Call for a new approach to the process of clinical trials and drug registration

The process of phases I-IV of clinical trials is the cornerstone of the drug registration and regulation process.¹ Yet expanding drug expenses, the slow process of development in critical areas of need, poor worldwide distribution of new therapeutics, and an emerging multibillion dollar industry receiving finances for clinical trials as a part of healthcare delivery, indicate that the process needs revision. Recent changes in testing by clinical trials, different methods of supervision of drug development by pharmaceutical companies, and new methods for the monitoring of drug safety suggest a direction for this revision. A fresh approach is proposed (see boxes, p 921). This article reviews the basis for considering this new approach.

Summary points

• The process of phases I-IV of clinical trials for drug registration is expensive, raises ethical issues of informed consent in phases III and IV, and has been expanded to try to answer too many questions

• Careful assessment of recent, well constructed safety data from phase II studies compared with data from the end of phase III is expected to show little new safety information gained from phase III

• A fresh approach to drug registration is needed

Suggested steps in new approach to drug assessment

1. Summarise innovative aspects of the drug (based on chemistry of the drug, animal toxicology, mechanisms of action, and initial clinical trials) to identify special approaches to drug development that should be considered

2. Summarise characteristics of the disease or diseases for which the drug is to be used, healthcare needs, and market pressures that indicate special risks or benefits regarding the drug's development

3. Review data about the drug collected from countries outside of the region in which registration is occurring, and summarise data about drugs marketed by competitors

4. Review patent issues, estimate the potential sales, and record the proposed drug price

5. Enter into special negotiations regarding the approach for drug development based on conclusions from the above analysis

A fresh approach to drug registration

Initiate flexible drug development based on the drug type and scientific database

Utilise decision making from interactive, pharmaceutical, and regulatory experts at the end of phase II to accomplish the flexibility

Integrate estimates for drug sales, prices, and worldwide distribution of the drug in these discussions

Ensure that all data from phases I and II are completely assessable to the healthcare community

Discontinue phases III and IV placebo controlled trials

Enhance the use of post marketing surveillance for drug safety assessment

Recent changes in the process of phases I-IV of drug development

The initial goal of clinical trials for drug registration was that a new drug treatment must be safer and more effective than no treatment for a specific disease. A plan to achieve this goal was outlined in the United States by the Kefauver-Harris Drug Amendments, approved in 1962.² The box on p 922 describes phases I-IV that have evolved.¹ This has become the central part of drug registration both by the Food and Drug Administration in the United States³ and by the European Commission.⁴

Improvements in the study design of phases I and II

In the past, phase I and II studies were often open label and non-randomised and done in small numbers of patients. As a result, neither efficacy nor safety data were considered reliable. Randomised, double blind, placebo controlled phase III trials were therefore considered justified and ethical.³ During the past decade, however, the importance of properly designed early trials (phases I and II) has led to dramatic changes in their design.⁵ These changes have included both proper randomised, double blinded designs and increased sample sizes. Although there is disagreement about how best to use data from phase II in the present process of phases I-III,⁶ there is little doubt concerning the high level of data available by the end of phase II. Because of the extensive data available, many phase III and virtually all phase IV placebo controlled trials are rendered redundant and potentially unethical.

Mistakes in study design versus drug inefficacy in phase III

Mistakes in the design of a drug trial are often made by designers in preparing and conducting phase III clinical trials. These mistakes are usually reported as drug failure rather than poor pharmaceutical company expertise, excessive marketing influence, regulatory micromanagement, or improper patient enrolment and follow up. Approaches in the design of phase III clinical trials have also led to incorrect recommendations regarding drug dose or duration in trials considered to have been successful. Examples of mistakes include requirements for inappropriate end points for evaluating the drug, study of a single drug dose or treatment, incomplete data for calculating sample sizes, over enthusiastic reports of occurrence rates of diseases by potential study centres, inadequate attention to patient inclusion or exclusion criteria, and incomplete follow up. The assumption has been that these are problems for the pharmaceutical companies to solve—the regulatory role is simply to identify them and reject failed studies. This is false. It is a problem created by the process of clinical trials, and it should be solved by the entire healthcare community.

Ethical problems with informed consent in phase III

Issues of ethics and possible conflicts of interest have been noted recently in studies done in developing countries,⁷ and potential conflicts of interest have been identified in the conduct of studies in the United States.^8–12 The ethical issues raised contribute to the global concern that activities carried out during the later stages (phases III and IV) of the present process of clinical trials are balancing on the edge of inappropriate activities, both by academic medical centres¹² and during physician-patient interactions.⁶,¹¹ These issues have been addressed by regulatory authorities in Europe by their request for comparative phase III trials of marketed drugs. But because comparative drug trials tend to require larger numbers of patients, as the goal is often to show equivalence, global pharmaceutical companies have sought approval in the United States with placebo controlled trials, then have used these to register in Europe. As long as authorities require or accept these trials the ethical problem remains.

