Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

17.1 Introduction

This book has provided an introductory overview of the immensely complex process of lifecycle drug development. Since various aspects of this process have been discussed individually in the preceding chapters, this final chapter takes a more global view. It brings together threads that have run throughout the individual chapters and presents an integrative summary of topics discussed to date. Following these summary discussions, several new topics are introduced.

Looking back across the previous chapters’ discussions of new drug development, several themes emerge

  • It requires that attention be paid to ethical considerations.

  • It requires that attention be paid to design, methodological, operational, and analytical considerations.

  • As well as statistical significance being a necessary aspect of drug development, clinical significance is also extremely important. Clinical significance is typically captured in terms of confidence intervals.

  • Pharmacokinetic and pharmacodynamic considerations are of interest at every stage of the process.

  • Many decisions have to be made throughout the process.

  • There is much more subjectivity in this decision-making than might be initially thought.

  • Benefit–risk decisions pervade the drug development process.

  • The ultimate goal of drug development is to provide approved (and marketed) drugs that can safely change patients’ biology for the better.

Once a drug is available to a physician to prescribe to a patient, we move into the full realm of Integrated Pharmaceutical Medicine (Turner, 2009b). While we have focused on drug development (and briefly on manufacturing since drug development requires drug products), the therapeutic use of drugs requires commercial-scale manufacturing and the involvement of many health professionals, including prescribing physicians, dispensing pharmacists, and nurses who administer drugs in in-patient and residential care settings.

17.2 Ethical Considerations

Ethical considerations are pervasive throughout new drug development. The need for ethical treatment of all subjects who are willing to participate in clinical research—an activity that is designed for the greater good, not for their individual benefit—is paramount. Also, since it is unethical to include subjects in a study where poor design and/or poor methodology and/or poor operational execution leads to less-than-optimum data and therefore less-than-optimum answers to the study’s research question, everyone involved in clinical research has the responsibility to act in an ethical manner. Just like Statistics, ethics are not simply something for “someone else” to worry about.

In the previous chapters, discussions of ethical considerations have occurred in many contexts, including

  • Designing a study in an ethical manner such that the design is capable of producing optimum quality data.

  • Subject recruitment, including providing informed consent.

  • Sample size estimation. A study design requires sufficient subjects but not an unnecessarily large subject sample.

  • Conducting all aspects of a study in the best manner possible such that the methodology is capable of producing optimum quality data.

  • Data monitoring committees face difficult ethical challenges, in particular deciding whether a clinical trial should be terminated early.

  • Authors have an ethical responsibility to report information accurately and fully in clinical communications, since these directly impact patient care.

As noted in Section 1.8.1, Derenzo and Moss (2006) captured these sentiments very well, and their quote is worth repeating here:

Each study component has an ethical aspect. The ethical aspects of a clinical trial cannot be separated from the scientific objectives. Segregation of ethical issues from the full range of study design components demonstrates a flaw in understanding the fundamental nature of research involving human subjects. Compartmentalization of ethical issues is inconsistent with a well-run trial. Ethical and scientific considerations are intertwined (p. 4).

17.3 Design, Methodology, Operations, and Analysis

Given the statements in Chapter 1 that design, methodology, operations, and analysis are central characters in this book, it is not surprising that they have been frequently encountered throughout the previous chapters. It has been noted that design, methodology, operational execution, and analysis are of importance across the entire spectrum of new drug development, including drug discovery and design, nonclinical research, and clinical development: they are also pertinent in postmarketing surveillance. A vast array of numerical information is collected and analyzed during these various stages of research.

In nonclinical research, attention to detail is every bit as important as in clinical development. As Gad (2006) observed, “the importance of nonclinical laboratory studies demands that they be conducted according to scientifically sound protocols and with meticulous attention to quality.”

Early-phase clinical studies involve relatively small numbers of subjects. However, this does not mean that design, methodology, and analysis are any less critical. Machin and Campbell (2005) noted that these early-phase clinical studies provide key information for the drug development process and that it is “essential that they are carefully designed, painstakingly conducted, and meticulously reported in full.”

While the nature of the analyses of data from therapeutic confirmatory clinical trials is relatively straightforward compared with those undertaken in earlier stages of clinical development, supreme care should again be taken in all aspects of design, methodology, execution, and analysis, since only then can optimum quality data be used to provide optimum answers to well-constructed research questions. By the time that therapeutic confirmatory trials are conducted, there should be a small number of precisely asked research questions that address the efficacy of the drug.

