PMCF studies and registries
A post-market clinical follow-up study is indicated in particular situations and takes one of four designs, one of them a review of registry data. Registry data support the follow-up only as far as their quality and specificity allow, and a study cannot be used to extend a device's indication.
When a PMCF study is indicated
A PMCF study can be indicated when possible residual risks related to the device have been identified, or when its long-term performance is uncertain. A study can also confirm the device's long-term performance, generate data specific to defined patient populations, or confirm its safety and performance in a more representative patient population. For example, for a relatively new device already investigated in an IDE study, a PMCF study with less restrictive inclusion and exclusion criteria can be considered.
Evidence from post-market studies is usually expected for implantable devices and class III devices.
Objectives and the investigation plan
A PMCF study, like clinical studies in general, requires a lot of time, resources and money, so it needs clear objectives with specific questions about the safety and performance of the device. As for an IDE study, its clinical investigation plan should clearly define the study population, the inclusion and exclusion criteria and a justification for the appropriateness of the study design. The plan also has to clearly define the study endpoints and the number of patients, with a statistical rationale for that number, and outline the duration of follow-up, the data to be collected and the statistical analysis plan.
The PMCF plan should include the proposed study design, sample size, endpoints, inclusion and exclusion criteria, and a statistical rationale.
The four designs
A PMCF study can extend the follow-up of patients already enrolled in a pre-market investigation, be a completely new clinical investigation, review data derived from a device registry, or review relevant retrospective data from patients previously exposed to the device.
Extended follow-up of a pre-market investigation
An extended follow-up suits, for example, a new device investigated in an IDE study with a short or mid-term follow-up.
For example, the clinical evidence for a recently CE-marked class III device used in cardiovascular surgery in combination with pacemakers derives from a pre-market clinical investigation with a follow-up of 36 months. As it is an implantable device used in cardiovascular surgery, notified bodies will expect a longer follow-up, for example up to 5 years, to provide evidence on its long-term safety.
The earlier study protocol, including its endpoints and inclusion and exclusion criteria, can be used to develop the follow-up study, which takes less effort to set up than a new clinical investigation. But the longer the study runs, the more likely patients are to be lost to follow-up, which can affect the validity of the results.
The protocol of a pre-market investigation can plan for its continuation. As a very rough estimate, a trial would never be approved with fewer than 100 patients, because patient numbers are calculated from a statistical sample size, but the protocol can state that the CE mark will be applied for after, for example, 50 or 60 patients, and that the study will continue to enrol, and its follow-up be extended, as a post-market follow-up study. The CE mark can then be applied for more or less whenever the sponsor chooses, with whatever degree of follow-up has accrued, and the continuing study becomes the post-market follow-up, or part of it.
A clinical investigation may eventually become a registry, and that is frequently how it is done. Many devices are required to provide information on patients' overall survival over many years.
A study planned before the CE mark
Where a manufacturer has planned a PMCF study and is writing its protocol before an MDR submission that will take time to receive its CE mark, one practitioner holds that a head start on data collection is always recommended, as the data can be leveraged once the submission goes through.
Registries
The registries that serve PMCF are either manufacturer-initiated or public. Manufacturer or national public registries on the device or the device group can be a good source of real-world clinical evidence. One practitioner holds that registry data are a very good opportunity to collect PMCF data, because once a registry is established it allows clinical data to be collected over the long term and from many patients.
Device registry studies usually take place over a long period of time and ask very simple questions; for an implant, whether the patient is alive one or two years after implantation. One practitioner holds that clinical trials could learn a lot from registry studies about asking simple questions, keeping things simple and getting data over a long period. In that practitioner's experience, many of the companies that conduct device clinical investigations have much experience of registry studies and less of clinical investigations, and the two are very different.
Manufacturer-initiated registries
A manufacturer-initiated registry includes a pre-specification of the quality and quantity of its data based on the risk of the device. A new manufacturer-initiated registry has the advantage of being device-specific, but it will not contain historic data and will take time to accumulate data on a large number of patients over a long period. If a new registry is initiated, the PMCF plan should include a description of it and a preliminary specification of the expected quantity and quality of its data.
Registries are one of the best ways of continuously collecting real-world clinical data. Initiating one requires time, money and appropriate endpoints, and the registry needs to be well maintained to generate high-quality data. PMCF is a continuous process, and once a registry has been set up, long-term data collection is possible.
Public registries
A public registry can be used where it holds data on the subject device or on an equivalent device with the expected quality and quantity. An existing national registry can be very useful if it contains historic data on similar devices from a large patient population, but it has the disadvantage of not being device-specific.
The data quality of public registries may be insufficient. Some of them, or most, do not collect device-specific data, so it can be quite difficult to retrieve data about the subject device. Many registries collect a great many data points, of which most, or at least some, might not be relevant to the device, and the data are often incomplete. A retrospective analysis of public registry data on the device also has to be aligned with data protection regulations.
Public registries are numerous, at least in some specialties. There are many registries for vascular interventions, where data can be collected on stents and stent grafts but also on PTA balloons or catheters, several for urological interventions, which may hold data on devices such as ureteral stents, and many in the orthopaedic field, for knee prostheses, hip implants and others.
Registries of well-established prostheses
A prosthesis, such as a hip or knee prosthesis, that has been on the market for more than 20 years is a common example of a well-established device. Many registries collect data on all implanted prostheses and produce yearly reports summarising those data, which are a real-world means of analysing the performance and safety of a particular prosthesis. These registries contain an impressive amount of data and represent real-life cases on all prostheses, including competitors' devices, which is useful when comparing safety and performance data across devices and over the long term. Their data are not influenced by the manufacturer, but negative results are likely to be reported more frequently than positive results. Different registries may present data in different ways and formats, and analysing all the data together may be challenging. Annual reports are created for some registries, but not all registries are publicly accessible.
Extending a device's indication
A PMCF study cannot be used to extend a device's indication. Data for an extended indication require a new clinical investigation.