Things I Didn’t Know About Clinical Innovation: My Encounter with Biodesign

So I’ve been busy this fall taking a course in ”Clinical Innovation and Design”, because the type of tech I’m working with often lends itself to healthcare related applications, and I feel I need to learn more about the complexities and perils of launching products into this domain.

What follows are some scattered thoughts and observations which were all sparked by attending some thirteen seminars and reading the 900+ pages brick of a book Biodesign : The Process of Innovating Medical Technologies.

Economics

Rapid expansion of middle class in India, China and Brazil has meant that sales of medical devices grow two to five times faster in those countries, compared to ‘the developed world’.
The compound annual growth rate of the medtech industry in India is expected to be 14 percent. That number in China: 26 percent. (Interesting though: among the top ten Chinese medtech players, seven are foreign companies or joint ventures.)
The notion of value is on its way to shift. Medtech companies traditionally pursued ‘optimal improvement’, regardless of its costs. More and more now the focus is to drive down costs instead. Sometimes while boosting quality, but sometimes also while accepting that quality takes a hit.
Healthcare will always, by definition, be full of unmet needs. As an innovator, the problem isn’t to find a need, but to find one where there’s an “opportunity ripe for value realization”. Telltale signs can include patters such as extreme variations in treatment costs across geographies. Economists will play just as important a role here, as technologists.
In this report, PWC predicts that the US will remain leaders of the medtech space, but that the ground is shifting to the advantage of regions with more relaxed regulatory frameworks. Europe and Israel is mentioned, but also countries like Brazil, India and China.
60 percent of the European medtech industry is concentrated to Germany, France, Italy and the UK, but Sweden belongs to a club of countries with a positive medtech trade balance, meaning it exports more than it imports.
The cost of healthcare as compared to GDP is ballooning everywhere in the world, but nowhere is the problem as apparent as in the US. Accounting for 7.2 percent of the Gross National Product in 1970, it has now grown to around 18 percent. That’s nearly twice as much as the average OECD country (the equivalent metric in Sweden is around 11 percent. Spain hovers just above 7 percent).
The US in particular stands out for getting poor value per dollar spent on healthcare. This has been confirmed in a number of major studies, one of which was conducted by WHO in 2000, placing US on 37th place on a list of 191 countries.
France is considered to give its citizens the same quality of healthcare as the US, at half the price per capita.
About half of the spending growth in healthcare (some say 1/3), can be derived from the introduction of new technologies. Much of this shiny tech has only marginal impact on patient healths. The reason it’s still being deployed, is a near complete lobotomy between therapeutic and economic decision-making. This is pretty much true for all countries, but the US present an extreme example.
Looking at entrepreneurship through the lens of medtech innovation makes it *so* clear, that not all problems are appropriate to address with technology. In other words: just because you’ve identified a need, doesn’t mean you’re ready to start a company. That’s why you want to pause early in the process, zoom out a bit, and take the WOMBAT test, short for “is this a Waste Of Money, Brains And Time? The answer very well might be yes. In fact I’d go so far as to hypothesise that solving a problem which turns out to be trivial—nobody’s real pain— is probably the leading reason that startups fail.

Regulatory

It’s generally less complicated to achieve CE-marking than it is to get FDA approval. This explains why many medtech companies chose to first launch their products in the EU, sometimes years before the same products gain traction in the US. (Although Swedish medtech investor Bengt Juhlander says that’s no longer the case, in this great interview).
The FDA oversees products that account for roughly 25 percent of all consumer spending in the US.
The Federal Food and Drug Act was passed in 1906. It was replaced in 1936 by the Food Drug & Cosmetics Act. Devices remain pretty much unregulated until this law was amended in 1976 by the Medical Device Amendment Act.
The FDA is tasked with “Advancing the public health by helping to speed up innovation“.
Cost-effectiveness plays no part in the FDA assessment (which sets it apart starkly from the Swedish equivalent process).
A device’s “regulatory pathway” depends largely on which of three risk-classes it belongs to, although there’s not a one-on-one correspondence between class and pathway. It’s complicated.
The pathways through FDA breaks down as so: Device Exemption, 510(k) and Premarket approval (PMA).
A devise that passes FDA scrutiny by way of 510(k) is said to have been “cleared”, whereas a Class III device that follows the PMA pathway is “approved”.
The company making a submission for either a 510(k) or a PMA is referred to as a “sponsor”.
This list of factual tidbits could go on to fill a book, but it gives me a bit of head ache so let’s move on.

