On the variability of EuroMOMO reporting of all-cause all-ages mortality z-scores

ObjectivesAssess the persistence of data reporting in EuroMOMO web site. Detect weeks with high variability due to amendments carried out after the consolidated publication. Study designRetrospective web examination looking for amendments of weekly all-cause all-ages mortality z-scores published in EuroMOMO site since the year 2018 until october 2023. MethodsPast published all-age all-cause mortality z-scores for all collaborating countries have been recovered monthly from the Internet Archive. The variance of each week reported z-scores is computed accross the recoverd Internet Archive samples. A Poisson distribution fitted to the variance time series allows to detect weeks with significant variance (p < 0.01) and extremely significant variance (p < 0.001) . ResultsOver the entire pool of countries, the study finds 179 weeks with significan variance in the reports, among them 61 weeks of extremely significant variance. France, UK, and Germany have the largest number of these weeks. ConclusionsVariance in reported z-scores due to amendments do not show systematic patterns that may be attributed to modeling updates. On the other hand, they appear to have some association with COVID-19 pandemic events. Full publication of data and models should clarify the sources of non-persistence and in the published reports.