Concerns over risky medicines are shutting down clinical studies and boosting demand for comparative drug analysis.


The growing alarm over harmful side effects related to a number of popular prescription drugs is affecting a range of issues of critical importance for clinical trial sponsors and investigators. The National Institutes of Health (NIH) recently halted important clinical trials due to fear that the painkillers under study increased risk for cardiovascular events, and more studies are under review. At the same time, support is growing for more comparative analysis of drugs and medical treatments, as well as expanded postmarket surveillance. 

Congress may take action to stiffen requirements for hospitals and physicians to report adverse drug events. Safety concerns also may slow down the campaign to legalize drug importing from foreign nations. Import advocates maintain that the Food and Drug Administration's failure to keep unsafe products off the market illustrates how foreign drugs really are not the problem. Amidst all these developments, FDA has been put on the defensive as policymakers propose to revise how the agency evaluates pre- and postapproval drug safety data.


An alarming result of recent drug safety disclosures is the growing nervousness of the research community about conducting clinical studies that involve pain medications or other high-risk treatments. 

In December, researchers halted an important NIH-sponsored cancer study designed to test whether anti-inflammatory drugs might help prevent colon cancer. The data from more than 2,000 patients in the APC (Adenoma Prevention with Celecoxib) study showed some evidence associating Pfizer's Celebrex with cardiovascular events. Pfizer executives decided to keep Celebrex on the market pending further review of the data, which the manufacturer believes is fairly inconclusive on the safety front and still may reveal beneficial anticancer effects. But Pfizer did comply with an FDA request to halt all its direct-to-consumer advertising and to provide more risk information to doctors pending further analysis of safety evidence. 

A few days later, NIH halted another prominent study due to surprising evidence that another anti-inflammatory might have harmful cardiovascular effects. The three-year-old Alzheimer's Disease Anti-Inflammatory Prevention Trial (ADAPT) had been testing naproxen (Bayer's Aleve) and Celebrex on 2,400 elderly patients to see if the medicines might help delay the onset of Alzheimer's disease in individuals with a family history indicating increased risk for the condition. NIH officials and study leaders put the trial on hold after detecting signs of increased risk of cardiovascular and cerebrovascular events in patients taking naproxen--but not Celebrex. Because the study participants were healthy, the sponsors decided it was unethical to expose them to an added risk without clear evidence of an offsetting benefit. Investigators also stopped enrollment in a number of NIH cancer trials seeking evidence of additional uses for commonly prescribed painkillers. Some 40 studies testing Celebrex for various purposes are now under review. 

The analysts acknowledged that they examined the adverse event data from these trials earlier than planned due to all the bad news about COX-2 safety problems. In the APC cancer study, the data is particularly confusing because the same group of cardiology experts reviewed both the NCI trial and a similar study conducted by Pfizer, which failed to exhibit the same heart attack risk evidence for patients taking Celebrex. It may be that additional side effects have surfaced in some cancer trials because they use much higher doses than what is recommended for arthritis patients. The reviewers felt, though, that even very slight risks could be dangerous because more than 20 million people have been taking these painkillers.

The unexpected bad news about naproxen cast a shadow over chronic use of other nonsteroidal anti-inflammatory (NSAID) drugs. FDA issued a statement advising patients using nonprescription naproxen to follow the label carefully and to stay within recommended doses. Some doctors believe that aspirin is the only painkiller with clear cardiovascular benefits, but note the higher risk of stomach bleeding. Everyone acknowledged that they were being ultra-cautious, and that a much broader review of the long-term impact of all anti-inflammatory painkillers is needed. FDA is working with the Institute of Medicine (IOM) on its drug safety study and plans to address COX-2 issues at a meeting of its drug safety advisory committee at the end of February.

Another drug safety response is increased interest in research to compare the effectiveness and safety of similar medicines. The usual objective of such analysis is to reduce the use of more costly drugs that lack clear benefits over older therapies. But experts also consider comparative data requirements a way to pressure sponsors to conduct the studies needed to prove superiority. 