Expansion of the questions asked during clinical trials

The initial goals of drug evaluation have been modified to include new questions directed at goals other than drug efficacy and safety. For example, testing a drug in a population representing the “real world setting” has become a major basis for phase III trials. This step has so altered the evaluation of the drug itself that potentially useful drugs—usually those with partial effects in a complex disease process—will never be made available. For example, in a recent review, over 75 different potential therapeutic agents were listed for use in the treatment of septic shock.¹³ None will be available because of the heterogeneity of patients with septic shock from the “real world” enrolled in phase III trials.

Present steps in clinical trials and their activities

Phase I

After a full review of preclinical data on drug efficacy and toxicology this phase tests various doses of the drug in healthy human participants to determine the dose related tolerability and safety in humans and drug absorption and distribution pharmacokinetics

Phase II

On the basis of the dose range and pharmacokinetics identified in phase I this phase evaluates the efficacy and tolerability of the drug over a range of doses compared with placebo in patients with the disease or diseases selected for treatment

Phase III

This phase uses a selected dose from phase II to evaluate efficacy and tolerability of the drug in a large number of patients compared with placebo or other similar marketed drugs. The results of these studies determine whether the drug can be licensed

Phase IV

This phase is started during the several years of the registration and licensing process and continues after marketing of the drug. It is designed to answer relevant ancillary questions regarding use of the drug, promote experience with the drug among leading doctors, and position the drug in the market place

Post marketing surveillance

This activity, also referred to as phase V, records adverse events of the drug submitted by doctors during use of the drug after it is marketed, and it sometimes includes large surveys of the profile of adverse events of specific drugs among tens of thousands of treated patients

Other new questions that have been asked are “how should a physician be advised to use the drug in her or his medical practice?” and “is the drug better or similar to a drug already available?” Pharmaceutical companies have asked for financial risk and benefit analyses and quality of life studies to accomplish successful registration and marketing of a drug. Obtaining answers to these questions has become a part of the phase III process of clinical trials. Clinical trials have moved from a role in drug assessment to physician education and competitive marketing. These changes have required a new management system dominated by calculations of sample size, goals for patient recruitment to accomplish large clinical trials, and the emergence of a new industry to organise clinical trials.¹¹,¹²,¹⁴ In retrospect it must be asked whether these are efficient and appropriate changes to accomplish better drug assessment and improved health care.¹⁵

Changes in supervision of drug development by pharmaceutical companies

Another major event has been the change in the management structure of pharmaceutical companies since the present structure of clinical trials began.¹⁶,¹⁷ The change in supervision of the drug development process by pharmaceutical corporations occurred, in part, because of regulatory requirements and because of corporate, marketing, and legal forces that emerged concurrently. For example, pharmaceutical companies began instituting a careful review of phase transitions during drug development, including a full summary of data and scientific presentation at the end of phase II within the company, and, in some cases, to regulatory agencies such as the Food and Drug Administration. This change has created ethical issues that have been largely ignored. With all the data to hand it is very difficult to defend the presentation of an informed consent form to a patient about to be enrolled in a phase III placebo controlled trial.

Safety assessment during phase III studies and the emergence of post marketing surveillance

Because safety data about common, severe, and dose related adverse events are now fully collected and reviewed at the end of phase II, very limited new information about the drug is identified in phase III. It has been stated that uncommon adverse events, such as those occurring at a frequency of less than 1 per 100 patients, cannot be identified in phases I-III.¹ Rare adverse events emerge only during post marketing surveillance. Proper evaluation of a drug's safety requires tens of thousands of treated patients and can only be done by careful post marketing surveillance.¹⁸ This type of surveillance has been particularly successful in the United Kingdom where the government's “yellow card” system of reporting all drug related adverse events has provided the healthcare community with an accurate picture of a drug's safety profile. Academic medical centres have developed effective research units for drug surveillance, and pharmaceutical companies now catalogue annually the pattern of adverse events of any marketed drug. These initiatives seem to be far more relevant to understanding the safety of drugs than are expanded numbers of phase III trials. What can be done to redirect the clinical process in a more efficient and financially reasonable direction?