In a fixed design trial, the analyses used to provide compelling evidence of efficacy are typically straightforward: They may be somewhat more sophisticated in group sequential studies and adaptive design trials. As emphasized throughout the book, however, while statistical evidence of efficacy is required by regulatory agencies at this time, the clinical significance of the treatment effect is ultimately the primary concern. By the time that therapeutic confirmatory trials are conducted, the sponsor will have a good idea of the drug’s efficacy from earlier clinical studies. These later trials will only be conducted if they are likely to confirm that the treatment effect is indeed clinically significant.

17.3.1 Reducing Bias and Improving Precision

Statistical methodology has two important goals: reducing bias and improving precision. The process of randomization reduces bias (the word reduces is deliberately used here, since total elimination is not a feasible goal), as does the procedure of blinding. Statistical inferences are based on the use of randomization to reduce bias to the greatest extent possible and ensure comparability of the treatment groups with respect to pertinent variables such as age, sex, and other important prognostic factors (Chow and Liu, 2004). Analysis of covariance can also be used in this context to address influences that could not be addressed by randomization.

Improving precision as much as possible is also a highly desirable attribute in a study design, and statistical methodology aims to improve precision in several ways, including reducing error variance. Better measurements will yield data that provide a more precise answer, e.g., narrower confidence intervals around the treatment effect obtained in a trial.

17.3.2 Our Definition of Statistics Revisited

This book’s definition of Statistics was first presented in Section 1.3. According to this definition, Statistics can be thought of as an integrated discipline that is important in all of the following activities:

  • Identifying a research question that needs to be answered.

  • Deciding upon the design of the study, the methodology that will be employed, and the numerical information (data) that will be collected.

  • Presenting the design, methodology, and data to be collected in a study protocol. This study protocol specifies the manner of data collection and addresses all methodological considerations necessary to ensure the collection of optimum quality data for subsequent statistical analysis.

  • Identifying the statistical techniques that will be used to describe and analyze the data in an associated statistical analysis plan, which should be written in conjunction with the study protocol.

  • Describing and analyzing the data. This includes analyzing the variation in the data to see if there is compelling evidence that the drug is safe and effective. This process includes evaluation of the statistical significance of the results obtained and, very importantly, their clinical significance.

  • Presenting the results of a clinical study to a regulatory agency in a clinical study report and presenting the results to the clinical community in journal publications.

This definition of Statistics may have seemed rather expansive when it was first encountered. By now, I hope it may seem much more appropriate. As has been noted many times, statistical awareness is essential throughout the entire drug development process, from designing a study to answer a research question right through to presenting the study results to regulatory agencies and the clinical community.

17.3.3 Numerical Representations of Biological Information

The data acquired in a clinical trial are not simply numbers: They are numerical representations of biologically important information. The number 9 is meaningful by itself (it is an integer between 8 and 10). However, in a clinical database, the digit 9 may represent many things, including a decrease of 9 mmHg seen in a subject’s SBP following the administration of an investigational antihypertensive drug for several weeks. In this context, the digit 9 is a numerical representation of biologically important information concerning a change in blood pressure. The employment of the discipline of Statistics is a means to an end here, and the end is producing a drug that is safe and has a beneficial therapeutic effect on a patient’s biology.

In this context, drugs prescribed in psychiatric care are considered to have a beneficial therapeutic effect on a patient’s biological well-being. While some sources differentiate between psychological well-being and physical well-being, and therapies such as cognitive behavioral therapy may also be used in psychological and psychiatric counseling, psychiatric pharmacological agents exert their influence via biological changes that contribute to the patient’s well-being.

17.3.4 Some Thoughts on the p-Value

The requirement to show statistical significance in therapeutic confirmatory trials places a certain importance on a priori hypothesis testing. However, as we have seen, there is much more to clinical research than p-values. Piantadosi (2005) commented on this issue as follows:

Not even a brief discussion of estimation and analysis methods for clinical trials would be complete without an appropriate de-emphasis of p-values as the proper currency for conveying treatment effects. There are many circumstances in which p-values are useful, particularly for hypothesis tests specified a priori. However, they have properties that make them poor summaries of clinical effects…In particular, they do not convey the magnitude of a clinical effect. The size of the p-value is a consequence of two things: the magnitude of the estimated treatment difference and its estimated variability (which is itself a consequence of sample size). Thus the p-value partially reflects the size of the experiment, which has no biological importance. The p-value also hides the size of the treatment, which does have major biological importance (p. 432).

