Patent power in biomedical innovation: technology governance in biomodifying technologies

Biomedical innovation is often rewarded by exclusive proprietary rights such as patents. In the case of gene editing, induced pluripotent stem cells, and three ‐ dimensional (3D) bioprinting (here described as biomodifying technologies), the limitations of the patent system come into stark relief, generating both technical and political doubts. Generally, political and technological limitations are supposed to be solved with so ‐ called good governance, based on some principles. We focus on three of such governance principles (participation, accountability, and transparency) to show how they have been weakened, instead of strengthened, by the current patent system. We demonstrate that although patent applications are submitted by both public and private players, the latter have imposed a growing dominance in gene editing, induced pluripotent stem cells, and 3D bioprinting, disseminating their aggressive and exclusive strategies. As a consequence, a logic of experimentality tends to prevail where the three fundamental governance principles fail to


| Governance and experimentality
Is it sufficient that therapies be effective and safe, as attested in clinical trials, or is it also important that they be originally developed and owned via inclusive and transparent schemes? Differently said, are the institutions and companies responsible for the development and intellectual appropriation of therapies exclusively subjected to technical and scientific standards or should they also comply with democratic mandates to be transparent, inclusive, and accountable? To explore these questions, this paper focuses on the patent system as it has evolved in three areas: gene editing, induced pluripotent stem cells, and 3D bioprinting. This analysis will be guided by the concept of experimentality, as proposed by Petryna. 1 Pharma and biomedical companies, in order to argue for the viability of their products, may conduct some studies whose design is controversial (because minimal superiority over alternative products has to be shown) and whose outcomes have limited publication (due to market secrets). Experimentality emerges when the logic of market secrecy leaves the realm of industries and laboratories. At such moment, the targets of experiments are not only products but also institutions, populations, and social relations.
In other words, experimentality is the logic disseminated by global schemes of biomedical innovation where a few market players subject populations and countries to an innovation process in which corporate rationales tend to prevail. For example, the formation of a global clinical trials industry has turned national regulatory frameworks into useful cogs in the commercial, globalised clinical trials system while turning national populations into potential research subjects. 2 Parallel to such logic of experimentality, there is a rationale frequently described by means of the concept of governance. Biomedical innovation, because of its growing complexities, requires inputs from a variety of institutions and companies. 3 It is frequently difficult to strike a good balance in this social interplay, as some players have greater capacity to impose their agendas on their collaborators. With its political legitimacy, the nation state is expected to intervene and foster the balanced relations associated with governance.
In his review of the literature on governance, Reich 4 verified that accountability and transparency are the principles discussed the most frequently. 'Mechanisms of good governance can include transparent, democratic institutions as well as efficient and effective public services'. 5 D'Orville provides us with an even more detailed picture: There are three features of governance. First, it has been defined as the rules of a political system established to solve conflicts between actors and adopt decision-this provides the legality.
Secondly, governance also describes the proper functioning of institutions and their acceptance by the public-in other words the legitimacy. Thirdly, governance invokes the efficacy of government and the achievement of consensus by democratic means, thus defining a participatory dimension. 6 However, governance is practised not only by the government; it is rather a social arrangement formed by the state, companies, research institutions, citizens, and so on. 7 In a widespread approach, also adopted in this paper, governance can be defined as a balanced framework where all relevant stakeholders contribute towards maintaining accountability and transparency in technology development. This definition is in line with the participatory governance approach 8 and, more broadly, with the tenets of the so-called deliberative democracy perspective. 9 It deviates, then, from another approach where governance is equated with management, leading to ideas such as 'private governance', which is said to occur 'when certain phenomena, such as the use of new biotechnologies, are regulated by private agents'. 10 From the social perspective adopted here, private management is certainly part of the governance landscape but is too narrow to fully characterise governance.
In the development of biomedical technologies, which may be funded by government agencies and whose clinical effects may reach large populations, it seems particularly important to promote transparency and accountability, thus reducing the weight of experimentality. For the control of these technologies can result in public health tragedies if realised in abusive ways. Examples of such abuses have been identified in the conduct of clinical trials in lax regulatory environments 11 ; in the murky relations between pharma companies and regulatory agencies 12 ; in the restrictive composition of research ethics committees 13 ; in confidential agreements leading to the diffusion of magnetic resonance scanners of controversial safety 14 ; among other issues.
