>> As part of an initiative for quality benchmarking of church-affiliated hospitals, the Quality Indicators for Church-Affiliated Hospitals (QKK), the challenges posed by the pandemic were considered early in the procedures [2]. Specific indicators for COVID-19 were integrated into the indicator set, the risk adjustment for hospital mortality (Hospital Standardized Mortality Ratio, HSMR) was reconsidered, and specific analyses for the member institutions of the QKK on the topic of COVID-19 were carried out and published. After the structure of care [3] was examined in the first wave and treatment concepts [4] in the second wave, the focus of the third wave was on assessing the course of treatment under changing conditions. For example, 741 of the individuals aged 80 and over surveyed in Germany had received at least one vaccination by mid-April 2021 [5]. The alpha variant dominated the third wave. In Germany, the alpha variant was found in one quarter of the tests in calendar week seven 2021 and in three quarters of the tests in calendar week 12 [6].
The study aimed to trace the development of mortality rates in inpatient treatment cases during the first three waves of the Corona pandemic.
material and methods
Routine data from inpatient admissions between January 1, 2020, and June 30, 2021, from 62 hospitals were used for the study. Only cases with DRG reimbursement and discharge up to and including June 30, 2021, were included. A COVID-19 case was identified using the codes U07.1! for "COVID-19, virus detected" or U07.5 (2020) or U10.9 (2021) for "Multisystem inflammatory syndrome associated with COVID-19, unspecified" from the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, German Modification Version (ICD-10-GM). Three waves of the COVID-19 pandemic were distinguished. The start of waves two and three was defined as the Monday of the calendar week with the lowest number of new COVID-19 admissions. The first case of Covid-19 was admitted on January 4, 2021. All cases were assigned to a wave based on the date of hospital admission.
• Wave 1: Recordings made between January 1, 2020 and June 28, 2020
• Wave 2: Recordings between June 29, 2020 and February 21, 2021
• Wave 3: Recordings between February 22, 2021 and June 30, 2021

The cases from the three waves were descriptively compared with regard to general characteristics and ventilation status. To compensate for the shorter observation period in the third wave, a matched-pairs analysis was performed. Matching was carried out across six age groups (under 1 year, 1 to 17 years, 18 to 45 years, 46 to 60 years, 61 to 75 years, 76 years and older), sex (male/female), length of stay in three categories (under 6 days, 6 to 19 days, 20 days and older), ICD-10-GM chapter for the primary diagnosis code, mechanical ventilation (ventilation hours greater than 0, yes/no), and presence of severe acute respiratory infection (SARI) [7], yes/no. Each case was included only once. The three groups were compared using analysis of variance (ANOVA) or the chi-square test. The impact of the third wave was also investigated using logistic regression with hospital mortality as the dependent variable, considering all cases treated for COVID-19. Additional covariates included age in years, DRG of the surgical partition, chapter of the main diagnoses according to ICD-10-GM, sex (male), comorbidity via a comorbidity score [8], emergency, SARI, and transfer. The inclusion method was single-source. The goodness of fit of the model was determined using the area under the curve (AUC) from the ROC analysis and Nagelkerke's R-squared.
The data were managed in Microsoft Access 2013 and analyzed using Microsoft Excel 2013 and IBM SPSS Statistics 25. Significance was assumed at p < 0.05. The terms treatment case, case, and patient are used synonymously.
Results
The study included 1,119,656 inpatient cases, of which 18,044 were due to COVID-19 (1.6%). Figure 1 illustrates the course of the pandemic from a healthcare perspective. While the number of cases and the relative mortality rate correlated in the first two waves, this phenomenon diminished in the third wave, with a smaller increase in COVID-19 deaths. The first wave was assigned 1,721 patients, the second 10,744 patients, and the third 5,579 patients with COVID-19.
Table 1 summarizes some information about the
three waves together.
Between 22.51 and 27.91 of the cases treated were cared for in an intensive care unit, and between 12.81 and 16.71 of the cases required mechanical ventilation. The proportion of patients receiving non-invasive ventilation, or at least undergoing a non-invasive ventilation trial, increased from 25.11 to 68.71 and 77.11. Concurrently, the duration of mechanical ventilation decreased from 384.1 ± 418.00 hours to 244.4 ± 320.45 hours and 226.2 ± 267.71 hours.
Table 2 shows the age shift in the third wave. The proportion of COVID-19 patients aged 76 and older was significantly lower in the third wave (30.11%) than in the first wave (42.41%) and the second wave (49.11%). The 18-45 and 46-60 age groups were more frequently represented in the third wave. Mortality decreased overall in the third wave, particularly in the oldest age group. There, it was more than 10 percentage points lower (62.61%) than in the first wave (75.11%) and the second wave (78.81%). In the third wave, almost 401% of those who died were younger than 76, compared to 251% and 211%, respectively, in the first two waves.
