Symptom Expectations and Subsequent Symptom Reporting Following COVID-19 Vaccination: A Prospective Naturalistic Study of Nocebo-Related Processes

Abstract Purpose. Response expectancies are central to nocebo models, but naturalistic studies cannot readily distinguish expectancy-related symptom reporting from biological reactions. We tested whether pre-intervention symptom expectations prospectively predicted subsequent reports. Methods. Adults awaiting AstraZeneca or Pfizer-BioNTech vaccination in Poland completed a pre-vaccination assessment in the first half of 2021 (N = 244); 197 completed telephone follow-up about two weeks later. The primary outcome was the number of reported symptoms. A multivariable negative-binomial model adjusted for vaccine type, sex, and age. Symptom-specific logistic models were adjusted for the same covariates and controlled using the Benjamini-Hochberg false-discovery rate. Attrition-weighted and alternative-model analyses assessed robustness. Results. The expected and reported symptom counts were positively correlated (rho = .34, 95% CI [.22, .46]). In the adjusted count model, each additional expected symptom was associated with an 11% higher expected number of reported symptoms (incidence rate ratio = 1.11, 95% CI [1.02, 1.20], p = .011). Vaccine type was not independently associated with symptom count (incidence rate ratio = 0.91, 95% CI [0.61, 1.36]). Of five symptom-specific models, only expected headache remained associated with reported headache after adjustment and false-discovery-rate correction (adjusted odds ratio = 4.06, 95% CI [1.52, 10.88], q = .027). Attrition-weighted analysis produced a similar estimate for expected symptom count. Conclusions. Pre-intervention symptom expectations were independently associated with later symptom reporting in a real-world medical context. The temporal ordering and adjusted analyses are consistent with nocebo-related processes, but the observational design does not establish an expectancy-induced nocebo effect.