UCL School of Life and Medical Sciences


CASMI Blog Posts

Emerging Challenges in Research Ethics |  Professor Sir Jonathan Montgomery


CASMI Fellow, Jonathan Montgomery, had an opportunity to share his thoughts on emerging challenges in research ethics with members of the Ministry of Defence Research Ethics Committee in January. He identified a series of trends that placed the traditional paradigm under pressure.

The closer integration of researchers and participants whereby projects were co-designed rather than imposed on ‘human subjects’. This goes beyond the previous shift to the language of participation into more equal partnerships. This suggested that ethics committee reviews of recruitment processes and consent forms would need to take into account that those being ‘recruited’ had been part of the design. There were particularly interesting challenges in relation to participant-led research as explored by an article in the Journal of Medical Ethics on which he and a number of CASMI Fellows were co-authors. It would also become necessary to reconsider the idea that ethics committees had to protect participants from research that was ‘too risky’. He suggested that the idea of a ‘fair offer’ as raised by the Nuffield Council on Bioethics (which Jonathan chairs) in its report Children and clinical research: ethical issues (2015).

Increasing expectations of personalisation, which Jonathan, along with CASMI Fellows John Tooke, Rob Horne and Morten Ravn have suggested will require a new social contract for medical innovation. Reflecting on this, Jonathan suggested the personalisation agenda would bring to the fore some of the intrinsic challenges of evidence–based medicine; including the way in which it tended to present inferences as if they were deductions, the mismatch between study cohorts and patients in clinics, and the uneasy relationship with ideas of autonomy whereby patients are entitled to take decisions based on non-scientific issues. Jonathan suggested that there was also an intrinsic paradox within the drive towards personalisation, in that it relied on good quality population data, from which the differentiation of individuals could be derived by analysis of variations. This linked to the next trend…

The dominance of data in analysis sits uneasily in a paradigm of research ethics that is based on addressing the risks of new pharmaceutical interventions (in the shadow of Thalidomide and other similar problems). That traditional paradigm had led to excluding various categories of data-based research from research ethics committee (REC) oversight – such as audit and service evaluation – and also inadequate governance for innovation. The governance of surveillance and the ethics of ‘big data’ would become increasingly important. Indeed, the Government had announced its intention to establish a Council for Data Science Ethics. Amongst the consequences of this was that…

There are a number of ‘new kids’ on the research block who are not familiar with the existing structures for research governance – as had been exposed by the late realisation that the collaboration between the Royal Free and Google’s Deep Mind needed regulatory approvals. This might not be as novel as it first appears; the idea that medical research is done by doctors has long been something of a myth. Clinical trials are industrialised in terms of scale, funding and production – using experts in relevant disciplines (e.g. statistics, bioinformatics) who are not subject to the regulatory sanctions that bind health professionals. Perhaps even more importantly, those newer players do not necessarily subscribe to the ethical commitments that developed through the World Medical Association and are enshrined in the Declaration of Helsinki. The multiplicity of professions and industries from which these players are drawn, together with the integration of research participants means that it may be necessary to re-establish, and possibly negotiate, the norms and regulatory authority of our research governance structures.


Description automatically generated

Research ethics has drawn a significant part of its strength from reflection on the differences between care and research, such as the conflict of interests between individual and society. However the line between care and research is under pressure. In different ways, the demise of the ‘blockbuster’ model of drug development, the needs of payers to assess value for money (as well as safety and clinical effectiveness), and the appetite of patients for accelerated access to novel treatments all bring the ‘real world’ closer to the clinic. This manifests itself in various ways; the assertion of the ‘right to try’ unproven therapies, the rebalancing of what is anticipated to be measured prior to market access being granted and what will be learnt from closer monitoring in clinical use.

Looking to the future, research ethics will need to pay closer attention to its contribution to the problem of lack of evidence. Existing structures and concepts are designed to cope with problems of wrong-doing. Norms (at least since Nuremberg) have developed in response to evil more than in pursuit of virtue. The gatekeeping functions of research ethics committees (and institutional review boards) seek to prevent ill-conceived studies being commenced. Research governance is aimed to ensure that the conduct of studies meets the necessary standards. This precautionary approach has made it difficult to research in some areas, for example medicines for children. The lack of the evidence we need to improve care and provide therapeutic options is a significant challenge that research ethics needs to address. We need a stronger ethical framework for the right innovation and research, not just a set of barriers that constrain poor practice.

