The U.S. Food and Drug Administration (FDA) has recently issued guidance for sponsors and investigators interested in submitting a non-interventional study, commonly known as an observational study, to contribute evidence of a drug’s safety and effectiveness. These drug studies aim to distinguish the drug’s effects from other factors like disease progression, placebo responses, or observer bias. Non-interventional studies, such as those using electronic health record data in cohort designs, may yield inaccurate conclusions due to confounding factors like noncomparable treatment groups or bias leading researchers to choosing specific types of participants thereby missing data. Addressing these issues is essential for obtaining accurate data and results. FDA’s guidance assists sponsors in recognizing and mitigating these challenges before finalizing study protocols and statistical analysis plans (SAP).

Understanding Non-Interventional Studies: Background

Non-interventional studies are invaluable for gaining insights into the actual safety and effectiveness of drugs in real-world settings. These studies involve patients receiving approved drugs as part of routine medical care, without specific interventions being dictated by a protocol.

There are various types of non-interventional study designs, including observational cohort studies, case-control studies, and self-controlled studies. These designs enable researchers to observe outcomes of interest without disrupting the natural course of treatment.

The reliability and relevance of real-world data (RWD) are crucial for the success of non-interventional studies. RWD further encompasses a wide range of information on patient health status and healthcare delivery from diverse sources. In these studies, the quality and consistency of RWD from different sources are pivotal for ensuring accurate conclusions are drawn.

Overview of Research Guidance

FDA has provided guidance to researchers regarding the:

1. Proposed approach;
2. Elements of a study design;
3. Selection of data sources; and
4. Analytic approach to reviewing data.

Summary of the Proposed Approach

Sponsors should first finalize their protocol by outlining the detailed plan for conducting the study. This protocol should encompass the study’s roadmap, covering aspects such as the research question, hypothesis, rationale for study design choice, selection of data sources, results of preliminary studies, approach to causal inference, and ethical considerations. Engaging with the FDA early is important for assuring alignment with expectations.

Study Design

Crafting a robust study design is the cornerstone of any successful research endeavor. The study design elements of the proposed research should include a schema outlining the overall study design and a causal diagram specifying the theorized causal relationship. Additionally, critical components such as the source population, eligibility criteria, conceptual and operational definitions for key variables, relevant covariates, index date determination, and start and end of follow-up periods should be concisely described in the protocol.

Data Sources

Selecting appropriate data sources is key for the success of a non-interventional study. The description of the proposed data source(s) should include details on how the data were originally collected and the rationale for choosing the data source(s). The description of the sources should also address the relevance to the drug-outcome association, information on confounding factors, data reliability, common data models used, timing and completeness of key data elements, appropriateness of coding based on operational definitions, alignment with the target patient population, and quality assurance activities. Additionally, plans for data quality assurance and potential links to other data sources should be outlined.

Analytic Approach

Once the study protocol is established, the focus shifts to the analytic approach. Sponsors must craft a detailed SAP that aligns with study objectives. This includes specifying the statistical approach for evaluating treatment effects, handling intercurrent events and censoring rules, and accounting for potential confounding factors, including assessment of unmeasured confounding factors.

The SAP also considers potential overadjustment of intermediate variables on the causal pathway, outlines approaches for subgroup analyses, addresses unequal detection of outcomes across compared groups, evaluates the possibility of reverse causality, describes methods for handling missing or misclassified data, and outlines strategies for managing multiplicity, such as analyzing multiple exposures or outcomes.

Conclusion

Non-interventional studies play a vital role in the evaluation of drug safety and effectiveness. FDA’s guidance offers recommendations to sponsors considering the submission of non-interventional studies to the FDA to support the demonstration of a drug’s safety and effectiveness in connection with RWD. The development of this draft guidance reflects the increasing interest among stakeholders in leveraging non-interventional studies to provide evidence of a drug’s safety and efficacy. By carefully considering the outlined recommendations at each stage of the research process, researchers can enhance the quality and reliability of their study outcomes.