Verified by 
Everyone wants to make money but no one wants to talk about it. The kind of research effort that pushes science toward practice is called translational research, and has many components. One component is commercialization, and commercialization is associated with impact: in order for something to save lives, it has to be used. This is an interesting problem. Looking at technology and science from the side of commercialization is a mind-twisting exercise. Recently my team and I have come up with the idea of X. We tested X. It seemed to have worked in the lab. I was immediately advised by a senior entrepreneur to not publish it or even give talks about it.
Given how much anxiety and sweat and blood go into fighting to publish first to avoid getting scooped in science, I find the whole idea of avoiding publication hilarious. I drafted an “Invention Disclosure” and the technology transfer office returned with an equally corporate “Marketing Strategy Report”. We met with the respectable lawyers and technology transfer officers.
If this all sounds awfully dry, it’s because it really is. Protecting intellectual property is a tedious, technical and long process. Traditionally, technologies that satisfy a set of complex formula are eligible for patenting. The patent office processes applications filed on behalf of the inventors, evaluates them and issues patents. Disagreements of patentability (especially with regard to “non-obviousness”) are common, and a complex set of rules governs how appeals can be made. While the lawyer explained the patent review process, my mind drifted off to the image of the young Albert Einstein walking down a dusty, dark corridor looking haplessly through a pile of German index cards.
Even if you are granted a patent, this only provides a legal basis for you to sue someone for selling the same product, not any promise of cash per se. A patenting decision is therefore a commercialization decision, not a legal one: is the amount of money that will be made by the patent litigation worth the cost and risk of carrying through the patent process and subsequent litigation?
From this criterion, it certainly seems X is not so patentable.
What about writing an algorithm that uses the data source X to make diagnoses or find an optimal treatment? This is the holy grail of Precision Medicine. It turns out the best way to commercialize an algorithm is not to patent it. The reason is how slow patenting is. By the time patents are granted the algorithms may already be out of date. This is one reason why large companies such as Google rarely patent their algorithms. Algorithms are “licensed”, not “patented”. One writes an algorithm so someone else puts it into their device and uses it. You charge per use, kind of like how Microsoft charges a license fee for putting your Windows 2000 Home Edition (my analogies date me) into your computer. So again the IP question becomes a commercialization question: who is going to license your technology and why would your algorithm add value to their product?
Let’s summarize: Precision Medicine, make me money! In a hypothetical scenario, we invented a Technology X and an algorithm Y, which yielded an optimal treatment Z. X + Y is not worth much commercially if getting that Z is not going improve outcome that justifies the cost.
Using this simple criterion, we see that expensive imaging modalities such as positron emission tomography (PET) have very little commercialization value unless they can be made much cheaper. Personalize antidepressants in a garden variety depression case also may not be worth it: first off, up to 80% of patients respond to any first line antidepressant; secondly, choosing an antidepressant is often driven not by efficacy, but by other factors such as preference and side effects. The value of X + Y that tells you citalopram will work 50% of the time, and sertraline will work 80%, may be not that high, and certainly would require extensive cost-benefit calculations prior to manufacturing and marketing.
In order for Precision Medicine to be worth it, Z should be a treatment decision that is expensive, high stakes, common, and effective in a potentially identifiable subset of patients. Expensive new medications, such as biologics, are one possibility. Expensive equipment and procedure (i.e. TMS, ECT) may be another. In substance abuse treatment, Vivitrol injection, which costs thousands, may very well be a Z that’s worth personalizing. Another is trying to identify those who would truly benefit from an inpatient detox-rehab.
I ask readers to suggest ideas (Z) in mental health and substance abuse that would make money. I will figure out the X+Y and we can start a company together.
