Using Fourier components analysis to examine variations in outcome scores for individuals with Meniere’s Disease

Problem

Meniere’s disease is an incurable, chronic disorder of the inner ear, with patients experiencing varying levels of severity in hearing loss, tinnitus, aural fullness and vertigo. It affects around 120,000 people in the UK, with the majority being of working age. The symptoms experienced by patients significantly impact on patients’ quality of life, with increased incidences of social isolation amongst sufferers. Rhythmic fluctuations of symptoms (also known as chronobiology) at different time-points have been previously reported for patients with chronic conditions. However, there has been a lack of literature on how time of the day affects Meniere’s symptoms. Focusing on the fluctuating patterns of Meniere’s symptoms may support clinical practitioners in better understanding, supporting and diagnosing patients.

Approach

A pre-existing dataset was provided which used a mobile app to collect data from individuals with Meniere's disease. The "Meniere's Monitor" app allowed individuals to monitor their symptoms on a daily basis collating information on their level of severity in dizziness, aura fullness, tinnitus, and hearing loss. In addition, other questions regarding stress, sleep quality and demographics were asked. The data collection period was between 2015 and 2017 and a total of 853 individuals provided data. Variability of symptom severity over a 24-hour period was assessed with a multivariate mixed-effects regression model using Fourier components. Time transformations, using sine and cosine functions, were created and added to the four main symptoms. Adjustments were made to differentiate trends from demographic factors.

Findings

The majority of participants were female (68.3%), and the mean age of the sample was 48.9 years. Over half of the participants were employed (58.3%), and from Europe (54.2%). There was variation in all four symptoms across a 24-hour period. Peak aura fullness occurred between 4pm and 8pm compared to midnight. Tinnitus severity peaked at three points in the day from early morning (6am), midday (12pm), and 5pm. Dizziness symptoms peaked at different points in the day, mainly 10am and 3pm.

Consequences

Usage of fourier analysis enabled variability of Meniere’s symptoms to be captured and analysed over a 24-hour period demonstrating peaks of symptoms at different times of the day. This analytic method would be useful for patients in better understanding and managing the disease ultimately affecting their overall wellbeing. More tailored interventions using the time of day information Fourier analysis provides could be designed and implemented by clinicians.