It is hoped that once good governance is in place, then knowledge, skills, industrial knowhow, and technologies can circulate so as to safeguard the public interest. This assumption, which puts the concept of participation at the core of the governance debate, 15 has been strengthened by analysts asking for more participatory schemes to enhance democracy. 16 Governance could then constitute a realm of transparent and publicised measures, in opposition to the convoluted and restrictive schemes disseminated by the principle of experimentality. However, in spite of the efforts made by the nation state, some barriers to governance can emerge as a result of advanced globalisation.
One example was seen in 2020 when national governments, called upon to deal with the urgent needs brought about by the COVID-19 pandemic, realised that a proper response depended on access to technologies strongly subjected to intellectual property rights and the rules of the market. The need for quickly developing a vaccine showed that market rationales are frequently at odds with the public interest. 17 Suddenly, it seemed impossible to conceal what Maskus 18 described as the tension between 'private rights and public problems'.
To understand the interplay between experimentality and governance, this paper focuses on the emergent landscape of patent protection of innovations in gene editing, induced pluripotent stem cells, and 3D bioprinting. It will be seen that even though patent applications are submitted by both public and private players, the logic of experimentality tends to prevail, shape the patent landscape, and defy governance principles.
To understand patent issues, it is, therefore, necessary to consider not only the patent law itself but other parallel aspects. In this way, whenever we talk about the 'patent system' in this paper, we also bear in mind factors such as the shape of biomedical markets, competition law, and the structure of intellectual property in each technology. 19 This paper is organised in four parts. After presenting the features of what we call biomodifying technologies, our research methods are outlined. We move on to analyse the patent landscape of our three case technologies in light of the governance principle of participation. Subsequently, we analyse some strategies of patent protection by considering the governance principles of transparency and accountability. The closing section brings some considerations related to governance as well as some policy recommendations.

| Biomodifying technologies
The analysis of gene editing, induced pluripotent stem cells, and 3D bioprinting can shed some light on the roles of good governance and experimentality as they are reflected in the patent regime. Such analysis is particularly important for pharmaceutical and biotechnology innovation, where patents have played a key role. As these three technologies imply the modification of basic life structures (cells, tissues, and genetic codes), we characterise them as biomodifying technologies. Due to their relative novelty, and also because they have been the object of constant investigation and patent submissions, they constitute a suitable empirical topic for the discussion proposed here.
Gene editing can be considered an upgrade on older genetic modification techniques, in terms of speed and accuracy. It consists of the alteration of DNA using molecular tools such as CRISPR/cas9, TALENS, and Zinc fingers.
Each of these tools contains a programmable 'targeting' domain (such as the 'CRISPR' part of CRISPR/cas9) that can be designed to attach itself to a particular sequence of genetic material in a living cell. Once the target sequence is bound, a 'molecular scissors' part of the tool (e.g. the 'cas9' enzyme) can cut out a particular piece of the DNA, replace it, or change its content.
Induced pluripotent stem cells technology takes mature cells of the adult body, usually of an accessible type such as skin cells (fibroblasts) or hair follicle cells (keratinocytes), and 'reprograms' them to an immature state, so that, like the cells of an early embryo, they can be directed to become any type of cell in the body. They have the same genetic material as the original donor and can be used for both modelling human disease and as a source of new material to replace diseased or damaged tissues in the human body.
In bioprinting, by means of software-controlled devices (bioprinters) that deposit successive layers of substrates containing cells (bioinks), it is possible to generate organoids or tissues that might potentially be used in clinical applications. 20 Although scientists can now bioprint only simple structures like skin or cornea, it is expected that sophisticated structures, and even functional organs, will be bioprinted in the future. 21 In these three domains, it is possible to take advantage of the human body's plasticity to generate therapies, which characterises the field of regenerative medicine. 22 The question that arises is: considering the ways in which biomodifying technologies have been developed and Three main research methods have been mobilised. First, an extensive literature review has been conducted, so as to identify the main topics being debated not only in scientific domains but also in social science analyses.
