Statistical Considerations for FDA COVID-19 Guidance

In response to the newly released U.S. Food and Drug Administration (FDA) guidance on COVID-19, the Statistics Department at Premier Research has committed to highlighting specific and necessary actionable considerations that directly address key FDA recommendations in the guidance. These considerations call attention to some general actions required to ensure the operational integrity and scientific robustness of clinical trials, especially as we face unusual challenges in trial conduct due to the continuing COVID-19 pandemic.

With the FDA guidance allowing some flexibility in trial procedures, including protocol modifications, study teams should engage the project statistician regarding these changes, in order to effectively assess the level of impact on the trial robustness pertaining to the primary and key secondary efficacy objectives, study power, and the analysis strategy.

Additionally, the guidance specifically notes that a separate listing of all trial participants who were affected by COVID-19 study disruptions will be required. Therefore, it is important to document any implemented contingency measures in a way that allows the statistician to use this information in analysis summaries. The study team must work together to decide how best to document these contingency plans consistently.

Following are excerpts from the FDA guidance with corresponding statistical recommendations. In an upcoming blog post, we will release additional practical measures to be implemented during individual trial conduct.

Access to treatment

FDA guidance: “In some cases, trial participants who no longer have access to investigational product or the investigational site, may need additional safety monitoring (e.g. withdrawal of an active investigational treatment).”

Statistical recommendations:

The project team should ensure that treatment start/stop dates are recorded in the study Case Report Form and flag missed visits that are due to COVID-19. This will facilitate assessment of treatment compliance.

Alternative processes

FDA guidance: “The implementation of alternative processes should be consistent with the protocol to the extent possible, and sponsors and clinical investigators should document the reason for any contingency measures implemented. Sponsors and clinical investigators should document how restrictions related to COVID-19 led to the changes in study conduct and duration of those changes and indicate which trial participants were impacted and how those trial participants were impacted.”

Statistical recommendations:

The project team should ensure appropriate documentation of changes to trial conduct. This allows for adequate reporting and analysis summaries. Changes may be documented in the study database or other data capture systems that can be accessed for analysis. Such changes directly affect statistical approaches for handling missing data, intercurrent events, and algorithms for constructing derived variables.

Handling missing information

FDA guidance: “Changes in study visit schedules, missed visits, or patient discontinuations may lead to missing information (e.g., for protocol-specified procedures). It will be important to capture specific information in the case report form that explains the basis of the missing data, including the relationship to COVID-19 for missing protocol-specified information (e.g., from missed study visits or study discontinuations due to COVID-19). This information, summarized in the clinical study report, will be helpful to the sponsor and FDA.”

Statistical recommendations:

  1. Proactive measures should be implemented in the trial execution plans to minimize missing data where at all possible. Such measures as implementation of enhanced risk-based monitoring practices that involve periodic evaluation of extent of missingness in the trial database should be considered.
  2. The trial design should allow the possibility of extending visit windows (as reasonable) so as to accommodate likely delays for some assessments. This minimizes occurrence of missing values and the corresponding loss of statistical power.
  3. Reasons for changing visit schedules, as well as reasons for failing to obtain efficacy assessments, should be documented in such a way that it can be used in the analysis. This information directly impacts the statistical approach to handling missing data in sensitivity analyses.
  4. Specifically, if data are missing due to COVID-19, the reason needs to be recorded electronically and reference COVID-19. This is particularly important for the reason for subject withdrawal.
  5. Following database lock, thorough assessment on the extent of missing data due to COVID-19 should be carried out by the project statistician, specifically as pertains to the primary and key secondary endpoints. This will inform decisions around the reasonability of imputing missing data and the overall credibility of the final available data.
  6. The statistical section of the protocol should include a high-level description of strategies for handling missing data due to COVID-19 contingencies but the details of missing data handling should be specified in the Statistical Analysis Plan (SAP). As such modifications to the SAP may be needed.

Modifications to study drug administration

FDA guidance: “If scheduled visits at clinical sites will be significantly impacted, certain investigational products, such as those that are typically distributed for self-administration, may be amenable to alternative secure delivery methods. For other investigational products that are normally administered in a health care setting, consulting FDA review divisions on plans for alternative administration (e.g., home nursing or alternative sites by trained but non-study personnel) is recommended. In all cases, existing regulatory requirements for maintaining investigational product accountability remain and should be addressed and documented.”

Statistical recommendations:

  1. The project team should ensure that relevant information regarding changes to treatment administration or treatment dispensation are captured electronically. Such information directly impacts the assessment of treatment compliance in the statistical analysis.
  2. Examples of relevant information include:
    Date of change
    Method of alternative administration/dispensation
    Relevant changes to compliance expectations

Efficacy assessments

FDA guidance: “With respect to efficacy assessments, FDA recommends consultation with the appropriate review division regarding protocol modifications for the collection of efficacy endpoints, such as use of virtual assessments, delays in assessments, and alternative collection of research-specific specimens, if feasible. For individual instances where efficacy endpoints are not collected, the reasons for failing to obtain the efficacy assessment should be documented (e.g., identifying the specific limitation imposed by COVID-19 leading to the inability to perform the protocol-specified assessment).”

Statistical recommendations:

  1. If considering alternative methods for capturing efficacy assessments, study teams must consult with the project statistician in order to assess the potential impact to analyses.
  2. The project statistician must be involved in the event of any protocol modifications that affect the collection of efficacy endpoints, whether primary or secondary. The statistician can advise on how best to document the reasons for missing information.
  3. The project statistician should evaluate whether modifications are required to the statistical section of the protocol and the SAP to address these changes.

Protocol deviations

FDA guidance: “If changes in the protocol will lead to amending data management and/or statistical analysis plans, the sponsor should consider doing so in consultation with the applicable FDA review division. Prior to locking the database, sponsors should address in the statistical analysis plan how protocol deviations related to COVID-19 will be handled for the prespecified analyses.”

Statistical recommendations:

The study team needs to document all protocol deviations, including those due to the impact of COVID-19, in an organized manner that can be captured for the analysis.

Central and remote monitoring

FDA guidance: “If planned on-site monitoring visits are no longer possible, sponsors should consider optimizing use of central and remote monitoring programs to maintain oversight of clinical sites.”

Statistical recommendations:

For individual studies, the Statistics and Data Management teams collaborate proactively in identifying critical variables to assist with determining what data need to be included in the alternative monitoring strategies. This list can be expanded by the broader project team to ensure data integrity across the trial.