key: cord-0282431-zm8md4pp authors: Baunez, C.; Degoulet, M.; Luchini, S.; Pintus, P.; Teschl, M. title: COVID-19 Acceleration and Vaccine Status in France - August 2021 date: 2021-09-22 journal: nan DOI: 10.1101/2021.09.18.21263773 sha: c5d7ea7e6b35cf5bcecd983ea9d9812f10b7752d doc_id: 282431 cord_uid: zm8md4pp Abstract: Objectives: This note provides an assessment of COVID-19 acceleration among groups with difference vaccine status in France. Methods: We assess viral acceleration using a novel indicator introduced in Baunez et al. (2021). The acceleration index relates the percentage change of tests that have been performed on a given day to the percentage change in the associated positive cases that same day. We compare viral acceleration among vaccinated and unvaccinated individuals in France over the period May 31st - August 29, 2021. Results: Once the state of the epidemic within each groups is accounted for, it turns out that viral acceleration has since mid-July converged to similar levels among vaccinated and unvaccinated individuals in France, even though viral speed is larger for the latter group compared to the former. Conclusion: Our results call for an increasing testing effort for both vaccinated and unvaccinated individuals, in view of the fact that viral circulation is currently accelerating at similar levels for both groups in France. Existing vaccines against SARS-CoV-2 currently on the market in Western societies are said to have a high efficacy in reducing the number of symptomatic cases, and also in cutting transmission (e.g. is believed to spread faster than other variants" (Planas et al. 2021 ) and the US-based Centers for Disease Control and Prevention (CDC) indicates for example that fully vaccinated people with the Delta variant can spread the virus, although it appears for a shorter period of time. 1 It is therefore of great importance for any public health authority to be able to observe correctly the viral spread, including among the vaccinated population. In France, we observe that at the end of August 2021, about 6 times more new positive cases are found among the unvaccinated individuals, compared to the vaccinated. This indicates that new infections are much larger for the unvaccinated but those absolute levels do not correct for the fact that unvaccinated people are tested much more each day (about 3 times more, as of August 29, 2021). The usual way to account for tests is to compute the ratio of the number of new positive cases generated on a given day to the number of new tests performed on that same day, the positivity rate. At the end of August, the positivity rate for the unvaccinated about is twice as large compared to that of the vaccinated. 2 This suggests that the virus is spreading faster among the non-vaccinated than the vaccinated population and one may conclude that vaccines have successfully contributed in slowing down viral transmission among the vaccinated group. The difficulty with this conclusion however is that a direct comparison relies on the implicit assumption that those positivity rates do not change with sample size. If, for instance, the number of tests of vaccinated would be increased to match those of non-vaccinated, their positivity rate is likely to differ. An alternative approach is to monitor whether the pandemic is decelerating or accelerating, that is, whether harm (here represented as the number of cases) is accelerating or decelerating (Taleb 2012 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint In Baunez et al. [1] we propose an acceleration index, which is the percentage change in cases given a percentage change in tests. The intuition is that if we increase tests by, say, one percent and find more than one percent of cases, we infer that the pandemic is accelerating. If we find less than one percent of cases, it is decelerating. The index is basically an elasticity, which measures the responsiveness or sensitivity of one variable (here cases) in response to another variable (here tests). Most importantly, it gives us a scale-free number, which can be compared between groups such that if of two groups one has a higher elasticity, this group has a higher acceleration/lower deceleration in comparison to the other one. As long as acceleration is greater than one, the pandemic is not under control, whether the daily positivity rate is small or not. In this paper, we calculate the acceleration index for the vaccinated and unvaccinated population since the start of the vaccination campaign in France until the end of current available data, which is August, 29 2021. What we find is that despite vaccination, the pandemic is still accelerating in both population groups. Moreover, despite lower absolute numbers among the vaccinated, viral spread accelerates similarly among the vaccinated and non-vaccinated populations. This finding suggests that vaccines cannot be the only mean to dampen the pandemic. Testing remains an important instrument