Digital Therapeutics: a bumpy road to commercialisation


29th May 2020

Timing is just about right: COVID-19 does not only act as a catalyst for telemedicine and triage solutions, but also for Digital Therapeutics (hereafter DTx) solutions as their practicability is immune to current COVID-19 related restrictions. Recent spikes in usage of different DTx solutions across geographies and indications prove this.

In the aftermath of COVID-19, any measures boosting human resilience to such external health shocks will increase in significance. One personal example: poorly managed Type 1 diabetics with consistently elevated blood sugar levels are at higher risk of developing serious complications from COVID-19 (Link). The same applies to adjacent indications such as Type 2 diabetes and obesity. Patients with often preventable conditions such as Type 2 diabetes and with poorly managed chronic diseases such as Type 1 diabetes are more prone to an adverse course of disease. This highlights the essence of chronic disease management, reversal (if possible) or prevention in the first place.

Before we kick things off, let’s provide some clarity about the actual meaning of Digital Therapeutics. According to the Digital Therapeutics Alliance (Link), DTx:

  • Increase remote access to therapies that are clinically demonstrated as safe and effective.
  • Provide care independent of a patient’s schedule and in the privacy and safety of their home environment.
  • Are easily scalable and often accessible through patient-owned devices (e.g., smartphones, tablets).
  • Generate actionable, real world insights that enable intelligent data-driven care management and clinical decision making.

While this definition still leaves some room for interpretation, it is narrower than most of us think. The majority of the 45k apps categorised as “Digital Health” in the iOS App store does not meet the specifications of a DTx, mostly because they cannot point to Randomized Controlled Trials to back up their claims and perform in a real-world environment, demonstrating efficacy.

From an investor perspective, the inherent question marks in the fields of DTx seem to revolve around a bumpy road to monetisation. At btov, we have first invested in the field back in 2012, into a company that nowadays offers mental health DTx called Happify. By the time the global Sanofi partnership was revealed in 2019, Happify seemed to have reached the Champions League of DTx. As so often with superb companies, it took a long way to get there, including numerous business model adaptions — initially focusing on B2C to form the foundation to move into B2B (corporate health management), B2G (health plans) and pharma partnerships. In fact, at the time we first invested Happify had not met all of the DTx classifications.

Setting expectations, DTx companies don’t tend to be SaaS companies generating revenues through paid pilots or co-development agreements from day one. Yes, country-wide regulatory liberalisation such as the Digital Care Act (“Digitale Versorgung Gesetz”) in Germany will shorten this commercialisation path, but I would argue that monetisation is often an unfair and unsuitable KPI to evaluate a DTx company, at least at an early stage, as early monetisation doesn’t necessarily serve as a good predictor of sustainable commercial success.

Instead, the fundamental “fuel” of a working business model in the digital health space is a strong practical value proposition (PVP), as opposed to a mere theoretical value proposition (TVP). We often see two types of companies:

  • Strong theoretical, often clinically validated value propositions but no PVP: Such solutions often do not consider all involved stakeholders, for example, clinicians. If the prescription of and treatment with DTx leads to a higher workload, they are unlikely to prescribe it (remember, in Germany they soon represent a powerful sales channel). Note that greater patient-doctor interaction does not always correlate with doctor value-add. Also, solutions delivering the science in a controlled environment do not suffice. Example: For Diabetes diaries to provide value-adding actionable insights they require comprehensive data entry, including blood sugar levels, physical activity, carbs consumption and insulin intake — all this being an absolute nightmare (I have tried most of them myself). Therefore, the more passive data generation is, the better. The TVP, assuming spot-on data entry, is high. The (real-life) PVP less so. If this practicability aspect is violated, user stickiness and retention will suffer.
  • Great UI/UX but insufficient TVP and PVP: Despite a carefully designed product integrated into the user’s daily routine, a sufficient medical value proposition is missing in the first place. This often leads to notable initial usage and early retention, however, diminishing significantly over time.

