The pharmaceutical industry needs more clinical pharmacologists

Clinical pharmacologists play an important role in drug development. They assess whether drug candidate molecules have the properties required of a successful medicine; contribute to the development of clinical development plans; take accountability for the design, conduct and analysis of phase I clinical trials including first‐in‐man studies; guide the choice of appropriate doses at all stages of development; undertake studies to determine how to guide prescribing in the real world; contribute much of the information in a summary of product characteristics and package leaflet/insert; write several sections of the filing documents submitted for approval by competent authorities; help answer questions from authorities; support the development of new formulations and new indications for approved drugs and continually monitor emerging information about approved drugs to ensure they continue to be prescribed in as safe and effective manner for as many patients as possible. Most pharmaceutical companies have departments of clinical pharmacology to undertake this work, in partnership with colleagues in contract research organizations and academia. The current challenges and opportunities the pharmaceutical industry faces provide more opportunities for clinical pharmacologists to contribute, and we may need new skills to best make that contribution.

Improving drug development success rates

Drug development is high risk and expensive with estimates for the out‐of‐pocket costs of bringing a drug to market of $0.9–1.4 billion (fully capitalised $1.5–2.6 billion) 1, 2, 3, while the chances of successful development of a new clinical candidate from the start of good laboratory practice (GLP) toxicology through to eventual regulatory approval are 5% 4. Almost 80% of the cost of bringing a drug to market is spent on the molecules that did not succeed; relatively small improvements in success rates, especially in expensive clinical development can substantially decrease the total cost of drug development 1. The principal reason for development attrition in clinical trials is lack of efficacy; it accounts for the majority of development failures in phase II, and even in phase III trials there is still a failure rate of one‐third to a half with over half of this being for lack of efficacy 5. When one considers that many projects terminated for ‘strategic’ or ‘commercial’ reasons are because there was inadequate efficacy relative to existing treatments, lack of adequate efficacy could account for as many as three‐quarters of all late phase failures.

Failure for lack of efficacy may be because the underlying molecular target is in fact unconnected to the disease pathogenesis. However, it is also possible that heterogeneity in the disease, the patients and the drug effects means that effective treatments are being missed in clinical trials. Suppose a new drug improves survival from 50% to 65% in an unknown 20% of the population but has no effect in the rest of the population, whose survival remains 50%. The effect in the whole population is an increase of just 3% (0.15 × 0.20 + 0.0 × 0.80), from 50% to 53%. To detect this improvement reliably (80% power, significance level 5%) would require a trial of more than 6000 patients and in a typical phase II trial of a few hundred patients, this drug effect would be missed. This is not just a hypothetical consideration. Mepolizumab was discontinued after it showed no efficacy in moderate asthma patients 6, yet over a decade later it has been approved for use in the subset of asthma patients with hypereosinophilia 7, 8. An anti‐amyloid antibody, gantenerumab, was discontinued for lack of efficacy in phase III until a retrospective analysis showed an efficacy signal in patients with more rapid underlying disease progression from mild cognitive impairment to Alzheimer's disease 9, along with a need for higher doses. ALZ‐801 is being developed for Alzheimer's disease patients who are homozygous for ApoE4/E4, after previous similar molecules failed in all‐comers but showed signals of benefit in this subpopulation 10. Lampalizumab, an anti‐factor D monoclonal antibody for geographic atrophy was only mildly effective in a phase II trial in all‐comers with much of the benefit in the subset (approximately half the patients) who were complement factor I positive 11.

The mission of clinical pharmacology is to understand variability in drug response and guide how best to use the drugs to overcome this variability. Understanding response variability is critical to addressing the problem of failing to identify effective drugs when useful benefits in subpopulations are masked by the lack of benefit in others. Thus clinical pharmacologists are necessary to improve success rates in drug development and to help identify the effective drugs that are currently being missed and that could continue to be missed in the future.

Precision medicine

Doctors have always tried to tailor their treatments for each individual patient. However, this can be difficult as most of the data available is based on studies in populations of patients, producing guidance on how to treat an average patient. For some well‐defined situations, such as extremes of age, organ failure and presence of interacting concomitant medication, clinical pharmacologists are already able to guide how prescribers should adjust dosing. But the reality is that many drugs remain ineffective or suboptimal for the majority of patients 12, 13. Precision medicine aims to use the increasing amount of molecular information available about patients, including genetics and biomarker profiles, plus knowledge of the patients' lifestyle, environment and past history, to better target the right drug at the right dose to each patient. Great success has been achieved in a few cancers and other genetic diseases, where treatment is directed only to those patients with specific genetic abnormalities and there are a few cases where genetics or other biomarkers can be used to guide when certain patients should not receive a particular treatment or the dose should be adjusted for them. Wider application of precision medicine has obvious health benefits for patients and is likely to improve cost effectiveness of healthcare.

