Exploring the impact of selection bias in observational studies of COVID-19: a simulation study

Analysis of Missingness Scenarios for Observational Health Data

Despite the extensive literature on missing data theory and cautionary articles emphasizing the importance of realistic analysis for healthcare data, a critical gap persists in incorporating domain knowledge into the missing data methods. In this paper, we argue that the remedy is to identify the key scenarios that lead to data missingness and investigate their theoretical implications. Based on this proposal, we first introduce an analysis framework where we investigate how different observation agents, such as physicians, influence the data availability and then scrutinize each scenario with respect to the steps in the missing data analysis. We apply this framework to the case study of observational data in healthcare facilities. We identify ten fundamental missingness scenarios and show how they influence the identification step for missing data graphical models, inverse probability weighting estimation, and exponential tilting sensitivity analysis. To emphasize how domain-informed analysis can improve method reliability, we conduct simulation studies under the influence of various missingness scenarios. We compare the results of three common methods in medical data analysis: complete-case analysis, Missforest imputation, and inverse probability weighting estimation. The experiments are conducted for two objectives: variable mean estimation and classification accuracy. We advocate for our analysis approach as a reference for the observational health data analysis. Beyond that, we also posit that the proposed analysis framework is applicable to other medical domains.

Neurocognitive Impairment in Post-COVID-19 Condition in Adults: Narrative Review of the Current Literature

The severe acute respiratory syndrome coronavirus 2 virus has, up to the time of this article, resulted in >770 million cases of COVID-19 illness worldwide, and approximately 7 million deaths, including >1.1 million in the United States. Although defined as a respiratory virus, early in the pandemic, it became apparent that considerable numbers of people recovering from COVID-19 illness experienced persistence or new onset of multi-system health problems, including neurologic and cognitive and behavioral health concerns. Persistent multi-system health problems are defined as Post-COVID-19 Condition (PCC), Post-Acute Sequelae of COVID-19, or Long COVID. A significant number of those with PCC report cognitive problems. This paper reviews the current state of scientific knowledge on persisting cognitive symptoms in adults following COVID-19 illness. A brief history is provided of the emergence of concerns about persisting cognitive problems following COVID-19 illness and the definition of PCC. Methodologic factors that complicate clear understanding of PCC are reviewed. The review then examines research on patterns of cognitive impairment that have been found, factors that may contribute to increased risk, behavioral health variables, and interventions being used to ameliorate persisting symptoms. Finally, recommendations are made about ways neuropsychologists can improve the quality of existing research.

Decoding Immuno-Competence: A Novel Analysis of Complete Blood Cell Count Data in COVID-19 Outcomes

BACKGROUND

While 'immuno-competence' is a well-known term, it lacks an operational definition. To address this omission, this study explored whether the temporal and structured data of the complete blood cell count (CBC) can rapidly estimate immuno-competence. To this end, one or more ratios that included data on all monocytes, lymphocytes and neutrophils were investigated.

MATERIALS AND METHODS

Longitudinal CBC data collected from 101 COVID-19 patients (291 observations) were analyzed. Dynamics were estimated with several approaches, which included non-structured (the classic CBC format) and structured data. Structured data were assessed as complex ratios that capture multicellular interactions among leukocytes. In comparing survivors with non-survivors, the hypothesis that immuno-competence may exhibit feedback-like (oscillatory or cyclic) responses was tested.

RESULTS

While non-structured data did not distinguish survivors from non-survivors, structured data revealed immunological and statistical differences between outcomes: while survivors exhibited oscillatory data patterns, non-survivors did not. In survivors, many variables (including IL-6, hemoglobin and several complex indicators) showed values above or below the levels observed on day 1 of the hospitalization period, displaying L-shaped data distributions (positive kurtosis). In contrast, non-survivors did not exhibit kurtosis. Three immunologically defined data subsets included only survivors. Because information was based on visual patterns generated in real time, this method can, potentially, provide information rapidly.

DISCUSSION

The hypothesis that immuno-competence expresses feedback-like loops when immunological data are structured was not rejected. This function seemed to be impaired in immuno-suppressed individuals. While this method rapidly informs, it is only a guide that, to be confirmed, requires additional tests. Despite this limitation, the fact that three protective (survival-associated) immunological data subsets were observed since day 1 supports many clinical decisions, including the early and personalized prognosis and identification of targets that immunomodulatory therapies could pursue. Because it extracts more information from the same data, structured data may replace the century-old format of the CBC.

Who is more vulnerable to effects of long-term exposure to air pollution on COVID-19 hospitalisation?

OBJECTIVE

Factors that shape individuals' vulnerability to the effects of air pollution on COVID-19 severity remain poorly understood. We evaluated whether the association between long-term exposure to ambient NO2, PM2.5, and PM10 and COVID-19 hospitalisation differs by age, sex, individual income, area-level socioeconomic status, arterial hypertension, diabetes mellitus, and chronic obstructive pulmonary disease.

METHODS

We analysed a population-based cohort of 4,639,184 adults in Catalonia, Spain, during 2020. We fitted Cox proportional hazard models adjusted for several potential confounding factors and evaluated the interaction effect between vulnerability indicators and the 2019 annual average of NO2, PM2.5, and PM10. We evaluated interaction on both additive and multiplicative scales.

RESULTS

Overall, the association was additive between air pollution and the vulnerable groups. Air pollution and vulnerability indicators had a synergistic (greater than additive) effect for males and individuals with low income or living in the most deprived neighbourhoods. The Relative Excess Risk due to Interaction (RERI) was 0.21, 95 % CI, 0.15 to 0.27 for NO2 and 0.16, 95 % CI, 0.11 to 0.22 for PM2.5 for males; 0.13, 95 % CI, 0.09 to 0.18 for NO2 and 0.10, 95 % CI, 0.05 to 0.14 for PM2.5 for lower individual income and 0.17, 95 % CI, 0.12 to 0.22 for NO2 and 0.09, 95 % CI, 0.05 to 0.14 for PM2.5 for lower area-level socioeconomic status. Results for PM10 were similar to PM2.5. Results on multiplicative scale were inconsistent.