to observe viral spread together with additional health policies, if necessary, such as, for example, contact tracing, quarantine, mask-wearing and social distancing. In Baunez et al. the cumulative numbers of positive and diagnosed persons up to date t. Importantly, variables P t and D t therefore define the state of the epidemic at date t, in terms of the current stocks of positives and tests, cumulated since date 1. In fact, those states reflect the history of each variable and it might differ across the groups of individuals that we consider below, due to vaccine status. The acceleration index, denoted ε T at date T , thus gives the percentage change of cases divided by the percentage change of testing as follows: where the latter equality follows from the very definition of cumulated variables, the variations of which are just daily flows. When ε T > 1 we say that the epidemic is accelerating (it is on the loose), 3 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted September 22, 2021. ; https://doi.org/10.1101/2021.09.18.21263773 doi: medRxiv preprint since a given growth rate of cumulated tests produces a larger growth rate of cumulated positive cases, while it is decelerating (the pandemic is under control) when ε T < 1. As a consequence, our indicator can tightly be linked to an arguably desirable objective of public health policy, which is to get proportionally less infected people when tests are increasing. Rearranging the terms of the latter equality, we see that the acceleration index relates to the daily and average positivity rates, in the following way: where the average positivity rate is here defined as the ratio of stocks at end date T . In sum, the acceleration index is the elasticity of cumulated positive cases to cumulated tests (that is, the ratio of their growth rates), which may also be decomposed as the ratio between daily and average positivity rates. It then follows that if the daily positivity rate is always constant, then ε T = 1 because current and average speeds are equal. If ε T > 1, daily positivity rate exceeds average rate, indicating acceleration, while deceleration prevails when ε T < 1. We compute the acceleration index We also estimate a Pearson correlation coefficient on a rolling window of 7 days between the acceleration index for the fully vaccinated and that of the recent incomplete vaccination, efficient incomplete vaccination, not vaccinated. This provides a simple measure of moment-by-moment local synchrony of viral propagation across groups with different vaccine status. Our main results are depicted in Figure 1 . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted September 22, 2021. ; https://doi.org/10.1101/2021.09.18.21263773 doi: medRxiv preprint expressed through the daily positivity rate with the "past" viral spread or average positivity rate. Intuitively this means that if today we find more cases per tests than we did on average on all previous days, then the pandemic is accelerating, if we find less cases per tests today than we did on average in the past, the pandemic decelerates. This is exactly what the acceleration index does and in panel (c) we report its level for the 4 groups. Our acceleration index is scale-free since, as we see in Equation (2), the daily positivity rate, i.e. the daily flows of positives and tests for a particular day, respectively p and d, are corrected by the average positivity rate, i.e. their stock values, respectively P and D for that particular day. We therefore see in panel (c) of Figure 1 that viral acceleration is essentially similar among fully vaccinated and unvaccinated individuals. This is because even though the unvaccinated have a larger daily positivity rate than the vaccinated, they also have a larger average rate: unvaccinated people are less protected against SARS-CoV-2 infection and therefore also have a larger P/D over the epidemic past. Although both groups currently have different viral speed (that is, daily positivity rates), their acceleration indices turn out to be very similar. What is also striking in panel (c) is how close the acceleration indices stay over the entire period. More precisely, the resurgence of the acceleration regime, at the beginning of the summer break in early July, starts a little bit earlier for fully vaccinated individuals compared to unvaccinated people in France. The level of acceleration is about 1.2 for the two groups on August 29, 2021. This means with 1% of additional tests (the total number of test, or its stock, is increased by 1%), we find 1.2% additional cases (the total number of cases, or stock of positive cases, increases by 1.2%) in both groups. Acceleration for both groups reached its peak during the second half of July ( a factor of about 3) and is steadily decreasing since the beginning of August. But the pandemic remains in an acceleration regime for both groups. The acceleration indices have been very much synchronized since the resurgence of the