Precision medicine is only necessary because there is variability in response between patients. The contribution of clinical pharmacology to understand which patients are most likely to respond has already been discussed. Determining the right dose is also important for development success and, in a future of precision medicine, industry clinical pharmacologists should lead the investigation of when greater dose individualization can improve efficacy and/or safety of new drugs 14. The goal here is to enable more patients to be treated safely and effectively by understanding who can respond well to lower doses and who needs, and can tolerate, higher doses. Precision medicine will increase complexity for patients who will usually need additional tests to help determine which drugs and doses are right for them, for healthcare providers and for prescribers who will need to integrate and interpret a lot more data when making individual prescribing decisions. Clinical pharmacologists are needed to help guide when the benefits achieved are sufficient to justify the additional complexity. They should also contribute to developing the systems and tools needed to support the implementation of precision medicine and make it something that all prescribers can use 15. Academic clinical pharmacologists are leading this today, with a focus on better prescribing for drugs that are already approved but that were not developed with a precision medicine approach. As new drugs are developed using precision medicine, industry clinical pharmacologists will be needed to ensure the right information is captured during development and the tools and systems developed to ensure effective translation into practice by prescribers, providers and payers.

Oversight of first‐in‐man studies and other phase I trials

Clinical pharmacology phase I clinical trials are generally safe with a very low incidence of serious, drug‐related adverse events 16. However, the first‐in‐man studies with TGN1412 in 2006 and in 2016 with BIA 10‐2474, in which one healthy subject died and another four suffered brain damage, serve as a reminder that phase I trials are not without risk. This remains true for all stages of development and even with non‐investigational medicinal products when not used appropriately. Two healthy subjects required intensive care when accidentally administered a 1000‐fold overdose of caffeine during a study of caffeine effects at Northumbria University in 2015 17.

The root cause of the TGN1412 toxicity was an inappropriately high starting dose that did not take adequate account of pharmacological principles. Industry clinical pharmacologists proposed an approach to dose setting based on estimating the dose associated with minimum anticipated biological effect level (MABEL) that could prevent recurrence of a similar event. This was subsequently adopted as a recommendation from the Expert Scientific Group 18 and included in the subsequent EMA guidance on design of first‐in‐man studies and its recent update 19. Clinical pharmacologists in sponsor organizations and at Contract Research Organizations, where most commercial phase I trials are performed, are essential contributors to minimizing risk in phase I trials while at the same time ensuring the trials will provide clear answers to necessary research questions.

Innovating drug development

Improving success rates has the largest impact on the costs and effectiveness of drug development. However, decreasing costs and/or the time needed to undertake clinical trials, without compromising safety, are also important 1. Clinical pharmacologists play an important role in developing new methods for investigating drug activity. The largest benefit is from increased use of mathematical modelling to provide insights into drug action and response variability and to enable useful predictions of effects in situations that have not been studied in a clinical trial. PKPD modelling is a required element of the development of all new drugs as a means to support dose selection and to identify when and how dose adjustments might be required to account for variability between patients. Often the results from such modelling are enough to support recommendations without additional clinical trials. Physiologically based pharmacokinetic modelling (PBPK) can predict human pharmacokinetics, and sometimes effects, based on knowledge of the physiological systems involved in drug absorption, distribution, metabolism and elimination and the physicochemical properties of the drug 20. This increases safety and effectiveness of early trials and can help decrease the number of subsequent trials needed to explore response variability. Mathematical modelling of diseases, including empirical models of disease progression, or meta‐analyses drug and disease effects from previous clinical trials and mechanistic systems pharmacology models based on the underlying biology and pathophysiology of a disease, are also contributing to better understanding and/or prediction of drug effects 21. All of these modelling methods are being further developed by clinical pharmacologists in industry and academia and refined to provide wider utility and better predictive power to support more effective and efficient drug development and to improve how drugs are used in the clinic.

Guiding prescribing

Between one‐third and one‐half of the prescribing information for a drug is generated by clinical pharmacologists. This includes justification of the approved doses, guidance when and how to adjust dosing to ensure efficacy and safety in different patients, guidance on potential risks and guiding when not to prescribe a particular drug or when to take specific precautions to minimize the risk of adverse events. The methods by which this information is obtained during development of new drugs will change over time using some of the technologies and approaches described above. For future drugs, it is to be hoped that there will be much greater understanding of how to decrease response variability and ensure that more patients can benefit from a given drug than is the case today such that drugs are used much more effectively and patients benefit to a much larger extent. It is likely that the package insert or other tools to support safe and effective prescribing will continue to be dominated by information provided by clinical pharmacologists.