This eloquent quote captures the essence of the respective contributions of assessments of statistical and clinical significance so well that it is worth reading twice and sharing with all of your clinical research colleagues.

Piantadosi also addressed a commonly expressed view when a researcher obtains a nonsignificant result in a study. It is often said that the estimated treatment effect obtained might have attained statistical significance in a larger sample. The implicit (or sometimes explicit) intent of the researcher is to convey that the result is still of (considerable) importance. The statement itself is true, but not helpful. First, it is true because any nonzero effect would attain statistical significance in a large enough sample. Second, focusing on the size of the estimated treatment effect and its clinical (biological) significance is what would be helpful. It is important to remember that p-values “incompletely characterize the biologically important effects in the data” (Piantadosi, 2005).

17.4 Confidence Intervals and Clinical Significance

Confidence intervals are extremely informative in clinical research since, unlike p-values, they focus on the magnitude of the estimated treatment effect and therefore facilitate consideration of its clinical significance. As Fletcher and Fletcher (2005) observed, confidence intervals “put the emphasis where it belongs, on the size of the effect.” The width of a confidence interval around an experimentally determined treatment effect point estimate, and hence the range of plausible values for the population treatment effect, provides very important information about the clinical significance of the treatment.

As we saw in Chapter 10, confidence intervals can also be used to deduce levels of statistical significance. While they do not yield precise p-values (the p-values can be calculated separately), they reveal whether or not a given level of statistical significance is achieved. More importantly, they are uniquely informative in assessing clinical significance. Therefore, confidence intervals offer a considerable advantage over p-values in the clinical context. They have become an important way of reporting the main results from clinical research studies (Fletcher and Fletcher, 2005).

17.5 Pharmacokinetics and Pharmacodynamics

An investigational drug’s pharmacokinetic/pharmacodynamic profile is of critical importance in determining its therapeutic usefulness, and assessment of this profile continues throughout the entire spectrum of new drug development. Pharmacokinetic issues are a major factor in a drug’s success after it has received marketing approval. A drug product certainly not only has to be safe and effective, but it also has to be convenient to use if it is going to be widely used and commercially successful.

Pharmacokinetic issues are being increasingly addressed in drug discovery. If a drug candidate has a pharmacokinetic profile that suggests potential later problems, it is better that the drug fails earlier in the discovery/development process rather than later. If the drug candidate looks promising, its pharmacokinetic profile will be evaluated in nonclinical studies. If this looks promising, the pharmacokinetic profile of the investigational drug will be evaluated in human pharmacology clinical studies. One of the most common reasons for not continuing with clinical development of a drug is an unsuitable pharmacokinetic profile, and it is therefore strategically important to evaluate a drug’s pharmacokinetic profile in early-phase drug development.

The extensive study of the pharmacodynamic (and toxicodynamic) potential and properties of a drug is also of interest throughout the entire development process. Maximizing interactions between the drug and its target receptor (and minimizing interactions between the drug and nontarget receptors) is of considerable interest in drug discovery and design. Therapeutic exploratory and therapeutic confirmatory clinical trials address the topic of the drug’s efficacy more formally.

17.6 Decision-Making

If forced to summarize the purposes of study design, experimental methodology, operational execution, and statistical analysis in one sentence each, the following might be suitable:

  • Study design: Designing a clinical trial to facilitate the collection of data, i.e., unbiased and precise numerical representations of biologically important information, that best answer the study’s research question.

  • Experimental methodology: Considering and implementing all necessary procedures that, if executed correctly, allow the acquisition of optimum quality data.

  • Operational execution: Conducting all operational and experimental tasks correctly and therefore actually acquiring optimum quality data.

  • Statistical analysis: Describing, summarizing, analyzing, and interpreting the data collected to answer the study’s research question.

Expanding on the last point, numerical representations of biologically important information facilitate answers to questions that arise during the process of new drug development and thus provide the basis for making the best possible decision at that time given the best evidence available at that time. (It is quite appropriate to use additional information to come to a new decision at a later time.)

