Guest blog: Joseph Lengfellner on the challenges and solutions for clinical trials – Part 2

Joseph Lengfellner on the challenges and solutions for clinical trials - Part 1

Guest blog: Joseph Lengfellner on the challenges and solutions for clinical trials – Part 2

Four things that would transform clinical trials – in part two of this blog series, Joseph Lengfellner details his pick of the most exciting new and emerging solutions to the challenges of running clinical trials.

In part one of this blog, I discussed the biggest challenges of managing clinical trials, and listed some of the new and emerging solutions to these problems. In my view, these have the potential to transform how we conduct clinical trials, making this a particularly exciting time to be in this field.

Here’s a little more on the solutions I touched on in part one:

Common data standards

Underlying much of the advancement we’re seeing in the use of data for clinical trials, is the development of common data standards like HL7 FHIR. With FHIR increasingly becoming a universally accepted and used data standard, healthcare technologies are being powered towards modernity. Not only does this mean increased ability to share data, but also to develop tools and interfaces that better serve clinicians, research teams and patients.

Healthcare professionals across the board have long found themselves using tools in their work life that lag behind the technology they’re used to in their personal life. Your typical smartphone will feel far more advanced than many healthcare systems – and certainly more user-friendly.

Now, SMART on FHIR applications, built on modern RESTful API web technologies are slowly allowing us to build modern tools, featuring user-friendly interfaces that support rather than hinder all aspects of healthcare data entry, use and re-use.

EHR-to-EDC

One of the solutions being driven by FHIR deserves its own mention. EHR-to-EDC technology is bringing the next major shift in how we manage clinical trials, by allowing for the swift and seamless re-use of electronic healthcare record (EHR) data.

Existing methods for transferring source data from EHRs into electronic data capture (EDC) systems for clinical trials typically involve the painstaking and laborious manual re-keying of data from one system into another. In my previous blog, I outlined some of the issues this causes in clinical trials, including data quality issues, the time and costs associated with the process, and an overburdening of valuable research coordinators with menial copy-and-paste work.

The reason my team at Memorial Sloan Kettering began a collaboration with IgniteData to provide data transfer integration between the EHR at MSK and the electronic systems of two clinical trial sponsors, is because the value of this is so clear. As things stand, the technology can automate the transfer of large amounts of data routinely collected in the EHR, including structured data like lab results, vital signs and medication data. This represents a significant percentage – around 50% – of clinical trial data that is no longer being manually re-entered into the clinical trial system. Not only does this significantly reduce the burden on research coordinators, but it ensures the highest quality data, direct from source, is being used. It reduces queries from trial sponsors and contract research organisations (CROs), and research coordinators can focus on the more valuable and challenging aspects of their role.

We can feel confident about these benefits, because we’ve tested out the concepts before. I personally have been working with eSource technology since around 2017 and have seen the successes of small-scale proof of concept studies that proved the benefits of using data direct from the source. What was missing from previous efforts was a solution that was truly scalable, that could work with multiple EHRs and multiple EDCs, and that introduced a seamless data pipeline, without the need for additional steps mid-way.

Now that advanced EHR-to-EDC technology like IgniteData’s Archer is solving these issues, we are finally at a tipping point. I see this very quickly becoming the way we manage studies. That, in turn, will fuel a major shift in clinical trials. It will speed up the execution of trials and drive down the overall cost, remove a huge burden from both sponsors and trial sites, and make it much easier for sites to engage with sponsors.

Artificial intelligence (AI)

AI, though not quite at the point of offering out-of-the-box solutions, is evolving rapidly. Clearly visible down the pipeline, is an ability to use AI tools, such as natural language processing (NLP) and machine learning, to deal with all the unstructured data we currently hold in EHRs.

In this way, I see AI bringing the next phase of EHR-to-EDC capability. While we’ve already established the ability of EHR-to-EDC to successfully map and transfer large amounts of structured data, there is also a lot of data that we need to access for clinical trials held in unstructured form in the EHR. Free text notes and imaging are prime examples that contain staging information for cancer patients.

Within the next one-two years, I can see us being in a position to use AI to bring the next iteration of EHR-to-EDC. To be able to extract concepts from unstructured notes, and identify elements and values that can then be used by research teams to populate research systems.

The benefits of this will be many-fold:

  • It will greatly reduce the data and documentation burden on our clinical teams, where so much time is currently spent sifting through unstructured data to extract meaning.
  • It will create exciting opportunities for second use of data, with the ability to take existing clinical data, bring structure to it, and complete further research.
  • Similarly, it will transform our ability to conduct real-world studies, with robust real-world data suddenly becoming easily accessible.
  • It will support recruitment of patients for clinical trials, since many of the inclusion/exclusion data for trials is held in unstructured notes. Bringing structured concepts to notes will make a huge difference to our ability to identify suitable patients for trials.

Modernised clinical trial documentation

A shift I hope to see in the near future is the modernisation of clinical trial documentation and protocols. This may not sound as exciting as using AI, but would dramatically improve our ability to get trials up and running more quickly.

Typically, the clinical trial documents that set out how a study will be done are produced as unstructured Word or PDF documents. The research team has the job of going through this 30-50 page document to glean everything from the science behind the treatment, to inclusion/exclusion criteria for patients, to safety reporting requirements. The process of translating this written protocol document into action (including creation of study tools, database design, contract and budget negotiation) adds to the time required to open a new study, and ultimately increases the overall time it takes to complete a study.

Whilst we are yet to see any major successes in tackling this problem, there is now increasing recognition of the need to modernise this process and the benefits of doing so are clear. If we are able to bring structure to clinical trial protocols, we can quickly identify inclusion/exclusion criteria, assess the feasibility of identifying a sufficient number of patients to participate in the trial and begin recruiting those patients immediately. We can structure the data elements we need to collect as part of the trial. We can automate the build of the EDC system. We can build a budget for the study more quickly. Ultimately, we could vastly speed up the time it takes to go live with a trial and remove much of the pain of this process.

Tangentially, if clinical trial documentation was more structured and easily interrogated, we could take a more critical look at how we design studies. In some cases, there is a tendency to copy and paste from one protocol to another, without full consideration of whether we need to collect certain data elements. This only exacerbates the increasing complexity and quantity of data we collect for clinical trials. It’s in the interests of everyone involved, from research teams to sponsors and CROs, to make efforts to tighten up how we design protocols, including the data we need to collect to ensure the safety and efficacy of treatments.

A vision for a new way of managing clinical trials

The solutions I’ve outlined are at varying stages of development. But we can also realistically see a world where they are all enacted. In the case of EHR-to-EDC technology, it certainly feels that the industry has waited a long time for this solution to be a real, scalable option. And we’re now finally on the verge of it being at the heart of a major shift in clinical trials.

Taken together, these solutions would offer a real transformation to our ability to conduct clinical trials. For the people working in clinical trials, they would improve job satisfaction significantly. At a programmatic level, the efficiencies they offer would allow research sites to be far more efficient and to conduct more studies. In turn, this would offer more trial opportunities for patients and more potentially life-saving treatments being made available. I, for one, am looking forward to seeing that day very soon.

The opinions offered during this interview reflect Joseph Lengfellner’s views only and do not represent those of MSK.