The Medicare Modernization Act of 2003 authorized the HHS Agency for Healthcare Research & Quality (AHRQ) to support more research on comparative effectiveness of medical interventions (including prescription drugs) for the 10 top conditions affecting the elderly. AHRQ's 2005 budget includes $15 million for this program--much less than the $50 million initially proposed, but enough to get the program started.

With that funding in hand, in December AHRQ issued a long-awaited list of priority conditions that warrant comparative review: ischemic heart disease, cancer, asthma, stroke, arthritis, diabetes, dementia, pneumonia, stomach ulcers, and depression.AHRQ will fund projects to survey and assess the scientific literature on the outcomes from various treatments for these conditions. AHRQ is not supposed to set national standards of care, but such analysis is likely to lead to more common agreement on the effectiveness and risks of certain interventions. The aim is to help physicians and patients make informed treatment choices by communicating the results to Medicare, Medicaid, health plans, prescription drug plans, and the public. 

Pharmaceutical companies maintain that any comparative drug analysis should focus on effectiveness more than costs, include drugs among a full range of treatment options, and recognize that patients respond differently to various treatments and products. PhRMA (Pharmaceutical Research and Manufacturers of America) made these points in a white paper on government-supported outcomes research issued in January 2004 in anticipation of increased involvement in this area by AHRQ, NIH, and FDA. Industry's main concern is that comparative studies could block access to products deemed less effective and keep new products off the market, even when those drugs offer benefits for certain patient populations.

Sponsors have long opposed any move by FDA to require comparative clinical studies of new versus existing drugs to bring a new drug to market. But charges of insufficient safety testing may build Congressional interest in such policies. Private insurers and payers are supporting more comparative studies of similar drugs, as seen in the growth of the Drug Effectiveness Review Project operated by Oregon's Center for Evidence-based Policy Support. It now is providing 11 states and other organizations with head-to-head comparisons of drugs that treat common health problems such as diabetes, migraine, and cholesterol. The prime objective of the research is to help public and private health plans select the most effective and cost-effective drugs for formularies. Such analysis is likely to pay more attention to the comparative safety of similar drugs in the wake of current concerns.			

View from Washington

Drug Safety Fears Threaten Research

Writing computer code isn't as difficult as you might think.


Warning! This column goes over the deep edge. If you aren't from the world of computers and information technology, it may even shock you. Why? Because I am going to suggest that you are probably already writing your own computer programs today, and should be routinely writing programs in the future. 

 Let's say that you were a "frequent buyer" on eBay, perhaps one with several hobbies. Each day after dinner you did a variety of complex searches and browsed through the results, carefully watching for the right item at the right price. 

If it was easy to do, it would be great to write a little utility program that automatically performed the searches you wanted each day while you were at work, and saved the results for you. Each night you would simply review the results and save the hour it took to do each individual research. In this case, there already IS a little program that does this task very well (http://www.timeblaster.com/tbeindex.shtml). The point is that there are many such tasks that could be automated, saving time and effort on a daily basis. The problem is that it is so darn difficult to program something like this.


It is worth reviewing some of the history of computer software to understand how programming came to be as complex as it is. Suffice it to say, this is going to be a rather limited and superficial treatment of a complex subject. 

As most people know, the computer is an electronic tool that works with pure binary information--1s and 0s. The only real information in a computer system is a string of 1s and 0s. For example, <hi> is encoded in ASCII as:

Text, numbers, and instructions to the computer are all "understood" by the computer in pure binary. A "machine language" understands segments of this binary stream as instructions to add the binary numbers (in groups of 8, 16, 32 or 64), compare one to another, or keep track of where they are in memory. Programming a computer is nothing more than providing the instructions to do this adding, comparing, and tracking. Over and over again, and very, very fast. 

