Navigating Combination Products: Device Complexity And Software Integration

By Samir A. Shah, Ph.D., Devices and Digital Therapeutics, CMC Regulatory Affairs, AstraZeneca

In today’s digital landscape, it’s easy to see the advantages of combining therapeutics, such as drugs and devices, and then pairing those with software and other electronic technologies. However, it’s equally important to consider the challenges and cross-functional collaboration necessary to develop products that combine drugs and devices, and those that are software-enabled. What should be considered when technical development relies heavily on communication between drug, device, and software teams?

In this Q&A, Life Science Connect’s Izzy Dininny catches up with Samir A. Shah from AstraZeneca about the communication constraints, regulatory implications, and strategies for navigating the development of drug/device combination products, including those that are software-enabled. The observations below reflect his experience working on combination products and are intended as general industry learnings, not commentary on any one company.

1. In drug/device combination product development, where do you most often see communication breakdowns between teams and what are the consequences?

I most often see communication breakdowns at the drug–device interface, where drug teams don’t always realize the significance of the device constituent/part requirements and the time it takes to generate the supporting evidence and documentation. Device activities such as risk management, usability/human factors, performance testing, and supplier collaboration may not be treated as submission‑critical until late, or teams assume they can be accelerated when they often can’t (a four‑week temperature cycling study still takes four weeks).

A key cause is the lack of harmonization across regions in how combination products and device constituents are defined and regulated. That reduces predictability, and teams can take different views of the true timeline risk. In Europe, for example, Article 117 of the MDR (applicable since 2021) introduced the need for Notified Body input to support marketing authorization applications for integral drug/device combinations, and Notified Body lead times and documentation effort can be significant. If this isn’t built into the plan early, it can become a critical path bottleneck and drive urgent, unplanned work. Earlier in development, there can also be situations where both medicine and device-related clinical applications need to be considered, so those discussions must happen early.

In addition, teams sometimes encounter country-level expectations that are not always explicitly addressed in public guidance and may only become clear through questions during review. For these reasons, combination product programs benefit from not only strong project teams but also dedicated capability to monitor regulatory intelligence and participate in public consultations (e.g. submitting public comments), so emerging expectations are translated into a practical plan before they become late-stage conflicts.

2. Where do you find teams underestimating the complexity of combination product development, in terms of technical or regulatory?​

Underestimation is most prevalent in evidence generation at the drug–device interface and in post-approval change management. Device development can sit in a separate lane, but regulators increasingly expect evidence that the drug and device work together as a system. For example, labeling elements such as the instructions for use typically need to be assessed through human factors/usability, which depends on understanding the user population, aligning on study design, and often engaging with regulators, activities that need to start well in advance of submission. It’s also easy to assume an on‑market device is already proven, but for a new formulation, presentation, or patient population, authorities may still expect device safety and performance to be demonstrated or robustly risk assessed in the new context of use.

The second area is planning for changes. Device constituents often have long supply chains where change is inevitable. For stand‑alone medical devices, many regions have clearer conventions for changes that don’t require authority interaction, but for drug/device combination products the expectations can be less predictable. That uncertainty drives debate and rework over what can look, on the surface, like a straightforward change.

Efforts to improve predictability, such as the application of ICH Q12 principles on changes to drug/device combination products, are therefore important. It also reinforces why teams need bandwidth for advocacy (commenting on proposals) and regulatory intelligence (tracking adoption and interpretation), so lifecycle plans are based on what authorities actually expect in practice.

3. What new challenges does the addition of software introduce when integrated into combination products?

When software is included in a drug/device combination, it adds a lot of challenges. I’ll just speak to a couple.

One way software is used with drug/device combination products is through digital health technologies (DHTs) to monitor patients and physiological parameters during clinical trials. A significant question that often comes up is whether the tool is being used as a consumer/general wellness tool or whether it needs to meet the higher bar associated with software as a medical device. The answer comes down to how it is used in the clinical protocol.