Methodology / entrepreneurship

Biodesign is the name of a methodology developed at Stanford, tailored for introducing technology based innovation into the healthcare sector. It’s sprung out of Design Thinking, but is far more rigorous.
According to Biodesign, it’s essential to study and define a certain need in isolation from possible solutions. At the most fundamental level, a need statement should address what change in outcome is required to resolve a stated problem, not how the problem will be addressed. This mindset reminds me of Behaviour Driven Design (/ Domain Driven Design), which is a methodology for organising software projects.
The most common problem with early drafts of a need statement—except for mixing in solutions—is that it’s either too broadly or too narrowly defined. A broad need statement is likely to give false positives during validation, while a narrow one will result in a customer segment that isn’t big enough to sustain a viable business.
The healthcare sector is highly specialised with each profession gaining an ever deeper expertise in their narrowly defined field. This is inevitable, the same tendency can be observed in engineering. But it also presents a challenge in terms of launching innovations, which often requires a broad understanding spanning many different disciplines.
Using stents instead of surgery to mitigate blood clots is a good example of this: an important innovation, where nevertheless the ‘need’ was initially hard to validate simply because the solution didn’t comply with the way specialists divided up their turfs between themselves. Diseases don’t read textbooks.
One way in which Biodesign’s heritage from Design Thinking is apparent, is its stance towards needs exploration. There’s *one* good way of doing it, and that way is modelled on methodology adopted from the field of anthropology. The budding entrepreneur is supposed to go into the field with a blank slate, keeping prejudices at bay and acting as a fly on the wall observer. I don’t agree with this stance. I argue my case in this post.
To their credit, the Big Bible of Biodesign includes some examples of heretics. One of them describes his need exploration phase like so:

“I tend not to just go and try to sit down with docs and just ask them a bunch of questions. Personally, I find it more useful to wait until I am very conversant about the anatomy and physiology that we’re talking about. Otherwise, we can’t really have a discussion, and I can’t challenge them. And a huge amount of us being successful at what we do is being able to challenge the docs on the dogma of the clinical practice in the area that we’re in. We say, ‘Why is it that way?’ and ‘Couldn’t it be this way?’ And you can’t do that unless you understand the system you are questioning. Before I reach out for interviews, I read everything I can.”
Mark Deem, interviewed in Biodesign : The Process of Innovating Medical Technologies.

What’s true for living organisms is also true for companies: there are a lot more dead than living ones. People tend to forget that when they study the competitive landscape. It makes a lot of sense to study the initiatives that tried to break into a certain area, but failed. What are their lessons learnt?
Stakeholder analysis can get particularly tricky in healthcare, where example of local optimisation abound. Take nephrology, for example. Nephrologists—kidney experts—”lose” patients who undergo kidney transplants. An article in the New England Journal of Medicine confirms that patients at (super lucrative) for-profit dialysis clinics were less likely to put on waiting lists for renal transplants, compared to patients at non-profit clinics, leading to higher mortality for paying patients. (Source: Effect of the Ownership of Dialysis Facilities on Patients’ Survival and Referral for Transplantation)
Language matters. Practically all innovation is about getting a cost-benefit tradeoff right, but it’s hard to sell a product or a service that “might not be as good as the gold standard, but at least it’s cheap”. You’re going to have to get creative and find other ways of presenting your value proposition. There’s an example in the book where someone develops a test which isn’t as exact as sending your blood sample to a remote laboratory, but has the advantage of being cheap and easy to perform in the field. The product is meant to be launched in poor rural areas of Africa where resources are strapped, so the value proposition was phrased to mirror those conditions:

“We framed the tradeoff not as between perfect and imperfect information, but instead between no information and really good information.”
Bill Rodriguez, founder of Daktari, quoted from Biodesign : The Process of Innovating Medical Technologies.

Speaking of tradeoffs: is being a small fish in a big pond preferable to dominating a narrow niche? This is the perennial question that every entrepreneur must ask themselves, and like a zen koan there will never be a definitive answer, as the following contrasting quotes from two very successful startup founders will attest:

“The need has to have a very large market… There’s only so much time in the day, and it’s really just as easy to develop a solution for a large market as it is for a small market”
Richard Stark, president of Silicon Valley medical device incubator Synecor

vs:

“I’m a big fan of niches – of not trying to take on an entire market at once – because there’s less resistance to innovation when you do it on a smaller scale.”
John Abele, co-founder of Boston Scientific

