening hypoxemia, and, in advanced disease, right heart failure. Management of remains challenging; most drugs approved for pulmonary arterial hypertension have demonstrated no benefit or potentially harmful effects in this population. A notable exception is inhaled treprostinil, which has demonstrated improvements in exercise capacity in patients with ILD-associated PH, representing a meaningful but limited advance in an area of substantial unmet therapeutic need. Medication non-adherence is associated with greater morbidity and mortality across chronic diseases and has been estimated to increase U.S. healthcare costs by over $170 billion annually. For patients with ILD and PH - who carry high treatment burden, significant medication side effects, and complex regimens - adherence represents a particularly critical and modifiable variable. The FORTISKAP™ platform (Cosmos Rx, Inc.) is a bottle-top electro-optical device that records time-stamped prescription bottle openings and estimates pill counts using proprietary image processing algorithms, providing an objective measure of adherence. A companion smartphone application delivers dose-deviation reminders to subjects, and a physician dashboard alerts clinicians to adherence trends. This study will generate the first prospective data on FORTISKAP™ performance in an ILD/PH population. Medication adherence is challenging for patients across the spectrum of medical disorders that are treated with chronic use of oral medications. Poor medication adherence can result in increased risk of complications,disease progression, increased healthcare utilization and poor therapeutic outcomes. While these therapies can be effective for certain ILD subtypes, they carry significant risks including cytopenias, nephrotoxicity, weight gain, adrenal suppression,etc and they can be expensive - creating a burden for the patient and the healthcare system. 4.3 Study Design/Statistical Analysis Plan Fifty subjects with a primary diagnosis of ILD, sarcoidosis, and/or pulmonary hypertension will be enrolled from the Sentara Pulmonary, Critical Care \& Sleep Specialists clinic and followed for 270 days (approximately 9 months) while using the FORTISKAP™ device on their primary oral ILD/PH medication. This prospective cohort will be compared to 50 propensity-matched historical controls drawn from the same clinical practice who received usual care without the device. The primary analytic approach is difference-in-differences (DiD), estimating the treatment effect by comparing within-group change over time between the intervention and control arms. Propensity score matching will be performed on clinically relevant baseline covariates including diagnosis (ILD, PH, or both), disease severity indices, age, sex, baseline functional status, and relevant comorbidities. Difference-in-Differences: 6-Minute Walk Test (6MWT) The 6MWT distance (meters) will be assessed at baseline and Day 270. The DiD estimator will be computed as: DiD = (FORTISKAP™ post - FORTISKAP™ pre) - (Control post - Control pre) This will be implemented using a linear mixed-effects model with fixed effects for time (baseline vs. end of study), group (FORTISKAP™ vs. control), and their interaction term (group × time), which directly yields the DiD estimate. Subject will be included as a random effect to account for within-person correlation. The interaction coefficient and its 95% confidence interval will be the primary reported effect estimate. A minimally clinically important difference (MCID) of 30 meters will be used as a reference threshold for clinical significance. Difference-in-Differences: Pulmonary Function Tests (FVC and DLCO) FVC (expressed as percent predicted) and DLCO (expressed as percent predicted) will each be analyzed using the same linear mixed-effects DiD framework described above, with separate models for each outcome. Given the known correlation between FVC and DLCO in ILD, a sensitivity analysis using a bivariate mixed-effects model may be considered to jointly model both outcomes and account for their within-subject correlation. An FVC decline of ≥10%8 and DLCO decline of ≥15%9 will be used as reference thresholds for clinical significance. Difference-in-Differences: Echocardiographic Parameters (TRV/sPAP and RV Size) TRV (m/s) and estimated sPAP (mmHg), along with a qualitative or quantitative measure of RV size (e.g., RV basal diameter in cm or a semi-quantitative grading), will each be modeled using the same DiD linear mixed-effects framework as FVC and DLCO. Objective vs. Self-Reported Adherence Agreement between objective and self-reported adherence will be evaluated by cross-classifying subjects according to two binary adherence definitions: Objective adherence (FORTISKAP™): Adherent if ≥80% of prescribed doses recorded via electronic bottle-opening monitoring. Self-reported adherence (ARMS): Adherent if ARMS score ≤16 The primary measure of agreement will be Cohen's kappa (κ). Sensitivity, specificity, and positive and negative predictive values will be calculated treating FORTISKAP™ electronic monitoring as the reference standard, given its objective nature. McNemar's test will be used to assess whether discordance between the two methods is asymmetric - specifically, to evaluate the expected directional hypothesis that self-report systematically overestimates adherence relative to electronic monitoring. Longitudinal ARMS Trends (Day 7, Day 90, Day 270) ARMS scores collected at three timepoints (Day 7, Day 90, and Day 270) in the FORTISKAP™ group will be analyzed using a linear mixed-effects model with time as a fixed effect and subject as a random intercept, to characterize the trajectory of self-reported adherence over the study period. If the ARMS is also collected in controls, a group × time interaction term will be added to compare trajectories between arms. Pairwise contrasts between timepoints (Day 7 vs. Day 90, Day 90 vs. Day 270, Day 7 vs. Day 270) will be tested with appropriate correction for multiple comparisons (e.g., Bonferroni or Tukey). Exploratory analyses will examine associations between ARMS trajectory and baseline demographic variables and social determinants of health (SDOH) using multivariable regression, with ARMS score change from Day 7 to Day 270 as the dependent variable. Sample size and power: With 50 subjects per group, this study is powered as a pilot/feasibility investigation. Formal power calculations for the DiD analyses should be conducted using estimates of within-group change and variability from the published literature in ILD and PH populations; these should be reported alongside observed effect sizes to inform the design of subsequent confirmatory trials. Multiple comparisons: Given the preliminary and exploratory nature of this study, p-value thresholds will be set at α = 0.05 for primary DiD outcomes, with secondary and exploratory analyses interpreted with appropriate caution. No formal multiplicity correction is proposed across the full set of outcomes at this stage, but this should be revisited in the confirmatory protocol.