New Narcolepsy Questionnaire Passes Its Toughest Test Yet in Phase 3 Trials

Narcolepsy has never been an easy disease to measure. The condition, marked by overwhelming daytime sleepiness, sudden muscle weakness called cataplexy, fragmented night sleep, and vivid hallucinations at the edges of sleep, has long been assessed with tools borrowed from other sleep disorders. The widely used Epworth Sleepiness Scale, for example, asks how likely a person is to doze in eight everyday situations, but it says nothing about how narcolepsy dismantles a person’s ability to work, socialize, cook a meal, or finish a sentence without losing their train of thought. A new study published in Frontiers in Neurology now reports that a purpose-built questionnaire called the Functional Impacts of Narcolepsy Instrument, or FINI, has passed a comprehensive battery of statistical tests in people with narcolepsy type 1, using data from two large phase 3 clinical trials of an investigational drug.

The FINI is a 28-item patient-reported outcome measure that asks people with narcolepsy to rate their experiences over the previous seven days. It was built from the ground up with input from patients, not just clinicians. During its development, qualitative interviews revealed that the functional impacts people with narcolepsy report most often relate to excessive daytime sleepiness, cataplexy, and cognitive difficulties: trouble walking, driving, cooking, and managing household chores; disruptions to work or school, including reduced attendance or abandoned career plans; and limits on social and leisure activities. Those interviews shaped the final structure of the instrument, which covers six domains: Tiredness, Cognitive Functioning, Cataplexy, Social Activities, Everyday Activities, and Everyday Responsibilities. The Cataplexy domain applies only to narcolepsy type 1, the subtype in which cataplexy is present.

Each FINI item uses a five-point response scale scored from 0 to 4, and the whole questionnaire takes roughly five minutes to complete. Domain scores are averaged and standardized to a 0-to-100 scale, where 0 represents the best health and 100 the worst. That brevity matters: lengthy questionnaires invite fatigue and abandoned surveys, particularly in a population whose defining symptom is sleepiness. The instrument was translated into 30 languages for the phase 3 program following international guidelines for translation and cultural adaptation, with cognitive debriefing conducted in each target language with people with narcolepsy before finalization.

To validate the FINI, the research team, led by scientists at Takeda Development Center Americas and Clinical Outcomes Solutions, pooled patient-reported data from two randomized, placebo-controlled, double-blind phase 3 trials of oveporexton, an oral orexin receptor 2 agonist being developed for narcolepsy type 1. The trials enrolled 167 and 105 participants respectively across 19 countries, from France, Italy, and Spain to Japan, China, Australia, and the United States. Participants were seriously affected: all had an Epworth Sleepiness Scale score of at least 11 at screening, and nearly 80 percent fell into the severe daytime sleepiness category. The FINI was administered at baseline and again at weeks 8 and 12 of treatment, alongside a suite of comparator measures including the Epworth Sleepiness Scale, the SF-36 quality-of-life survey, the EuroQol-5D health status measure, the Narcolepsy Severity Scale for Clinical Trials, and a cognitive complaints inventory.

The first question was whether the six-domain structure actually holds together statistically. Using confirmatory factor analysis, a technique that tests whether observed responses fit a hypothesized underlying architecture, the team found that the data fitted the six-domain model well. The comparative fit index reached 0.96 and the non-normed fit index 0.95, both comfortably above the conventional 0.90 threshold, while the root mean square error of approximation and the standardized root mean residual both came in below 0.08, the standard ceilings for acceptable fit. Nearly every item loaded strongly onto its intended domain, with one modest exception: a nap-related item in the Everyday Activities domain loaded at 0.47, just under the 0.50 cutoff, but was retained for its theoretical importance and its grounding in patient interviews.