Virtually any product or service can be improved by addressing its specific shortcomings. Incremental improvement opportunities are plentiful. Yet once a stakeholder’s basic needs are met, they tend to be resistant to upgrading or accepting new features that incrementally enhance the performances of an offering until those improvements achieve a threshold of innovation meaningful enough to motivate a behaviour change. Most solutions fail because the innovators underestimate this threshold within their target market.
Or otherwise put: assessing how much value (in terms of improved performance or lowered cost) is necessary in order to differentiate a new product, is key.
Apart from underestimating this, the other most common mistake entrepreneurs tend to make is to overestimate how big of a market share they can realistically capture in the early years. As a rule of thumb: don’t expect to gain more than one percent market share in the first year, and no more than an eventual maximum of 15 – 30 percent.
Also, while still on the topic of market analysis: never underestimate what products your competitors have in their R&D pipelines. Those products are what you’re really up against, not what’s on the market today.
Many companies in the healthcare domain strive for a blended product/service model and may frequently sell a product at or near cost in order to generate service revenue, where the margins can be high and the revenue is recurring. From the customer point of view, a model like this usually leads to a strong lock-in effect. (this makes me think of the printing industry, where the machines themselves are typically leased at affordable prices, but the manufacturer’s make their margins on the so called click fee, which is based on the number of times per time period in which the machine is operated).
Treat business modelling as another design challenge, a parameter to tweak and perfect. In doing so, constantly ask yourself: Could a competitor deliver the same solution with a different model, and if so, would that be a threat?

Design & prototyping

Break your concept down into smaller blocks corresponding to its different functions, then build prototypes for each of those blocks.
Make a conscious decision as to whether you should develop said prototypes sequentially or in parallell.
There aren’t just one type of prototype. Think of them in terms of the job-to-be done and you get, at least: Works-like models, Feels-like models, Is-like models, Looks-like models and Looks-like-feels-like-models.
If you try to build a prototype that does all of the above, it will slow you down
Your speed of iteration is probably one of the single most important metrics to keep track of.
Don’t just think in terms of products. It’s just as important to prototype workflows.
While I personally tend to see design as an art, there are many proponents of a more scientific view. One hallmark name here is Stuart Pugh, who suggested a structured framework for requirements- and concept selection in his 1981 classic Total Design. The framework rubs me the wrong way, but having said that I do actually think there’s a lot to be gained by taking a (more) disciplined approach to design. Particularly of course if you’re designing devises that can turn out to make life-or-death differences.

Intellectual Property Rights

First of all, on the note of ‘language matter’, I think the following is a very eloquent form of defining IP:

“Any product of the human mind which has some value in the marketplace and, ultimately, can be reduced to tangible form.”

I never really stopped to think about this, but it actually makes sense that there are two reasons why governments are willing to grant exclusivity by means of patent. There’s the obvious one, which is to incentivise innovators to invest time and effort into bringing an innovation to the market, but there’s also the less obvious one, which is to incentivise all other players to get creative and try to design superior alternatives.
The most common type of patent sought for medical devices in the US, are so called utility patents, meant to protect “an apparatus or method wherein the novel invention is useful for accomplishing a specific end result“.
There there are also design patents, meant to protect “the unique ornamental, visible shape or design of non-naturally occurring objects.”
When our book went into print 2015, there were some eight million patents granted in the world. Now eight years later that number has more than doubled!
Here’s a great resource from WIPO, keeping track of lots of useful IPR-related metrics.
The America Invents Act, which was passed 2011, fundamentally changed US IP law from a “first to invent” to a “first to file” strategy, harmonising it in this respect with much of the rest of the world.
This I already knew: when filing a patent the inventor must “describe the invention in a clear, unambiguous and definite way so that a person in the field would be able to make and use it“. But this I didn’t know: the inventor is also “required to include a description of the best mode for the invention“, meaning the most appropriate embodiment or way of practicing, in the opinion of the inventor at the time of filing. So for example, if you’ve come up with a certain type of stent, and you know that stainless steel is the optimal material, then you’re required to state that. (But: If at a later point experience shows that cobalt-chromium is superior to stainless steel, then that’s also ok, it won’t invalidate the patent).
I also didn’t know about the “reduction to practice” requirement. It means you have to show that “the idea is being diligently pursued“, or to put it in other words: If you have no clue how to implement your own invention, you ultimately won’t be able to patent it.
PCT is short for (the) Patent Cooperation Treaty. It’s what allows you to file once and gain protection in (almost) the entire world.
You can and should be smart with filing. There are both tactical and strategic aspects to IPR and the required expertise aught not to be underestimated. Just as an example: this article outlines pros and cons with three different approaches to provisional patent filings.
The non-provisional patent application must have at least one inventor in common with the inventors named on the provisional application, in order to claim benefit of that application’s filing date.
Style-wise a patent application needs to be “intelligible to a bright reader who is not necessarily versed in the field.” Think in terms of someone who subscribes to Scientific American.
It’s a common misconception that the provisional patent application process was developed to enable inventors to file “quick and dirty” applications. In fact, the process was put into place to give inventors in the US the same rights as inventors overseas with respect to patent term. The 20-year patent term is based on the date of a non-provisional patent application is filed. However, inventors outsider the US have one year from their ex-US filing date to file a patent application covering the same subject matter in the US.
Think of the claims in a patent application lika a deed to a piece of real estate – they specifically describe the boundaries of what is owned.
There are two types; dependent and independent claims.
When listing their claims, inventors often emphasise structural elements, while failing to describe their invention from a value-first point of view. Rather than claim such as “A devise comprising X, Y and Z“, try something like: “A devise for use in X procedure wherein the devise may be fully operated with one hand…” or “A devise used in X procedure wherein the devise enables simultaneous deployments of elements…” This way, the invention will likely be protected in a way that makes it more difficult for competitors to copy it, and the inventor is also free to improve upon the device as it evolves over time, while still staying within the protected scope.
In 2011, China overtook the US in terms of numbers of patent applications filed.
There are three types of patents in China; invention patents, utility model patents and design patents.
Confusingly, a Chinese “invention patent” is the closest to what’s known in the US as a “utility patent”.
Foreign companies active in China tend to file for invention patents, while the majority of domestic companies and individuals file for utility model– and design patents.
The Chinese government incentivise people to file for patents by giving them tax-breaks. Many feel this has led to a quantity-over-quality bloat.
One of the best assets you can have if you end up defending your patent in litigation, is an innovation notebook where you keep track of your process. Much like in book-keeping, there are a number of requirements you have to stick with in order for the entries in such a notebook to be legally defensible.
The default mode is that IP generated by academics belong to the university where they’re employed. When it comes to students, policies vary across institutions.
Sweden is the exception in this regard. Since the 1950’s there’s a law that says researchers own the IP they generate. (Which in my humble opinion explains why this country has become such a fenomenal producer of innovative startups).
Be aware that even casual conversations with colleagues are considered legally as a form of disclosure.
So called “confidential” editorial review processes are—at best—a legal grey zone from an IPR point of view, so be sure to file a provisional patent application before submitting a manuscript.
Conversations with IP attorney’s are confidential whether there’s an NDA or not, thanks to the attorney-client privilege.
VC firms will routinely abstain from signing NDA’s. All the more reason to clearly mark all material shared with them as “confidential”.

