By Lauretta Ihonor, MBBS
The past 10 years have brought a remarkable shift in the way data is collected and stored in healthcare. From clinical trial statistics to drug reaction rates and patient hospital records, the collection of so-called “Big Data” is big business. And while the benefits of gathering and analysing data in these conventional settings are apparent, there is also much potential for Big Data to revolutionise the medtech industry.
Gathering the Data
Safety and reliability of medical devices is a key area of concern in medtech worldwide. Indeed the US Food and Drug Administration (FDA) cited “advancing methods to predict clinical performance of medical devices and their materials” as a key regulatory science priority for 2016. Given the potential of capturing and feeding back real-life medical device performance to transform the safety of newer medical products, it is clear that Big Data can, at least in theory, improve medical device safety and predictability.
The FDA’s Sentinel Initiative – a US-wide electronic system for medical product safety surveillance – officially launched in February 2016, after several years of preliminary testing.
The Sentinel System utilises electronic healthcare data from patients treated with regulated medical products to monitor adverse reaction rates. Its ultimate aim is to improve the overall safety of medical products and public transparency about the potential risks and benefits associated with each product. However, while such an aim is to be applauded, its effective execution has its challenges – challenges that span obtaining consent, storage of data, maintenance of confidentiality, and obtaining data that has real life relevance.
The latter point is a key focus of US-based non-governmental organisation (NGO), The Pew Charitable Trusts. The NGO has long been vocal about the importance of collecting relevant high-quality medical data that can be used to improve patient care, and has set up a Medical Device Initiative dedicated to doing just that.
Ben Moscovitch (pictured left), Officer for The Pew Charitable Trust’s Medical Device Initiative, says: “Medical device development and regulation increasingly relies on a total product lifecycle approach, where data are gathered on device safety and performance during device development and after the product is marketed.”
He believes that the first step to improving the quality of data available is: “Adding Unique Device Identifiers (UDIs) to electronic health records and insurance claims. This will ensure that these sources of real-world data contain information on the products used in care.”
The Role of Tracking Numbers and Registries
Moscovitch is not alone in his advocacy of UDIs – device-specific codes that allow users to retrieve identification data for a given device – in improving device monitoring. Its use has been championed outside of the US, forming a key focal point in the almost decade-long discussions surrounding the reworking of the European Medical Device Directive (MDD) into the new European Medical Device Regulation (MDR).
Identified as an effective method of tracking devices through distribution and use, UDI implementation has the potential to monitor device performance and thus improve patient safety and postmarket device surveillance.
And as noted by the Agency for Healthcare Research and Quality in a 2014 publication, Registries for Evaluating Patients Outcomes, the usefulness of UDIs can be further bolstered by the presence of robust patient data registries. This is because it is these registries that provide much-needed clinical context about how a given device has been used and how it fares in the real world.
However, Professor Bruce Campbell (pictured right), recent Chair of the National Institute of Clinical Excellence (NICE) Advisory Committees on Interventional Procedures and Medical Technologies, highlights that while there are now some well-established registries in the US, UK and Europe, there is still work to be done.
He says: “Registers can be hugely valuable in showing the performance of surgeons and hospitals, but they vary in terms of the specificity of data they give about particular devices. They often provide information on device classes because of the kinds of operations people are reporting, but not necessarily on particular branded devices.”
But, he adds that recent years have brought progress in this area, primarily owing to agreement by the majority of the medical and surgical community that registries have an important role to play in improving patient outcomes.
Challenges to Overcome
Nonetheless, the establishment of device-specific registries has its challenges. A 2015 report by Avalere Health LLC identified difficulty in sharing information between electronic health record platforms, a lack of commonly adopted data standards and concerns about data security and patient privacy as key obstacles that are yet to be addressed.
Cardiac, vascular and orthopedic registries in the US, UK and Australia already collect detailed information on devices implanted during surgical procedures. And they have demonstrated that successful surveillance of device safety is possible via registry implementation. For example, the US Data Extraction and Longitudinal Trend Analysis (DELTA) network study monitored 23,805 adults who received coronary devices, such as stents, over five years. A pre-agreed acceptable level of adverse events among all patients was set and a safety alert triggered if this level was exceeded. The continuous monitoring of data collected from the network of hospitals participating in the study led to two safety alerts being triggered, allowing the devices to be closely reassessed and confirmed safe for continued use.
In short, the study demonstrated that post-market real-time safety surveillance is feasible and provides an opportunity for the type of data-driven intervention that maximises product safety.
Putting it All Together
Just as the use of UDIs has limited efficacy without registries to keep track of the context under which each device is used, the full potential of registries cannot be realised unless used in combination with other sources of patient data.
Moscovitch explains: “The establishment of an integrated system that leverages registries, electronic health records, and insurance claims can help generate better data on the performance of medical devices.
“Each data source provides information that the other data sources may lack; when used together, these data sources can provide more robust information on device quality over time.”
Indeed, hospital electronic health records provide in-depth information on patients during a particular episode of care, but they often lack data on patient care after discharge – a crucial source of context when analysing factors influencing device performance outside of a hospital setting. In contrast, insurance claims have data on adverse events that arise from a procedure or device, but less information on clinical care and the specifics of the device used.
Tying together different data sources will provide an ultimate Big Data set that could transform device safety and innovation, but the focus must now turn to how to pool these distinct databases, while navigating the problems of confidentiality, data ownership and patient consent for their data to become more widely disseminated than ever before.
This is not a mere pipe dream.
The Pew Charitable Trusts reports that expert groups convened by the FDA are currently assessing how a national device evaluation system can be developed. But while work continues on developing systems that maximize the use of Big Data in device safety and innovation, one key player in data-based safety assessment should not be overlooked – the manufacturer.
As Prof Campbell, explains: “It’s so important that medical device manufacturers understand the type of data evidence they will need to demonstrate the safety and efficacy of their product, and that they learn how to best generate it. This is not only for meeting regulatory demands, but also to have a very clear idea about the outcomes it will produce when used in daily clinical practice.”