This includes, for example, collecting exploratory data versus influencing clinical decisions (such as safety monitoring) versus supporting a study endpoint. Each of those uses can drive very different expectations for classification, documentation, oversight, and planning. And because protocols evolve and can be lengthy, it’s important to flag the relevant sections to device/software regulatory experts. This way the regulatory position in the regions of interest can be reassessed and remains aligned to how it is classified for the study.

FDA’s recent request for information and comments on advancing DHT use in clinical investigations shows the field is evolving and FDA is actively seeking input to inform future guidance and activities. This is another reason teams benefit from having a dedicated regulatory intelligence capability and, where appropriate, participating in public consultations, including submitting comments to improve clarity and predictability.

A second challenge is how much validation is needed, especially for commercial off‑the‑shelf technologies. Teams sometimes mix up validating the technology (sensor performance, algorithm accuracy, software verification/validation) with validating the clinical measure (whether the measure is clinically meaningful and interpretable as a drug effect). This needs to be planned up front, because regulators’ expectations will depend on what is ultimately claimed, and the responsibility for demonstrating fit‑for‑purpose sits with the drug sponsor, not the technology vendor.

4. Looking ahead, where do you see the biggest opportunities for innovation at the intersection of drug delivery and software?

I see the biggest opportunity in using software with drug delivery to expand access and participation in clinical trials, while also enabling the collection of more meaningful clinical evidence remotely. Done well, this can help studies recruit faster across a wider geography, assess outcomes that better reflect real-world functioning, and generate high-quality evidence so therapies can reach patients sooner.

First, software can enable virtual or hybrid participation, reducing reliance on large academic centers and allowing participation from patients who are geographically distant or are in rarer populations where traditional site-based recruitment is a bottleneck. Telehealth, wearables, and other software-enabled tools can lower practical barriers to participation and make assessments more feasible outside the clinic.

Second, software creates the opportunity to collect novel remote data that goes beyond snapshot site visits. Continuous or more frequent measurements can provide a more complete view of patient status and day-to-day function. This can improve statistical efficiency and support endpoints that are more meaningful to patients and clinicians, assuming the tools are fit for purpose and the clinical interpretation is clearly defined (the challenge I described in the last question).

If we get those foundations right, the intersection of drug delivery and software becomes a means for faster development and more efficient regulatory submissions. This step forward can help bring therapies to the patients who need them more quickly.

5. What skills and capabilities do teams need most to succeed in drug/device combination products, particularly those that are software-enabled, and how can leaders build them?

Leaders can equip their teams by recognizing that success in combination products, and especially those involving software, requires more than strong science and engineering. It also requires strong influencing skills, because drug, device, and software teams are working across different frameworks and expectations. This means people need to be able to listen well and ask the right questions so they can understand what matters to other stakeholders and where assumptions differ.

Credibility also matters. Teams build credibility by being responsive, communicating clearly, and following through. Since so much collaboration happens digitally now, it helps when scientists and engineers can get to the point quickly and communicate in a way that is both competent and warm. Even simple habits, like adding context when you reach out to someone you do not know, can make collaboration more effective. That combination makes it easier to surface issues early, align on expectations, and avoid last-minute surprises.

A further point is to encourage teams to find ways to monitor regulatory intelligence, so they don’t miss opportunities to make public comments. Whether through internal processes, vendor tools, or other approaches, staying aware of emerging policies and contributing feedback where appropriate is important in these niche areas.

When leaders invest in these capabilities, they reduce the communication breakdowns that are common at the interfaces and help teams navigate the real complexity of software and drug/device combination products. Ultimately, that improves planning and execution and supports getting high-quality therapies to patients sooner.

About The Author

Dr. Shah guides teams on regulatory strategies to achieve worldwide approval of drug/device combination products, including those connected with software and mobile apps. His approvals include integrated smart/digital inhalers, as well as inhaler and autoinjector products. He currently works at AstraZeneca in Regulatory CMC, Devices and Digital Therapeutics, and previously worked at Teva Pharmaceuticals. Prior to joining regulatory, he was a formulation scientist at Merck / Schering-Plough for respiratory and nasal products. Dr. Shah earned his B.S. from Case Western Reserve University in Polymer Science Engineering and a PhD in Biomedical Engineering from Wake Forest University School of Medicine/Virginia Tech.