These trends require a concerted effort to build a constructive research culture that addresses the widespread distrust in what many perceive as an enterprise driven by profit and competition rather than searching for sustainable medical innovation. Research needs to be more collaborative and to demonstrate more clearly why people can trust results; transparency, reproducibility, management of competing interests, and responsiveness to community needs can all be enhanced. Such concerns are inadequately integrated into our current thinking about research ethics.

– Jonathan Montgomery

CASMI Academic Agenda: A focus on ‘pull’ factors | Professor Sir John Tooke


CASMI’s mission is to help make the medical innovation process more efficient and sustainable, recognising that there is much unmet clinical need to address and the development and adoption of new drug solutions are slow and costly. Since inception we have seen the medical innovation process as a circular one (Figure 1), which starts and ends with the patient – from recognising the patient’s clinical challenge that needs to be addressed to ensuring that the patient has access to and is supported in adhering to the resultant treatment as appropriate.

Conventionally academic effort has been principally focused on the biomedical quadrants of the medical innovation circle – for example formulating the research question, conducting discovery science that may lead to new insights into disease mechanisms, the identification of potential drug targets and proof of principle, and of course clinical trials of resultant products, often in conjunction with industry. These essential contributions, at which incidentally the UK is very good, help ‘push’ new products towards the market place. It is critically important that such work is done as efficiently as possible with greater attention paid to the forging constructive alliances between academia and industry which clearly recognise the contribution made by both parties, and earlier, more definitive proof of concept studies to limit late stage attrition. But CASMI has recognised that greater attention also needs to be paid to what might be referred to as the ‘pull’ factors – the reception, purchasing, adoption and diffusion, and ultimately taking of and adherence to new medical solutions. CASMI’s Medical Innovation Academic Consortium (MIAC) has a wide range of research interests in this area best summarised as questions relating to the perception of risk and value at the level of the individual (citizen, patient and prescriber), system (e.g. the NHS), and the regulator.


Description automatically generated with medium confidence

Figure 1 – The Innovation Cycle

One important facet of this ‘pull’ agenda is how we all judge scientific evidence on the potential benefits and harms of new medicines. Triggered by concerns about the suggestion that treatment thresholds for the use of statins to prevent cardiovascular disease should be lowered, the Academy of Medical Sciences has been involved over the last eighteen months in trying to make sense of this complex agenda and come up with recommendations that improve decision making and the confidence of those making decisions in the future. Several CASMI Fellows have played important parts in the various work streams which have involved i) consideration of the methods used to generate evidence and their relative value and limitations, ii) how we might enhance trust in the resultant evidence, and how it is best communicated to both healthcare practitioners and of course patients – all underpinned by extensive public engagement including surveys and deliberative dialogue. It has been my role to chair the Oversight Group that will attempt to bring this all together, easily the most complex exercise of its type I have been involved in! We plan to publish in the Spring of 2017 and one of the outputs that I hope it will stimulate is the generation of a research agenda that reflects important ‘known unknowns’ and the commitment of funders to support efforts to address it.

It has always struck me that if the focus of academia remains firmly centred on ‘push’, the early stages of development of a new medicine and resultant trials, we are only contributing to half the job. Our contribution needs to be holistic if the efficiency and sustainability of the innovation cycle is to be optimised, drawing on expertise from well beyond the conventional biomedical fold. The late social historian Roy Porter at the end of his treatise on medical advances over the centuries concluded: ‘Medicine has led to inflated expectations which the public eagerly swallowed. Yet as those expectations were unlimited they were unfillable: Medicine will have to redefine its limits even as it extends its capacities.’ To defy that somewhat gloomy prediction and reap the full rewards of modern science for patient benefit the academic community will have to help resolve the many factors that impact on the take up of new medicines by society – the ‘pull’ to complement the ‘push’.