What skills and capabilities will the clinical pharmacologist of the future need?

In the last 25 years there has been a huge increase in the use of mathematical modelling in clinical pharmacology and this is likely to become even more important in the future. Models are essential to understand the relationships between dose and concentration and between concentration and effect, and to understand the factors that contribute to variability in these relationships between patients. To date, much of the focus has been on the relationship of dose to concentration and how to adjust doses to ensure different patients achieve similar concentrations. But often this is only a minor contributor to the total variability in drug effect 22. Future clinical pharmacologists will need to better understand when and why the relationship of concentration to effect is different between, and indeed within, patients and how to overcome that. Modelling approaches will be essential tools and all industry clinical pharmacologists will need an adequate grasp of how modelling can help. In turn the expert modellers will need better understanding of the underlying biology of drug effects and also the diseases being treated as they will increasingly need to consider how best to include disease heterogeneity in their models. Perhaps the key role for those clinical pharmacologists who are not expert modellers will be to provide a link between expert modellers, who may come from backgrounds such as mathematics, physics or engineering, and the biologists and clinicians, who will still lead the activities to understand disease pathophysiology and identify suitable drug targets. Pharmacists and other biological scientists with training in quantitative skills as well as pharmacology and physiology could be at a premium. Physician clinical pharmacologists will undoubtedly need better training in quantitative methods than is provided in most medical schools today. Even those who do not become clinical pharmacologists will need this so they can better implement precision medicine and the decision support tools that will support it.

Future clinical pharmacologists must have a good grasp of emerging drug effect measures such as all the various ‘‐omics’, imaging methods, FACS, cytokine panels and others. New data analytical approaches such as machine learning, to detect patterns in complex data, will be important to make sense of which parameters actually matter and how they connect to each other. Future industry clinical pharmacologists will collaborate closely with experts in informatics and should have an understanding of the main methods they use. One goal will be to link machine learning that detects patterns in complex data with modelling that can provide insights into what the patterns mean and provide the capability to make predictions for subgroups or even individual patients based on their specific patterns.

Critical for the success of precision medicine will be for clinical pharmacologists to find the simplest ways to characterize as much response variability as possible and to guide how the added complexity of prescribing is justified by additional benefit to the patients, prescribers and payers. Their understanding of response variability and how to manage it will be the foundation on which clinical decision support and other tools are developed to aid prescribers. They will need better links to the diagnostics industry for development of the devices and software needed to allow prescribers to implement the precision medicine the clinical pharmacologist has characterized.

In many pharmaceutical companies today the scientific leadership of, and decision making for, clinical development projects is often by therapeutic or disease area aligned clinicians. Clinical pharmacology can often find itself as a support function. With the need to better understand the relevance of disease heterogeneity in order to improve drug development success and precision medicine, I don't expect this organizational model will change much in the next few years. Consequently future clinical pharmacologists will need very effective influencing and communication skills.

Concluding remarks

Clinical pharmacologists in the pharmaceutical industry can now reshape how drugs are developed and how they should be used differently in the future. They will continue to ensure prescribers have the information necessary to choose the best drugs for their patients. They should become important contributors to improving success rates for drug development; to increasing the impact of and benefits from precision medicine and to increasing its uptake by patients, prescribers, providers and payers.

The industry clinical pharmacologist of the future should not justify their existence by the tasks they perform such as first‐in‐man, drug interaction or bioequivalence studies or developing population pharmacokinetic pharmacodynamic (PKPD) models. They will continue to perform these activities, and many more, and they must be done well. But they are not themselves the end goal. That remains to understand the relationship of dose to exposure to effects and why it varies between people. The goal for industry clinical pharmacologists should not be only to do what is needed to gain Food and Drug Administration (FDA) or European Medicines Agency (EMA) approval. However, if we understand dose, exposure and response we will be able to meet the requirements of health authorities for approval, write informative and useful prescribing information, and provide the tools for patients and prescribers to improve the benefit–risk ratio from the medicines they use.

Competing Interests

The author is an employee of, and holds shares and share options in, F. Hofmann la Roche Ltd.

The author thanks Professor Sir Munir Pirmohamed for the conversation that triggered writing this article and Professor Max Parmar for conversations about the challenge of identifying responder subgroups within a population of non‐responders.

Peck, R. (2017) The pharmaceutical industry needs more clinical pharmacologists. Br J Clin Pharmacol, 83: 2343–2346. doi: 10.1111/bcp.13370.