The Netherlands, 09 February 2026

Ubiquitous in real-world data (RWD) projects, Electronic Medical Record (EMR) and Electronic Health Record (EHR) data is making its way slowly but surely into clinical trials, and the possibilities appear endless… what’s the current situation, what are the pros and cons, and what are the practical steps to using this data? We sat down with resident expert Prof Rick Grobbee to get his thoughts.

What is the current situation?

The use of EMR data in traditional clinical trials is in its infancy but I believe, as do many of my counterparts, that this is going to take off. Currently we are seeing some proof-of-principle studies (e.g. Text-mining in electronic healthcare records can be used as efficient tool for screening and data collection in cardiovascular trials: a multicenter validation study – ScienceDirect) and some studies are now in place with a hybrid approach, with data collected through EMR and registries alongside more standard clinical trial methods.

What EMR data lends itself to clinical trials?

EMR data that typically lends itself to clinical trials include demographics, medical history and lab/imaging reports. This data can be used to identify patient candidates for clinical trials (identifying patients with rare disease or typically high screen failure rates such as MASH, speeding up recruitment), or used to complete part of the study eCRF, thereby freeing up site staff. Pre-study, EMR can be an invaluable feasibility tool, from finessing study design, to which country(ies) to consider, through to which sites can most efficiently enrol participants.

How is EMR data collected and incorporated into the trial?

In a typical trial the sponsor (or CRO) would also be responsible for setting up the import of EMR data into the trial database. There are also companies, especially in US, that have direct EMR access and act as an intermediary between the sponsor/CRO.

We will continue to see the integration with AI and machine learning to improve predictions of for example patient eligibility, with a combination of data that are standardized (such as certain diagnosis or lab values) and “prediction” of variables that are not standardized but predictable based on combinations of notes/terms using NLP or, increasingly, LLM’s and AI algorithms. Our data science team is at forefront of these developments and can, for instance, translate complex inclusion and exclusion criteria into computable logic that can be applied directly to EHR data to identify eligible patients. They also support the creation of external control arms by matching trial populations to real-world patient cohorts, improving feasibility and reducing reliance on placebo or comparator groups (interesting read here: Development of a reflection paper on the use of external controls for evidence generation in regulatory decision-making – Scientific guideline | European Medicines Agency (EMA).

Is privacy a concern?

This depends a bit on what is being collected. For example, in patient recruitment large patient data sets are being mined to find eligible participants. This should be done in a privacy protected manner eventually aiming to provide the investigator with a listing of patients to contact for trial participation. Once participants have entered the trial, the situation is different as they must give informed consent which should include the use of their EMR data.

This is different again from data mining on pseudo-anonymised data where consent is not necessarily required in the case of many secondary data use projects.

Using EMR in trials requires upfront attention to GDPR, HIPAA, and local privacy regulations—this must be built into study design.

How available is EMR globally?

EMR data is readily available in USA, and in the EU, I estimate more than 80% of countries have some form of EMR system, with about half boasting excellent records based on systems such as Epic that could be used in clinical trials (https://www.oecd.org/en/publications/progress-on-implementing-and-using-electronic-health-record-systems_4f4ce846-en.html). Globally this is more heterogenous; for example, in Asia many but not all countries have good access, and in Africa, new hospitals being built have EMR infrastructure in place.

What are the pitfalls of using EMR data in a clinical trial?

A major limitation is missing data, although not all missing data is meaningful, and can in part be compensated for by increasing sample size. Also, random noise can be created, although we can often mitigate for this. Some data may not be recorded in the EMR system but is needed for trial requirements e.g. certain safety data and patient reported outcomes (PRO). A last limitation of EMR is that it’s often restricted to a hospital unit/clinic, while the trial may require data from other sources, such as a participants’ general practitioner. Linkage between the various care providers exists in some instances but would require review on a case-by-case basis.

Issues like missing data, interoperability, and limited data models (CDMs) remain ongoing barriers in EMR-supported trials.

What are some long-term benefits we should start to see?

Once the infrastructure is in place within the various actors (hospitals but also service providers), projects can scale up with efficiencies being gained and leading to a reduction in the cost of clinical trials. The use of EMR data should also facilitate decentralization of clinical trials and reduce the burden on participants as well.

Longer term, EMR-driven recruitment and real-world data integration can help reduce timelines, ease site monitoring, and support more participant-friendly studies.

If you are thinking of including EMR data in your next trial, we would love to hear more, connect with us for a conversation about the possibilities!