Today’s industries are being disrupted by software-driven innovations, as evidenced by Amazon’s impact in retail markets, Netflix’s success over traditional cable providers, and the upheaval in hospitality industry from online marketplaces such as Airbnb. It is hence no surprise that the healthcare industry is seeing a surge in the development of software driven products to drive new healthcare services or augment and complement existing ones.
One of the most notable trends is the area of digital therapeutics, otherwise known as “DTx”. Digital therapeutics, as defined by the Digital Therapeutics Alliance, deliver software-generated interventions directly to patients to prevent, manage, or treat a medical disorder or disease [1]. DTx, like other therapeutic interventions, must also demonstrate robust, measurable clinical outcomes, coverage by payers, and regulatory approval.
In a recent paper, published by the IBM Institute for Business Value, we interviewed three digital health professionals from UCB, Novartis, and Roche for their perspectives on the topic. We also conducted a survey of fifty professionals from across the healthcare landscape for their perceptions and familiarity with the topic of DTx to better understand the potential and challenges of software delivered therapeutic interventions [2]. The survey results show very different awareness levels among healthcare and life science professionals. Where 93% of all Life Sciences professionals stated to have a basic awareness of the topic, only 57% of all Healthcare professionals can claim the same. [Figure 1].
Figure 1: “Have you come across the term digital therapeutics”?
IBM mini pulse survey of healthcare and life sciences professionals in Europe and the US
In addition to Life Sciences companies, investors are clearly attracted to the topic having poured $12.5 billion into digital health ventures in 2018 [3]. Compared to 2013 this constitutes an increase of 250% where the average deal size grew by 67% [4]. Our survey respondents also reflect optimism that DTx will positively impact healthcare through increased benefits of personalization and the opportunity for greater patient engagement. [Figure 2]
Figure 2: “What main benefits do you think will be realized by digital therapeutics?”
IBM mini pulse survey of healthcare and life sciences professionals in Europe and the US.
Digital therapy is not a direct replacement for pharmaceutical intervention, or meant to provide a placebo effect, but maybe a useful adjunct to traditional therapy. However, digital platforms like mobile apps can help encourage and enhance behavioral change with improved access to, engagement with, and quality control of a therapy. Doctors might include a prescription for, say, a mobile app in addition to medicine [5].
As computational costs continue to decrease, better access to healthcare data for analysis and improved sensor technologies will likely drive further adoption of standalone DTx and opportunities to augment existing therapies. It is hence no surprise that the healthcare industry sees the largest potential of DTx in the coming years in chronic disease management and data gathering [Figure 3].
Figure 3: “For which application area do you see the largest potential for digital therapies?”
IBM mini pulse survey of healthcare and life sciences professionals in Europe and the US.
Consider the management of a metabolic disease, such as diabetes as an example. To be able to replace the function of a pancreas, sensing technologies are needed to quantify food intake, measure metabolic activity and continuously monitor blood sugar levels. Technologies are required to contextualize and analyze data in real-time and on an individual patient level to allow for timely and accurate insulin dosage advice. Diabetes constitutes a typical example of DTx augmentation as it directly affects the insulin therapy. The development of such products is complex and typically requires clinical trials to collect an initial set of data to start the development of the algorithms. Like most AI-powered services DTx algorithms will improve over time. They follow the “virtuous cycle of AI” which means that initial services will attract end-users, which will generate new data and can be used to improve the existing services or create new ones [6].
Unlike today’s search engines, DTx cannot afford to provide the patient with bad recommendations at any point of the product life cycle. This poses a large conundrum as real-world experience or “evidence” is essential for the development of such products.
Clinical trials provide strongly controlled settings, patients naturally feel observed leading to biased behavior. As a result, they only provide the algorithms with a part of the truth. In a recently published study, Roche Diabetes care and IBM successfully illustrate the limitations of predictive algorithms for the chronic kidney disease developed on clinical trial data [7]. The Roche/IBM algorithm which was trained on real-world data strongly outperformed algorithms that were developed on clinical trial data in a clinical as well as a real-world setting.
To create an environment that supports the development of AI-powered medical products along with ensuring the patients’ safety, regulators and software-as-a- medical device providers need to jointly progress the current regulatory landscape. For the very first time, the Food and Drug Administration (FDA) and European Medicines Agency (EMA) are confronted with medical products that continuously evolve over time. The FDA has already reacted to this changing digital landscape by introducing the Digital Health Software Precertification (Pre-Cert) Program with a strong emphasis on real-world performance and a framework that allows the authority to react quickly to changing products [8].
Despite these promising advances of the authorities, regulatory concerns still rank highly among most healthcare professionals when asked about the largest risk of Digital Therapeutics adoption. Responders of the survey agree with us that we have reached technological maturity to bring these services to bare.
Figure 4: “What are the greatest challenges for the adoption of digital therapies?”
IBM mini pulse survey of healthcare and life sciences professionals in Europe and the US.
Every new approach and innovation involve risks. A clear focus, close collaboration, and partnership between technology providers, pharmaceuticals and regulating authorities will hopefully reveal the unprecedented potential of digital therapeutics.
Sources
1. “Digital therapeutics: Combining Technology and evidence-based medicine to transform personalized patient care.” Digital Therapeutics Alliance. October 2018. https://www.dtxalliance.org/wp-content/uploads/2018/09/DTA-Report_DTx-Industry-Foundations.pdf
2. “How digital therapy benefits patients, providers, and the health ecosystem: Evidence-based, personalized treatment can increase engagement and improve outcomes.” IBM Institute for Business Value, April 2019. https://ibm.co/digital-therapy
3. Megan Zweig, Halle Tecco and Michelle Huang, “2018 Midyear Funding Review: Digital Health Déjà Vu in Yet Another Record Breaking Half.” Rock Health, accessed June 24th, 2019, https://rockhealth.com/reports/2018-midyear-funding-review-digital-health-deja-vu-in-yet-another-record-breaking-half/
4. Ibid
5. “How digital therapy benefits patients, providers, and the health ecosystem: Evidence-based, personalized treatment can increase engagement and improve outcomes.” IBM Institute for Business Value, April 2019. https://ibm.co/digital-therapy
6. Andrew Ng “AI for everyone – The definitive starting block for AI novices.” MAIEI. March 2019. https://montrealethics.ai/andrew-ngs-ai-for-everyone-the-definitive-starting-block-for-ai-novices/
7. Stefan Ravizza et al. “Predicting the early risk of chronic kidney disease in patients with diabetes using real-world data.” Nature Medicine. January 2019. https://www.nature.com/articles/s41591-018-0239-8
8. “Digital Health Software Precertification (Pre-Cert) Program.” FDA, May 2019. https://www.fda.gov/medical-devices/digital-health/digital-health-software-precertification-pre-cert-program