Abstract
OBJECTIVE: The PaCUDAHL randomized clinical trial evaluated an HPV self-sampling device provided by the family doctor to female patients not participating in the usual opportunistic cervical screening program from 2016 to 2019. Reliable data on the Hawthorne (observer) effect (HE) in clinical trials were lacking. This nested study aimed to verify whether there was a significant difference between participating and non-participating general practitioners (GPs) in the trial, and to measure whether there was an HE in the female patients of participating GPs. RESULTS: We carried out an analytical retrospective cohort study involving 332 GPs and their 70,983 female patients, aged 25-65, registered with the Health Insurance Fund of Flanders, using claims database for the three-year periods 2012-2015 and 2016-2019. Statistical analyses were performed using a linear generalized hierarchical mixed model with geographic level as a random effect. The patients of the 24 participating GPs did not have a cervical cancer screening rate different from that of the non-participating GPs, either before recruitment (p = 0.24) or during the PaCUDAHL trial period (p = 0.15). There were significant increases in cervical cancer screening rates over four years regardless of the group considered (p < 0.0001). In conclusion there was no observer effect but a significant cohort effect.