Most tinnitus programs describe their results in adjectives. We publish our definitions before we publish our numbers, and we decided what would count as success before seeing any data. This page is the methodology behind every outcome figure we quote.
What we measure. Every patient in our tinnitus management program completes the Tinnitus Functional Index — a validated 25-item questionnaire scored 0–100 that quantifies tinnitus impact on sleep, concentration, emotional wellbeing, and quality of life — at intake, at twelve weeks, and at twenty-four weeks. Between visits, patients submit a weekly tinnitus rating on a 0–100 numeric rating scale — framed as "over the last week, on average, how much has your tinnitus bothered you?" — collected by text message or in-app and analyzed as six-week block averages to smooth day-to-day variation.
What counts as improvement. Both thresholds were pre-specified — fixed before any patient data was reviewed, and not adjusted since:
- Tinnitus Functional Index: a reduction of 13 points or more from intake, the accepted benchmark for clinically meaningful change on this instrument.
- Weekly rating: a reduction of 15 points or more between six-week block averages. Unlike the Tinnitus Functional Index benchmark, this single-question threshold is a pragmatic pre-specification rather than a published validated minimum — and we are calibrating it: program patients complete both measures, so the single question's threshold will be anchored to the validated instrument as the dataset grows. We state this openly because a methodology page that hides its weakest link isn't one.
How we report. We report responder rates: the percentage of program patients who meet a pre-specified threshold at the measurement point, alongside the percentage who do not. We publish the failure percentage with the same prominence as the success percentage, because a responder rate is only as credible as the honesty of its denominator. We do not report averages alone, and we do not quote program-level figures until the dataset is large enough that they are stable.
Program design notes. Patients are not assigned to treatment versus no-treatment groups; every patient receives standard care, and program elements are added in sequence, so no care is withheld. This is quality-improvement measurement of a clinical service, not a controlled trial, and we describe it as such. Hearing aid use is recorded and accounted for as a covariate rather than treated as a comparison group. Patients presenting with red-flag features — pulsatile tinnitus, marked asymmetry, sudden onset, or neurologic accompaniment — are referred for medical evaluation before program entry.
Where the numbers come from. Booth-calibrated pitch and loudness matching, minimum masking level in decibels sensation level, full audiologic evaluation, and the instruments above. When our consumer app contributes self-measurements, they are labeled as such and never substituted for clinic measurement.