Second, by using two platforms (Google Patents and The Lens), we conducted a patent analysis for the three technologies focused on in our projects. As these technologies were developed in different moments, it was necessary to 'normalise' the analysis in terms of historical focus. The year 2006 was taken as the starting point of our analysis. This is when the technique for engineering iPSCs became available. 25 The first phases of gene editing and bioprinting can be traced back to the 1990s and even before. However, these two techniques only acquired modern forms in the 21st century with the use of bioinks and the foundation of specialised companies for bioprinting, and the discovery of techniques such as TALENS and CRISPR for gene editing. In this way, it seemed appropriate to take 2006 as a starting point rather than using time warping algorithms that process data in 'aggressive' ways.
The patent search, conducted in April 2020, is summarised in Table 1.
With Boolean operators, we tried to exclude as many patents related to vegetals and animals as possible, thus focusing on human health. For gene editing, it was seen that a search based on the generic expression ('gene editing') tends to overestimate the number of patents, going beyond the limits of gene editing. It was then decided to use keywords related to the main gene editing techniques. For data processing and data visualisation, the R programming environment 26 was used (more specifically, the following libraries: stringr, dplyr, readr, ggplot2, and reshape2). In the following section, the analysis of our findings emphasises the governance principle of participation.

| PATENT LICENSING AND THE PRINCIPLE OF PARTICIPATION
As already explained, governance depends, to a large degree, on the nation state's capacity to oversee the innovation system and steer its evolution in socially acceptable ways. In more theoretical terms: 'The constitutional structure of the political system is preserved only if government officials hold out against corporate bargaining partners and maintain the asymetrical position that results from their obligation to represent the whole of an absent citizenry […] '. 27 However, the advancement of globalisation and harmonised rules of free trade often weakens the efficacy of regulations, widening the space for corporate experimentality. 'People […] lack resources to address voids in national regulation and international policy, voids that enable the uncharted expansion of experimentality […]'. 28 In recent decades, corporations have even had the power to bring national governments to court under international investment agreements. 29 The evolution of the patent landscape reflects these trends.
In the three technologies focused on here, there has been a steady expansion in the number of patents filed, as shown in the following chart.
The expansion has been particularly impressive for gene editing, especially after 2012 when Crispr, a gene editing technique that would gain rapid diffusion, was invented, as we described elsewhere. 30 As seen in Chart 1, bioprinting has had relatively modest performance. Indeed, the generation of patentable bioprinting products and techniques has been slower due to two main factors. First, scientists cannot yet bioprint vascular systems, a limitation that restricts the size of bioprinted tissues. 31 Second, techniques allowing the bioprinting of more complex tissues, particularly laser-based bioprinting, are still expensive and therefore less widespread. 32 However, when one considers the pace at which patents have been filed, bioprinting does not occupy the lowest position, as seen in the following chart (Chart 2).
Considering the 1301 bioprinting-related patents we found, around 17% were filed in 2018, and almost 25% in 2019. For gene editing the proportions were around 22% in 2018, and almost 30% in 2019. From this more telling point of view, iPSCs are the technology with the lowest expansion. This is arguably signalling greater anticipated difficulty in commercialising applications of iPSC technology beyond its current use as a tool in preclinical drug screening. The multiplication of patent applications is accompanied by an increase in the number of stakeholders filing patents. Interestingly, all three domains are marked by a concentration in terms of patent holders. Let us consider the ten main patent applicants/holders in each technology, as summarised in Table 2.
The first 10 players hold a substantial proportion of the overall patent portfolio, especially for gene editing (with almost one quarter of patents filed by the main players) and bioprinting (22.2%). Furthermore, private companies have dominated the patent landscape, as seen in the following chart.
In the three domains, companies hold over 50% of the patents, a proportion that reaches around 65% for bioprinting.
In 2007, in their analysis of the stem cell patent landscape, Bergman and Graff 33 concluded that public institutions were as important as private companies. Therefore, the dominance of companies illustrated in Chart 3 is very recent. For example, 33.4% of the iPSCs patents held by companies were filed in the 4 years from 2016 through 2019. The reduction in the relative patent power of universities might be explained by decreasing government investments in academic research. It is certainly explained by two phenomena.
On the one hand, innovations made in universities are frequently taken over by nonacademic players who end up securing the patent. For example, Martinez 34 showed that in most European countries, over 50% of academic inventions have been patented by companies. Furthermore, many companies in possession of promising innovations prefer to create spin-out companies to patent and explore them, a strategy that has been frequent, for instance, in the United Kingdom. 35 On the other hand, the shifts in the commercial exploration of biomodifying technologies have to be considered. Until the beginning of the 21st century, those technologies were mainly explored by small biotech firms with innovative capacity but modest financial resources. 36 Nowadays, as can be noted in Table 1 | 479 tends to be gained by few players (as seen in Table 1), the evolution of biomodifying technologies depends on decisions taken by this small group, in two ways.
