Vaccination Rates Are Associated With COVID-19 Incidence in Austria

Health authorities around the world are striving to increase COVID-19 vaccination rates in order to curb infections. Although it has been shown that approved COVID-19 vaccines are effective in preventing symptomatic SARS-CoV-2 infection and severe outcomes (1), less is known about their efficacy in inhibiting the spread of the virus causing COVID-19. A recent study in the USA suggests that the spread of COVID-19 is unrelated to vaccination coverage (2), whereas a German study (preprint) maintains that the “current COVID-19 crisis is driven mainly by the unvaccinated” (3). To clarify this issue, we conducted an ecological study predicting COVID-19 incidence rates and their increase on the basis of vaccination rates across all 94 districts of Austria during two phases of infection dynamics: a phase with rising infection rates and a phase when infection rates remained stable.

Figure 2.

Vaccination rates and the increase of the COVID-19 seven-day-average incidence over Austrian districts during a two-week period; every data point represents one discrete district

Method

For the first analysis with rising infection rates, we examined the association between average COVID-19 7-day incidence rates per 100 000 inhabitants (20 October 2021) and the percentage of the population with a valid COVID-19 vaccination certificate (i.e., vaccinations in accordance with the European Medicines Agency) 3 weeks previously (29 September 2021), using the aggregate data of all 94 Austrian districts with a total of 8.9 million inhabitants (source: www.data.gv.at/covid-19/, last accessed on 10 Nov 2021; for Vienna, a city of 23 districts, only summary data were available). This lag time was chosen because protective immunity through the COVID-19 vaccines is typically established about 2 weeks after full vaccination. In addition, we analyzed the change of incidence rates over a period of 14 days (20 October to 3 November). For the second analysis with constant infection rates, we investigated incidence rates on 22 September and their change up to 6 October with the vaccination rates from September 1, using exactly the same design. To assess the associations between COVID-19 incidence and vaccination rates, we calculated correlations using Spearman’s rank correlation coefficients.

Results

For the first analysis (rising infection rates), the median COVID-19 incidence rate was 194.5 (interquartile range [IQR] 140.5, range 0–654.1) cases/100 000 inhabitants, with a median vaccination rate of 59.5% (range 49.1–71.6%). The subsequent change in incidence rates amounted to a median increase of 262.3 (IQR 219.2, range 64.3–1203.5). The results show a moderate negative correlation between vaccination rate and COVID-19 incidence rate (r = -0.60, p < 0.001) across Austria’s 94 districts with their highly varying vaccination rates and COVID-19 incidence rates (figure 1). Vaccination rates were also moderately negatively correlated with the observed increase in COVID-19 incidence rates during this period (r = -0.51, p < 0.001; Figure 2). For the second analysis (stable infection rates), the median incidence rate was 106.4 (IQR 72.4, range 11.7–332.4) cases/100 000 inhabitants, with a median vaccination rate of 57.3% (range 46.9–69.6%). Here, the median increase in incidence rates amounted to a mere 9.4 (IQR 79.0, range -149.2 to 154.3). This analysis again revealed a slight to moderate negative relationship between vaccination rates and COVID-19 incidence rates (r = -0.45, p<.001), but no correlation between vaccination rates and the change in incidence rates (r = -0.14, p = 0.186).

Figure 1.

Vaccination coverage and COVID-19 incidence (7-day average) across Austrian districts; each data point represents a district

Discussion

In contrast to a previous study (2) we found a negative correlation between vaccination and COVID-19 incidence rates both when the incidence was stable and when it increased. However, in accordance with that study (2) we did not find an association between vaccination rates and the change in incidence during a phase of stagnating COVID-19 incidence rates. Therefore, the lack of correlation between COVID-19 vaccination coverage and incidence reported in this previous study may result from the inclusion of data during a period of low infection dynamics. However, more studies are needed to gain insight into the association between vaccination rates and the spread of the COVID-19 infection.

The limitations of our study were the analysis of group-level data, which yields results that are not necessarily valid for the individuals within the groups, and the fact that variables such as age may at least partly influence the associations found. A more complex statistical approach might have achieved a better insight into the data.

Apart from non-pharmacological interventions such as physical distancing and the wearing of masks covering the mouth and nose (4), vaccination remains a cornerstone in the reduction of the spread of SARS-CoV-2. However, COVID-19 infection rates may also depend on factors such as the infectivity of the respective virus variant (e.g., Delta or Omicron), and the time at which COVID-19 vaccine immunity wanes and booster vaccinations become necessary, as described in a recent preprint (5). Future studies will have to take these factors into account.