– John Tooke

Unmet needs in disease classification | Dr Stuart Calimport


Fundamental to the innovation process is the identification and assessment of unmet needs. Numerous areas with unmet needs have already been identified and assessed as high priority including therapeutic development and manufacturing, biomarkers, clinical tools, and focusing on patient populations and patient outcomes. However, there is currently one unmet need, which could be considered high priority and high impact that it might be reasonable to take the view that it has been somewhat neglected, globally[1]. This is the area of disease classification, including the classification of syndromes, particularly relating to ageing and ageing-associated pathologies. Currently, disease classification and maintenance falls primarily on the World Health Organization, medical societies and the pharmaceutical industry, and it is not currently a widely studied or grant-funded area in academia. Importantly, disease and syndrome classification is the foundation of clinical trials and clinical practice, given its importance in clinical trial regulation, diagnosis, treatment and reimbursement.

One might take the view that in order for potential clinical solutions to be better targeted, success or failure of therapeutics to be measured, and patient outcomes to be measured and maintained, that improved accuracy of the characterisation, definition and measurement of pathology is required, and for pathologies to be adequately formally recognised. Where characterisation and measurement of ageing-associated pathology appear to be widely appreciated challenges in academia and medicine, it might be reasonably considered that insufficient and inaccurate definition and subsequent classification of ageing-associated diseases and syndromes may cause numerous issues. A lack of accurate, rationalised, systematic and comprehensive ageing and ageing-associated disease and syndrome classifications may stymy the potential of medical innovation, prove an innovation blocker, and indeed risk stultifying the fields of preventative medicine, gerontology, pathology and epidemiology, whilst risking further delay to the promise of personalised medicine.

In terms of delivering solutions to this challenge there is a need for ageing and ageing-associated metabolic and accumulative degenerative change to be formally classified as pathologies, syndromes and diseases. As to whether ageing and ageing-associated metabolic and accumulative degenerative change can be accurately defined in disease terms, although some progress has been made, it has been disjointed and irregular[2], proceeding in a somewhat limited and non-coherent manner, lumping and splitting the ageing pathology as stakeholders related to a particular clinical specialism saw fit at the time, based on their knowledge, expertise and judgement[3]. Sometimes ageing and ageing-associated metabolic and accumulative degenerative change is classified at the level of the organ, sometimes due to internal or external causes, sometimes on the basis of symptoms, sometimes through coverage by a broad syndrome, however the majority of the time not at all. In this way, the World Health Organization International Classification of Diseases (ICD) codes as they exist provide a selection of proven comparators for what could reasonably be seen as a high priority and high impact ‘filling in the gaps and rationalisation’ exercise. It may also be the case that some ageing and ageing-associated accumulative and degenerative change already classified has been over-lumped together as one syndrome when splitting them out into further classifications may assist diagnosis, treatment and access to care. As it stands one could diagnose and reverse some or all ageing and ageing-associated degenerative change in one organ, and get reimbursed for doing so, but not another, unless doing so indirectly.

It might be thought reasonable that it should be a keen area of academic interest and work, as it offers opportunities to publish, collaborate with clinicians, directly affect patients globally, and make a historically-important contribution. If we wish to consider the future of medicine, and what that might hold, one could consider the coverage and resolution of disease and syndrome codification to be a lynchpin. If we can reach a stage in medicine whereby ageing and ageing-associated degenerative change can be fully coded up, in combination with high-resolution disease-code linked digital twins, we can then start to consider what the future of complex multi-modal interventions and prescriptions might look like. One might consider oncology coding, which is comprehensive and organ specific down to the cellular level, with accompanying precision and personalised therapeutics, as a look forward as to what the area of healthy longevity medicine could look like with improved classification and codification in the near term.

Given our knowledge of ageing[4], and previous ageing and ageing-related classifications in the ICD, it may be apt to consider that it is entirely feasible to embark upon the systematic and comprehensive definition, classification and codification of ageing and ageing-associated degenerative change. Further, given the current and forecast challenges of ageing populations and the burdens of ageing-related frailty and diseases, such activity could be seen as timely from a healthcare and economic perspective as well as from a position of scientific and clinical knowledge, in addition to any moral case.

Many times within the innovation process questions arise such as ‘Is this even a problem?’ ‘Is this problem worth addressing?’ and ‘Is the problem timely to solve?’ In relation to the innovation challenge of classifying ageing and ageing-associated degenerative change I hope to have sufficiently contented these questions and stimulated interest from the Collaboration for the Advancement of Sustainable Medical Innovation community and beyond.