First, as patent holders have exclusive rights to control the use of a certain technology, they become gatekeepers for that experimental area. An example was given by one of our interviewees, a patent specialist (IP3): Second, as we showed elsewhere, 40 patent holders have the power to select other players which will be included in, or excluded from, the development of particular biomedical areas. This can be done by imposing huge prices for the use of patents, thus excluding companies with little funding. 41 It is also frequently done by means of patent licensing, an agreement between the patent owner (the licensor) and an interested player (the licensee), allowing the latter to access the technology, explore it, improve it, and glean the patent's financial benefits. A license can be granted to several licensees at the same time but biomedical and pharma companies prefer exclusive licences, those in which only one company is allowed to explore the patent. As claimed by a patent specialist (IP1):  43 These examples serve to temper the exalted discourses about the heroic inventiveness of companies, reminding us that biomedical experimentation is increasingly dependent on joint efforts by public and private players. 44 The most resourceful universities, by means of their technology transfer offices, consider these strategic relations at early stages of research projects, with an eye to what can generate revenue through licensing. 45 The practice of patent licensing has been almost completely subjected to private agreements. Thus the principle of participation, which lies at the core of governance, 49 is seriously undermined by a system where it is up to companies, by means of unpublished agreements, to decide who will participate in the development of technologies. Eventually, the system tends to favour exclusion, instead of participation, and the logic of experimentally is diffused. 'Regulated state withdrawals make it possible, favouring economic interests over empowered systems of regulatory oversight-even if only for a certain time period'. 50 If the principle of participation is in jeopardy, it is worth asking how the patent system impacts on the principles of transparency and accountability.

| THE BORDERS OF PATENTABILITY AND THE PRINCIPLE OF TRANSPARENCY
In order for a technique to be patented, the scientific and technical characteristics of the innovation, compared against previous/current techniques (prior art/state of the art), are assessed. Different patent offices may adopt different approaches in this process. In the United States a decisive aspect of patent evaluations is 'nonobviousnes' while the European Patent Office works with the concept of 'inventive step'. 51 Notably, the European Patent Office adopts a 'problem and solution' approach to assessing inventive steps, requiring technical features and technical problems to be established.
For a socially relevant process (such as a patent application process) to be transparent, some fundamental requirements in disclosure should be satisfied. Particularly, the following criteria should apply: (1) The players responsible for conducting the process should acknowledge the limits of certainty involved in their patent claims, and (2) these players should make public how the process is being conducted, and such publication must be understandable and meaningful for as many interested players as possible. In the domain of patents for biomodifying technologies, both criteria fail to hold. Given biology's complexity, the outcome of any given experiment is increasingly uncertain.
Experimental trial and error-more than design in the 'dry' engineering fields-is critical to research in biology. This complicates courts' and patent offices' obviousness analyses, because even standard combinations of elements in the field routinely yield unpredictable results". 53 Moreover, patent applications are submitted before the conduct of clinical trials, when researchers have good ideas and reasons to believe such ideas but little evidence showing that therapies will be clinically effective.
Obviously, patent applicants do not disclose their uncertainties about their patent claims. A very widespread strategy used by patent attorneys is to write very broad initial claims. As the technology evolves and clinical data are collected, claims are generally narrowed down when the patent application is supplemented. According to a patent attorney: At the start you've got a lot of flexibility. At the first filing stage […] we might have quite broad claims […] We might deliberately draft them too broad but the art is to have… even if you've got something broad, that you've got something narrow underneath, that covers the product or what they're going to do and will stand up. That's the art.
The metaphor of art was also mobilised by a policymaker to explain that initial patent applications are usually ambitious: In this art, in the biological art, in the organic chemical art, in the pharmaceutical art, you would often have in a single application a claim to the products, a claim to a method for making the product or the process, and quite often then a claim to its use and therapy.
