Quantifying and communicating the burden of COVID-19.

BACKGROUND: An essential aspect of preventing further COVID-19 outbreaks and to learn for future pandemics is the evaluation of different political strategies, which aim at reducing transmission of and mortality due to COVID-19. One important aspect in this context is the comparison of attributable mortality. METHODS: We give a comprehensive overview of six epidemiological measures that are used to quantify COVID-19 attributable mortality (p-score, standardized mortality ratio, absolute number of excess deaths, per capita rate, z-score and the population attributable fraction). RESULTS: By defining the six measures based on observed and expected deaths, we explain their relationship. Moreover, three publicly available data examples serve to illustrate the interpretational strengths and weaknesses of the various measures. Finally, we give recommendation which measures are suitable for an evaluation of public health strategies against COVID-19. The R code to reproduce the results is available as online supplementary material. CONCLUSION: The number of excess deaths should be always reported together with the population attributable fraction, the p-score or the standardized mortality ratio instead of a per capita rate. For a complete picture of COVID-19 attributable mortality, quantifying and communicating its relative burden also to a lay audience is of major importance.

Absolute mortality risk assessment of COVID-19 patients: the Khorshid COVID Cohort (KCC) study.

BACKGROUND: Already at hospital admission, clinicians require simple tools to identify hospitalized COVID-19 patients at high risk of mortality. Such tools can significantly improve resource allocation and patient management within hospitals. From the statistical point of view, extended time-to-event models are required to account for competing risks (discharge from hospital) and censoring so that active cases can also contribute to the analysis. METHODS: We used the hospital-based open Khorshid COVID Cohort (KCC) study with 630 COVID-19 patients from Isfahan, Iran. Competing risk methods are used to develop a death risk chart based on the following variables, which can simply be measured at hospital admission: sex, age, hypertension, oxygen saturation, and Charlson Comorbidity Index. The area under the receiver operator curve was used to assess accuracy concerning discrimination between patients discharged alive and dead. RESULTS: Cause-specific hazard regression models show that these baseline variables are associated with both death, and discharge hazards. The risk chart reflects the combined results of the two cause-specific hazard regression models. The proposed risk assessment method had a very good accuracy (AUC = 0.872 [CI 95%: 0.835-0.910]). CONCLUSIONS: This study aims to improve and validate a personalized mortality risk calculator based on hospitalized COVID-19 patients. The risk assessment of patient mortality provides physicians with additional guidance for making tough decisions.

Methodological evaluation of bias in observational coronavirus disease 2019 studies on drug effectiveness.

BACKGROUND AND OBJECTIVE: Observational studies may provide valuable evidence on real-world causal effects of drug effectiveness in patients with coronavirus disease 2019 (COVID-19). As patients are usually observed from hospital admission to discharge and drug initiation starts during hospitalization, advanced statistical methods are needed to account for time-dependent drug exposure, confounding and competing events. Our objective is to evaluate the observational studies on the three common methodological pitfalls in time-to-event analyses: immortal time bias, confounding bias and competing risk bias. METHODS: We performed a systematic literature search on 23 October 2020, in the PubMed database to identify observational cohort studies that evaluated drug effectiveness in hospitalized patients with COVID-19. We included articles published in four journals: British Medical Journal, New England Journal of Medicine, Journal of the American Medical Association and The Lancet as well as their sub-journals. RESULTS: Overall, out of 255 articles screened, 11 observational cohort studies on treatment effectiveness with drug exposure-outcome associations were evaluated. All studies were susceptible to one or more types of bias in the primary study analysis. Eight studies had a time-dependent treatment. However, the hazard ratios were not adjusted for immortal time in the primary analysis. Even though confounders presented at baseline have been addressed in nine studies, time-varying confounding caused by time-varying treatment exposure and clinical variables was less recognized. Only one out of 11 studies addressed competing event bias by extending follow-up beyond patient discharge. CONCLUSIONS: In the observational cohort studies on drug effectiveness for treatment of COVID-19 published in four high-impact journals, the methodological biases were concerningly common. Appropriate statistical tools are essential to avoid misleading conclusions and to obtain a better understanding of potential treatment effects.

COVID-19 in-hospital mortality and mode of death in a dynamic and non-restricted tertiary care model in Germany.

BACKGROUND: Reported mortality of hospitalised Coronavirus Disease-2019 (COVID-19) patients varies substantially, particularly in critically ill patients. So far COVID-19 in-hospital mortality and modes of death under state of the art care have not been systematically studied. METHODS: This retrospective observational monocenter cohort study was performed after implementation of a non-restricted, dynamic tertiary care model at the University Medical Center Freiburg, an experienced acute respiratory distress syndrome (ARDS) and extracorporeal membrane-oxygenation (ECMO) referral center. All hospitalised patients with PCR-confirmed SARS-CoV-2 infection were included. The primary endpoint was in-hospital mortality, secondary endpoints included major complications and modes of death. A multistate analysis and a Cox regression analysis for competing risk models were performed. Modes of death were determined by two independent reviewers. RESULTS: Between February 25, and May 8, 213 patients were included in the analysis. The median age was 65 years, 129 patients (61%) were male. 70 patients (33%) were admitted to the intensive care unit (ICU), of which 57 patients (81%) received mechanical ventilation and 23 patients (33%) ECMO support. Using multistate methodology, the estimated probability to die within 90 days after COVID-19 onset was 24% in the whole cohort. If the levels of care at time of study entry were accounted for, the probabilities to die were 16% if the patient was initially on a regular ward, 47% if in the intensive care unit (ICU) and 57% if mechanical ventilation was required at study entry. Age ≥65 years and male sex were predictors for in-hospital death. Predominant complications-as judged by two independent reviewers-determining modes of death were multi-organ failure, septic shock and thromboembolic and hemorrhagic complications. CONCLUSION: In a dynamic care model COVID-19-related in-hospital mortality remained very high. In the absence of potent antiviral agents, strategies to alleviate or prevent the identified complications should be investigated. In this context, multistate analyses enable comparison of models-of-care and treatment strategies and allow estimation and allocation of health care resources.

