Registering observational studies

In 2017, registering an observational study was not mandatory, which left observational studies vulnerable to publication and reporting bias. ClinicalTrials.gov accepted observational studies and a European register accepted post-authorisation studies, but the number of observational studies registered was still considered low, and how to register them was not settled.

Why registration is not required but still worthwhile

In 2017, registering an observational study before the study started was not mandatory. ¹ Under the requirements of the US Food and Drug Administration Amendments Act (FDAAA) 801, an observational study, unlike a randomised trial, does not have to be registered on ClinicalTrials.gov. ² Observational studies are therefore quite vulnerable to publication and reporting bias, owing to selective reporting, misinterpretation of analyses and the lack of regulations on their registration and reporting. ² This vulnerability undermines the overall validity of observational studies and provides a rationale for registering them. ² Registering their studies is to the advantage of researchers and sponsors, because the ultimate aim is to improve patient healthcare and thereby benefit society. ¹

Registers that accept observational studies

In 2017, ClinicalTrials.gov allowed observational studies to be registered and provided specific data elements to fill in for them. ² In Europe, the European Union electronic Register of Post-Authorisation Studies (EU PAS Register) was publicly available for registering post-authorisation studies, to improve the transparency of observational research. ²

How many are registered

By 2017 the number of observational studies registered each year had increased over the past few years, and observational studies represented about 15% of all studies on ClinicalTrials.gov. ² Around half of these studies were from North America, a fifth from Europe and 13% from Asia, and 85% were funded by non-industry sources. ² The number of observational studies registered was still considered low, which exposed them to reporting bias. ²

Where registration falls in a study

The ideal process for designing, conducting and publishing an observational study places registration after ethics approval and before the study is conducted: ¹

1. Identify the data.
2. Prepare the protocol.
3. Get ethics approval.
4. Register the observational study.
5. Conduct the study.
6. Analyse the data.
7. Prepare the study report and manuscript.
8. Publish.

¹

Open questions

In 2017 there were no standard guidelines on registering observational studies. ¹ Challenges in registering them included: ²

• Most registered studies are prospective, and methods need to be established for registering other types of study, such as retrospective studies.
• The timeframe for registering an observational study needs to be defined, along with the attributes the registration should state.
• Whether or how to register sub-studies or secondary studies that use the same prospective data is undecided.
• The data elements for reporting different types of observational study need to be defined.

²

These issues need to be properly addressed to ensure the complete transparency of observational studies, and that requires discussion among all stakeholders, including sponsors, regulatory authorities and the public. ²

Standardised guidelines on registering observational studies, on their ethics committee approval and on disseminating their data would help to improve the transparency of research and validate the findings of observational studies. ¹ Such guidelines would help to avoid the duplication of data and the misinterpretation of results, and would contribute to the worldwide spread of knowledge. ¹