The final goal of a biomedical discovery is to reach clinical routine in patient care. But the pathway from the flash of a technological invention until its use as a medical device in every day care is tedious, burdensome and expensive. Consequently, an accelerated innovation process would save time, lower costs and keep the entrepreneurial risks manageable . Since the safety of medical products has to be kept as high as possible, the desired acceleration must not only focus on finding shortcuts to quickly deliver products to the market . In fact both the speed of innovation and product safety as opponents have to be balanced ,  within a better management planning of the innovation cycle on an improved understanding and better informed basis.
To make matters worse, the complexity of recent biomedical discoveries is heavily increasing and further slowing down the pace of corresponding innovations, especially for cell therapies . Apart from resolving the complexity problem, case studies note the most important challenges in the translation of cell therapies in the scalability, manufacturing and regulatory hurdles , and they recommend an early focus on commercialization even in academic environments. All these issues are among others subject of the field of Translational Research  that focusses generally on promoting knowledge from basic science to enhanced patient treatment, their quality of life, the speed and the progression of developing medical innovations. Due to the vast diversity of medical means as drugs, devices or treatment options, domain-specific recommendations were designed, e.g. for neurotechnology  or –even more –for cancer cure , , .
Surprisingly, the type of technology seems to only have a minor impact on the time-to-market. Several papers report consistently that the time span from the first invention to market entrance usually reaches a decade or even more , [10; 11; 12]. Special emphasis in research on the time lag is given to the pace of pharmaceutical innovation i.e. drug development. The findings range from 5 to 28 years from chemical synthesis (filed as a patent) until FDA approval : Sternitzke denoted 12.61 years for a sample of 64 drugs , Chandy and co-workers calculated an average of 9.47 years for 603 drugs . When broadening the scope from drug development to health intervention in general, a review yielded an even larger and confusing diversity of time lag values . Due to the inconsistent methods in use, the authors stated a general lack of knowledge about time lags in translating discoveries to clinical practice.
This seems to be predominantly the case for medical device’s time lags as very little is known about their specific time spans . Systematic approaches and commonly accepted definitions (e.g. starting- or endpoint) which could overcome the limitation of case studies are missing. A current PUBMED search revealed only seven matches for publications in the field of ‘translational medical research’ (Mesh-term) with the term ‘medical device’ in title/abstract. None of those seven hits contained the search terms ‘time’ or ‘speed’ in the title/abstract.
Despite of the small knowledge base, the threat of an ever increasing time lag is already understood, not only by companies complaining about the growing burden of regulatory requirements. Thus the FDA Center for Devices and Radiological Health (CDRH) launched the Medical Device Innovation Initiative in 2011 in order to shorten the time-to-market , . Despite such efforts which are not limited to the US but have as well been undertaken by European ‘medtech’-nations such as Germany  and Switzerland , some still see the innovation system for medical devices in crisis .
Therefore, this paper aims to contribute to the early and somewhat discrepant empirical results on the specific time lag of medical device innovation to improve the understanding of this specific ecosystem within the biomedical R&D.
For adequate comparisons to prominent paradigms in drug development research, we investigated the time lag between priority date of the first patent application (having in mind that different from the pharmaceutical sector with a patent rate of 80% , by far not every medical device is based upon a patent) and CE mark approval of the corresponding product. According to  we try to differentiate between the risk class of the products given by the CE Mark assuming that the time lag increases with an increase of the medical device’s risk class.
2 Material and methods
To enable comparison with the literature-based drug development data, we define the time lag as the number of days between the priority date of the patent and the registration date of the CE Mark approval of the corresponding product. When transformed to years, a basis of 365 days was used throughout. Leap years were not considered.
Filed patents (‘B3’ and ‘B4’ kind-of-document codes) classified as A61 (IPC-International Patent Classification) published between 2004 and 2014 assigned by German legal entities were selected from the data provided by the German Patent and Trademark Office (DPMA). To facilitate the tracing of the patents to the corresponding CE-Mark records, the search was limited to those companies which assigned no more than five records of either only B3 or only B4 documents (as pairs with the prior applications). The resulting subset embodied 107 companies with 118 B3-patents, and accordingly 337 firms with 501 B4-documents used for the CE Mark retrieval by company name at the German Institute of Medical Documentation and Information DIMDI. A set of 278 different CE Marks was identified; in case one company has announced several CE Marks, only the three top ranked records ordered by ascending registration date were included into the further procedure.