Three groups of 1,620 patients each were formed using matching. According to the matching criteria, the distributions of sex and chapters of the primary diagnosis according to ICD-10-GM were identical. The proportion of patients requiring mechanical ventilation was 16.01 TP3T (258 cases), and the proportion of patients with SARI was 71.01 TP3T (1,150 patients). The age at admission did not differ (p=0.709, ANOVA). The length of stay in the third wave remained shorter at 14.8 ± 12.74 days compared to the second wave (15.2 ± 14.32 days) and the third wave (16.2 ± 16.54 days) (p=0.025, ANOVA). However, the length of stay in the intensive care unit did not differ (p=0.181). There was also no difference in comorbidity (p=0.231, ANOVA). The mean duration of mechanical ventilation in the second and third waves was approximately 300 hours, about 100 hours less than in the first wave (403 hours) (p<0.001, ANOVA). Non-invasive ventilation attempts were more frequent in the second and third waves (741 TP3T and 831 TP3T, respectively) than in the first wave (p<0.001, chi-square test). Mortality among ventilated patients remained high in the third wave at 43.01 TP3T (p=0.617, chi-square test). Hospital mortality had now converged: wave 1 – 16.71 TP3T, wave 2 – 18.11 TP3T, wave 3 – 15.61 TP3T (p=0.169, chi-square test).
In the multivariate logistic regression, wave 2 and wave 3 showed no statistically significant effect (reference was wave 1, p=0.059 and p=0.957 respectively). Age (p<0.001), sex (p<0.001), emergency (p=0.006), referral (p=0.035), DRG of the surgical partition (p<0.001), SARI (p<0.001), and comorbidity quantified via a comorbidity score [8] (p<0.001) increased the risk of death in patients with Covid-19. Among the chapters of the primary diagnosis, neoplasms (Chapter II of the ICD-10-GM, p=0.005) and respiratory diseases (Chapter X, p=0.012) increased the risk of death, while diseases of the digestive system (Chapter XI, p=0.040), diseases of the musculoskeletal system and connective tissue (Chapter XIII, p=0.023), and injuries, poisonings, and certain other consequences of external causes (Chapter XIX, p=0.003) reduced the risk (reference was Chapter I with infectious and parasitic diseases). In the ROC analysis, the AUC was 0.884 (95% confidence interval 0.878–0.890). Nagelkerke's R-squared was 0.431.
discussion
The framework for inpatient care of COVID-19 patients changed significantly during the third wave with the advent of vaccinations against SARS-CoV-2 infection. A lower number of infected individuals was expected, assuming unchanged preventative measures. Infected individuals would likely experience milder cases of COVID-19, and the age structure of those infected would shift due to the age-dependent vaccination strategy. Therefore, a comparison of the third wave with the first and second waves was likely to reveal inconsistent patterns. The results from 62 church-affiliated hospitals belonging to the QKK e.V. confirm this expectation. The relative mortality rate from COVID-19 decoupled from the absolute number of inpatient cases. This suggests not only a lower risk of death during the third wave but also a change in the healthcare system.
The mortality rate of non-ventilated COVID-19 patients had almost halved to 6.61 TP3T. In contrast, the disease course was fatal for 4 out of 10 ventilated patients in the hospital. With a consistently high proportion of ventilated patients at approximately 161 TP3T, the decrease in overall mortality was probably not attributable to an improvement in treatment strategy. This is underscored by the results after matching, which showed no statistically significant difference in mortality across all three waves. The proportion of patients requiring intensive care also remained nearly constant at one-quarter. Despite the consistently poor outcomes of hospitalized COVID-19 cases, significant adjustments in ventilation strategy were observed, which have also been reported elsewhere [9]. The guideline-recommended trial of non-invasive ventilation became established in waves two and three. Three out of four ventilated patients initially or exclusively received mechanical ventilation via a mask. The duration of mechanical ventilation was almost halved, averaging 13 days. Whether the surviving patients benefit from these changes remains unclear. No reduction in mortality was observed. When interpreting the variable "third wave" as a proxy for the treatment strategy, logistic regression also did not suggest any improvement in mortality due to new treatment options. The high goodness of fit of the risk model compared to studies on other serious illnesses such as sepsis [10] supports an accurate interpretation of the results.
The observed decline in mortality is therefore most likely explained by increasing immunization among the older population, with a subsequent change in the hospitalized patient population. In particular, the patients were younger and thus had a lower risk of requiring mechanical ventilation. No evidence of improved treatment options was found. This calls into question the idea of using information on hospital resources as the primary indicator in the fourth wave instead of the frequency of infections, the incidence [11]. If population-based infection control measures no longer focus on preventing the healthcare system from becoming overloaded, but rather on reducing it until available resources are fully utilized, a further increase in the proportion of younger patients in hospitals must be expected. On the one hand, these patients have a lower risk of severe illness and death in hospital compared to the older patients of the first and second waves. On the other hand, the absolute number of younger people who die in hospitals will increase, as will the proportion of younger patients among the deceased. Therefore, basing population-based infection prevention measures on hospital occupancy will increase the disease burden among the younger generation and lead to more frequent experiences with illness, hospitalization, and mechanical ventilation.