Psychometric validation of a cognition and social participation bolt-on for the EQ-5D-5L in SARS-CoV-2 infected German healthcare workers.

PURPOSE: COVID-19 can result in long-term impairments, including cognitive difficulties and restrictions in social participation, which may not be fully captured by EQ-5D-5L. This study examined whether adding cognition (CO) and social participation (SP) bolt-ons improves EQ-5D-5L’s measurement properties in German healthcare workers (HCW) with SARS-CoV-2 infection.

METHODS

N = 3335 HCW with self-reported occupational COVID-19 completed an online survey including EQ-5D-5L, two candidate bolt-ons (CO, SP), and validated self-report instruments (e.g., Post-COVID Syndrome PCS-Score, PHQ-4, PTSD screening, SSD-12, WAI, WHODAS). Psychometric analyses covered distributional characteristics (response pattern, missing values, ceiling), construct (convergent and divergent) validity, known-groups validity, and explanatory power.

RESULTS

Both bolt-ons showed acceptable distributional properties; adding CO modestly reduced overall ceiling effect, while adding SP resulted in negligible change (‘11111’ = 18.8% vs. ‘111111’CO = 14.8% and ‘111111’SP = 17.8%). Construct validity was supported by expected correlation patterns, i.e. rCO,PCS-Score = 0.63, rSP, WHODAS = 0.71. Known-groups validity improved with the inclusion of bolt-ons, as reflected by higher or comparable relative efficiency (RE) compared with EQ-5D-5L in most group comparisons for CO (11/13, RE = 0.94–1.22), whereas this was observed in approximately half of group comparisons for SP (6/13, RE = 0.89–1.18). In multivariate models, adding CO to the EQ-5D-5L resulted in a small to moderate increase in explained variance for PCS symptom severity (Δ adj. R2CO = 0.04–0.07), whereas adding SP had a negligible impact (Δ adj. R2SP ≤ 0.01).

CONCLUSION

While CO improved ceiling, construct, and known-groups validity of the EQ-5D-5L in SARS-CoV-2 infected HCW, the added value of SP appeared limited.