Many decisions during the process of new drug development concern whether or not to proceed to the next step in the process. Adequate evidence needs to be obtained, and documented, to permit careful consideration of the benefits and risks of proceeding. Given a finite amount of resources and, particularly in the case of larger pharmaceutical companies, a choice of drug candidates upon which to focus these resources, it is financially prudent to proceed only if there is a reasonable chance of success (the definition of “reasonable” being unique to each sponsor and drug candidate). While business driven, the choice to pursue development programs that are likely to yield successful drugs is arguably in the best interests of patients: Pursuing development plans for drug candidates that are likely to fail reduces the sponsor’s ability to work on drugs that may get approved and help patients.

17.6.1 The Subjective Nature of Many Decisions

The title of this section may be surprising at first. The process of science, one may think, produces clear-cut answers, and scientists pride themselves on the objectivity inherent in their disciplines. Clinical science, clinical research, and clinical practice, however, require a combination of objective information and informed judgment. Since all judgment is subjective, subjectivity is an integral part of clinical science, clinical research, and clinical practice.

In this context, the word subjective does not carry the potentially negative connotations that may accompany it in other realms. All of us would likely welcome the medical opinion of a very experienced and well-informed clinician when making a decision concerning several possible therapeutic regimens. The opinion offered would be the clinician’s best clinical judgment based on the best available evidence at that time. In the context of study design, Piantadosi (2005) made the following comment:

It is a mistake not to recognize the subjectivity that is present, or to design and interpret studies in formulaic ways. We could more appropriately view experimental designs as devices that encapsulate both objective plans along with unavoidable subjectivity (p. 131).

Consider two examples from previous chapters that illustrate this. In Chapter 11, discussions of sample size estimation emphasized that the process is indeed one of estimation rather than pure calculation. A calculation is certainly executed, but the values that are placed into the appropriate formula are chosen by the sponsor. On each occasion, the sponsor must consider the influences of the choices that are made and make the most appropriate decision in the specific context of that trial. In Chapter 13, equivalence and noninferiority designs were discussed. In addition to the calculations that are involved using the data collected in a trial, equivalence or noninferiority margins must be established before the trial commences. Their choice is a clinical choice, not a statistical choice, and subjectivity is necessarily involved in this choice. Thus, the discipline of Statistics certainly involves using informed judgments. Statistics really is an art as well as a science, a sentiment expressed well by Katz (2001):

No degree of evidence will fully chart the expanse of idiosyncrasy in human health and disease. Thus, to work skillfully with evidence is to acknowledge its limits. All of the art and all of the science of medicine depends on how artfully and scientifically we as practitioners reach our decision. The art of clinical decision making is judgment, an even more difficult concept to grapple with than evidence (p. xi).

Consider also the decisions that must be made by regulatory agencies. From many perspectives, the role of regulatory agencies is far from easy. For example, they have to decide if it is appropriate to allow a sponsor to commence clinical testing based on data submitted in an investigational new drug (IND). It has been noted several times that no animal model is a perfect predictor of human responses to an investigational drug, and so the decision to allow a sponsor to commence clinical trials requires a judgment call. An enormous amount of information has to be provided to regulatory agencies to allow them to make this decision, but it is still a judgment call, albeit a very well-informed one.

The same is true when a regulatory agency evaluates the evidence presented in an NDA or biologic license application (BLA). Again, a tremendous amount of information is presented following the conduct of clinical trials, but these data cannot guarantee that serious adverse drug reactions will not be seen once the drug is approved and taken by a large number of patients. The agency has to evaluate all of the data in the marketing application and consider the benefit–risk ratio of approving the drug, an unenviable responsibility. Therefore, as noted in Chapter 16, while the randomized controlled trials that we have discussed in this book remain the gold standard for evaluating the efficacy of a new drug in preapproval clinical trials, and they do provide (some) safety data, they cannot be regarded as the sole source of safety data or as a guaranteed predictor of the drug’s effectiveness in the target population. In a real sense, marketing approval of a new drug can be regarded as the beginning of its true evaluation.

17.6.2 Determining Thresholds of Regulatory Concern

When employing a model to prospectively exclude unacceptable risk, we need clinical science, regulatory science, and statistical science. Regulatory scientists must consider the clinical evidence and then determine the thresholds to be employed. Choosing these thresholds may be a better expression than determining them: there is no precise formulaic manner to facilitate determination. Therefore, there is a degree of subjectivity in this process.

17.7 Benefit–Risk Considerations

As we have seen, the FDA’s Sentinel Initiative provided a useful definition of drug safety in terms of benefit–risk assessment:

Although marketed medical products are required by federal law to be safe for their intended use, safety does not mean zero risk. A safe product is one that has acceptable risks, given the magnitude of benefit expected in a specific population and within the context of alternatives available.