acceleration regime, starting early July. This is confirmed in panel (d), where we report the pairwise Pearson correlation coefficients on a rolling window of 7 days. In particular, the correlation between acceleration indices for fully vaccinated and unvaccinated individuals has reached its highest level mid-July and is still close to 1 at the end of the sample. In sum, our acceleration index shows no large difference across vaccinated and unvaccinated individuals since early/mid July. The main result that we document in Figure 1 is that in France, COVID-19 acceleration has been of similar magnitude among both the vaccinated and the unvaccinated population since early July 2021, despite the daily positivity rate being larger among the latter group. Positivity rate is a scaledependent indicator that roughly informs us about viral spread or speed, but what is important to know is whether viral spread is accelerating or decelerating. Our acceleration index is a scale-6 . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted September 22, 2021. ; https://doi.org/10.1101/2021.09.18.21263773 doi: medRxiv preprint independent measure that indicates the percentage increase in cases following a percentage increase in tests and thus captures better whether the pandemic is under control or not. Our indicator can also serve as a basis for public health policy to curb COVID-19, that, if successful, would be one that sees ever fewer positive cases as testing efforts increase. In such a case, our indicator would clearly indicate deceleration. The fact that acceleration persists even among the vaccinated population is broadly consistent with recent biological evidence that vaccinated and unvaccinated individuals have similar viral loads where the SARS-CoV-2 delta variant is most prevalent and that vaccine breakthroughs are seen more often in comparison with previous variants (see Brown regime. Indeed, research shows that test sensitivity is secondary to frequency for efficient COVID-19 screening (Larremore et al., 2021) and that there should not be a "one size fits all" testing strategy, primarily based on PCR-screening (Mina and Andersen 2021). The decision of the French authorities for example to concentrate mainly on vaccination to combat COVID-19, to make access Antigen-testing difficult and costly for individuals, to focus on PCR-testing by prescription only and to impose a vaccine passport ("pass sanitaire") that gives certain advantages to vaccinated people, notably to travel and to visit social gatherings without prior testing is questionable on the grounds that it does not guarantee to be effective to help curbing sustainedly the COVID-19 acceleration that is still observed in France. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted September 22, 2021. ; https://doi.org/10.1101/2021.09.18.21263773 doi: medRxiv preprint Tracking the Dynamics and Allocating Tests for COVID-19 in Real-Time: an Acceleration Index with an Application to French Age Groups and Départements Outbreak of SARS-CoV-2 Infections, Including COVID-19 Vaccine Breakthrough Infections Estimating the extent of asymptomatic COVID-19 and its potential for community transmission: Systematic review and meta-analysis Centers for Disease Control and Prevention (CDC) (2021): Science Brief: COVID-19 Vaccines and Vaccination Transmission of SARS-CoV-2 Delta Variant Among Vaccinated Healthcare Workers, Vietnam. SSRN working paper Virological and serological kinetics of SARS-CoV-2 Delta variant vaccine-breakthrough infections: a multi-center cohort study Comparing SARS-CoV-2 natural immunity to vaccine-induced immunity: reinfections versus breakthrough infections. MedRχv working paper Test sensitivity is secondary to frequency and turnaround time for COVID-19 screening Impact of BNT162b2 vaccination and isolation on SARS-CoV-2 transmission in Israeli households: an observational study. MedRχv working paper Viral infection and transmission in a large well-traced outbreak caused by the Delta SARS-CoV-2 variant Delta variants of SARS-CoV-2 cause significantly increased 8 vaccine breakthrough COVID-19 cases in Houston, Texas MedRχv working paper COVID-19 vaccine efficacy and effectivenessâthe elephant (not) in the room Reduced sensitivity of SARS-CoV-2 variant Delta to antibody neutralization Vaccination with BNT162b2 reduces transmission of SARS-CoV-2 to household contacts in Israel SARS-CoV-2 variants of concern and variants under investigationin England Comparison of two highly-effective mRNA vaccines for COVID-19 during periods of Alpha and Delta variant prevalence. MedRχv working paper Shedding of Infectious SARS-CoV-2 Despite Vaccination when the Delta Variant is Prevalent -Wisconsin The indirect effect of mRNA-based Covid-19 vaccination on unvaccinated household members Ineffective neutralization of the SARS-CoV-2 Mu variant by convalescent and vaccine sera It is made available under a perpetuity.is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint