# Validity of EQ-5D-5L breathing and cognition bolt-ons in non-hospitalized patients after COVID-19

## TL;DR

This study evaluates the usefulness of adding breathing and cognition questions to a health questionnaire for non-hospitalized patients recovering from COVID-19.

## Contribution

The study validates the addition of breathing and cognition bolt-on items to the EQ-5D-5L for assessing post-COVID-19 outcomes.

## Key findings

- Breathing bolt-on items showed moderate correlation with pulmonary function tests and dyspnea scores.
- Cognition bolt-on items had similar patterns to existing questionnaires but with slightly higher ceiling effects.
- The breathing bolt-on added 2% to the explained variation in health scores, while the cognition bolt-on contributed little.

## Abstract

The EQ-5D has been criticized for lacking dimensions for breathing and cognition. In acute COVID-19 and Long COVID, breathing problems and cognitive complaints are common. This study evaluated the inclusion of EQ-5D-5L (bolt-on) items that assess these dimensions.

In a follow-up of community-based non-hospitalized patients with COVID-19 in 2020 (n = 450), we used respiratory and cognitive bolt-on items alongside other questionnaires about 30 months later (n = 220). We assessed data quality and construct validity for the bolt-ons, including comparison with concurrently used instruments’ domain and item scores. Bolt-on scores were correlated with pulmonary function tests and tablet-based cognitive tests about 1 year earlier. Finally, we assessed the contribution of the bolt-ons to variability of the EQ VAS.

Most patients had none or slight problems with breathing and cognition. The breathing problems bolt-on had an 8% lower ceiling effect, but otherwise a similar response distribution to the mMRC. Cognition bolt-on and four DSQ-SF item scores relating to frequency and severity were similar, except 2–13% higher ceiling effects for the former. The breathing bolt-on had a rank correlation of 0.54 with mMRC scores and higher correlations with Dyspnoea-12 scale scores. EQ-5D-5L dimensions explained 54% of variation in EQ VAS scores; the breathing bolt-on contributed a further 2%, but the cognition bolt-on contributed very little.

The results support the validity of the two bolt-ons in follow-up of non-hospitalized patients after COVID-19. Adding items of respiratory and cognitive symptoms may enhance the EQ-5D's appropriateness in respiratory and cognitive research and clinical practice.