The publication on the European Medicines Agency (EMA) website of the “Discussion paper: Use of patient disease registries for regulatory purposes – methodological and operational considerations” has opened the public consultation phase towards the final submission of the document to the relevant EMA committees by the end of 2019. The consultation in open up to 30 June 2019 and interested parties may forward their comments and suggestions using the template available on the Agency’s dedicated webpage.
A useful source of clinical information
EMA’s final goal is to identify how to better use for regulatory purposes the great amount of data available within patient registries and to discuss methodological and operational aspects for the future implementation of the final document, once approved. The current draft has been prepared by the Cross-Committee Task Force on Registries established by the EMA Patient Registries Initiative.
Patient registries are frequently used as a source of data for the post-authorisation monitoring of medicinal products. Their use has been required by the EMA Committee for Medicinal Products for Human Use (CHMP) as a condition of the marketing authorisation for 30 (9%) of the 335 centrally approved products in the period 2005-2013. At least one registry was also included in the risk management plan (RMP) of 43 (37%) products out of the 116 new drugs approved by the CHMP from 2007 to 2010 and, out of these, it was imposed as an obligation in 20.9% of cases.
According to EMA, the requested registries are often created by companies as new “product registries” where inclusion is determined by exposure of the patient to a specific medicinal product. Different is the case of “patient disease registries”, where inclusion is determined by occurrence of a specific disease. This last type of registry often includes many clinical information which is currently not used for regulatory purposes, but that might represent an important additional source of information to better assess the benefit-risk of medicines under the real world evidence scenario.
The collection of high quality data, based on the future availability of “good registry practices” is another target of the Patient Registry Initiative, as it may allow for the structured and uniform building of dataset capturing the evolution of the clinical conditions and pharmacological treatments received by patients. Disease registries are more specific than the products’ ones, which might prove useful for selected groups of patients; a typical example are patients normally excluded from clinical studies, e.g. older ones.
From the regulatory point of view, says EMA, patient registries are preferable as they offer much information on the outcomes of the different treatments received by patients. A further benefit of patient registries is the possibility to support wider possibilities for the design of clinical studies.
Patient disease registries
Patient registries are particularly useful to collect data on rare or orphan disease, or on older people, and they may be used for drug utilisation studies (DUS), post-authorisation efficacy studies (PAES), or post-authorisation drug safety studies (PASS). An effective collaboration between all involved parties is needed in order to early identify registries relevant for a certain regulatory procedure and evaluate the adequacy and quality of the data collected. Governance of registries also requires a strict collaboration, and the reaching of agreement on principles of data sharing and protection between registries, marketing authorisation applicants/holders (MAAs/MAHs) and regulators. The document discusses in detail the different roles of the registry coordinator, the pharmaceutical companies and the regulators.
EMA’s document discusses the details of features typical of the different kinds of registries in order to make clear the differences and to highlight the best practices for their management. In the case of patient disease registries, a critical point is the selection of patients to be included, to avoid any selection bias based on the characteristics of the single person. A clear conceptual definition of the target population, its translation into an operational definition, the availability of a process whereby the patients enrolled in the registry would be representative of this operational definition and the complete follow-up of enrolled patients and completeness of information collected to allow accounting for confounding factors and investigation of effect modifiers are the fundamental items to be considered, says the Discussion paper.
The draft document also offers a detailed table of time elements that should be always recorded as essential components of all registries. Core data elements should also be always included, using a standardised terminology in order to allow for the comparison, evaluation and interpretation of data from different registries. This harmonised terminology should refer to all data that might be entered in a registry, i.e. diseases, diagnostic tests, symptoms, medicinal products, active substances, adverse events and other relevant data. A reference to international terms should be available in the case of local or national terminologies are used.
The management of the registries should follow strict quality framework and procedures in order to guarantee the final quality of data. The provision of quality planning, assurance, control and improvement activities should be run routinely to reach this target. Consistency, completeness, accuracy and timeliness are the intrinsic characteristics desired to support data quality. The Discussion paper also offer some suggestions of the possible measures to be adopted to improve data quality at the management or operational level. Examples of indicators of data quality are also provided.