Again, we approach the hidden political upshots of the patent system. In theory, legal provisions disentangle what can be patented from what is excluded. However, the practices of the patent system often defy such delineations. For example, computer software cannot be patented in the European Union but patents would apply when the inventor shows that the software package has a 'technical effect', a vaguely defined concept. 54 As a consequence, some software packages, such as the ones used in computer-assisted diagnosis, have been patented in the Union. This may bring about opportunities for bioprinting companies whose software packages, in association with bioprinting devices, may in the future be considered as generators of an innovative 'technical effect'. Indeed, as we showed elsewhere, 55  In this way, the patent system, tinkered with by experienced patent attorneys, pushes the boundaries of patentability, as explained by an attorney: '[…] a lot of the development in the patent world has been pushing at those margins, where you're on the boundary. There are some things you can push into the patentable side, as it were, and that's developed with the case law'. To be sure, such expansion of the frontiers of the patent system has important social consequences, which may remain hidden until the outbreak of the next public health crisis or epidemic.
The second criterion of transparency (widespread and clear publication) also fails to hold. There are several ways in which innovations remain removed from the public's eyes. For example, a report by the Hinxton Group concludes that some iPSC cell lines have been produced but the achievement is frequently not published. 'This has created a situation in which even a diligent stem cell researcher or entity that wishes to respect IPR [intellectual property rights] will face considerable uncertainty and enormous costs if they try to survey the IPR landscape'. 57 Due to this and similar practices, claims have been made for clearer systems where information on innovative biomedical technologies and granted patents would be disseminated more largely. 58 Every patent document must be published but publication takes some years to materialise. Having recourse to our data set, we calculated the difference between each patent's priority year (when the patent is initially filed) and year of publication. For gene editing, the average difference is 2.72, that is, gene-editing patents take, on average, 2.72 years to be published. For iPSCs, the average difference is 3.67; for bioprinting, 2.55.
However, not all patent documents obtain the same level of recognition, as some of them are consulted and cited more frequently than others. We then divided the patent documents in four groups, from those which have rarely been cited by other patent documents to those with high levels of citation. The outcome is seen in Chart 4.
Chart 4 shows that for all three technologies, patent documents cited very frequently (100 or more citations in other patent documents) have publication delays bigger than that of the documents with very low citation levels (from 0 to 9). Therefore, the more frequently patents have been consulted and mentioned, the longer they have taken to be initially published. This pattern is not manifested for bioprinting, arguably because of the relatively low total number of patents (1301), yielding four small groups whose average result is distorted by extreme cases. Patent laws have disclosure requirements according to which patents can only be granted if the inventor accepts to clearly disclose the nature of the invention and how replication can be achieved. 60 However, these requirements can be abated by the possibility of delaying patent publication when a substantial amendment to the original filing is realised. As a result, disclosure requirements become another argument and instrument to be used in the framework of experimentality whereas from the viewpoint of governance, the relevant practice (i.e., the one that promotes transparency) is patent publication.
Even when publication finally happens, it is made in a legal language that few people understand. 61 Given the uncertainties associated with developing biomodifying technologies, the usual approach is to focus on technicalities, expecting that patent attorneys, who master the language of patent claims, proposals, and litigations, will engineer innovation into applications. As explained by a patent attorney (IP4): '[…] you try to frame things in terms of problems, technical problems and solutions to those problems, or a particular technical problem when you're drafting something, as an attorney'.
Gradually, the dialogues involved in patent applications have gained a technical tone which most citizens would find obscure, even though everybody might eventually need products made expensive by patent protection. As patent granting has been institutionally separated from health technology assessment, every supposedly wellconducted patent assessment tends to disregard points that may be crucial for most citizens, such as technology

| HIDDEN PUBLIC HEALTH STRATEGIES AND THE PRINCIPLE OF ACCOUNTABILITY
According to the principle of accountability, actors should be made responsible for the strategies they adopt and the decisions they make, as long as those strategies and decisions have social implications. It may seem that patent applicants and offices are dealing with technical matters with little social effect. However, a more detailed analysis reveals that patent strategies have amounted to hidden public policies to which multiple players contribute but for which nobody is made accountable. 'Decentralized and diffused in character, this experimentality does not lend itself easily to prevailing tools of accountability'. 63 This is manifested in at least three ways.
