Yesterday I delivered a presentation on the regulation of stem cell-based interventions in Australia, focusing on the theory and practice of risk-based regulation and on some new developments in this regulatory environment.
The talk was a part of the Cellular Horizons research grant that has been awarded to a team of scholars of which I am a member, as a chief investigator. The research grant is funded by the Commonwealth Department of Health and is part of the Medical Research Future Fund. Our project, Cellular Horizons, seeks to improve patient decision-making about accessing stem cell interventions.
My role in the research group is a legal scholar tasked with wading through the complex regulatory frameworks that govern the administration of innovative cellular therapies in Australia and elsewhere. Part of this role involves analysing regulation; and, more broadly, looking for ways in which patient-focused improvements can be made to the quality of, and access to, clinical information about cellular therapies, on the basis that these improvements will lead to health-maximising and more just outcomes for those who may wish to access innovative medical therapies.
I sometimes describe what I am doing as ‘working towards regulatory justice in medicine.’ The analytical work seeks to demystify and then criticise the operations of complex regulation and propose, where appropriate, the creation of new regulations and legislation — laws that might enable access, reliability (safety and efficacy), trustworthiness, flexibility, fairness, and other features to health consumers.
The theoretical aspect of the talk looked at regulatory theory. There are many species of regulation and multiple ways of describing these types. Sometimes they are described in terms of styles:
- command-and-control: strict rules and penalty enforcement
- responsive: proportionate (least intrusive), flexible, coordinated
- smart: usually minimalistic and ‘in-built’ into existing structures: eg, self-regulation, or an emissions trading scheme; and, finally,
- risk-based: identifies and assesses (or ranks) risks, and prioritises ‘high’ or ‘real’ risks over low or unverified risks, and shapes regulatory tools (inspections, enforcement actions, education, and so on) in response to the risk assessment
In addition to regulatory styles, there are ways of describing regulation in terms of its regulatory mode. A mode of doing regulation may be described as, for instance:
- collaborative: open dialogue and consultation;
- market-oriented: designs regulations around market forces: eg, consider the ‘market mechanism’ often discussed around carbon trading schemes;
- state-centered or centralised: a byword for centralised governmental control through law and the antithesis of self regulation;
- subsidiarised: a mode of regulation that follows subsidiarity principles: ie, where decision-making and regulation occurs closest to the point of activity by those who are closest engaged with the activity; and
- laissez-faire: minimal government intervention; similar to market-oriented, although still prevalent in non-market states (eg, China)
As can be seen, often the mode of regulation is expressed in terms of its operations within the market economy. This, of course, is part in parcel with the historical development of the ‘regulatory state.’ That is, the creation of different regulatory modes has coevolved with the development of modern economic states; and so it should be no surprise that most of these modes are intrinsically tied to orthodox economic theory as it is practised today. That, of course, is also a problem if orthodox economic theory is dysfunctional (eg, high inflation, economic instability), for it means that these regulatory modes will wear the taint of that dysfunction. Indeed, regulation may contribute to or shape the economy. This is trite in terms of financial regulation; however, in terms of therapeutic goods, there are also important effects and impacts (consider, for example, subsidies and rebates for ‘approved’ or registered therapies).
Finally, regulation is also characterised more granularly in terms of its instrumentation. Regulatory theorists thus speak of regulatory ‘tools,’ such as:
- economic tools: taxes, quotas, prince controls, subsidies, competition rules
- authorising tools: licences, accreditations, quality assurance symbols or verifications: eg, medial, pharmacy, and hospital licences
- informational tools: product labelling, required disclosures (eg, certain contractual terms)
- structural tools: physical-environmental changes, rules affecting structures (eg, bans on M&As)
- rules created by traditional legal instruments/tools: these include parliamentary enactments (laws), rules, regulations, even guidances and guidelines (also known as ‘soft law’)
Therapeutic goods regulation: An example of explicit risk-based regulation
When it comes to the regulation of therapeutic goods in Australia, it is obvious that the TGA adopts a risk-based regulation (RBR) approach. Evidence of this fact is exhibited throughout all of the TGA documentation — both in its policy documents and its published guidances and explanatory notes. And the same approach has been adopted by the FDA (US), Health Canada, the UK’s Medicines and Healthcare Products Regulatory Agency (MHRA) and the European Medicines Agency (EMA).