Finally the matching of patents to the corresponding CE-Mark objects was done manually by inspection of title and description, double checked by two authors. Consentaneously ruled cases were taken directly into the analysis, inconsistent ruled cases were judged by a third examiner for the final vote. Since one CE Marked product can base upon several patents (case 1) and vice versa (case 2) the problem of ambiguity arose during the matching procedure. For case 1, the patent with the earliest priority date was chosen, for the case 2, the foremost registered CE Mark record was selected.
The final dataset comprises 61 matched pairs of patents-CE-Mark records as shown in Table 1.
2.2 Statistical procedure
As the assessment results proved the subsets of the sample to be not normally distributed, the Kruskal-Wallis test, a non-parametric ANOVA-equivalent test for skewed distribution, was applied to check for significant mean differences between time lags [years] according to the three risk classes. Negative time lags were deducted from the analyses. Due to the small and heterogeneously distributed sample size, the significance level was set to p = 0.1.
The median values indicate a mean time course for class 1 and 2 products of approximately 6 years whereas class 3 products show a mean value of 10.4 years. However, the variance in all classes is remarkably high (see Table 2).
The statistical analysis reveals a significant difference in the time span for class-3 medical devices compared to medical products assigned to risk class 1 (see Figure 1). This result confirmed the initial assumption on the expected impact of the risk class on the time lag partly, because the pairwise comparison matrix revealed no further significant differences.
In contrast to the often stated average time course of 17 years  from research evidence to clinical practice, our analysis revealed, that it takes approximately 6 years for medical products assigned to risk class 1 and 2, whereas class 3 products need 10.4 years to receive their CE Mark approval. The values of the last named group were in the similar range of 10 to 12 years that was recently reported in literature on the time lag of drug development , . It could be assumed, that either the increase of regulatory requirements for class 3-products or the type of an innovation (incremental vs. technological breakthroughs), the newness of technology or even the used knowledge sources could be responsible for the enlarged time course between the risk groups or for the remarkable variance within the groups. Sternitzke  followed this approach for pharmaceutical innovations but could not confirm an impact on the time course. The differentiation of the innovation relied on the FDA classifying of drugs, e.g. the novelty of the chemical substance. For medical devices another approach seems to be promising: evaluating forward and backward citations of a patent to assess both the basicness of knowledge and its impact on developing future technologies . This should be subjected to future work.
Additionally, in order to identify possible effects in a more detailed way, the pathway of innovation should be subdivided in different subsequent stages  such as i) early research/discovery and pre-clinical studies, ii) clinical studies until approval. As a first step into that direction, the time course for filing the patents (priority to filing date) was calculated for the given sample, but no differences between the risk class groups could be detected. The resulting average time taken to file the patents accounts for 4.3 years (median) with again a vast variation.
Considering the enormously varying data, the uneven number of cases per risk class and the diversity of products in medical technology, a substantial increase of the sample size is mandatory to confirm and specify the preliminary results of this paper (which therefore should be rather considered as trends than as proved significances). But this requires an automated retrieval and - even more important – matching procedure that is capable to safely identify the pairs of corresponding objects stemming from different data sources such as patents, CE Marks, trademarks, clinical trials etc. Performing such a text-mining procedure will be presumably most challenging to resolve the ambiguity of the related data objects. First attempts to create an appropriate and smart retrieval and matching model were encouraging that a text mining approach will work.
Additionally this will contribute to overcome the given bias of our sample towards assignees with only a few patents, which were selected to optimize the traceability. Thus the sample does not truly represent the general structure of German biomedical patents.
Since little is known about the time course of translating discoveries to become a medical device in clinical practice we conducted an empirical analysis of the time taken from patent priority to CE Mark approval of 61 cases. Similar to the development of novel drugs, devices assigned to the highest risk class 3 needed a decade (10.4 years) to get approved which was significantly more than class 1 and 2–products with a consistent time lag of approx. 6 years. To overcome the limitation of the small and uneven sample size for future work a text-mining approach is proposed to especially resolve the major challenge of ambiguity of the related landmarks along the pathway of biomedical innovation.
The support of DIMDI and DPMA in providing CE Mark and patent data is gratefully acknowledged.
