How do regulatory requirements within the medical device industry impact a database build?
What are the important items to be aware of? In this blog, we will look at these and other questions, with a focus on the data management/database build work within medical device studies for the US and EU.
Medical devices are an important part of healthcare. They run the spectrum from the common (e.g., bandages), to diagnostic tests and complex surgical interventions and drug delivery mechanisms. Before they can be put on the market, they need to be shown to be safe and effective. This is accomplished through testing based on regulatory requirements.
Depending on the ‘risk’ of the device and the geographic area (US, EU) of approval, the FDA has categorized this risk into 3 categories:
| Classification | Risk | FDA requirements | Examples |
| Class I | low | General controls | Bandages, hand-held surgical instruments |
| Class II | moderate | Special controls, such as Premarket Notification 510(k) | CT scanners, powered wheelchair, arthroscopic repair |
| Class III | high | Premarket Approval (PMA) | Implantable pacemaker, implanted prosthetics |
Together, the FDA and EU requirements present special challenges for clinical trial processes and gathering of necessary data for analyses to support the approval of medical devices.
Importance of database build for medical device studies
A database-build directly influences the quality of the study data and is an important part of the medical device study. The regulatory informational needs from both the FDA and EU need to be included in the overall strategy for building a quality and effective database.
What is a database?
The term ‘database’ often refers to both the collected data as well as the database management system (DBMS) — or software that interacts with the end-users, applications, and data itself.
Types of data include:
- eCRF/Electronic Data Capture (EDC) for data collection (structured data), including device specific identifiers (UDI-DD)
- Other electronic sources and documents (can include unstructured data – e.g., x-ray, CT scan, MRIs, device data, ePRO, eCOA)
Database management systems may deal with:
- Web-based data entry
- Query management
- Data hosting (e.g., Cloud platform), data transfer and data control access
What is a database build?
The set-up of a database plans for all data streams and may include strategies for skip logic, intelligent guided entry, and pop-ups to simplify the entry of data points following the ALCOA principles (Attributable, Legible, Contemporaneous, Original and Accurate) and to maintain the GCP principles for data reporting and data integrity. It must also be able to handle large amounts of data, found in many post-market surveillance device studies.
Build Challenge
The ability to integrate the various sources of data, from the eCRF data to the electronic sources, into an effective clinical solution that is accessible and supports regulatory reporting requirements, can be a challenge.
Choosing the Right EDC System
Finding the right EDC system to meet the study needs and requirements at the right price point is needed for both the satisfaction of the sponsor as well as the data management team. Will the trial have decentralized monitoring? Are advanced analytics and visualizations needed? Working knowledge of the various EDC systems as well as the technology to support needed capabilities is important.
Device-Specific Information Challenges
Specific information on the devices is needed for reporting purposes.
- While device trials tend to have smaller sample sizes than drug trials, the growing presence of post-marketing studies can dramatically increase the data needs for the database based on long term studies.
- The FDA in 2013 published the UDI System Final Rule, requiring manufacturers to label marketing devices with a UDI.1 The UDI-DD identifies the manufacturer and model as well as the lot number, serial number and expiration date.
- This increases the traceability of the device, implant identification and supports reporting for AEs, recalls, and other surveillance measures for reporting.
Some FAQs on Medical Device Database Builds
What kind of device trials have you supported?
As a CRO specializing in medical devices, we have supported database builds for Class I, II and III trials. The needs for Class I trials are the least of all trials, requiring only a study protocol and source document for the few PRO (patient-reported outcome) forms. A Class III trial (PMA) is the most extensive effort of the three with systems designed to capture secondary and tertiary endpoint information.
| Class 1 | Class 2 | Class 3 |
| Low Risk devices: Surgical Instruments, Bandages, etc. | Moderate Risk devices: Syringes, Sutures, etc | High Risk devices: Pacemakers, Shoulder rotator cuff, etc. |
Needed information for Database Build:
|
|
|
What kind of targeted assistance is needed with a device study Database Build?
Concentrating on the uniqueness of device trials, rather than the typical efforts that contribute to a database build would include:
- Expertise on needed datapoints to include based on regulatory requirements of FDA, EU or other regulatory agency.
- Pre-planning the Database Design for long-term studies which can also support multiple protocol amendments, often found in these types of studies.
- Customizing the visibility/entry of the eCRF based on the special role (e.g., Independent Assessor is often found in device trials) in the study.
- Customizing the design for Bi-lateral cases would support the site’s ability to enroll the subjects in bilateral cases. i.e., if the subject has an indication in both legs, then both legs will be involved in the study with the space to collect different sets of data with the same subject number.
- Providing an end-user friendly database allowing the site to select and enter data easily instead of by a more tedious manual entry.
- Device number entry (Catalogue numbers/lot numbers/device identifiers) can be entered. There is catalogue or schema with device numbers or lot numbers that is fed into a IRT system, which through API’s are integrated to EDC. So, every time a device is assigned to a subject in the trial, site select the number in the IRT system which gets integrated into the database.
- Lab Set-up
- eCRF forms can be created to make the collection of lab data easier where there may be a list of Analytes, Units, and Reference ranges for each lab across all sites.
- Local lab results are entered into the system for each lab analyte, then pre-configured conversion factors normalize the data into a common standard unit, in real-time.
- Normal reference range values for each lab test, with gender and age attributes can be maintained, and the system can send an alert for out-of-range values.
In what areas can cost efficiencies be found?
- In most EDC Systems, safety data Trackers can be built for AE and Device Deficiency reporting, eliminating the need for additional efforts.
- Email notification for multiple scenarios can be programmed within an EDC system, which can be used for the immediate reporting of AEs/DDs to regulatory bodies.
- Efficiency can be gained from previous study designs and data trends, which will yield faster and less expensive studies.
How is the Database design for different franchises handled?
- Handle the different franchise designs by using the Global Library and creating different CRF versions for the same forms based on the different franchises.
- Library of electronic patient reported outcomes (ePROs) is available to draw from; this can be segregated as region wise.
- Our Database supports multiple ePROs site wise.
- Single Database can be designed which can covers multi sites with different regulatory standards
These challenges are different from the processes and data needs of a clinical trial for a drug and are best met with the expertise of professionals working with device trials based on relevant regional experience.
Veranex is a company that delivers end-to-end integrated services with a wealth of expertise in Medical Device Solutions, including Data Management and Analytics services.
References:
1. Drozda JP Jr, Graham J, Muhlestein JB, Tcheng JE, Roach J, Forsyth T, Knight S, McKinnon A, May H, Wilson NA, Berlin JA, Simard EP. Multi-institutional distributed data networks for real-world evidence about medical devices: building unique device identifiers into longitudinal data (BUILD). JAMIA Open. 2022 May 25;5(2):ooac035. doi: 10.1093/jamiaopen/ooac035. PMID: 35663113; PMCID: PMC9154019. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9154019/