Federated causal inference based on real-world observational data sources: application to a SARS-CoV-2 vaccine effectiveness assessment.

INTRODUCTION: Causal inference helps researchers and policy-makers to evaluate public health interventions. When comparing interventions or public health programs by leveraging observational sensitive individual-level data from populations crossing jurisdictional borders, a federated approach (as opposed to a pooling data approach) can be used. Approaching causal inference by re-using routinely collected observational data across different regions in a federated manner, is challenging and guidance is currently lacking. With the aim of filling this gap and allowing a rapid response in the case of a next pandemic, a methodological framework to develop studies attempting causal inference using federated cross-national sensitive observational data, is described and showcased within the European BeYond-COVID project. METHODS: A framework for approaching federated causal inference by re-using routinely collected observational data across different regions, based on principles of legal, organizational, semantic and technical interoperability, is proposed. The framework includes step-by-step guidance, from defining a research question, to establishing a causal model, identifying and specifying data requirements in a common data model, generating synthetic data, and developing an interoperable and reproducible analytical pipeline for distributed deployment. The conceptual and instrumental phase of the framework was demonstrated and an analytical pipeline implementing federated causal inference was prototyped using open-source software in preparation for the assessment of real-world effectiveness of SARS-CoV-2 primary vaccination in preventing infection in populations spanning different countries, integrating a data quality assessment, imputation of missing values, matching of exposed to unexposed individuals based on confounders identified in the causal model and a survival analysis within the matched population. RESULTS: The conceptual and instrumental phase of the proposed methodological framework was successfully demonstrated within the BY-COVID project. Different Findable, Accessible, Interoperable and Reusable (FAIR) research objects were produced, such as a study protocol, a data management plan, a common data model, a synthetic dataset and an interoperable analytical pipeline. CONCLUSIONS: The framework provides a systematic approach to address federated cross-national policy-relevant causal research questions based on sensitive population, health and care data in a privacy-preserving and interoperable way. The methodology and derived research objects can be re-used and contribute to pandemic preparedness.

Comparing the incidence of SARS-CoV-2 across age groups considering sampling biases - use of testing data of autumn 2021 in Belgium.

BACKGROUND: To design efficient mitigation measures against COVID-19, understanding the transmission dynamics between different age groups was crucial. The role of children in the pandemic has been intensely debated and involves both scientific and ethical questions. To design efficient age-targeted non-pharmaceutical interventions (NPI), a good view of the incidence of the different age groups was needed. However, using Belgian testing data to infer real incidence (RI) from observed incidence (OI) or positivity ratio (PR) was not trivial. METHODS: Based on Belgian testing data collected during the Delta wave of Autumn 2021, we compared the use of different estimators of RI and analyzed their effect on comparisons between age groups. RESULTS: We found that the RI estimator's choice strongly influences the comparison between age groups. CONCLUSION: The widespread implementation of testing campaigns using representative population samples could help to avoid pitfalls related to the current testing strategy in Belgium and worldwide. This approach would also allow a better comparison of the data from different countries while reducing biases arising from the specificities of each surveillance system.

Homologous and Heterologous Prime-Boost Vaccination: Impact on Clinical Severity of SARS-CoV-2 Omicron Infection among Hospitalized COVID-19 Patients in Belgium.

We investigated effectiveness of (1) mRNA booster vaccination versus primary vaccination only and (2) heterologous (viral vector-mRNA) versus homologous (mRNA-mRNA) prime-boost vaccination against severe outcomes of BA.1, BA.2, BA.4 or BA.5 Omicron infection (confirmed by whole genome sequencing) among hospitalized COVID-19 patients using observational data from national COVID-19 registries. In addition, it was investigated whether the difference between the heterologous and homologous prime-boost vaccination was homogenous across Omicron sub-lineages. Regression standardization (parametric g-formula) was used to estimate counterfactual risks for severe COVID-19 (combination of severity indicators), intensive care unit (ICU) admission, and in-hospital mortality under exposure to different vaccination schedules. The estimated risk for severe COVID-19 and in-hospital mortality was significantly lower with an mRNA booster vaccination as compared to only a primary vaccination schedule (RR = 0.59 [0.33; 0.85] and RR = 0.47 [0.15; 0.79], respectively). No significance difference was observed in the estimated risk for severe COVID-19, ICU admission and in-hospital mortality with a heterologous compared to a homologous prime-boost vaccination schedule, and this difference was not significantly modified by the Omicron sub-lineage. Our results support evidence that mRNA booster vaccination reduced the risk of severe COVID-19 disease during the Omicron-predominant period.

Evaluating methodological approaches to assess the severity of infection with SARS-CoV-2 variants: scoping review and applications on Belgian COVID-19 data.