One clear example of the TGA’s RBR approach involves the distinction between the way in which cellular treatments are regulated. In medical terms, there is a very real distinction between allogeneic stem cell transplants, which involve a donor of cells and a recipient of the donated cells, and autologous cellular treatments, where the recipient receives their own cells, which were ‘donated’ at an earlier point in time so that they could be reinfused or re-engrafted later. In terms of the regulation of these treatments, however, there is also a very clear regulatory distinction, which is premised on the relative risks or dangers posed by each intervention. Generally speaking, allogeneic treatments are regarded as more dangerous than autologous treatments. This is readily explained by the medical evidence, as follows.
When a donor provides their human cells to a recipient, there is always a risk of the recipient’s immune system rejecting the cells. This is known as graft versus host disease (GvHD). GvHD arises due to a mismatch between the donor’s human leukocyte antigens (HLA), which are genes that help encode proteins that sit on the surface of cells to detect threats to the human immune system. HLA genes are something like the body’s own car alarm system.
If the donated cells are incompatible with the HLA genes of the recipient, then the immune system will effectively dysfunction; the genes responsible for ensuring the body’s immunovigilance will have been compromised. It might be like trying to install an ‘add-on’ to your car alarm system — something that is said to make it work better. However, if this ‘add-on’ is incompatible with the system, then not only does the system not recognise the ‘add-on’ but it may also stops working altogether, making the car vulnerable to would-be car thieves. GvHD effectively creates these kinds of difficulties for the human immune system, and the symptoms of GvHD can involve multi-organ syndromes of tissue inflammation and fibrosis.
Given the high risk of GvHD, allogeneic cellular therapies are regarded as more risky than autologous therapies. Since a cell recipient’s own cells are perfectly matched to their own HLA genes, there is no risk of GvHD if a person receives an autologous engraftment or reinfusion of their own unaltered, and previously extracted, cells. The HLA genes will be the same.
In recognition of this distinction, allogeneic and autologous therapies are regulated differently, according to their different risks. The diagram below makes it clear.
In a nutshell, the higher risk of GvHD posed by allogeneic cellular engraftments is managed, in a regulatory sense, through the creation of several regulatory tools known as ‘orders,’ which are a species of regulation (usually called subordinate legislation) authorised to be created by the primary piece of legislation regulating therapeutic goods (specifically, the Therapeutic Goods Act 1989 (Cth) s 10). This is the regulatory tool. In terms of the approach or style adopted by this order, we can call it a ‘risk-based’ approach, because it distinguishes or ‘carves out’ allogeneic treatments from autologous treatments.
This order, known as TGO 108 (full title: Therapeutic Goods (Standard for Human Cell and Tissue Products – Donor Screening Requirements) (TGO 108) Order 2021), contains requirements for the donation of human cell and tissue products, including requirements about how the donor’s cells should be ‘screened’ or tested, and how the recipient’s cells can be tested, before the transplant or engraftment can occur.
But how does this instrument carve out allogeneic from autologous treatments? It does so in its ‘Application’ section, which is contained in section 6 of the instrument. It states as follows:
As can be seen, the instrument is disapplied to autologous cellular treatments that are used in medical practice. Specifically, section 6(c), disapplies an HCT (a human cell and tissue) product that is manufactured ‘for autologous use only’ through a process of materials collection (harvesting) that is undertaken by (or supervised by) a medical or dental practitioner for administration to a patient who is treated by a medical or dental practitioner registered in an Australian state or territory.
As an aside, it is perhaps notable that the practitioner identified in subsection (c)(ii) — the collector — does not have to be the same practitioner as the one who administers the treatment — the treating practitioner — under subsection (c)(iii). If the wording of the roman sub-subsections were not enough to demonstrate this, then the pluralised use of ‘practitioners’ in subsection (c)(iv) makes it undoubtable.
