Innovations in Phase IV Clinical Trials for Change in Health Care

Co-authored with Dr. Michael Proschan

A Phase IV trial, otherwise called post-marketing surveillance, monitors the safety and efficacy profile of the drug or device once regulatory permission is granted for sale via prescription or over the counter. Post-marketing studies can detect adverse effects that occur rarely or because of extended intake, because administration occurs in a larger timeframe and study population than in the phase I-III clinical trials.

The reasons that Phase IV studies are generally not done include the belief that they are hard to organize and costly to accomplish. But the growth and development of information technology (IT) make it very feasible. One way to advance Phase IV correlation studies on a medication and its side effects in general practice is the passage of legislation mandating the release to neutral, responsible bodies such as several schools of public health of up to 8 years of information stored with insurance companies, HMOs, and other entities that reimburse doctors and hospitals for services rendered to patients. Naturally, personal details such as social security number and the patient's name are scrubbed before the files would be periodically downloaded to the entities specialized in analysis and research in complex clinical trials, such as schools of public health. This is the model that we are presenting for discussion to move forward the much-called-for but rarely conducted phase IV clinical trials.

Why We Need Such Effort, and a Model

The Vioxx (rofecoxib) debacle, for instance, highlighted why phase IV trials matter. The cost for patients and even the company may have been far less had there been a system in place from the very beginning, immediately after FDA approval for sale, to correlate upticks in cardiovascular incidents for patients on Vioxx over the baseline. It seems to us that far from designing a costly new system that relies on telephonic or face-to-face surveys of doctors and patients, that will be seen as yet another bureaucratic maneuver to expand the authority of the Food and Drug Administration (FDA) or the government, making better use of existing data now stored in the servers and digital tapes of private sector entities such as insurance companies and HMOs is an effort to achieve better efficiency, transparency and accountability. Hence our focus on analyzing data in secure facilities within schools of public health, in a manner parallel to the release of blinded, sensitive clinical trials information to data and safety monitoring boards (DSMBs) that are sworn to confidentiality until the time if/when they must act.

Currently, citing patient confidentiality, HIPAA and other reasons, entities that hold vast terabytes of data that can benefit patients, doctors, pharmaceutical companies and indeed society at large, are unable to do much with it. Ideally, Congress should enable the release of scrubbed data to secure facilities in schools of public health so that correlation studies and analysis can be undertaken in a serious, thoughtful yet speedy manner that can serve as electronic, real-time surveillance for most medicines on the market. Once a drug is approved for use in the community, it can take years before reports from doctors across the country back to the pharmaceutical companies' medical staff can be translated into specialized studies, and even then, the results for company-sponsored and other studies were presented amid acrimony between and within academics, regulatory agency officials and pharmaceutical executives in the case of Vioxx. Furthermore, debates in the Congress indicate that society is increasingly unwilling to solely depend on industry to self-regulate and self-report.

If the analysis is not done responsibly, at arm's length, the public may be concerned that undue pressure was brought to bear by overwhelmingly powerful industry. Yet, the federal budget is now increasingly stretched, with newly authorized phase IV trials remaining unfunded by Congress and only the permission having been granted to charge companies for the costs. Generating primary data through interviews of doctors and patients will be prohibitively expensive and organizationally challenging, thus dampening enthusiasm for such studies all around.

Thus, we seek only the ability to access and analyze existing data, already collected at a cost and currently exclusively used by doctors and hospitals for reimbursement purposes from insurance companies and government entities such as the Centers for Medicare and Medicaid services. In justifying each reimbursement, documentation is sent within fields indicated in the accompanying table by doctors and hospitals. Those are needed since each bill runs into the hundreds and sometimes tens of thousands of dollars. Every prescription incurs a cost to the insurance company or HMO and therefore must be stored electronically, and indeed for accounting purposes even a prescription that results in costs of a few cents must be accounted for. Prescribed medicines databases thus exist for all those who receive any medication and this will give the denominator to calculate rates.

Absence of detailed information in the required fields while seeking reimbursement would be tantamount to fraud, and therefore the willingness of doctors and hospitals to provide the data in the required fields has never been in doubt. Legally too, at present, those fields are required to be filled. But their use is at present for securing reimbursement. It makes little sense to compile the data for a financial reimbursement purpose and not use it for patient-safety. Needless to say, safeguards should be built in not to prematurely report possible adverse events, to prevent the problem of stimulated reporting - or reports of adverse events being generated just because there are reports in the media of adverse events to a certain drug. The modalities to be put in place can replicate similar safeguards established in Data and Safety Monitoring Boards that have the responsibility to evaluate whether alleged adverse events are sufficiently compelling to stop a trial. And undoubtedly, the same data can be downloaded to more than one School of Public Health so that they can independently verify whether the data are sufficient to take a particular action. However, care should be taken to ensure that having more than one such center does not lead to leak of data - but those sorts of protective measures are taken by multi-centric trials.

An Additional Purpose for Early Phase IV Studies:
Expanding Compassionate Use of Medicines Still in Clinical Trials

For dying patients with no therapy that is working, compassionate use programs exist to provide drugs in phase II or III clinical trials when the patient was ineligible to participate in the trial. But those programs are still very small, and could easily be expanded if the companies could be compensated above cost, and the patients could be monitored closely. Many cases have been cited in compelling testimony presented to various forums including think tanks and the Congress of dying patients who might have received an extra few months or years had they benefited from drugs on their way to approval but who were denied. Those patients and the companies concerned could agree to, in effect, early phase IV studies. Such specialized studies can actually be started even before phase II or III trials are completed to enable patients with little other recourse and who were denied entry into clinical trials, to receive investigational new medicines in phase II and above for compassionate purposes.

A Compact for Patients, the FDA and Industry

Such expanded, compassionate use programs with monitoring using the system we suggest could be the nidus of a transformation in the willingness of patients, companies, the FDA, Congress and society at large to support phase IV trials, albeit in innovative more cost-effective new ways, such as what we suggest. Once it is clear that industry, the FDA and patients' advocacy organizations are supporting changes that make it possible for early phase IV studies to be conducted to support compassionate use, a much greater number of phase IV trials could be done later as well, and the resulting widespread studies will help make medicine safer and more effective for all.

Dr. Chacko is Professor of Health Sciences (Adjunct) at Canada's Simon Fraser University, Vancouver, and a Consultant in public health & medicine and finance. Dr. Proschan is Mathematical Statistician, US National Institutes of Health, and Advisory Panel Member, US Food and Drug Administration.