Research funding: This work was funded by Klaus Tschira Stiftung gGmbH, Heidelberg, Germany. Conflict of interest: Authors state no conflict of interest. Material and Methods: Informed consent: Informed consent is not applicable. Ethical approval: The conducted research is not related to either human or animal use.
Bers JA, Dismukes JR, Miller LK, Dubrovensky A. Accelerated radical innovation: theory and application. Technol Forecast Soc. 2009;76:165–77. Google Scholar
Krucoff MW, Brindis RG, Hodgson PK, Mack MJ, Holmes DR Jr. Medical device innovation: prospective solutions for an ecosystem in crisis adding a professional society perspective. JACC-Cardiovasc Interv. 2012;5:790–6. Google Scholar
Curfman GD, Redberg RF. Medical devices — balancing regulation and innovation. N Engl J Med. 2011;365:975–7. Google Scholar
Dodson BP, Levine AD. Challenges in the translation and commercialization of cell therapies. BMC Biotechnol. 2015;15:70. Google Scholar
Polese F, Capunzo M. The determinants of translational medicine success - a managerial contribution. Transl Med UniSa 2013;6:29–34. Google Scholar
Leuthardt EC. Developing a new model for the invention and translation of neurotechnologies in academic neurosurgery. Neurosurgery. 2013;72(Suppl 1):182–92. Google Scholar
Abernethy A, Abrahams E, Barker A, Buetow K, Burkholder R, Dalton WS, et al. Turning the tide against cancer through sustained medical innovation: the pathway to progress. Clin Cancer Res. 2014;20:1081–6. Google Scholar
Epstein RJ. Unblocking blockbusters: using boolean text-mining to optimise clinical trial design and timeline for novel anticancer drugs. Cancer Informatics. 2009;7:231–8. Google Scholar
Lee LJ, Harris JR. Innovations in radiation therapy (RT) for breast cancer. Breast. 2009;18:S103-11. Google Scholar
Sternitzke C. Knowledge sources, patent protection, and commercialization of pharmaceutical innovations. Res Policy. 2010;39:810–21. Google Scholar
Roos A, Lindstrom M, Heuts L, Hylander N, Lind E, Nielsen C. Innovation diffusion of new wood-based materials - reducing the “time to market”. Scand J Forest Res. 2014;29:394–401. Google Scholar
Morris ZS, Wooding S, Grant J. The answer is 17 years, what is the question: understanding time lags in translational research. J R Soc Med. 2011;104:510–20. Google Scholar
Chandy R, Hopstaken B, Narasimhan O, Prabhu J. From invention to innovation: conversion ability in product development. J Marketing Res. 2006;43:494–508. Google Scholar
Center for Devices and Radiological Health U.S. Food and Drug Administration. CDRH Innovation Initiative; 2011. Available from: URL: http://www.fda.gov/downloads/AboutFDA/CentersOffices/CDRH/CDRHInnovation/UCM242528.PDF [cited 2016 Mar 29].
Die Bundesregierung, BMBF, BMG, BMWI. Nationaler Strategieprozess - Innovationen in der Medizintechnik: Abschlussbericht; 2012. Available from: URL:http://www.strategieprozess-medizintechnik.de/sites/default/files/Schlussbericht_NSIM.pdf [cited 2016 Apr 1].
Eidgenössisches Departement des Inneren EDI. Massnahmen des Bundes zur Stärkung der biomedizinischen Forschung und Technologie; 2013. Available from: URL:http://www.bag.admin.ch/themen/medizin/14583/ [cited 2016 Apr 1].
Dornbusch F, Neuhaeusler P. Composition of inventor teams and technological progress - The role of collaboration between academia and industry. Res Policy. 2015;44:1360–75. Google Scholar
Hanney SR, Castle-Clarke S, Grant J, Guthrie S, Henshall C, Mestre-Ferrandiz J, et al. How long does biomedical research take?: Studying the time taken between biomedical and health research and its translation into products, policy, and practice. Health Res Policy Syst. 2015;13:1. Google Scholar
About the article
Published Online: 2016-09-30
Published in Print: 2016-09-01
Citation Information: Current Directions in Biomedical Engineering, Volume 2, Issue 1, Pages 599–602, ISSN (Online) 2364-5504, DOI: https://doi.org/10.1515/cdbme-2016-0132.
©2016 Robert Farkas et al., licensee De Gruyter.. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. BY-NC-ND 4.0