The involvement of a ‘collecting practitioner’ and an ‘administering practitioner’ is probably commonplace in practice; and, notably, in the case of the former Dr Ralph Bright — a medical practitioner who is now deregistered for various departures from professional practice — multiple parties were involved in the collection and preparation of the autologous cells that his clinic reinfused into patients. Indeed, according to the coronial report relating to one of Bright’s patients, who died from blood loss during a cellular therapy process, Bright and his wife worked together as the treating practitioner and supervised manufacturer, respectively:
The liposuction concluded at about 10.10am and Sheila was taken to recovery. During this time Dr Bright’s wife, Pelin Bright, processed Sheila’s extracted fat in the clinic’s laboratory and returned the product of that process to Dr Bright, who administered it intravenously to Sheila. A report written by Ms Bright states that 1.5 billion stem cells had been administered to Sheila.
It appears that Bright’s practice was apparently to isolate stromal vascular fraction (SVF) from fat cells (adipocytes in adipose tissues). To do this, liposuction was performed on a patient (usually the flanks of their hips) and then the lipoaspirate (the fat tissues) were treated, most likely through cavitation or sonication, as was reported in the methods section of Bright’s published report on cellular treatments for migraines.
SVF is said to consist of a heterogenous mix of stem and stromal cells, progenitor cells, endothelial cells (ECs), monocyte/macrophages, smooth muscle cells, and pericytes. SVF is more heterogenous than simply a product consisting of mesenchymal stem cells. However, SVF may, and probably will, contain some mesenchymal stem cells too. in tbhis regard, it is notable that the material extracted from the patient who died was the subject of a written report (a form of screening) in which it was stated that ‘1.5 billion stem cells’ had been administered. But the point is that this kind of treatment, if SVF was extracted, would not be well described as a ‘stem cell treatment’ because there are other materials present in the SVF preparation. Indeed, some of these other cellular materials are thought to form part of the therapy.
Even through the dangers associated with injecting SVF back into a patient autologously are largely unidentifiable at this point (after all, clinical data is sparse; but here is one relatively recent study), the fact that the engraftment of adipocyte-derived SVF will not generally cause GvHD is enough to mean that the TGA is satisfied that a lower risk of harm is posed by the material. And this lower risk of harm means that the screening requirements under TGO 108 will apply to allogeneic cellular treatments but will not apply to these autologous cellular treatments. And so this is one clear way in which the TGA has ‘managed out’ the regulation of autologous cellular treatments, and specifically ‘managed out’ the cell screening requirements that would otherwise apply to human cell and tissues (biologicals) used in medical practice. And the reason for this risk-based regulatory distinction is that the risk of autologous cell transplants has been deemed ‘not high.’
Criticism of RBR
Of course, risk-based regulation generally has been much criticised since the early 2000s. When you say things are not subject to regulation by exempting or excluding them from the general rules, you basically reduce the size of governmental or regulator oversight. Needless to say, many policy theorists regard this as a good thing. Below are a few key quotes that explain the theory and thinking behind RBR and hint at why it might generate governance problems:
One of the problems raised by RBR critics relates to the fact that, in ‘managing out’ things that are perceived as ‘low risk’ — phenomena pre-assessed as unlikely to give rise to harm — governments effectively vacate a space and put on record that they are not responsible for regulating that particular thing. This might be called deresponsibilisation.
And, of course, the original risk assessment could be erroneous and the unregulated activity may, in reality, cause great harm and do so often. That is one pitfall of RBR: getting the risk assessment wrong. In one sense, it is a kind of failure of regulation. It involves the following steps:
- a risk plausibly within the purview of a regulator is assessed by lawmakers or regulators to be low probabilistically — both in terms of risk of incidence and degree of harm;
- because of this assessment, the risk is explicitly determined to be outside a government’s or agency’s purview (after all, efficiency demands that resources should not be directed towards managing such low risks);
- that risk assessment, however, is later shown to have been unjustifiably ‘managed out’ — either because it is actually a ‘high risk’ (dangerous in substance — causing a high degree of harm) or is actually likely to occur (has a high risk of incidence)
If this is one foreseeable consequence of RBR, then clearly the risk assessment has to be robust. But, even if that is so, failures may still arise. And they have done. And so the hypothetical question then becomes: how has RBR been justified when it is prone to fail? The answer is insightful. As critics such as Henry Rothstein and even the likes of Niklas Luhmann have argued (in different ways), when governments and regulators ‘manage out’ certain activities as low risk, they actually resolve a longstanding dilemma of regulation: namely, imperfect control.