Future medical innovation: technological advances alone are not enough, we must address health inequalities | Professor Sir John Tooke & Professor Robert Horne


The rapid development and production at scale of effective vaccines for Covid-19 is a testament to the efficacy of the medical innovation eco system. The speed of translation from laboratory to the clinic, including the successful integration of academic research with industrial innovation, shows that much can be achieved when innovation and investment are appropriately geared.

Responses to Covid-19 also provide encouraging examples of how different elements of the medical innovation ecosystem can work in synchrony to deliver innovation at pace. This included parallel processes for development, production, and approval, representing innovation in regulatory science, as well as impressive public health programmes to engage the public and deliver timely vaccination at scale, reducing death rates.

However, despite these evident successes, the Covid vaccination story has also highlighted a major fault-line in the medical innovation system: inequality of access, both between countries and within national populations. It is clear from death rate figures for individual nation states that there is huge disparity between high vs low vaccination rate countries and that vaccination maps onto wealth1

For Covid this has negative potential consequences for poor, low vaccination-rate counties as the virus lingers for longer, but also for wealthier countries. For high vaccination rate countries, the risk of emergence of new variants elsewhere and their onward transmission prolongs the health and economic costs of the pandemic.  In this respect we truly are ‘all in it together’, and as many authorities have commented advanced economies cannot relax until the global threat is diminished

Whereas it is perhaps human nature to seek and prioritise advantage for self and one’s own group, the Covid-19 pandemic illuminates the limitations of this approach. The capacity for ‘those that have’ to gain and retain the full benefits of ‘having’ is limited by inequality of access For wealthy countries to invest more in addressing health inequalities is not only a humanitarian imperative but also a pragmatic necessity or as expressed by Jeremy Farrar, Director of the Wellcome Trust: ‘enlightened self-interest, to sustain and deliver the health benefits for their citizens2.

It is well recognised that health inequalities, between and within countries, significantly reduce the health, well-being and lifespan of individuals3,4. The costs are huge in humanitarian, but also economic terms.  This is illustrated by an analysis of the differential effect of the Covid-19 pandemic in England between areas of high versus low economic deprivation5. Covid-19 accentuated the relative deprivation of the North of England. An extra 57.7 people died per 1000,000 and the Northern region was hardest hit in terms of mental and financial well-being and loneliness. It is estimated that the impact of reductions in mental wellbeing alone could cost the UK economy up to £5 billion in reduced productivity.

The NHSA report is a stark reminder that the origins of health inequalities reflect the major role played by the social, behavioural and environmental determinants of health that map onto deprivation indices, and not just access to technology. Deprivation was linked to higher COVID-19 infection rates, cases, case severity and deaths before the development of vaccines. Inequality increased vulnerability, susceptibility, exposure, and transmission. For example, the higher burden of pre-existing health conditions increased vulnerability, immune systems weakened by adverse environmental conditions, increased susceptibility, inequality in working conditions increased exposure and inequality in housing increased transmission rates5.

Beyond Covid, in countries that lack Universal Health Care or where access is limited by an individual’s financial means, inequality of healthcare provision is perpetuated across the full spectrum of physical and mental ill health. Despite providing some of the best healthcare in the world for those that can access it the lack of Universal Health Care in the U.S.A. is associated with substantial health disparities, with low socio–economic status segments of the population subject to decreased access to quality healthcare and increased risk of non-communicable chronic conditions such as obesity and type II diabetes, among other determinants of poor health6.

Despite impressive gains from medical innovation, our current ecosystem is not well placed to address the health inequalities that limit its impact on population health and wealth. Current trends in medical innovation might even accentuate health inequalities. The pharmaceutical industry, which deserves great credit for ‘stepping up’ in the Covid crises, is a major force in medical innovation. However, economic models that stimulate a focus on narrower disease indications and subsets risks the availability of cheap, deployable solutions for global health problems.

Moreover, future developments have the potential to further increase the health inequalities. Combined advances in genomics, data analytics and e-health technologies offer the potential for P4 medicine that is predictive, pre-emptive, personalised and participatory. Technological advances will improve our ability to predict the emergence of disease even before symptom development, identifying individuals at risk and targeting pre-emptive treatment tailored to personal variations in drug responsiveness. In the P4 concept, participation is effected by the capacity of digital technologies to empower and involve individuals. However, achieving equality of access is a fundamental challenge to the P4 Utopia. If personalised therapies have a more discrete market size, how do we ensure that inequality in access to P4 medicine does not follow on from, and add to, existing health inequity? One solution alongside population level initiatives is to extend the need for ‘precision’ to the means of accessing high risk populations and individuals however deprived or marginalised and availing them of approaches that reflect their personal circumstances7.