Suggestions for new approaches to drug registration

Assess the safety data from phase II and phase III clinical trials

The issue of whether new safety and efficacy data emerges during phase III clinical trials needs to be further studied. Comparison of profiles of adverse events at the end of phase II with those at the end of phase III would provide this information. Data from the end of phase II are available in the files of pharmaceutical companies and regulatory agencies, but this information is not usually, or systematically, available in the literature. By removing the need for safety data from phase III trials, then removing the use of placebo controlled phase III and phase IV trials altogether, most of the ethical and conflict of interest issues would disappear. Needless to say, careful review of all clinical trials by properly constituted ethics committees must continue. What then would be a substitute for the present reliance on phase III trials for drug registration?

Recognise the need for flexibility for drug development

Firstly, we must agree that different types of drugs and different diseases require entirely different approaches to clinical trials. It is illogical that a protective vaccine or a modulator of immunity to cancer would follow the same pathway for registration as a drug for treating constipation, sepsis, or dermatitis. Drugs developed on a basis of sound theoretical models and animal trials should not follow the same route as a compound without such a basis. Therefore flexibility in drug assessment would be a required part of the new process.

The process of clinical trials should begin, as at present, with randomised safety studies in healthy and ill participants and randomised, dose ranging, placebo controlled studies in carefully selected groups of patients with the disease or diseases to be treated. The results would determine the drug's profile of dose related adverse events and effectiveness.

Establish an interactive group for review of drugs

To decide how best to proceed after evaluating these studies the areas listed in the box would be assessed by appropriate experts within the regulatory agency in consultation with experts both within and outside the pharmaceutical company. This medical advisory panel could be similar to that recently recommended to evaluate the pharmaceutical industry.¹⁹

The first three tasks listed in the box would be directed at understanding special features of the drug and disease. Item 4 ensures full review of the data. Item 5 is important but only possible if government supervision of patent, price, and sales could be included during the licensing process. Negotiations could include issues such as patent extension, cost, distribution to poor areas of the country (or other countries), or use of the drug in special disease processes such as HIV infection, tuberculosis, or diseases that are preventable by vaccine. By this discussion the risks and incentives of pharmaceutical companies versus both national and international healthcare needs related to the drug could be addressed.²⁰

For most drugs, registration would proceed at this point. In some cases, additional studies of dose ranging or safety would be needed. In most situations, placebo controlled phase III studies would be avoided. Drug licensing would be based on complete presentation of data from phases I-II. Approaches that could be included in such a flexible system are described in the table.

Gain information on drugs after the licensing process

Use of the drug in patient populations from the “real world” and comparisons to similar available drugs would be left to the judgment of investigators and the goals of marketing departments after the drug is licensed. The main vehicle for drug safety would be the expansion of post marketing surveillance systems. Carefully conducted observational studies could be used to answer emerging questions about efficacy.²¹,²² A new role for experts within pharmaceutical companies would be to ensure that all data on dose ranging safety and efficacy done in phases I and II were published and that complete data listings were available on the internet. This would require new resources within pharmaceutical companies, but the activity would be far more useful to the healthcare process than is the case with the present phase III trials. The direction outlined in this article could provide needed vigour and innovation in drug development, faster access to advances in medical practice, and reasonable healthcare costs.

Table.

Flexible approach to drug development for registration

Drug type	Preventive drugs*	Innovative/good basics†	Innovative/poor basics‡	Redo/good basics§
Process
Phase I (random, double blind volunteers or patients)	Dose range studies in healthy subjects (placebo)	Yes (placebo)	Yes (placebo)	Yes (placebo)
Phase II (random, double blind dose and efficacy in patients)	Challenge studies in healthy subjects (placebo)	Yes (varied doses and placebo)	Yes (varied doses and placebo)	Yes (varied doses, historical or competitor controls)
Phase III	Rare (historical control)	No	Yes (placebo)	No
Post marketing surveillance	Yes	Yes	Yes	Yes
Decision issues	• Use support studies from other countries • Add trials to negotiate cost, patents, and risk sharing by government • Use phase III only when challenge data not possible	• To confirm data from preclinical trials in patients with specific disease • Safety issues resolved in post marketing surveillance • Use documented surrogate end points when available	• Phase III justified by absence of basic data • Use homogeneous patient populations • Added work would tend to discourage this approach	• Same as innovative • Historical controls preferred over competitor drug • Avoid placebo controls • These drugs may reduce cost of drug class

^*Drugs such as vaccines that have potential for challenge studies or that have surrogate end points. 

^†Innovative new entities for which there is a high level of theoretical basis and good animal and in vitro models. 

^‡Innovative new entities for which either mechanism of action, disease pathogenesis, or preclinical testing has not been well documented. 

^§Drug that is in a previously known compound or class of compounds, with previous good basic data.