First, patent strategies can block a certain technological field, shaping opportunities for market exploration and investment. One of these strategies is the division of industrial know-how into a combination of patents and confidentiality agreements. Another strategy is so-called patent thicket, whereby a certain company (or partner companies) files a series of patents targeting different subdomains of the same technology, preventing competitors from getting hold of areas of the technological territory. 64 The conditions of a patent thicket have been verified in the field of stem cell research. 65 Another example comes from the field of engineered meganucleases where most of the existing patents are owned by two companies: Cellectis (which appears in Table 1 Second, the high prices paid by those engaging in the patent system creates restrictive geographic strategies. Indeed, only very large corporations are economically able to file and maintain patents in various countries. A patent attorney (PAT02) disclosed the advice given to patent applicants: '[…] if people haven't got a lot of money, at the end of the day the essential thing is we will say "Keep the US, let everything else go." […] What we would normally recommend is Europe, US, and China'. The outcome of this approach is illustrated on Map 1, which shows the jurisdiction of patents, as well as the location of patent holders, for the field of gene editing.
As companies are generally not willing to take their key products to countries where the underlying technology is not protected, world patent strategies become associated with the uneven distribution of cuttingedge therapies. For example, CAR-T cell products, a gene-edited therapy first approved in 2017 to fight rare and resistant cancers, are now available only in the United States, China, Europe, Canada, Japan, Australia, New Zealand, Israel, and Singapore. 67 Most of these regions also appear, on Map 1, as those with the most intense patent activity for gene editing. To be sure, the coincidence does not mean that the patent system is the only reason why the vast majority of countries have failed to access CAR-T cell and other gene-edited products.
However, the costs associated with filing and maintaining patents are among the reasons leading to this global health failure.
Finally, the ways in which biomodifying therapies have been developed, as well as the associated development costs (including investments into securing and maintaining patents), result in products of skyrocketing prices. 68 For example, if gene therapies were developed to treat 1% of the American population, the costs would be equivalent to the yearly amount spent on all healthcare in the country. 69 If a proliferation of gene therapies has 'the potential to push the US healthcare system to the breaking point', 70 their impact on other, less robust national systems would obviously be even more drastic.
In a sense, patents can be considered as a delayed funding system whereby society, by purchasing the patented made by society can be very high, as the prices of some therapies are mostly well above the product's marginal cost, excluding some consumers willing to pay prices which are not so high but would still secure profits for the producer. 71 In this way, the patent system underpins economic strategies that may impose heavy burdens on health systems or on the shoulders of citizens not supported by health reimbursement schemes.
Therefore, patent holders have used the patent system to indirectly design policies having impacts on: technology governance and access to markets; the global distribution of therapies; and the costs to be faced by health systems and patients. However, nobody is made accountable for these strategies and decisions because they are not seen as actual policies, but as issues of intellectual property. When the need is felt, by either the nation state or citizens, for having quick access to some therapies made vital, all doors may be closed but it may be impossible to identify the players responsible for such closure in the convoluted landscape of patents, licences, and commercial agreements.
As claimed in a United Nations document: 'Accountability is a key theme running through and underpinning many aspects of governance […]'. 72 However, when patent issues hide governance issues, a situation emerges where key decisions depend on a techno-scientific dialogue between attorneys and patent offices. Thus the patent system remains impervious to social concerns that have led, for example, to the concept of Responsible Research and Innovation, which the European Union has tried to promote. 73 If such policy impacts can be verified, then it would be reasonable to expect that the assessments made by patent offices could transcend technical concepts such as 'nonobviousness', 'inventive step', and 'technical effect'.
More comprehensive assessments can indeed be envisaged by means of some changes, some of which are sketched in the concluding section.

| TOWARD MORE PARTICIPATIVE, ACCOUNTABLE, AND TRANSPARENT PATENT ASSESSMENTS
In this paper we have examined the current governance of biomedical innovation, focusing on three biomodifying technologies (gene editing, iPSCs, and 3D bioprinting). More specifically, we have analysed how subfields of these technologies have been appropriated through patent protection. It was seen that technology development and appropriation are supposed to be guided by governance principles which, nevertheless, have not always been effectively incorporated in the existing patent system. Frequently, what prevails is the logic of experimentality whereby social and political issues are overshadowed by a market rationale. Although academic researchers might, in theory, put in practice alternative rationales, current state and university policies on funding, performance evaluation, and technology transfer, especially in terms of licensing, tend to usher academic research into market logics.