BACKGROUND: Differences in the genetic material of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants may result in altered virulence characteristics. Assessing the disease severity caused by newly emerging variants is essential to estimate their impact on public health. However, causally inferring the intrinsic severity of infection with variants using observational data is a challenging process on which guidance is still limited. We describe potential limitations and biases that researchers are confronted with and evaluate different methodological approaches to study the severity of infection with SARS-CoV-2 variants. METHODS: We reviewed the literature to identify limitations and potential biases in methods used to study the severity of infection with a particular variant. The impact of different methodological choices is illustrated by using real-world data of Belgian hospitalized COVID-19 patients. RESULTS: We observed different ways of defining coronavirus disease 2019 (COVID-19) disease severity (e.g., admission to the hospital or intensive care unit versus the occurrence of severe complications or death) and exposure to a variant (e.g., linkage of the sequencing or genotyping result with the patient data through a unique identifier versus categorization of patients based on time periods). Different potential selection biases (e.g., overcontrol bias, endogenous selection bias, sample truncation bias) and factors fluctuating over time (e.g., medical expertise and therapeutic strategies, vaccination coverage and natural immunity, pressure on the healthcare system, affected population groups) according to the successive waves of COVID-19, dominated by different variants, were identified. Using data of Belgian hospitalized COVID-19 patients, we were able to document (i) the robustness of the analyses when using different variant exposure ascertainment methods, (ii) indications of the presence of selection bias and (iii) how important confounding variables are fluctuating over time. CONCLUSIONS: When estimating the unbiased marginal effect of SARS-CoV-2 variants on the severity of infection, different strategies can be used and different assumptions can be made, potentially leading to different conclusions. We propose four best practices to identify and reduce potential bias introduced by the study design, the data analysis approach, and the features of the underlying surveillance strategies and data infrastructure.

Clinical Severity of SARS-CoV-2 Omicron Variant Compared with Delta among Hospitalized COVID-19 Patients in Belgium during Autumn and Winter Season 2021-2022.

This retrospective multi-center matched cohort study assessed the risk for severe COVID-19 (combination of severity indicators), intensive care unit (ICU) admission, and in-hospital mortality in hospitalized patients when infected with the Omicron variant compared to when infected with the Delta variant. The study is based on a causal framework using individually-linked data from national COVID-19 registries. The study population consisted of 954 COVID-19 patients (of which, 445 were infected with Omicron) above 18 years old admitted to a Belgian hospital during the autumn and winter season 2021-2022, and with available viral genomic data. Patients were matched based on the hospital, whereas other possible confounders (demographics, comorbidities, vaccination status, socio-economic status, and ICU occupancy) were adjusted for by using a multivariable logistic regression analysis. The estimated standardized risk for severe COVID-19 and ICU admission in hospitalized patients was significantly lower (RR = 0.63; 95% CI (0.30; 0.97) and RR = 0.56; 95% CI (0.14; 0.99), respectively) when infected with the Omicron variant, whereas in-hospital mortality was not significantly different according to the SARS-CoV-2 variant (RR = 0.78, 95% CI (0.28-1.29)). This study demonstrates the added value of integrated genomic and clinical surveillance to recognize the multifactorial nature of COVID-19 pathogenesis.

The association between area deprivation and COVID-19 incidence: a municipality-level spatio-temporal study in Belgium, 2020-2021.

BACKGROUND: In Belgium, current research on socio-economic inequalities in the coronavirus disease 2019 (COVID-19) crisis has mainly focused on excess mortality and data from the first epidemiological wave. The current study adds onto this by examining the association between COVID-19 incidence and area deprivation during the first five wave and interwave periods, thus adding a temporal gradient to the analyses. METHODS: We use all confirmed COVID-19 cases between March 2020 and June 2021 in Belgium, aggregated at the municipality-level. These data were collected by the national laboratory-based COVID-19 surveillance system. A level of area deprivation was assigned to each Belgian municipality using data of three socio-economic variables: the share of unemployed persons in the active population, the share of households without a car and the share of low-educated persons. The spatio-temporal association between COVID-19 incidence and area deprivation was assessed by performing multivariate negative-binomial regression analyses and computing population attributable fractions. RESULTS: A significant association between COVID-19 incidence and area deprivation was found over the entire study period, with the incidence in the most deprived areas predicted to be 24% higher than in the least deprived areas. This effect was dependent on the period during the COVID-19 crisis. The largest socio-economic inequalities in COVID-19 infections could be observed during wave 2 and wave 3, with a clear disadvantage for deprived areas. CONCLUSION: Our results provide new insights into spatio-temporal patterns of socio-economic inequalities in COVID-19 incidence in Belgium. They reveal the existence of inequalities and a shift of these patterns over time.

