Can you patent an algorithm?
Can you patent an algorithm? If you assume the answer is no, you could well be missing out on commercially valuable protection
How often have you heard people say “You can’t patent software” or “It is impossible to get patents for algorithms”? I hear it all the time, often from people who should know better!
The reality is that there are many thousands of patents granted in Europe, the US and elsewhere for inventions based on new algorithms (or even new uses of old algorithms), including many in the medtech field. If you listen to the naysayers you are potentially missing out on the opportunity to obtain the strong, commercially valuable protection that a patent can provide.
Patents protect technical innovation
I am now going to say something that will appear at first sight to contradict the introduction above. Stick with me on this and all will become clear (or at least clearer!).
Although the laws of different countries express the concept of patentable inventions in different ways, we can broadly summarise the position by saying that patents protect technical innovation.
The laws of many countries also make it clear that abstract ideas, mathematical methods, business methods and various other categories of subject-matter are not considered to be technical innovations. European patent law, for example, expressly states that “mathematical methods”, “programs for computers” and “methods for doing business”, amongst other things, “shall not be regarded as inventions”.
“So,” I hear you say, “surely that means you can’t patent algorithms or the software that implements them?”
Well. Yes and no.
Certainly that is what the letter of the law seems to suggest, which (directly or indirectly) leads many innovators to wrongly jump to the conclusion that if their innovation is based around an algorithm then they can rule out patents. They don’t even get as far as seeking expert advice. They “know” or have “heard somewhere” that you can’t patent algorithms, so dismiss the idea and move on.
But, as I have already said above, many thousands of patents have in fact been granted for inventions based on algorithms.
Confused? Let me explain …
Algorithm based innovations can be patented
The problem arises from a misunderstanding about the way in which the legal exclusions from patentability are applied in practice.
In particular, we must draw a distinction between the legally excluded subject-matter itself (e.g. mathematical methods, programs for computers, business methods, etc) and patentable inventions that make use of the excluded subject-matter.
In other words, whilst an algorithm itself, absent context, is too abstract to be patented, a practical application of the algorithm may well be patentable. And this can be the case even where the only difference between what is already known and your new innovation is the algorithm itself.
That is not to say that all practical applications of an algorithm will be patentable. The innovation as a whole still needs to qualify by being “technical” in nature.
So, for example, the practical application of a new algorithm to provide an improved method for classifying academic documents is unlikely to be patentable (because document classification is unlikely to be considered “technical”). On the other hand, application of the same algorithm to classify sensor information for fault diagnosis in a machine may well be patentable (because fault diagnosis is “technical” in nature).
There are also “novelty” and “inventive step” (also referred to as “non-obviousness”) hurdles that must be cleared, as with any invention in any field, but these are topics for another day.
Still a bit unclear? Let’s look at some real world examples.
Examples
The patenting of algorithms and other software based innovations is not a new thing.
As early as 1986, the Appeal Board of the European Patent Office confirmed that even where the idea underlying an invention is a mathematical method, a patent can be granted for a technical process that makes use of the method. The case in question related to a new algorithm for image processing.
Around the same time, Koch & Sterzel secured a patent in Europe (and successfully defended the patent against oppositions by Siemens and Philips) for a new X-ray apparatus. The innovation lay solely in the algorithm that controlled the apparatus to ensure optimum exposure combined with adequate protection against overloading of the X-ray tubes.
The two examples above illustrate that algorithms can be patented when they are used for a specific technical purpose. Other examples of technical purposes that have supported the patentability of an algorithm include:
digital image or video enhancement or analysis, e.g. detecting persons in a digital image
speech recognition, e.g. mapping a speech input to a text output
determining the energy expenditure of a subject by processing data obtained from physiological sensors
deriving the body temperature of a subject from data obtained from an ear temperature detector
providing a genotype estimate based on an analysis of DNA samples, as well as providing a confidence interval for this estimate so as to quantify its reliability
providing a medical diagnosis by an automated system processing physiological measurements
It should be apparent from the above examples that there will be many examples of algorithms applied in the medtech field that will be capable of attracting patent protection.
Don’t miss out …
I hope that, if nothing else, you will take away from the above that it would be wrong for you to simply assume that you cannot secure patent protection just because your innovation lies in an algorithm.
Not all algorithms will be the basis of patentable inventions.
But many will.
Especially if you are creating algorithms that have a practical application in the healthtech field, you should be exploring the possibility of patent protection before the innovation is made public, otherwise you risk missing out on the commercially valuable protection that a patent can offer.
Hardian Health is clinical digital consultancy focussed on leveraging technology into healthcare markets through clinical strategy, scientific validation, regulation, health economics and intellectual property.
