Partner Introduction (Public Health Agency of Canada): 4-min video.

• Summary: Dr. Sarah Viehbeck (Chief Science Officer and Vice-President Data, Surveillance and Foresight at the Public Health Agency of Canada) discusses initiatives at the agency that intersect with pragmatic clinical trials.

NIH Pragmatic Trials Collaboratory (Living Textbook): Section 9 - The Research Question Drives the Data Requirements (Chapter: Design - Using Electronic Health Record Data in Pragmatic Clinical Trials): 1-page website.

• Summary: The secondary use of routinely collected health data* (e.g., administrative data, electronic health records, disease registries, epidemiologic data from surveillance systems) - rather than prospectively collecting new primary data explicitly for a trial - generally entails trading control for efficiency. That loss of control often involves some sacrifices, which are sometimes acceptable to the research team, but sometimes not. To determine what trade-offs are acceptable, the trialist must begin with a well-defined research question, and then assess how much bias would be introduced by leveraging existing data (rather than by attempting to collect new data). (*Note: Routinely collected data are also referred to as “real-world data” as they are typically collected within routine care.)

Hemkens LG. How Routinely Collected Data for Randomized Trials Provide Long-term Randomized Real-World Evidence. JAMA Netw Open. 2018 Dec 7;1(8):e186014. (3-page commentary)

• Summary: Summarizes benefits and barriers related to the use of routinely collected health data in randomized trials. e.g., Use of data collected in routine care avoids artificial follow-up procedures by using existing infrastructures, decreasing cost and increasing feasibility of doing the trial. However, such promises come with novel challenges, where issues related to data quality/completeness are among the most pressing concerns. (It is argued, however, that while reasonable data quality is a challenge for any type of real-world evidence, this is less of an issue for randomized trials [vs. observational studies] as trials typically do not require complex statistical models based on accurate measures of all potential confounders.)

NIH Pragmatic Trials Collaboratory (Living Textbook): Section 5 - Specific Uses for EHR Data in PCTs (Chapter: Design - Using Electronic Health Record Data in Pragmatic Clinical Trials): 1-page website.

• Summary: Describes areas where routinely collected data can be used to support five major activities related to pragmatic trials: (1) Preparation (e.g., estimating the number of eligible participants, estimating rates of outcomes, examining correlation within clusters [if applicable] to inform sample size calculations); (2) Enrolment/recruitment; (3) Assessing baseline characteristics of the sample; (4) Implementing and monitoring the delivery of the intervention; (5) Measuring study outcomes. (Note: While the Collaboratory focuses on electronic health records [EHRs], which is one type of routinely collected health data, the EHR-related concepts generalize to other types of routinely collected data as well.)

Using Routine Health Data to Support Cardiovascular Clinical Trials (HDRN Canada Pragmatic Trials Training Program): 13-min webinar & 17-slide presentation.

• Summary: Dr. Craig Rodrigues discusses the use of routinely collected health data to support cardiovascular clinical trials. The limitations of traditional randomized trial methods are overviewed (e.g., the resources required for the manual identification of clinical events from multiple sources and rigorous review by a clinical end point committee), as well as the benefits of integrating routinely collected data into trials (e.g., reduction of trial costs, increased sample size and follow-up). A study of the accuracy of routinely collected health data for outcome ascertainment in cardiovascular trials is presented as well.

NIH Pragmatic Trials Collaboratory - Assessing Data Quality (July 17, 2020): 9-min webinar (14-slide presentation; slides 39-52)

• Summary: Dr. Rachel Richesson emphasizes that routinely collected health data are an imperfect measure of clinical phenomena (e.g., measures of outcomes may be captured, often imperfectly, through queries of diagnostic codes in health records). Therefore, unlike settings where the researchers are directly measuring whether a participant is living with a given condition or experiencing a particular event, reliance on what is already available in the routinely collected data requires careful consideration. Therefore, several recommendations are provided: (1) Craft research questions that can be robustly answered using existing data, and design trials to minimize new data collection; (2) Engage data experts when defining endpoints and outcomes; (3) Budget for data and systems experts; (4) Clearly define endpoints and outcomes for transparency and reproducibility; and (5) Develop a robust data quality assessment plan.

Emerson SD, et al. Secondary use of routinely collected administrative health data for epidemiologic research: Answering research questions using data collected for a different purpose. Int J Popul Data Sci. 2024 Nov 19;9(1):2407. (12-page paper)

• Summary: The use of routinely collected administrative health data for research can provide unique insights to inform decision making. However, because these data are primarily collected to administer healthcare service delivery, challenges exist when using such data for secondary purposes (i.e., epidemiologic research). Many of these challenges stem from the researcher’s lack of control over the quality and consistency of data collection, and, furthermore, a lessened understanding of the data being analyzed. This article presents considerations derived from experiences analyzing administrative health data in a Canadian context. Considerations were organized around four themes: (1) Know the data and their primary use; (2) Understand classification and coding systems; (3) Transform data into meaningful forms; (4) Recognize the importance of validity when defining analytic variables.

Health Data Research Network Canada (HDRN Canada) - The Data Access Support Hub (DASH) Data Assets Inventory (2025): 1-page dashboard.

• Summary: The Data Access Support Hub (DASH) is a one-stop data access service portal for researchers requiring multi-regional health and health-related administrative data in Canada. DASH streamlines the data access process through its centralized data access request form and harmonizes processes to request data from multiple data repositories across the country. A resource provided by DASH is the Data Assets Inventory, a tool which can be used to familiarize yourself with the types of routinely collected data sources available for research across Canada.