Four essential considerations for launching digital diagnostic tools
Pharmaceutical companies interested in joining the global digital health revolution increasingly face difficult decisions about how to take advantage of innovative technology. Is it better to provide a low-risk digital tool that, for example, improves the patient experience, or to invest in a long-term digital technology, such as the development of a diagnostic tool with the potential to help improve clinical outcomes? While the answer will greatly depend on patient needs and the business model, here are some important considerations for success in the digital diagnostic space.
1. What is at risk?
Software functions1 that meet the definition of a device or software as a medical device (SaMD), can be deployed on mobile platforms, other general purpose computing platforms or in function or control of a hardware device. Digital technologies that are regulated by the United States Food and Drug Administration (FDA) include:2
- Mobile applications that use the built-in functionality of a mobile platform such as light, vibration, camera, or other similar sources to perform the functions of a medical device (for example, the mobile medical applications used by a licensed practitioner to diagnose or treat a disease).
- Software functions that control the operation or function (eg, changing parameters) of an implantable or body worn medical device.
- Software functions used in active patient monitoring to analyze patient-specific medical device data.
The FDA regulates based on risk, so companies often think the best bet is to create a Class I digital tool to avoid going through the much more rigorous Class II authorization and approval processes or class III. However, going the Class I path usually means launching a digital tool with limited functionality to meet a treatment goal or solve serious healthcare needs. It will also generally have limited opportunities for differentiation, as consumers can already choose from hundreds of thousands of simple health-related apps. In comparison, investing in a high-risk digital diagnostic product with proven clinical utility can provide clinicians with a more impactful tool and pharmaceutical companies a path to monetization.
2. What is your use case?
Basically, digital involves the pursuit of one of two main revenue models:
- Generate indirect income for supportive therapy
- To create a stand-alone direct income stream
Many companies lead with the first option, but often inadvertently end up performing a mishmash of the two. This can arise when there is a lack of clarity on the business case or a significant change in accompanying therapy (FDA approval, clinical data, etc.). For example, various pharmaceutical companies have invested in accompanying digital diagnostics, only to find the main asset will fail later. In such cases, many companies let inertia move the digital product forward without reassessing the investment thesis and business model. Establishing a stand-alone revenue stream for a digital product requires a radically different investment thesis, level of risk, and time horizon. In this case, the investment / risk ratio may signal the need for a different strategy, including a partnership to sublicense the product or discontinue the idea.
When developing a digital product, pharmaceutical companies should consider direct and indirect revenue pathways early in the product development process, especially when entering new markets or seeking FDA approval. for accompanying therapies. Although regulatory clearance is more effective than the approval of a traditional medicine, digital products can face significant barriers to post-market adoption, as well as commercial and marketing risks. refund. Moving from an indirect revenue model to a direct revenue model can be difficult and may require a complete reset of product design and functionality, regulatory strategy, and clinical planning. By mapping direct and indirect revenue pathways, companies can be better placed to assess product development tradeoffs and make strategic decisions with a clear end goal.
To fully understand your best option, perform a strategic analysis of each path taking into account your business’ market opportunities, positioning, assets, and risk tolerance. Consider the market need and what type of digital diagnostics will best meet that need, and whether there is a path in which you get a potentially greater ROI, for example by pursuing a reimbursable diagnostic tool. Also, keep in mind that the results of a strategic analysis might reveal that your best option is not to proceed. Time, opportunity cost, investment (clinical and otherwise) and risk – without guaranteed income – can be excessive.
3. Can you make the work of clinicians easier?
Some companies aim to take advantage of artificial intelligence (AI) technologies to automate clinical work. As physicians are typically the end customer, the prospect of automating their work is sensitive and is likely to introduce additional barriers to adoption, barriers to commercialization, and legal considerations. In contrast, it stands to reason that a digital health product that complements clinical decision making, leading to improved utility and outcomes, will gain popularity (and increase the chances of reimbursement). As you develop your digital diagnostic strategy, consider how the product can help improve a physician’s practice and / or improve clinical outcomes.
4. Is your development process FDA ready?
Traditionally, software developers have been accustomed to working with agility through development sprints, with the goal of getting a minimum viable product (MVP) to market as quickly as possible, and knowing that they can make adjustments, fix bugs and implement upgrades in later versions if necessary. While developers typically verify and validate these products along the way, the level of rigor required is usually much less intense than for a high-risk Class II or Class III medical product that has a direct impact on the lives of patients. .
Navigating your way through the FDA’s strict regulatory approval process requires both a thorough understanding of compliance as well as personal patience. This means following regulatory design controls: maintaining detailed documentation throughout the development process, as well as throughout the rigorous testing, evaluation and validation processes throughout a product’s lifecycle to ‘to obsolescence. And, while products can be updated and improved upon after approval, the initial FDA-approved version must be proven to fulfill its clinical purpose and perform at an exceptionally high level.
Also, while all machine learning (ML) model developers should perform some form of model validation, the due diligence threshold is much higher for medical products. If the validation is done incorrectly, the results of the clinical trial will reveal that it does not perform reliably in a real environment and you will return to research and development. In technical terms, this means using best practice validation, such as nested cross-validation, which is appropriate given the size of the sample, the patient, and the characteristics of the disease. It also means throughout the validation process to avoid “target leaks” – that is, the subtle validation errors that will cause a model to appear effective during the development process but ultimately fail when it does. will be deployed in the real world. This is essential for the results of clinical trials to provide the expected evidence.
Prepare for digital health success
It all comes down to matching the digital health tool strategy to your business case and market opportunity. For many companies, it will be a question of creating a complementary and strategic digital diagnosis of a therapeutic asset. But without doing your due diligence to determine your best option, you might miss opportunities to monetize digital health innovation, or at least ensure optionity throughout the product development cycle. To introduce valuable digital health technology, it is essential to consider and understand the needs of your target market and your therapeutic landscape. This includes understanding the physician’s point of view, the incentives and barriers to adoption. It also means considering creating a digital diagnostic that truly adds value for the patients and clinicians who rely on you for treatment options, even if it means a more intense development process and under regulatory approval.
Bill Woywod is associate director of health; Jim Williams is associate director in life sciences; and Jthatob Graham is partner of the Life Sciences practice, all at Guidehouse.