Benefit–risk decisions are made by regulatory agencies whenever approving (or not) a drug for marketing and when considering removing an approved drug from the market. Physicians and their patients also make benefit–risk decisions when deciding whether or not a patient should receive a drug. However, the natures of these decisions are different. Regulatory decisions are made at the public health level, while physicians consider treatment decisions one patient at a time.

17.8 Biological Considerations Pervade Our Discussions

We noted at the beginning of this chapter that the ultimate goal of drug development is to provide physicians with drugs with which to change patients’ biology for the better. In this section we will consider how our increasing knowledge of molecular biology is providing the basis for the practice of precision medicine.

17.8.1 Biological Underpinnings of Precision Medicine

Individual variation in responses to the same drug is a considerable (arguably the greatest) problem in clinical medicine. When a given drug is administered to different patients for whom, based on the best available diagnostic evidence, it is appropriate, many of them will safely experience a therapeutic benefit. However, there are other possible outcomes that may be experienced by relatively small numbers of patients:

  • Individuals may not show a beneficial therapeutic response (nonresponders).

  • Individuals may show an undesired excessive therapeutic response (e.g., becoming hypotensive instead of normotensive following the administration of an antihypertensive agent).

  • Individuals may show relatively serious undesired effects (adverse responders).

Precision medicine has the promise of identifying how an individual patient is likely to react to a given drug before the drug is actually administered, since it is predicated on knowledge of a person’s biology.

The term precision medicine is deliberately used here instead of other common terms such as personalized medicine and individualized medicine. Clinicians will argue, very reasonably, that they have always practiced personalized medicine to the limit of currently available knowledge. Thoughtful clinical care of a patient has always involved, and always will involve, using all available evidence concerning an individual patient’s unique set of circumstances and knowledge of all available treatment options to tailor a course of treatment accordingly. The major difference in the context of present discussions is that, in the future, clinicians will likely have access to detailed information about the biological make-up of individual patients under their care and will be able to consider this information before embarking on any treatment plan.

Genetic differences between individuals are directly related to individual differences in response to drugs. The expression, lack of expression, or overexpression of certain genes in individual patients influences how they will respond. Interactions can also occur between an individual’s genetic information that codes for drug responses and various environmental factors. However, while environmental factors such as nutrition, use of tobacco and alcohol, concomitant prescription medication, and disease can influence how an individual responds to a given drug, the predominant factor in overall individual variation is genetic variation, specifically variation in the structure of the target receptor and in pharmacokinetics (Primrose and Twyman, 2006).

17.8.2 Pharmacogenetics

Pharmacogenetics is concerned with the contribution of genetic variation to the variation in response to pharmacotherapy. Kalow (2005) noted that the discipline can be regarded as focusing on “person-to-person differences in drug metabolism and response.” Protein manufacture is under direct genetic control, and two factors are of particular relevance here. First, the precise structure and function of proteins functioning as drug receptors (both target and nontarget receptors) vary between individuals. Since the structure and function of the protein are directly related to how the drug molecule will interact with that protein, individuals’ responses to the drug will vary. Second, there are genetic variations in proteins functioning as metabolic enzymes and hence in metabolism. Such variation can be associated with altered distribution, metabolism, or uptake of the drug and may result in enhanced drug clearance, impaired drug clearance, or inactivation of the drug. These observations provide a direct molecular genetic link between an individual’s biology and their drug responses.

17.8.3 Pharmacogenomics

An organism’s genome is the collection of all genes within that organism. At the time of writing, a reasonable estimate of the number of human genes is somewhere shy of 25,000 (a dramatic decrease from estimates of 100,000 not too long ago). These genes comprise the human genome. The term pharmacogenomics “reflects the evolution of pharmacogenetics into the study of the entire spectrum of genes that determine drug response, including the assessment of the diversity of the human genome and its clinical consequences” (Meyer, 2002). Rothstein (2003) noted that pharmacogenetics addresses “the role of genetic variation in differential response to pharmaceuticals” and pharmacogenomics addresses “the use of genomic technologies in assessing differential response to pharmaceuticals.” Genotype information, information about a person’s whole genetic make-up, permits the possibility of pharmacological therapy targeted at particular individuals based on this knowledge.

