How long does it take to bring an innovation to clinical practice?

5-20 years, but clinical input can help significantly.
Trying to get discoveries made through basic scientific research to the point where they can actually help patients is, no surprise, really hard.
Commonly referred to as "translational research" or "bench to bedside," the goal of this work is to end up with new drugs, novel biomedical or diagnostic devices, or a new system or process that greatly improves some aspect of healthcare. More recently, algorithms and associated AI systems are falling into the realm of translational research, though are often integrated into other devices or processes.
So how long does it normally take to bring a laboratory discovery to the patient bedside?
This question comes up pretty frequently in our work and is essentially impossible question to answer specifically, as every discovery and innovation is a unique case. But what about more generally? We all know that it will take some time, but what is a ballpark estimate?
For drug discovery, it can take nearly 20 years to bring a drug to market.[1] More recent improvements in parts of the drug discovery process, from changes in regulatory rules, to changes in the way the pharmaceutical industry identifies good leads has led to significant reductions in the time it takes to bring a drug to market down to about 12 years and in special cases as short as 6 years.[2,3] While this is a significant improvement, the failure rate is significant and the costs massive, on average costing around $2 billion to bring a drug to market.[1] In addition, it is still a pretty fuzzy as to how long it normally it takes to identify a good drug target in the lab, before work is done to make a develop the associated drug. So still many opportunities to both better understand this process and make improvements that help patients.
The translational process for medical devices and diagnostics (the "Medtech" space) is similarly challenging and in many ways more difficult to define. Every potential drug is different, but there is a fairly well-defined process through which any drug must go through to make it to market. The same is largely true for devices, but the range of technologies - hardware, software, algorithms, etc. - and the setting in which it may be deployed, make the specific hurdles that any given device has to make it through extra uncertain. In particular, the specific regulatory hurdles that have to be surpassed vary significantly between devices, making it particularly challenging to estimate.
As a result, there is surprisingly little data available as to how long it takes to bring a new technology into the clinic. However, after much digging, I did manage to find a study focused on the translation of new surgical devices into the operating room.[4] This study estimates that it takes around 5-7 years to bring a surgical device to the point where it impacts patient care. Probably not quite as long as a drug, but still a considerable time.
Really interesting is that the authors of this paper estimate the chance that a device will make it to a first-in-human study over time. While only estimated at three time points, their results show that the chance of achieving this milestone is less than 12% after 20 years. Not the best odds.

Another piece of information that really stood out in this paper was that devices were 6 times more likely to be translated to first-in-human studies if they were accompanied with early clinical involvement (input from clinicians).
How long does it take for AI technologies? I have yet to find a good estimate for this - I'll let you know if I find out.
People may not realize just how much of the early stages of this translation process, whether it be discovery of new drug targets or the identification of unmet clinical needs that require new device innovation, actually happens within academic research institutions. In the case of Medtech, the clear improvement in success rates when clinicians and engineers worked more closely together clearly points to the importance of building better systems and processes for these groups to interact and work with each other to solve these important problems.
There are really good reasons that it takes time to develop and test a drug or device to make sure it actually works and also doesn't cause harm to patients. Identifying areas where the speed of this processes can be improved should be a focus for anyone interested in trying to bring innovation to patients more effectively.
References
[1] Austin CP. 2021. Opportunities and challenges in translational science. Clinical and translational science. DOI: 10.1111/cts.13055.
[2] Mohs RC, Greig NH. 2017. Drug discovery and development: Role of basic biological research. Alzheimer’s & dementia: the journal of the Alzheimer's Association 3:651–657.
[3] Stewart DJ, Stewart AA, Wheatley-Price P, Batist G, Kantarjian HM, Schiller J, Clemons M, Bradford J-P, Gillespie L, Kurzrock R. 2018. The importance of greater speed in drug development for advanced malignancies. Cancer medicine 7:1824–1836.
[4] Marcus HJ, Payne CJ, Hughes-Hallett A, Gras G, Leibrandt K, Nandi D, Yang G-Z. 2016. Making the Leap: the Translation of Innovative Surgical Devices From the Laboratory to the Operating Room. Annals of surgery 263:1077–1078.