The governance principle of participation has been marginalised in a system where the patent power of some companies has increased, consolidating mechanisms of exclusion. As seen in Section 3, small groups of companies have secured the largest proportion of patents, with weak participation of government institutions that can be involved in national health policies. The principle of accountability is threatened because the complexification of the commercial and intellectual property landscape makes it increasingly difficult to identify the actors responsible for the potential impacts of therapies of increasing prices on public health systems. In Section 5, it was seen that key decisions have been taken in the framework of secret private contracts and agreements. Eventually, it gets very hard to identify the players responsible for the current configuration of the pharmaceutical and biomedical domain.
Finally, the principle of transparency is threatened by the delayed publication of patents, as well the secretive stance of companies that frequently withdraw information on their most innovative achievements, as part of trade secrecy. This point was also made in Section 4, which drew attention to the technical dialogue between patent attorneys and patent offices, without consideration for the social implications.
To be sure, the current patent system has some characteristics that make room for social considerations. For example, by means of an opposition mechanism called 'observations by third parties', the European Patent Office, unlike its American counterpart, considers some arguments voiced by citizens on why a certain patent should not be granted. 74  as the creation of more efficient systems for patent publication, 77 the establishment of an intellectual property clearinghouse mechanism 78 or the clarification of the contexts in which the targeted technology has been developed. 79 Considering the three governance principles highlighted here (participation, accountability, and transparency), the following draft measures are proposed to enforce them in the patent system. Second, the fees practised by patent offices could be re-examined. To a large degree, they have been a deterrent for smaller companies and researchers based in marginal economic countries. A new system could be envisaged where reduced fees would apply for some groups of companies and countries. In addition, more collaboration between patent offices can be encouraged, aiming at streamlining the procedures of less strategic offices, so they can become more attractive to innovative players. These measures would foster participation at both the national and international levels. | 487 with few manufacturing units producing for the world, 83 which implies a concentration of expertise and skills.
Moreover, as seen in the previous section, some cutting-edge therapies have failed to reach the vast majority of countries.
Finally, as the views of the lay public and patient organisations have been taken into account in the conduct of health technology assessment, 84  We are aware that social concerns, when translated into institutional mandates like those proposed above, may lead to mere ceremonial procedures whereby companies tick boxes to show their corporate social responsibilities.
For some analysts, this has been the case, for example, for the assessments made by research ethics committees. 90 However, institutional changes may be helpful by at least triggering social debate where debate is either nonexistent or, as in the case of patents, restricted to a few elite stakeholders. It can be said that governance mirrors the evolving nature of the nation state, which 'does not represent a finished structure but a delicate and sensitive-above all fallible and revisable-enterprise, whose purpose is to realise the system of rights anew in changing circumstances'. 91 The measures sketched above could evidence the need for the 'supportive collaborative-based discursive relationships among public and private sectors' referred to by Fischer. 92 They could help make the patent system undergo the move described by Nowotny, 93 whereby scientific and technical activities transcend the scope of limited expertise to become grounded on socially robust knowledge. It has been claimed that the patent system fails to stimulate innovation and development in biomedicine and pharmaceuticals, 94 and creates a bias towards shortterm gains. 95 Such discussion is out of the scope of this paper but the proposed measures would contribute toward more transparency and accountability even if economic and scientific outcomes are poor.
Currently, the patenting of biomodifying technologies is strongly shaped by the logic of experimentality, which creates weaknesses and gaps that may only be noticed in the future, in contexts of pressing health needs.
'Ultimately, this experimentality underwrites a research agenda that does not necessarily provide the most valid and relevant medical outcomes, and it introduces added risks'. 96 Among these risks, we identify not only the biological risks to which future patients may be subjected but also the risks imposed on weakened health systems, as well the political risks faced by nation states overwhelmed by corporations' economic and scientific power. Indeed, the issues dealt with here are particularly relevant for nation states, considering that most patent offices have a national mandate.
When states decide to enforce austerity measures, finding a balanced governance strategy becomes even more complex. 97 Then, political wisdom may be replaced with political desperation. London and Kimmelman 98 warned that sometimes, the eagerness with which some countries welcome international clinical research '[…] might represent the desperation of host communities more than their zeal for advancing the research agenda […]'. Equally, the willingness to receive patent applications might be reflecting countries' desperation about their scientific relevance or patent offices' anxiety to attract fees paid by innovative players. If the principles of governance are not shielded and put into practice, then the patent system is likely to turn into a generator of such political fragilities, which tend to be worsened by the primacy and globalisation of experimentality.