Stuff I didn’t even get to

As any reader who made it this far might be able to tell, I turned out to be embarking on quite some journey when I committed myself to take this course. I’ve certainly learnt a lot, although perhaps not mainly in the fields I thought I would. More importantly however, I’ve gained an appreciation of all that I don’t know.

IP strategy is one of those fields and regulatory is another. I usually take the approach that efficient entrepreneurship requires just barely good enough domain knowledge, since if you knew what you will be up against you most likely would never take a stab at it. I still feel that way, but I now realise that when it comes to clinical innovation, regulatory and IP are exceptions that prove the rule. Furthermore, I have a feeling that these two areas are so vast and complicated that it’ll probably always be worth it to find people who are experts, rather than trying to become one myself. In other words, these two blind spots have transitioned from being unknown unknowns to known unknowns.

Here’s another important take-away: I’ve previously felt that it’s impossible or even counter-productive to try and teach someone entrepreneurship. To a very large extent, I still feel this way: It seems to me that people who try to enter the startup world by means of taking a course in entrepreneurship are at a disadvantage.

Again however, I think I’ve found an exception to the rule. Because taking a course like this after already having founded three companies, has undeniably been quite enriching. I’ve often felt like a street fighting urchin who’s put in a karate class; instinctively rebellious against any meddling with what’s already battle-tested. But again and again, I’ve had to accept that going back to first principles and carefully building up a new set of habits, let alone conceptual framework, has been worth the effort.

I think probably there are a lot of areas outside of deeptech in general and medtech in particular where it’s still optimal to go with the instinctual whatever-works approach to entrepreneurship, but those areas have now transitioned in my mind from being the default, to becoming the ‘exception to the exception’.

What seems to be the problem, doc?

If there’s *one* insight that I prize above others, it has to do with compartmentalisation. The Biodesign process does this religiously; every step is delineated and has its own succinctly defined patterns. Again, I used to think that this view is incompatible with actual entrepreneurship. Again, I still think so, but what I now realise is that just because reality is continuous and messy doesn’t mean that has to be reflected in your approach.

Engineers love to make “Root cause analysis” to figure out what initially caused something to go wrong, but the same line of reasoning can also be applied to understand success factors. If I’m trying my hand on that, I’d end up with the importance of isolating an observed problem from the infinite possible solutions that might address that problem. The Biodesign process helps you to do this in a very rigorous way, and I feel it’s probably at the root of its many benefits.

One last thing; it’s good to have company

Last but not least, I’ve rediscovered how encouraging and creative it can be to be surrounded by people—fellow students as well as brilliant teachers—who are on a similar journey. It’s been a privilege to share paths with all of you!