Harmonizing Heterogeneous Endpoints in Coronavirus Disease 2019 Trials Without Loss of Information.

OBJECTIVES: Many trials investigate potential effects of treatments for coronavirus disease 2019. To provide sufficient information for all involveddecision-makers (clinicians, public health authorities, and drug regulatory agencies), a multiplicity of endpoints must be considered. The objectives are to provide hands-on statistical guidelines for harmonizing heterogeneous endpoints in coronavirus disease 2019 clinical trials. DESIGN: Randomized controlled trials for patients infected with coronavirus disease 2019. SETTING: General methods that apply to any randomized controlled trial for patients infected with coronavirus disease 2019. PATIENTS: Coronavirus disease 2019 positive individuals. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We develop a multistate model that is based on hospitalization, mechanical ventilation, death, and discharge. These events are both categories of the ordinal endpoint recommended by the World Health Organization and also within the core outcome set of the Core Outcome Measures in Effectiveness Trials initiative for coronavirus disease 2019 trials. To support our choice of states in the multistate model, we also perform a brief review of registered coronavirus disease 2019 clinical trials. Based on the multistate model, we give recommendation for compact, informative illustration of time-dynamic treatment effects and explorative statistical analysis. A majority of coronavirus disease 2019 clinical trials collect information on mechanical ventilation, hospitalization, and death. Using reconstructed and real data of coronavirus disease 2019 trials, we show how a stacked probability plot provides a detailed understanding of treatment effects on the patients' course of hospital stay. It contributes to harmonizing multiple endpoints and differing lengths of follow-up both within and between trials. CONCLUSIONS: All ongoing clinical trials should include a stacked probability plot in their statistical analysis plan as descriptive analysis. While primary analysis should be on an early endpoint with appropriate capability to be a surrogate (parameter), our multistate model provides additional detailed descriptive information and links results within and between coronavirus disease 2019 trials.

Statistical Analysis of Clinical COVID-19 Data: A Concise Overview of Lessons Learned, Common Errors and How to Avoid Them.

By definition, in-hospital patient data are restricted to the time between hospital admission and discharge (alive or dead). For hospitalised cases of COVID-19, a number of events during hospitalization are of interest regarding the influence of risk factors on the likelihood of experiencing these events. The same is true for predicting times from hospital admission of COVID-19 patients to intensive care or from start of ventilation (invasive or non-invasive) to extubation. This logical restriction of the data to the period of hospitalisation is associated with a substantial risk that inappropriate methods are used for analysis. Here, we briefly discuss the most common types of bias which can occur when analysing in-hospital COVID-19 data.

Joint analysis of duration of ventilation, length of intensive care, and mortality of COVID-19 patients: a multistate approach.

BACKGROUND: The clinical progress of patients hospitalized due to COVID-19 is often associated with severe pneumonia which may require intensive care, invasive ventilation, or extracorporeal membrane oxygenation (ECMO). The length of intensive care and the duration of these supportive therapies are clinically relevant outcomes. From the statistical perspective, these quantities are challenging to estimate due to episodes being time-dependent and potentially multiple, as well as being determined by the competing, terminal events of discharge alive and death. METHODS: We used multistate models to study COVID-19 patients' time-dependent progress and provide a statistical framework to estimate hazard rates and transition probabilities. These estimates can then be used to quantify average sojourn times of clinically important states such as intensive care and invasive ventilation. We have made two real data sets of COVID-19 patients (n = 24* and n = 53**) and the corresponding statistical code publically available. RESULTS: The expected lengths of intensive care unit (ICU) stay at day 28 for the two cohorts were 15.05* and 19.62** days, while expected durations of mechanical ventilation were 7.97* and 9.85** days. Predicted mortality stood at 51%* and 15%**. Patients mechanically ventilated at the start of the example studies had a longer expected duration of ventilation (12.25*, 14.57** days) compared to patients non-ventilated (4.34*, 1.41** days) after 28 days. Furthermore, initially ventilated patients had a higher risk of death (54%* and 20%** vs. 48%* and 6%**) after 4 weeks. These results are further illustrated in stacked probability plots for the two groups from time zero, as well as for the entire cohort which depicts the predicted proportions of the patients in each state over follow-up. CONCLUSIONS: The multistate approach gives important insights into the progress of COVID-19 patients in terms of ventilation duration, length of ICU stay, and mortality. In addition to avoiding frequent pitfalls in survival analysis, the methodology enables active cases to be analyzed by allowing for censoring. The stacked probability plots provide extensive information in a concise manner that can be easily conveyed to decision makers regarding healthcare capacities. Furthermore, clear comparisons can be made among different baseline characteristics.