Methodologies for Observational Studies Comparing Inpatient COVID-19 Treatments

In this study we seek to evaluate current methodologies for observational comparative studies of inpatient COVID-19 treatments [Overall Study Objective]. To support this overall study objective, we have defined additional supporting objectives related to the research process [Process Objectives] as applied to an illustrative example of an observational study to evaluate the comparative effectiveness of inpatient COVID-19 treatments [Illustrative Example].

Process Objectives:

Characterize differences in study results when an intentional multiphase approach (IMA) to diagnostics and contingencies is applied to real world data (RWD) analysis compared to an approach that pre-specifies all covariates and statistical approaches without consideration of whether key statistical assumptions hold (Single-phase Prespecification Approach, SPA).

Illustrative Example:

Our illustrative example will use a large population-based US claims data source to emulate a hypothetical target trial to assess the comparative effectiveness of IL6Ri (TCZ or SAR) versus JAKi (BAR or TOF) added to systemic corticosteroids of interest (CSI). The underlying hypothetical target trial of interest consists of patients hospitalized and requiring respiratory support for COVID-19 and receiving a CSI who are assigned to receive either an IL6Ri or JAKi in addition to CSI within 4 days after hospital/ICU admission. The patients are followed for up to 28 days for the outcomes of inpatient mortality and progression to IMV/ECMO. The Illustrative Example is separated into two sub-objectives based on the different COVID-19 severity populations of interest.

Illustrative Example - Objective I aims to characterize the risk of inpatient mortality [Primary Outcome] and progression to IMV or ECMO [Secondary Outcome] up to 28 days after IL6Ri or JAKi initiation among patients hospitalized with COVID-19 who initiate a CSI and require supplemental O2/NIV/HFO (but not IMV/ECMO).

Illustrative Example - Objective II aims to characterize the risk of inpatient mortality [Primary Outcome] up to 28 days after IL6Ri or JAKi initiation among patients admitted to the ICU at hospital admission with COVID-19 who initiate a CSI and require IMV/ECMO.

Hazard ratios (HR) and corresponding 95% confidence intervals (CI) will be estimated and reported for all outcome risks in Illustrative Example objectives.

Inclusion criteria

• Hospitalized from June 16, 2020 to February 01, 2022 with an international classification of diseases (ICD)-10 diagnosis code of U07.1 in an any admitting diagnosis position (Day 0)
• Initiate either IL6Ri or JAKi within 4 days after hospital admission (Days 0 to 4)
• Receiving systemic CSI on day of IL6Ri/JAKi initiation (Day T)
• Receipt of at least one respiratory support procedure (oxygen supplementation at a minimum) from admission to IL6Ri/JAKi initiation (Days 0 to T)

For Illustrative Example - Objective I only:

-maximum modified World Health Organization (mWHO) disease severity of O2/NIV/HFO from admission to IL6Ri/JAKi initiation (Days 0 to T)

For Illustrative Example - Objective II only:

• Admission to ICU at hospital admission (Day 0) and mWHO disease severity of IMV/ECMO from hospital/ICU admission to IL6Ri/JAKi initiation (Days 0 to T)
• Continuous medical claims enrollment (60-day gaps permitted) during the 183 day baseline period prior to and including hospital or ICU admission (Days -183 to 0) to minimize the potential for misclassification of baseline covariates

Exclusion criteria

• Exclude patients without continuous medical claims enrollment (60-day gaps permitted) during the 183-day baseline period prior to and including hospital/ICU admission (Days -183 to 0) to minimize the potential for misclassification of baseline covariates.
• Exclude patients if COVID-19 hospitalization (Day 0) begins >14 days after initial COVID-19 diagnosis. Patients will be excluded if any COVID-19 diagnosis is recorded from 90 days to 15 days before admission (Days -90 to -15) to exclude patients with possible long-term COVID or post-acute sequelae while still permitting prior infections recorded more than 90 days pre-admission.
• No age, sex, or geographic region recorded on hospital admission (Day 0)
• Age less than 18 years at hospital admission (Day 0)
• Evidence of a prior COVID-related inpatient hospitalization in the previous 14 days (Days -14 to -3), with a two-day buffer to permit brief inpatient utilization directly proceeding transfer to a chargemaster hospital (i.e., inpatient utilization permitted on Days -2 to -1)
• Systemic CSI use (dispensing or remaining supply) in the previous 14 days (Days -14 to -3), with a two-day buffer to permit CSI use beginning in the emergency department or other healthcare setting immediately prior to chargemaster hospital admission (CSI use permitted Days -2 to -1)
• Any recorded use of IL6Ri or JAKi (dispensing or remaining supply) in the 90-day washout period before IL6Ri or JAKi initiation (Days -90 to T-1), to satisfy new use definition
• Baseline receipt of oxygen or ventilation support for non-COVID conditions (e.g., supplemental oxygen for COPD), defined as two or more respiratory support procedures recorded on different days from Days -90 to -15 via procedure codes from medical claims, chargemaster, and where available, oxygen supplies recorded in pharmacy settings.
• Death or discharge occurs before or on the day of treatment initiation (Days 0 to T)
• Recorded use of both IL6Ri and JAKi on Day T