Comparative analysis of lethality in hospitalized patients during the waves in contrast to the interwave period

We used Fisher exact test to evaluate whether the proportion of survivors and deaths among hospitalized patients showed differences between the periods evaluated. Since the ratio between survivors and deaths is proportional to the CFR, we concluded that the CFR among hospitalized patients changed during the periods of greater transmission and greater demand for hospitalization.

The following contrasts were made to assess the significance of the prior changes: 1) comparison between hospitalized patients during either of the two waves and in the interwave period and 2) comparison between hospitalized patients in the first peak and the second peak. The comparison was made using Fisher exact test, considering a contingency table (see Table S1 in the Online Supplementary Document).

The P-values of the multiple comparisons were corrected using the False Discovery Rate (FDR), and an FDR value <0.05 was considered significant.

To avoid Simpson Paradox issues that might obscure differences, these comparisons were performed in subpopulations defined by age group and the combinations of the three comorbidities studied; additionally, the subpopulation of patients not presenting with any of the three specified comorbidities but with any other recorded comorbidity was studied separately from the subpopulation with no recorded comorbidity. We additionally, compared the interwave period with the two waves to gain a better notion of changes in lethality associated with periods of increased transmission and associated with greater demand for hospitalization. Thus, one of our objectives was to evaluate whether there were changes in lethality between the interwave period and the second wave. The rationale was that while in the first wave knowledge of the disease was scarce and detection programs were being developed, by the interwave period, several interventions had been adopted, including increased testing programmes. Given that lethality increases during periods of higher transmission and demand for hospitalization, we calculated the Pearson correlation between weekly CFR for hospitalized patients and hospital occupancy as reported [20].