Guest Column | February 18, 2022

When & How To Develop A Novel Digital Endpoint In Clinical Research

By Michelle Crouthamel, head of digital science, AbbVie

Making Decision GettyImages-1286473067

Digital health technologies (DHTs), including mobile health apps, wearable sensors, telehealth, personalized medicine, and health information technology (as defined by the FDA), represent the convergence of technologies, health data, and advanced analytics and offer potential opportunities to transform drug development and improve healthcare access and care delivery.

Many biopharmaceutical companies are leveraging DHTs to improve existing clinical endpoints or uncover new endpoints (novel digital endpoints) that could not be measured before. For instance, researchers can now use DHTs to investigate the burden of sleep disturbances in patients with various disease conditions in the comfort of the patient’s own home. These novel digital measures of health can have a significant impact on a broad spectrum of drug development, providing critical insight around efficacy and safety during drug development phases as well as post-marketing real-world effectiveness.

 “Tools for remote data gathering (such as mobile electronic Clinical Outcome Assessment (eCOA), novel sensors, actigraphy, camera, voice, and video) are increasingly being validated, establishing standards for their broader use. Digital endpoints are being used more and more as primary endpoints, accounting for 63 of 225 digital endpoints crowdsourced by the Digital Medicine Society (DiMe) as of September 2021,” according to recent McKinsey research.

Despite significant investment and efforts, progress with the development of novel digital endpoints (NDEs) has been slow, which can be attributed to a variety of reasons:

  • Over the past 10 years, many attempts to develop NDEs were conducted in a siloed and fragmented manner. The library of NDEs published by the DiMe highlighted this exact problem. The silos fragmented learning, data, and algorithms across industry, academics, and vendors and significantly delayed the advancement of NDEs. We need cross-industry leaders to recognize this barrier and come together to address this issue head on.
  • Although the FDA emphasized that the NDE development should follow the patient-focused drug development (PFDD) process as well as the eCOA development guidance, due to the lack of successful NDEs qualified by the agency, the industry overall still has limited understanding of the evidentiary requirements, resulting in many NDE attempts failing to meet the basic criteria of “meaningfulness to patients.”
  • The development and validation of NDEs is a multifactor, multidiscipline, and multiyear endeavor that requires the full alignment of many elements, e.g., the right drug, the right indication, the proper choice of NDE, the readiness of fit-for-purpose DHT, the clinical operation, and regulatory alignment, and all need to work in concert to achieve success. Many companies struggle to have internal alignment of proper know-how, right-size risk management, and long-term commitment.

However, efforts are being made to improve education and support the development of NDEs.

Growing Support For NDEs

As part of the 21st Century Cures Act, the FDA formalized the drug development tools (DDT) process and established three milestone stages, Letter of Intent (LOI), Qualification Plan (QP), and Final Qualification Package (FQP), to qualify tools such as biomarkers, clinical outcome assessments, and animal models to facilitate drug development. DHTs that enable passive monitoring of health outcome are reviewed under the COA office.

In addition to the DDT pathway, the FDA also offers reviews of DHT-derived endpoints under the Critical Path Innovation Meetings (CPIM) or the IND pathway.

Earlier this year, the FDA launched a pilot program called the Innovative Science and Technology Approaches for New Drugs (ISTAND), designed to encourage the creation of drug development tools that are out of scope for existing qualification programs, such as the use of novel DHTs for patient assessment, digital photography, and AI-enabled tools. Additionally, the FDA also offers DDT research grants to encourage tools development.

There has been some headway with gaining regulatory approvals. For instance, six DHT-Passive Monitoring COAs have an accepted Letter of Intent under the DDT program. In addition, the FDA accepted the use of actigraphy, which measures moderate to vigorous physical activity via a small sensor, usually worn on the wrist, as a primary endpoint for investigational product development in patients with pulmonary hypertension associated with interstitial lung disease.

TransCelerate Biopharma’s Patient Technology Initiative has developed a variety of resources available to accelerate the adoption of patient-facing digital technologies (PT) in clinical trials, including a toolkit to assist the ecosystem with the implementation of PT in clinical trials. Our recently published paper, “Developing a Novel Measurement of Sleep in Rheumatoid Arthritis: Study Proposal for Approach and Considerations,” walks through the process and key considerations for developing and navigating regulatory acceptance of digital tools for NDE implementation in a clinical trial, using a case study example.

So, When Does It Make Sense To Pursue An NDE?

NDEs are not necessary for every drug program. The following questions are a good starting point to assess whether it’s worth considering:

  • Do you have a new concept related to a disease that is complex to measure by clinicians yet meaningful to patients, e.g., nocturnal scratching and sleep disturbance in atopic dermatitis?
  • Is there clear demand for new drug development tools due to the unmet medical needs or rare disease areas where a robust clinical assessment tool is missing or has significant limitations, e.g., ALS, Duchenne muscular dystrophy, or Parkinson’s disease?
  • Will the tool add value to patients or prescribers or payers in the process of drug development or care delivery?
  • Do you have a reasonable hypothesis to believe that the experimental drug can improve the symptoms you intended to measure using a DHT?
  • Do you have a fit-for-purpose DHT that’s mature enough to measure what you want to measure?

If the answer is “yes” to at least a few of these questions, it’s worth exploring the possibilities of developing an NDE.

Developing An NDE: Key Considerations

Developing an NDE requires a multidisciplinary, multi-study approach with strategic planning and a regulatory-guided pathway to achieve regulatory and clinical acceptance. What follows are some of the key considerations at each stage of the process:

1. Assess functional impact and meaningfulness to the patient.

Clinically meaningful endpoints that directly measure how a patient feels, functions, and survives are the basis for approval of new drugs.