The field of pharmacogenomics involves the use of genomic technologies in assessing differential responses to drugs. As Monkhouse (2006b) commented

Pharmacogenomic studies promise to revolutionize medicine by providing clinicians with prospective knowledge regarding the likelihood of an individual patient’s response to a particular medication and, ultimately, the identification of patients who might benefit from targeted dosing of the drug or alternate drug therapy (pp. 26–27).

17.8.4 Pharmacoproteomics

The number of proteins in humans is much larger than the number of genes in our genome. These proteins comprise the putative human proteome. This phenomenon is the result of “the simple although not widely appreciated fact that multiple, distinct proteins can result from one gene” (Holmes et al., 2005).

The journey from genome to proteome is not a straightforward one. It can be represented as a multistep process, starting with a gene of interest in the genome:

  • DNA replication results in many gene forms.

  • RNA transcription leads to pre-messenger RNA.

  • RNA maturation results in mature messenger RNA.

  • Protein translation results in an immature protein.

  • Protein maturation results in a mature protein in the proteome (post-translational modifications are possible here).

The tremendous diversity of proteins in the proteome is facilitated by multiple possible means of protein expression. At each stage in the multistep process just described, alternative mechanisms can produce abnormal variants of the normal variant protein. The combination of possible variations in the multistep process results in an enormous potential diversity in the resulting proteome. Soloviev et al. (2004) captured the nature of the developing field of proteomics as follows:

Characterization of the complement of expressed proteins from a single genome is a central focus of the evolving field of proteomics. Monitoring the expression and properties of a large number of proteins provides important information about the physiological state of a cell and an organism. A cell can express a large number of different proteins and the expression profile (the number of proteins expressed and the expression levels) vary in different cell types, explaining why different cells perform different functions (p. 218).

Pharmacoproteomics is of interest in drug therapeutics since, as we have discussed several times, so many drug receptors for both small molecule and biologicals are proteins, as are many metabolic enzymes.

17.9 Integrated Pharmaceutical Medicine

Drug development, drug manufacturing, and pharmacotherapy are components of integrated pharmaceutical medicine (Turner, 2009b). We have addressed the first two topics, and the third is not really in this book’s scope. Nonetheless, the drugs whose development and manufacturing we have discussed will be prescribed by physicians, dispensed by pharmacists, and administered by nurses, along with other caregivers and patients themselves. It is this author’s contention that an understanding of the drug development process is of considerable benefit to these and other health professionals.

Once a drug has been marketed the term medication safety becomes applicable. One component of this that receives occasional sensationalist coverage, as opposed to a more sustained acknowledgment and vigilance, is medication errors. The occurrences and costs of medication errors are stunning (see Institute of Medicine, 2007, for a very readable introduction, and Cohen, 2007, for a more detailed coverage).

Another component of integrated pharmaceutical medicine is adherence to a prescribed drug regimen, or rather the lack of adherence, either non-intentional or intentional. Given that it takes 10–15 years to develop drugs, and physicians and pharmacists spend years of professional training to get them to patients, the number of patients who do not take their medication as prescribed is disturbingly high.

17.10 Concluding Comments

One major goal of this book has been to illustrate the central role of numerical representations of biologically relevant information in the process of new drug development. Accordingly, it has highlighted the roles of study design, experimental methodology, operational execution, and statistical analysis in the areas of drug discovery, nonclinical research, preapproval clinical trials, and postmarketing surveillance. Emphasis has been placed on the roles of optimum quality design, methodology, and execution in providing optimum quality data for analysis, interpretation, and use in decision-making. In addition, new drug development requires attention be paid to ethical, intellectual, scientific, biological, clinical, organizational, regulatory, financial, legal, congressional, social, and political considerations, and this list is likely far from exhaustive.

A second goal has been to illustrate that all of the activities described in this book are ultimately conducted to improve patients’ health and well-being by changing their biology for the better. It is appropriate to remind ourselves frequently that our work has a very real impact on patients’ lives. New drug development is a very complicated and difficult undertaking, but one that makes an enormous difference to the health of people across the globe. It is a noble pursuit.

If you work in this field, might be interested in doing so in the future, or have read this book because of your interest in clinical medicine, I hope it has helped you to have an appreciation of design, conduct, and analysis in new drug development and the importance of every professional’s contribution to this process. I also hope the book has served to illustrate the nature of new drug development, the critical role of numerical representations of information in making informed decisions, and the central importance of biological and clinical considerations.