The old problem might have been to control risks perfectly by making the right rules and allocating the right powers to government. But, under the RBR framework, only those rules that are necessary to manage the high risks will be implemented. The responsibility to control all risks is relinquished, and governments essentially vacate the space when it comes to activities assessed as low risk.
In short, government are able to defend their imperfect control of risk in society by saying, ‘Not my problem!’ when something inevitably goes wrong. And things will ‘inevitably’ go wrong because the risk was never eliminated to begin with and it still exists — to the knowledge of regulators. Instead, the risk was just ‘managed out.’ It is, according to the instruments of RBR, ‘someone else’s problem.’
Rohstein captures something of this analysis:
Now, in the case of therapeutic goods, this may mean, for example, that the TGA, the official federal regulator, ends up not regulating autologous cellular treatments. And one way in which it does thus delegate or ‘manage out’ its regulatory duties is captured in the example of the TGO 108 furnished above.
The reason RBR seeks to ‘deregulate’ in this way (or, better put, to ‘deresponsibilise‘) has a lot to do with its origins. As a legal and historical phenomenon emerging in the 1990s and 2000s, risk-based regulation initiatives sought to achieve policy goals that were fiscally conservative (aiming for economic efficiency), generally deregulatory (removing administrative burdens), and opposed to so-called compensation culture’ (that is, seeking to limit liability and provide certainty to businesses and government actors). There is ample evidence to confirm these political-institutional origins of risk-based regulation, which could be summed up as a more practical description of what it means to effectuate ‘small government.’
In the US, the Reinventing Government and New Public Administration movements of the 1990s spearheaded RBR; and, in terms of therapeutic goods, these initiatives clearly led the FDA biologics division, CBER, to adopt rules that reduced reporting obligations for the manufacture of biologics, as was reported in proceedings of the regulation of medicine at the time. In the UK, the publication of the Hampden Report in 2005, which led Tony Blair to promise an end to ‘compensation culture’ and to remove regulation seen to burden those engaged in low-risk activities. And in Australia, the publication of the Ipp Report of 2002 and the consequent enactment of the Civil Liability Acts, significantly limited the liability of those who committed negligent acts and harmed others. All of these documents and historical reforms illustrate the increasing adoption of risk-based approaches to regulating harms in civil society, including in terms of professional practices such as medical services and the provision of therapeutic goods.
RBR pervades the therapeutic goods law in several shapes and forms
In addition to the TGO 108, there are four other essential ways in which the therapeutic goods legislation and other subordinate instruments risk manages cellular interventions (ie, biologicals) in Australia. While I will not identify them in full here, they may be summarised as follows:
- One instrument ‘artificially’ nominates some ‘safe’ human cellular products that meet the definition of a biological to be treated as therapeutic goods (ie, medicines) and thus excludes them from the framework for biological regulation;
- The primary Act requires ‘approved’ products (those that are registered on the ARTG for marketable supply) to be registered in one of four classes, and sets out, in subordinate guidances, different evidence (data) requirements for each class of biological, according to their risk;
- A schedule to the Therapeutic Goods Regulations 1990 (Cth) provides that certain therapeutic goods are, as low risk products, exempted from the framework for biologicals on certain conditions: namely, if they are used in a single (one-off) clinical procedure for a single clinical indication; and, finally
- One ‘determination’ — another form of subordinate legislation, like an ‘order’ — provides that certain products, including products that would otherwise be medium risk products according to the registrable classes of biologicals (Class 3), may be administered in ‘hospitals’
Obviously, for the purposes of this blog post, I will not go into the details of each of these regulatory carve outs. All of this will be covered in a detailed academic article. Suffice to say that the regulation of autologous cellular therapies is riddled with exceptions that allows medical practitioners to continue to administer them without much oversight from the TGA in certain circumstances.
The reason that the TGA gives for not regulating autologous treatments — or for ‘excluding’ those treatments — is articulated in various places, but most authoritatively in the Australian Guidelines for Biologicals (ARGB) — a soft law guideline for sponsors/manufacturers — where the it is explained as follows.
