Impact of the COVID-19 on the Seasonal Mortality Pattern in Spain: Unexpected Change in Trend as an Objective and Early Marker of Health Alert

Background: The first death by COVID-19 in Spain was on February 13th. However, on that precise date the tracking of positive cases of infection was mainly limited to subjects at high risk of contagion. The aim of this study was to determine whether unexpected changes in trend of the all-cause mortality curve, could serve as an indicator to provide early preventive interventions. Method: To analyze all-cause mortality data a Poisson distribution was selected with a Log function. The goodness of fit statistics was studied, and both the predicted value of mean of response and residual were obtained. For daily mortality, a seasonal decomposition was performed by weighted moving averages method.Finding: The results showed a constantly increasing risk from the ninth week (Sunday 23 rd February). Three weeks before lockdown. Adjusted RR: 0·403 (Week 8), 0·406 (W9), 0·408 (W10), 0·446 (W11), 0·583 (W12), P < 0·001 in all cases. In males the risk increased from the same week. Adjusted RR: 0·372 (W8), 0·386 (W9), 0·388 (W10), 0·429 (W11), 0·578 (W12), P<0·001. In contrast, for females the risk started to increase from week 11. By groups of age there were again two weeks of difference between the population under 65 years (W 11) vs over 74 years of age (W 9). Six days were earned using baseline years with similar seasonal trend pattern. Interpretation: Facing an unknown viral process, epidemiological surveillance must pay attention to subtle modifications of seasonal trend patterns to take appropriate preventive measures as early as possible.Funding: None.Declaration of Interests: No competing interest. Ethics Approval Statement: Non-applicable.