Assuming you have a PVP, attractive engagement metrics such as stickiness and retention will follow over time, providing the foundation for longitudinal datasets on a large scale. Longitudinal datasets then again are crucial for the full quantification and clinical validation (if applicable) of the PVP. We call this chain reaction initiated by a strong PVP the “Monetisable State”.

Monetisable State:

PVP → stickiness & retention → longitudinal datasets → quantified/ clinically validated PVP

The quicker this state is obtained, the shorter the road to reimbursement — the gateway to monetisation paradise.

The ultimate component of the Monetisable State is the quantified and/or clinically validated PVP. The clinical proof requirements depend on the medical device classification. As Digital Health Applications (“DiGa’s” according to the German Digital Care Act) generally bear lower risk and are less costly, there is often no need for a genuine clinical trial. Instead, companies have to conduct a comparative study, including case reports, expert opinions and comparable evidence, to prove clinical efficacy.

There are two ways to conduct large scale data collection facilitating the quantification and (clinical) validation of value proposition. While both of them act as enablers for reimbursement, they often simultaneously demonstrate a viable business model even before qualifying for reimbursement, assuming that a PVP has been obtained.

1) B2C: Generating a willingness to pay for DTx on an end-user level is deemed challenging, especially in Europe where users are not used to paying for health-related services. However, assuming that a PVP is obtained, why shouldn’t they pay as much as other personal wellbeing enhancing solutions?

2) B2B2C (Corporate Health Management): B2B2C models provide an attractive gateway to larger user groups, because they act as an effective way of “outsourcing” sales and adoption. The inherent alignment of interests for all stakeholders (employer, employee and health insurances, if applicable) contributes to the viability of this approach. Ideally, the pricing for B2B2C models is based on a combination of usage and fixed pricing, incentivising employers to assure usage. However, low usage rates due to stigmatisation, i.e. employees not willing to reveal their health issues to their employers, often lead to challenging unit economics. Different user progress tracking initiatives often including a gamification element followed by an intact reward system have proven to be successful.

This quantification process is deemed rather challenging because feedback cycles on sufficient positive effects in patient care are often very long. Also, note that real world data is often regarded as noisy data, especially with small sample sizes. Early adopters tend to have a higher propensity to use the product in the right manner, leading to a sample bias.

Putting the different components of commercialisation together, we obtain a so-called “Road to DTx monetisation”:

  • Not all, but many roads lead to Rome, Rome being guaranteed commercialisation. The ones leading to Rome must include the PVP. Reimbursement guaranteeing commercialisation is conditioned to an intact PVP. Quantification and clinical validation of value proposition without having obtained a PVP will not qualify for guaranteed commercialisation.
  • An excessive focus on clinical trials often detracts attention from the PVP. A controlled environment does not equate to real-life. Controlled trial participants mostly have a higher propensity to use the product in the right manner, leading to a sample bias. While real-life data is often perceived as noisy data (and yes, to some extent it is), we deem real life data more relevant in obtaining a PVP in the first place. The PVP shall never be compromised. Otherwise, commercialisation is at risk.
  • The road to Rome (and in the first instance to PVP) can be bumpy and long. Short-notice detours and unexpected stops are part of the game. The acceleration from zero to > 120 km/h is often a matter of years — ask our friends from Happify. We as an early-stage investor do not expect founders to have solved this complicated commercialisation puzzle at initial investment, however, a destination and a vehicle capable of navigating to this destination need to be visible.

Looking forward, we believe that multimodal solutions consisting of a Diagnostics and Therapeutics component will become more common. This combination will eventually form the basis for a fully integrated and customised therapy. However, in practice, therapeutics solutions do not fall under the umbrella of the Digital Care Act (as they are classified as risk class III opposed to IIa/b). Thus, no regulatory frameworks (neither Germany or the US) allow for full reimbursement, leading to an obligatory (partial) self-payment. If prerequisites of revenue generation are met (remember the Monetisable State), it is only a matter of time until reimbursement and regulatory bodies follow with approval.

In light of the inherent complexity of DTx commercialisation, we are particularly interested in your (contrary) views. The more feedback the better. Either reach out to me here on Medium or through

This blogpost was originally published on Medium on May 29th, 2020.

Press enter or esc to cancel