Association between COVID-19 Primary Vaccination and Severe Disease Caused by SARS-CoV-2 Delta Variant among Hospitalized Patients: A Belgian Retrospective Cohort Study.

We aimed to investigate vaccine effectiveness against progression to severe COVID-19 (acute respiratory distress syndrome (ARDS), intensive care unit (ICU) admission and/or death) and in-hospital death in a cohort of hospitalized COVID-19 patients. Mixed effects logistic regression analyses were performed to estimate the association between receiving a primary COVID-19 vaccination schedule and severe outcomes after adjusting for patient, hospital, and vaccination characteristics. Additionally, the effects of the vaccine brands including mRNA vaccines mRNA-1273 and BNT162b2, and adenovirus-vector vaccines ChAdOx1 (AZ) and Ad26.COV2.S (J&J) were compared to each other. This retrospective, multicenter cohort study included 2493 COVID-19 patients hospitalized across 73 acute care hospitals in Belgium during the time period 15 August 2021-14 November 2021 when the Delta variant (B1.617.2) was predominant. Hospitalized COVID-19 patients that received a primary vaccination schedule had lower odds of progressing to severe disease (OR (95% CI); 0.48 (0.38; 0.60)) and in-hospital death (OR (95% CI); 0.49 (0.36; 0.65)) than unvaccinated patients. Among the vaccinated patients older than 75 years, mRNA vaccines and AZ seemed to confer similar protection, while one dose of J&J showed lower protection in this age category. In conclusion, a primary vaccination schedule protects against worsening of COVID-19 to severe outcomes among hospitalized patients.

One year of laboratory-based COVID-19 surveillance system in Belgium: main indicators and performance of the laboratories (March 2020-21).

BACKGROUND: With the spread of coronavirus disease 2019 (COVID-19), an existing national laboratory-based surveillance system was adapted to daily monitor the epidemiological situation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in the Belgium by following the number of confirmed SARS-CoV-2 infections, the number of performed tests and the positivity ratio. We present these main indicators of the surveillance over a one-year period as well as the impact of the performance of the laboratories, regarding speed of processing the samples and reporting results, for surveillance. METHODS: We describe the evolution of test capacity, testing strategy and the data collection methods during the first year of the epidemic in Belgium. RESULTS: Between the 1st of March 2020 and the 28th of February 2021, 9,487,470 tests and 773,078 COVID-19 laboratory confirmed cases were reported. Two epidemic waves occurred, with a peak in April and October 2020. The capacity and performance of the laboratories improved continuously during 2020 resulting in a high level performance. Since the end of November 2020 90 to 95% of the test results are reported at the latest the day after sampling was performed. CONCLUSIONS: Thanks to the effort of all laboratories a performant exhaustive national laboratory-based surveillance system to monitor the epidemiological situation of SARS-CoV-2 was set up in Belgium in 2020. On top of expanding the number of laboratories performing diagnostics and significantly increasing the test capacity in Belgium, turnaround times between sampling and testing as well as reporting were optimized over the first year of this pandemic.

Conceptual causal framework to assess the effect of SARS-CoV-2 variants on COVID-19 disease severity among hospitalized patients.

BACKGROUND: SARS-CoV-2 strains evolve continuously and accumulate mutations in their genomes over the course of the pandemic. The severity of a SARS-CoV-2 infection could partly depend on these viral genetic characteristics. Here, we present a general conceptual framework that allows to study the effect of SARS-CoV-2 variants on COVID-19 disease severity among hospitalized patients. METHODS: A causal model is defined and visualized using a Directed Acyclic Graph (DAG), in which assumptions on the relationship between (confounding) variables are made explicit. Various DAGs are presented to explore specific study design options and the risk for selection bias. Next, the data infrastructure specific to the COVID-19 surveillance in Belgium is described, along with its strengths and weaknesses for the study of clinical impact of variants. DISCUSSION: A well-established framework that provides a complete view on COVID-19 disease severity among hospitalized patients by combining information from different sources on host factors, viral factors, and healthcare-related factors, will enable to assess the clinical impact of emerging SARS-CoV-2 variants and answer questions that will be raised in the future. The framework shows the complexity related to causal research, the corresponding data requirements, and it underlines important limitations, such as unmeasured confounders or selection bias, inherent to repurposing existing routine COVID-19 data registries. TRIAL REGISTRATION: Each individual research project within the current conceptual framework will be prospectively registered in Open Science Framework (OSF identifier: https://doi.org/10.17605/OSF.IO/UEF29 ). OSF project created on 18 May 2021.