In mature democracies awareness of the prevalence and impact of inequality is increasing to the point where intolerance of inequality should be reflected in national policy provision across domains. Again, on the international front, it could be argued that it is also ‘enlightened self -interest’ to adopt policies which limit the resultant capacity for social unrest and health and economic migration.

The challenge therefore for medical innovation is to consider holistically how it impacts the unacceptable persistence of health inequalities. This should involve developing methods for closing the equalities of access gap that do not diminish the likelihood of producing timely technological advances at scale in an environmentally and economically sustainable manner. It also requires a rebalancing of the precision medicine construct better to recognise the wider determinants of health and approaching them with the same precision, energy and ingenuity that has been afforded to genetic determinants.

The nature of the challenge and the need for a global solution is perhaps analogous to that required to deal with the climate emergency. Effective solutions will require multidisciplinary, pan-national approaches. To date, our history suggests that our ability to create technological advances greatly exceeds our ability to produce the social and economic systems necessary to achieve equality of access and address social determinants of health. We need to act now to reverse this trend and to develop a medical innovation system that plays its full part in reducing health inequalities as well as securing technological advances. We call on all those involved in medical innovation to respond to this challenge and debate to ensure that the medical innovation ecosystem evolves in a way that embraces the learnings from the Covid-19 pandemic.


1.            Kirk PGA. Vaccine inequality: how rich countries cut Covid deaths as poorer fall behind. The Guardian. 2021 28 June 2021.

2.            Looi M-K. Jeremy Farrar: Make vaccine available to other countries as soon as our most vulnerable people have received it. BMJ 2021; 372: n459.

3.            Marmot M. Social justice, epidemiology and health inequalities. Eur J Epidemiol 2017; 32(7): 537-46.

4.            Marmot M. Social determinants of health inequalities. The Lancet 2005; 365(9464): 1099-104.

5.            Bambra C ML, Alexandros A, Barr B Brown H, Davies H, Konstantinos D, Mason K, Pickett K, Taylor C, Taylor-Robinson D, Wickham S. Covid-19 and the Northern Powerhouse: takling inequalities for UK health and productivity, 2021.

6.            Zieff G et al. Universal Healthcare in the United States of America: A Healthy Debate. Medicina 2020; 56 (11): 580.

7.            Horne R, Bell JI, Montgomery JR, Ravn MO, Tooke JE. A new social contract for medical innovation. Lancet 2015; 385(9974): 1153-4.


Our Aim

Create a diverse collective of emergent and established leaders in medical science, health innovation policy and practice to begin the process by:

Raise awareness of the relationship between health inequalities and medical innovation Improve our understanding of the nature and effects of inequality of access and the capacity of medical innovation to both exacerbate and address health inequalities Develop solutions to inform a medical innovation ecosystem that improves health inequalities

Health Inequalities and Medical Innovation – A Response | Professor Rosalind Raine


This piece eloquently highlights the hollowness of the mantra that we are ‘all in it together’, even in countries with universal health care. Professors Sir John Tooke and Rob Horne helpfully refer to the “major role played by the social, behavioural and environmental determinants of health inequalities” and the need to achieve “equality of access”. If we genuinely wish to reap the benefits of P4 medicine across the population, then we need to unpack these statements.

It is fundamentally important to recognise the scale of the barriers faced by socially disadvantaged groups in achieving health equity and the entrenched nature of these barriers.  To focus on behavioural blocks may be politically attractive but it misses the point. The reality is that choices are constrained and that “men make their own history, not of their own free will, but under given circumstances with which they are confronted”[1]. Multiple, intertwined structural, financial and institutionally discriminatory obstructions interact with community (e.g. cultural) and individual (e.g. psychological and informational) barriers. Tackling these needs action at (inter)national policy, regional and individual levels.