CONCLUSIONS

Long-term exposure to air pollution had a larger synergistic effect on COVID-19 hospitalisation for males and those with lower individual- and area-level socioeconomic status.

Prospective study design and data analysis in UK Biobank

Population-based prospective studies, such as UK Biobank, are valuable for generating and testing hypotheses about the potential causes of human disease. We describe how UK Biobank's study design, data access policies, and approaches to statistical analysis can help to minimize error and improve the interpretability of research findings, with implications for other population-based prospective studies being established worldwide.

Time-sensitive testing pressures and COVID-19 outcomes: are socioeconomic inequalities over the first year of the pandemic explained by selection bias?

BACKGROUND

There are many ways in which selection bias might impact COVID-19 research. Here we focus on selection for receiving a polymerase-chain-reaction (PCR) SARS-CoV-2 test and how known changes to selection pressures over time may bias research into COVID-19 infection.

METHODS

Using UK Biobank (N = 420,231; 55% female; mean age = 66.8 [SD = 8·11]) we estimate the association between socio-economic position (SEP) and (i) being tested for SARS-CoV-2 infection versus not being tested (ii) testing positive for SARS-CoV-2 infection versus testing negative and (iii) testing negative for SARS-CoV-2 infection versus not being tested. We construct four distinct time-periods between March 2020 and March 2021, representing distinct periods of testing pressures and lockdown restrictions and specify both time-stratified and combined models for each outcome. We explore potential selection bias by examining associations with positive and negative control exposures.

RESULTS

The association between more disadvantaged SEP and receiving a SARS-CoV-2 test attenuated over time. Compared to individuals with a degree, individuals whose highest educational qualification was a GCSE or equivalent had an OR of 1·27 (95% CI: 1·18 to 1·37) in March-May 2020 and 1·13 (95% CI: 1.·10 to 1·16) in January-March 2021. The magnitude of the association between educational attainment and testing positive for SARS-CoV-2 infection increased over the same period. For the equivalent comparison, the OR for testing positive increased from 1·25 (95% CI: 1·04 to 1·47), to 1·69 (95% CI: 1·55 to 1·83). We found little evidence of an association between control exposures, and any considered outcome.

CONCLUSIONS

The association between SEP and SARS-CoV-2 testing changed over time, highlighting the potential of time-specific selection pressures to bias analyses of COVID-19. Positive and negative control analyses suggest that changes in the association between SEP and SARS-CoV-2 infection over time likely reflect true increases in socioeconomic inequalities.

The Covid-19 pandemic as an accelerator of economic worries and labor-related mental health polarization in Germany? A longitudinal interacted mediation analysis with a difference-in-difference comparison

Objectives

Labor-related mental health polarization refers to exposure to low-paid employment and unemployment decreasing mental health. Previous research identified economic worries as a key mediator. Against this background, the Covid-19 pandemic is often assumed to have accelerated already existing processes and affected vulnerable populations the most. Our study sought to investigate whether the Covid-19 pandemic accelerated the mediation by economic worries between employment type and mental health.

Method

Using the German Socioeconomic Panel (GSOEP) from 2016 onwards, we created a pre-Covid-19 sample (N = 8266) and a per-Covid-19 sample (N = 7294), with each having a t0 wave (2016/2018) and a t1 wave (2018/2020). We applied the mediational g-formula for longitudinal mediation with exposure-mediator (XM) interaction between employment type (X) and economic worries (M). We decomposed the total effect into a direct, indirect, and interacted effect of employment on mental health and provided a difference-in-difference comparison of the effects.

Results

During the Covid-19 pandemic, economic worries increased, and mental health decreased. However, the mediation by economic worries reduced by approx. 18.0% (e.g., from 35.0% to 28.9%). A decreased indirect effect caused the reduction in mediation, while the direct and interacted effect remained rather stable. We also found stark gender differences towards males having a higher total effect but a lower mediated effect during the Covid-19 pandemic.

Conclusion

Our results highlight that mediators competing to economic worries must have emerged during the Covid-19 pandemic. Such mediators could be the risk of infection at the workplace, the possibility of remote work, and gender-specific mediators. Our study is also the first to extend the mediational g-formula with the difference-in-difference approach. It can be used as a blueprint for researchers interested in evaluating the impact of events, such as the Covid-19 pandemic, on preexisting processes.

Long-term exposure to air pollution and severe COVID-19 in Catalonia: a population-based cohort study

The association between long-term exposure to ambient air pollutants and severe COVID-19 is uncertain. We followed 4,660,502 adults from the general population in 2020 in Catalonia, Spain. Cox proportional models were fit to evaluate the association between annual averages of PM_2.5, NO₂, BC, and O₃ at each participant's residential address and severe COVID-19. Higher exposure to PM_2.5, NO_2, and BC was associated with an increased risk of COVID-19 hospitalization, ICU admission, death, and hospital length of stay. An increase of 3.2 µg/m³ of PM_2.5 was associated with a 19% (95% CI, 16-21) increase in hospitalizations. An increase of 16.1 µg/m³ of NO₂ was associated with a 42% (95% CI, 30-55) increase in ICU admissions. An increase of 0.7 µg/m³ of BC was associated with a 6% (95% CI, 0-13) increase in deaths. O₃ was positively associated with severe outcomes when adjusted by NO₂. Our study contributes robust evidence that long-term exposure to air pollutants is associated with severe COVID-19.