The instrument also distinguished reliably between people of different symptom severities, a property known as known-groups validity. Participants reporting severe disease on a global severity scale scored significantly higher, indicating worse function, than those with moderate disease across every FINI domain, with large standardized effect sizes ranging from 0.67 to 0.94 and all comparisons significant. The same pattern held for comparisons based on the Epworth Sleepiness Scale, with moderate effect sizes between 0.34 and 0.68. Meanwhile, convergent and divergent validity were supported: correlations between FINI domains and conceptually aligned reference measures averaged 0.54, while correlations with measures of unrelated constructs averaged only 0.29. The strongest links were exactly where theory predicted, such as the Cognitive Functioning domain against a cognitive complaints inventory, with a correlation of 0.81, and Tiredness against the SF-36 Vitality subscale.

Reliability and responsiveness, the two remaining pillars of questionnaire validation, also came through. Internal consistency, which measures how coherently items within a domain hang together, exceeded the 0.70 threshold for five of six domains, with Cronbach’s alpha values from 0.78 to 0.92. The Everyday Activities domain fell short at 0.63, which the authors attribute to its deliberately broad scope, mixing items about watching television, reading, sleep attacks, and napping; they argue the heterogeneity was an intentional trade-off to preserve content that patients identified as important. Test-retest reliability was excellent, with intraclass correlation coefficients of at least 0.79 for all domains among participants whose condition was stable between screening and baseline. Responsiveness, the capacity to detect real change, was exceptional: correlations between FINI change scores and reference measure change scores exceeded 0.35 for all domains, and effect sizes in substantially improved patients were frequently more than double the conventional benchmark for a large effect.

Perhaps the most consequential output of the study is a set of thresholds for meaningful within-patient change, the score changes that patients themselves perceive as a genuine improvement. Deriving these numbers is notoriously difficult, and few studies have followed the methods recently outlined by the US Food and Drug Administration. The team triangulated evidence from anchor-based analyses, in which score changes are mapped against patients’ global impressions of improvement, with distribution-based statistics and novel forest-plot methods that pool estimates across multiple anchors. The resulting thresholds, on the 0-to-100 scale where negative change means improvement, were 35 points for Tiredness and Everyday Activities, 30 points for Cognitive Functioning and Everyday Responsibilities, 25 points for Social Activities, and 20 points for Cataplexy. These values will allow researchers to run responder analyses, calculating the proportion of patients in each treatment arm who achieve a change that patients themselves would consider meaningful, rather than relying on statistically significant but clinically opaque averages.

The authors are candid about the limits of the work. Because the validation sample came from clinical trials that required elevated sleepiness for entry, people with well-controlled symptoms are underrepresented, and the study could not examine relationships with depression or anxiety, since severely depressed individuals were excluded from the trials. The FINI’s performance in narcolepsy type 2, which would use five of the six domains, and in idiopathic hypersomnia remains to be established. Still, the study concludes that the FINI is a robust, valid, and highly sensitive measure of what matters most to people living with narcolepsy: whether they can function in their daily lives. Beyond trials, its modular six-domain design could support shared decision-making in clinics, letting patients and physicians track progress on the specific aspects of life that narcolepsy disrupts most.

Subject of Research: Validation of the Functional Impacts of Narcolepsy Instrument, a patient-reported outcome measure, using phase 3 clinical trial data in narcolepsy type 1

Article Title: Psychometric properties of the novel patient-reported Functional Impacts of Narcolepsy Instrument (FINI) using data from phase 3 clinical trials in narcolepsy type 1

Article References: Romero, H., Savva, Y., Crawford, S., Howerter, A., Dave, S., & Doll, H. (2026). Psychometric properties of the novel patient-reported Functional Impacts of Narcolepsy Instrument (FINI) using data from phase 3 clinical trials in narcolepsy type 1. Frontiers in Neurology, 17, Article 1884901. https://doi.org/10.3389/fneur.2026.1884901

Image Credits: AI Generated

DOI: 10.3389/fneur.2026.1884901

Keywords: narcolepsy, narcolepsy type 1, patient-reported outcomes, FINI questionnaire, psychometric validation, cataplexy, excessive daytime sleepiness, clinical trials, oveporexton, orexin receptor agonist, meaningful change threshold, quality of life