Clearly, the risk of these treatments is managed out to ‘other bodies,’ as the second paragraph notes. But these other bodies are not specified. They would have to include those bodies responsible for dental or medical practitioner governance, including AHPRA (and the Boards), LHDs and LHNs (including their governance and ethics committees), state-based complaints entities such as NSW’s HCCC or the Queensland Health Ombudsman, and the ACCC. The question then becomes whether these bodies are sufficiently well-equipped to regulate the risk of these cellular treatments.
And, of course, the answer to that question may depend on what bodies are involved and at what stage of the activity.
- If an Local Health District is involved, what is their governance arrangement? And, where no LHD is involved at all (for example, LHDs do not generally have oversight of private clinics or small private hospitals), then what arrangement, if any, are in place to oversight treatments in these clinics?
- Is the state regulator of medical or dental practitioners (such as, in the case of NSW, the Medical Council) equipped to discipline any practitioner ‘experimenting’ on patients with cellular therapies who might cause harm? For example, could they take immediate action to suspend the practitioner under s 150 of the National Law? Or, outside of NSW, could the relevant state Medical Board take immediate action under s 156 of the National Law?
- And this question may in turn depend on whether the case law permits the relevant practitioner regulator to exercise these disciplinary powers broadly or narrowly. For instance, our analysis of a 2022 decision of the NSW Supreme Court’s Court of Appeal showed that the courts may limit the scope of powers exercisable by the practitioner regulator in some circumstances, limiting their ability to regulate.
- Finally, can the consumer complaints entities prosecute the practitioner as a health professional (eg, HCCC) or prosecute the practitioner’s business for false and misleading advertising (ACCC)?
When it comes to registered or approved biological products that might be described as ‘proven’ autologous cellular therapies in light of a TGA assessment, there are essentially none. That is, there appear to be no autologous stem cell therapy products registered on the ARTG. AT the moment, only 44 biologicals are registered in total, and these comprise either low-risk (Class 2) products (eg, musculoskeletal tissue) or high-risk (Class 4) products that involve an artificial function: ie, human cellular materials that have undergone functional manipulation and a form of cellular ‘programming’ through the introduction of an adenoviral vectors (AVVs) — otherwise known as gene therapies.
There are no cellular products that are registered that have undergone any lesser form of manipulation than this ‘high-risk’ and ‘artificial function’ manipulation. In other words, there are no cellular therapies on the register of therapeutic goods (ARTG) that involve the isolation of stromal vascular fraction (SVF) or the culturing of mesenchymal stem cells. For many years, these were often described in advertisements by small clinics and in medical news stories as innovative ‘stem cell treatments’ or ‘regenerative medicine.’ If any such treatment was to be registered as an approved product, it would presumably be a ‘medium risk’ product registrable in Class 3.
It is perhaps not unthinkable that ‘off the shelf’ treatments will appear one day. But it would require a manufacturer to work out a way to scale-up autologous cells in such a way that it would be a viable (or profitable) product that was also sufficiently supported as a safe and effective treatment for the purposes of an application dossier to be submitted to the TGA pursuant to registration on the ARTG. There are business in the US that are trying to enable such products to scale-up or scale-out through improving cell extraction technology. It is probably inevitable that the evidence base for potential therapies in this category will advance to the point where registrable products are manufactured.
Until then, however, the only way in which we can detect when and why unapproved cellular therapies that appear to posed a medium risk are administered, probably lawfully, in hospitals and in one-off clinical procedures, is when something goes wrong. This is by design, because the risk-based approach of the TGA means that they are not responsible for these treatments. And, by the same token, hospitals and clinics have no apparent responsibility to report on when these treatments are administered, other than in the case of adverse events.
Another aspect of the risk-based regulation of cellular therapies is the strict ban on advertising biologicals. As of 2018, any advertising of biologicals is effectively prohibited. There are no loopholes or conditions here, because even the exemptions and exceptions are conditional upon there being no advertising. As soon as a biological is advertised, in any context, it is returned to the regulatory purview of the TGA — even if the biological would be otherwise exempt from TGA oversight.
However, if patients are not subject to advertising, and they are persuaded in a consultation to try an alleged stem cell treatment (or, as appears to be more common, a cellular treatment involving the engraftment of SVF isolated from autologous cells), those patients will remain vulnerable to practitioners who may wish to advance their research or exploit patients financially.