It’s my personal view that not every drug development program needs an NDE. When initiating an NDE project, drug developers should carefully select which drug development programs will benefit from NDEs. In some programs, patient reported outcomes (PRO) are well suited to assess whether the drug helps the patient feel better. On the other hand, DHTs can measure certain daily functions objectively. The most critical question to start with is: What functional improvement would be clinically meaningful to the patient and to demonstrate efficacy?

For instance, many patients with atopic dermatitis suffer severe skin itching. A significant reduction of itching would be a critical endpoint to prove the treatment benefit of the drug. However, itching is a subjective sensation that is difficult to assess objectively. One may propose measuring scratching as a surrogate measure of itching. Before jumping into the development of scratching measures, it is important to seek the patient’s input and gain regulatory alignment. Similarly, in the paper mentioned above, patients with rheumatoid arthritis often suffer severe joint pain with daytime fatigue and nighttime sleep disturbance. There’s pathophysiology evidence linking misregulation of inflammatory cytokines and sleep disturbance in inflammatory diseases. One may hypothesize that a significant improvement of joint pain, as well as reducing fatigue and sleep disturbance, could be important and meaningful to patients. Therefore, developing novel measurements suitable for assessing fatigue and sleep disturbance in RA patients could be valuable. Again, before initiating the NDE effort, seeking patients’ and regulators’ or payers’ alignment is highly recommended.

2. Identify the fit-for-purpose tool.

Once you determine what functional improvement is important to the patient, it’s critical to identify the right tool and evaluate how it will be used. When measuring sleep disturbance in RA, it is important to consider details such as: which sleep parameters are important to the patient, i.e., is it the number of awake hours after sleep onset, sleep efficiency, and/or is the DHT capable of measuring those sleep parameters accurately and reliably? In addition, consider the age of the patient, patient comfort, length of time required to measure the endpoint, how the device will be worn, how frequently data will be collected, and how to mitigate potential issues around data collection (e.g., uncharged devices, missing data).

3. Study design and execution are critical.

Depending on the maturity and the validation of the DHT that is selected, additional validation may be necessary. Using sleep as an example, sleep quality and quantity are normally assessed by polysomnography (PSG) as it is the clinical “gold standard” and regulators expect proper validation conducted in the intended patient population.

The broadness of your claim will affect the design and scope of the study. Historically, the FDA required PSG data to substantiate claims related to sleep improvement. When developing NDE, particularly when there’s an established gold standard, it is critical to seek regulatory alignment early to avoid wrong assumptions and costly do-overs. Operational and data privacy laws are constantly evolving and vary by country. Patient consent forms should clearly explain how data will be collected, used, analyzed, and shared — and who will have access to it. The principles of GCP and minimal data collection still apply in the NDE development.

4. Determine how data will be analyzed.

Invest ample time up front to determine how data will be analyzed. Digital endpoints require digital signal processing before endpoint analysis — and that requires a special skillset that is relatively new to the pharmaceutical industry. Spend time thinking through your threshold criteria for patient compliance. It is inevitable that some patients won’t wear their device for the required number of hours per day or may miss a day.

Once the data analysis is complete, it will need to clearly show how the new endpoint correlates with disease activity, the variability, and the difference or benefit of using the novel approach to enable decision-making.

5. Engage early for regulatory acceptance.

Do not wait until the very end to reach out to regulators. As mentioned above, engaging early and obtaining regulatory alignment is an essential factor to increase the probability of success. Get involved in different engagement pathways with regulators, such as critical path innovation meetings (CPIM, FDA) and Innovation Task Force (ITF, EU), to garner early alignment around the concept. Alternatively, you can submit a briefing book and request meetings or written feedback through IND pathways. Importantly, circle back with regulators after the data is collected to have a robust conversation about the findings. Firm up their support so you can use the information in your confirmatory study and justify the potential label claim.

Success Is A Collaborative Effort

Standardization of outcomes assessment is a big topic of discussion surrounding NDEs, and it is crucial for regulatory agencies and payers when they compare how different drugs perform. Many pharmaceutical companies agree that the drug molecule (e.g., how the drug performs in safety and efficacy) should be the focal point for industry competition, not the clinical endpoint itself.

Industry collaboration is crucial to standardizing endpoints and the consistency of measurement solutions. Harmonization and alignment with stakeholders are key to avoiding multiple measurement solutions for similar concepts. A variety of public-private partnerships exist to identify NDEs that assess fatigue, sleep, and other activities related to daily living with various indications. For example, The Critical Path for Parkinson’s Consortium is a group of pharmaceutical companies that are working together to develop novel digital endpoints and new treatments for Parkinson’s disease; developing novel endpoints would otherwise be challenging due to the complex nature of the disease.

We are excited to see the draft DHT implementation guidance published by the FDA recently. While the path to NDEs can be difficult, with the right support and guidance it can be an exciting opportunity to set a new bar that will have a major impact on drug development programs and improve patient lives.

About The Author:

Michelle Crouthamel, head of digital science at AbbVie, is an industry thought leader in digital health, with a broad spectrum of R&D experience. She currently leads the digital health strategy as well as develops digital biomarkers and endpoints. In her previous roles, she has led drug discovery, clinical development, project management, and digital-pharma strategy and execution. She is very passionate about transforming the pharmaceutical industry to become truly patient-centric.