# Associations Between Daily Symptoms and Pain Flares in Rheumatoid Arthritis: Case-Crossover mHealth Study

## TL;DR

This study uses mHealth data to show that emotional and sleep changes are linked to pain flares in rheumatoid arthritis patients.

## Contribution

The study introduces a case-crossover mHealth approach to identify factors associated with pain flares in rheumatoid arthritis.

## Key findings

- Pain flares were common, with 88.7% of participants experiencing at least one monthly.
- Worsening mood and anxiety variability over three days increased the likelihood of pain flares.
- Sleepiness variability was also linked to increased flare risk, while sedentary time had no significant effect.

## Abstract

Mobile health (mHealth) technologies, such as smartphones and wearables, enable continuous assessments of individual health information. In chronic musculoskeletal conditions, pain flares, defined as periods of increased pain severity, often coincide with worsening disease activity and cause significant impacts on physical and emotional well-being. Using mHealth technologies can provide insights into individual pain patterns and associated factors.

This study aims to characterize pain flares and identify associated factors in rheumatoid arthritis (RA) by (1) describing the frequency and duration of pain flares using progressively stringent definitions based on pain severity, and (2) exploring associations between pain flares and temporal changes in symptoms across emotional, cognitive, and behavioral domains.

Our 30-day mHealth study collected daily pain severity and related symptoms (scores 1-5, higher are worse) via a smartphone app and passively recorded sleep and physical activity via a wrist-worn accelerometer. Pain flares were defined using a 5-point scale: (1) above average (AA): pain severity > personal median, (2) above threshold (AT): pain severity > 3, and (3) move above threshold (MAT): pain severity moves from 1, 2, 3 to 4 or 5. A case-crossover analysis compared within-person variations of daily symptoms across hazard (3 days before a pain flare) and control (3 days not preceding a pain flare) periods using mean and intraindividual standard deviation. Conditional logistic regression estimated the odds ratio (OR) for pain flare occurrence.

A total of 195 participants (160/195, 82.1% females; mean age 57.2 years; average years with RA: 11.3) contributed 5290 days of data. Of these, 88.7% (173/195) experienced at least 1 AA flare (median monthly rate 4, IQR 2.1-5). Nearly half experienced at least 1 AT or MAT flare (median monthly rate 2, IQR 1-4). These pain flares lasted 2 days (IQR 2-3) on average across definitions, with some extending up to 12 days. Worsening mood over 3 days was associated with a 2-fold increase in the likelihood of AT flares the following day (OR 2.04, IQR 1.06-3.94; P<.05). Greater variability in anxiety over 3 days increased the likelihood of both AT (OR 1.67, IQR 1.01-2.78; P<.05) and MAT flares (OR 1.82, IQR 1.08-3.07; P<.05). Similarly, greater variability in sleepiness (OR 1.89, IQR 1.03-3.47; P<.05) also increased the likelihood of AT flares. Sedentary time (%) consistently showed almost no influence across all definitions. Similarly, the simplest definition of AA demonstrated no significant associations across all symptoms.

Pain flares were commonly observed in RA. Changes in sleep patterns and emotional distress were associated with pain flare occurrences. This analysis demonstrates the potential of daily mHealth data to identify pain flares, opening opportunities for timely monitoring and personalized management.