Safety analysis is a fundamental activity to be run within patients’ disease registries and it should follow the national requirements for the management of safety data in the case it is conducted by the academia or medical research associations. National or regional pharmacovigilance systems should also be used for the reporting of suspected adverse reactions. Should a MAH wish to manage or fund its own active data collection system in a disease registry referred to its medicinal products, this must follow the regulatory framework for PASS. The Discussion paper indicates that disease registries are generally not suitable for a rapid statistical analysis of new safety signals. Instead, they collect data useful for the monitoring and characterisation of known or suspected adverse reactions. Adverse events of special interest (AESI) can be also integrated in the routine data collection system.
Registry-based studies (registry studies) are something completely different, and the Discussion paper goes in details in explaining such difference as it has to be properly understood in order to correctly approach the set up and use of new registries for regulatory purposes. A table detailing the differences between a “registry” and a “registry study” is also provided.
Registry studies are usually established by MAHs for post-authorisation safety (PASS) and efficacy (PAES) studies, upon request of the regulatory authority. Legal obligations and responsibilities apply to MAHs in such instances, which are responsible to identify the benefit-risk ratio of the product under exam and communicate to the authority any new information which might influence it. A registry study has to be included in a public database, and regulators can exert the right to audit, inspect and verify the respect of all obligations.
The planning by the MAAs of a new registry study should start very early during the regulatory procedure, ideally at the stage of scientific advice, and it should see also the participation of the registry coordinator. The protocol of the study should follow the best methodological standards applicable to pharmacoepidemiological research and it should describe the measures identified to account for bias and confounding factors. It should also address very early the decision to collect primary data (directly from patients as they come to the attention of the investigator) or secondary ones (where the data for the study are already available and extracted from a dataset). Recommendations of the Good pharmacovigilance practice (GVP) Module VIII and the technical guidance on the format and content of the protocol for non-interventional PASS also apply.
A joint registry study based on a single protocol, with the participation of all concerned MAHs, is suggested by EMA for studies addressing several products. A common protocol would also be preferable for multi-site studies, in order to use common core data elements and a common design, with specification of the study size and the feasibility of attaining this sample size within the registry using conservative assumptions. Milestones to complete the different phases of the study should be also provided by MAHs.
The study population is selected starting from the registry patient population on the base of inclusion and exclusion criteria; it may also include new users to assess the safety or effectiveness of a new treatment in a disease registry. EMA’s document also discuss the available possibilities to include incident patients (patients newly diagnosed with the disease and who received a first prescription) or prevalent patients (already included in the registry to whom the new treatment is prescribed), as this choice might impact on data analysis and interpretation. Prospective recruitment is also discussed, as it is important to include in the study of all eligible patients treated in the individual centres.
The registry study should only include those data strictly needed to provide valid results, as determined by the registry coordinator. Data needed for sensitivity analyses as outlined in the study protocol and statistical analysis plan (SAP) should be also collected. The quality measures routinely applied to the registry defines the need for additional measures for data quality control in a registry study. Data source verification for a minimum of 10% of randomly selected patients registered in individual study centres is suggested by EMA as an adequate verification acceptable from the regulatory point of view.
MAHs are free to choose the statistical method more appropriate to the design of each specific study in order to address the scientific question of interest. Pharmacovigilance activities and reporting should clearly distinguish between studies with a design based on primary data collection from health care professionals or consumers (which should dispose of an electronic system to collect, analyse and report information on adverse events) and studies based on secondary use of data. Investigators of registry studies should also be informed of how to report adverse effects or suspected adverse reactions through the national spontaneous reporting system.
Reporting of registry studies follow legal obligations in the case of imposed PASS and PAES studies, and the regulatory authority may request the MAH for additional information or clarifications. The scientific aspects of such request should be addresses under the responsibility of the lead investigator of the study.