(Inter)national action is bound up with politics, but there are actions we can commit to now to address the authors’ challenge.  The first of these is to shift the debate from inequality to equity. Equity is about fairness and justice and indicates that everyone should have an equal opportunity to attain their full potential for health or for the use of health care[2]. The notion of equity transcends equality which can lead to unfair access to care. If we agree that achievement of equity is key, then a flurry of subsequent questions arises.  For a start, different ideologies (libertarian, liberal and collectivist) lead to different definitions of fairness and even if these definitions can be agreed (generally at societal level), consensus around one’s ‘need’ for health care will be difficult to realise.  There is also the thorny question of achieving both horizontal and vertical equity. Horizontal equity refers to the equal treatment of those with equal needs and vertical equity recognises that people with greater clinical needs should have more intervention. But whether meeting such needs requires positive discrimination is a contentious issue. Transparent debate and a clear understanding about how decisions are made by those in power are both essential.

So we need to listen: to citizens and patients; practitioners, planners and policy makers; front line staff and senior leaders; industry and the public sector. They have different perspectives about what needs to be done, what their priorities are and the value they place on these. Only by working hand-in-glove, in a sustained way with all these stakeholders, can we begin to design and implement a system to enhance equity.  We must also commit long term and sufficient resources, in terms of time, money, data system development and, crucially research, with the power to measure and monitor inequalities in access, use and outcomes across social groups, geography and conditions. In addition, we need to invest in our workforce so that it reflects the populations it serves (and therefore is better able to productively engage and to understand their needs and beliefs); has the skills to deliver the ‘P4 Utopia’ from bench to bedside to community; and (with particular reference to building capacity in LMICs), feels sufficiently valued to remain in their community rather than taking their valuable skills elsewhere.

We can only address inequalities fairly once we have measured them. There are (at least) three issues to think about here. First: the domain with which we are concerned. Inequalities exist between medical conditions (and between health care and population health interventions), in part because of the differential focus pharmaceutical industries, governments, research funders, providers and citizens place on different conditions and on health care/public health. I suggest that we should take care to avoid prioritising P4 medicine on those conditions which benefit from effective lobby groups (who may operate as ‘patient groups’ but funded by industry) rather than those which could achieve the greatest population impact. Furthermore, whilst work on socio-demographic inequalities appropriately highlights socio-economic, racial, gender and age inequalities, people with disabilities have been woefully neglected. Finally, geographic inequalities in access to both care and to clinical research need to be tackled at both national (e.g. rural and coastal versus metropolitan) and global levels.

Second, I also propose that we shift from addressing socioeconomic (or age) gaps in access to precision medicine, to ensuring that we tackle the inequality gradient.  This is because an inherent danger of drawing a line around those ‘in most need’ is that a larger group of those who are also in need may be excluded, resulting in a general widening of inequalities. It is hard to reduce inequalities overall without compromising uptake in any socio-economic group and harder to reduce inequalities more in the most socially disadvantaged groups than in the most privileged groups. But we have shown that it is possible[3]. 

Finally, we need to recognise that equity of access to P4 medicine is insufficient. We need a comprehensive approach which maps out relevant health care pathways to ensure that equity prevails in the identification of risk, as well as in each management and treatment component across hospital, primary care and community settings.

This takes committed leadership, time and a lot of work. But our efforts will not be wasted because it is the most marginalised in society who have the greatest needs for health care and population health and in whom the greatest impacts can therefore be achieved.


Professor Rosalind Raine



1.  McKee M, Raine R. Choosing health? First choose your philosophy. Lancet 2005;365:369-71.

2.  Wardle J, Von Wagner C, Kralj Hans I, Halloran S, Smith  S, McGregor  L, Vart G, Howe R, Snowball J, Handy G, Logan R, Rainbow S, Smith S, Thomas M, Counsell N, Morris S, Duffy S, Hackshaw A, Moss S, Atkin W, Raine R. Effects of evidence -based strategies to reduce the socioeconomic gradient of uptake in the English NHS Bowel Cancer Screening Programme (ASCEND); four cluster randomised controlled trial. Lancet 2015; DOI:https://doi.org/10.1016/ S0140-6736(15)01154

3.  Raine R, Or Z, Prady S, et al. Evaluating health-care equity. In: Raine R, Fitzpatrick R, Barratt H, et al. Challenges, solutions and future directions in the evaluation of service innovations in health care and public health. Southampton (UK): NIHR Journals Library; 2016 May. (Health Services and Delivery Research, No. 4.16.) Essay 5. Available from: https://www.ncbi.nlm.nih.gov/books/NBK361257/ doi: 10.3310/hsdr04160-69