Obviously, human manipulation of the 1s and 0s can be overwhelming for even the tiniest of problems. So, the manipulations were changed from base 2 (binary) to base 16 (hexadecimal) where numbers, letters, and instructions could be more easily read and manipulated. Our <hi> becomes 3C 68 69 3E. As you can see, even this doesn't dramatically improve the readability of anything. Therefore, a language known as assembly was developed to convert the machine language into "readable" chunks. For example:

fld 32real [0001BEC8]
fchs
fadd 32real [0001E000]
fdiv 322 real [0001BEC0]
fstp 32 real [ebp-08]


Assembly language is far more readable than machine language, binary or hex, but it is still a long way from being useful for complex computing. However, in the early 1950s, machine and assembly language were the only ways to program a computer. In those early years, a fanciful project was started that was intended to be used for the "automatic programming" of computers. The FORTRAN group (standing for Formula Translator) developed a language that is still in (modest) use today for the programming of scientific problems. FORTRAN condensed large numbers of assembly language instructions into a single, much more readable statement. For example, our assembly language instructions in FORTRAN would read: 

part of a FORTRAN program for converting temperatures from Celsius to Fahrenheit. FORTRAN created a revolution in programming, allowing for dramatic advances in the development of computers. Computer languages developed for specialized purposes since that time; one particular development was oriented towards the creation of a language that was even more like natural language.

COBOL was developed towards the end of the 1950s as the Common Business Oriented Language. It was truly a language developed by committee. This committee was made up of government (civilian and military) and industry (the major computer manufacturers--IBM, Burroughs, Honeywell, RCA, Sperry Rand, and Sylvania). The intention of the language was for it to be easily readable, to essentially allow programming in English for business and financial uses. A COBOL instruction might read:

     SUBTRACT 32 FROM FAREN AND
DIVIDE BY 1.8 GIVING CELSIUS


The language was hampered by its cumbersome nature, although a more symbolic expression of the language could be used. It is still in heavy use in the financial and business community, mostly in legacy applications. (The vast majority of the famous Y2K problem involved COBOL programs.) Although it intended to bring programming to the masses, COBOL never achieved its promise.

Structured programming continued to develop in the ensuing years, with languages such as LISP, Algol, Pascal, and C developing to meet specialized and general needs. Another high-level language similar to FORTRAN was Basic--developed as a simple, easy-to-learn language for college students who weren't programming oriented. Basic was the right language at the right time. It was included with the operating system on the first IBM PCs, and its use skyrocketed. For the most part, however, Basic required considerable education, skill, and commitment.

Computer languages have continued to develop. Two programming trends, object-oriented programming and the virtual machine, have led to the development of a pair of commonly used professional programming languages--Java and C#. Both of these are challenging languages for the novice or casual user. They are typically developed in a "visual" environment in which the programming of the user interface is often automated. But they're quite distant from the ease-of-use hallmark that had been programming holy grail since the earliest days.


The toy manufacturer Lego has been selling a robotics set called Lego Mindstorms (http://mindstorms.lego.com/) for over five years now. Using the kit, older children can construct a robot out of Lego blocks, small engines, and light and touch detectors that can perform pretty sophisticated tasks. The "brain" of this robot is a programmable Lego block (called an RCX control module) that can operate the motors and get feedback from the detectors. It is programmed by downloading a program from the computer. The programs are built by clicking together "blocks" on the computer screen. Each "block" does an action. 

Can real programming be as easy as programming a Lego Mindstorms robot? Many believe it can, and there are some excellent examples of so-called Visual Programming that currently are in use. One of these is known as LabView. This very sophisticated software package allows lab scientists to control lab instruments and to acquire and analyze data from these instruments using a symbolic manipulation of objects, in a fashion reminiscent of the Mindstorm blocks. The package can also allow scientists to perform simulations of instruments and many other tasks.

There are scattered uses of visual programming in clinical trial technologies as well. Clinical simulation tools for clinical trials allow patients, measurements, drug dosing, and other objects to be snapped together graphically in a similar fashion to that of LabView, creating a program for simulation of clinical trial results. In addition, some workflow applications for the processing of clinical trial data allow the creation of complex workflows through "snap in" manipulation of icons, leading to actual transfer/approval workflow of documents. 

In the case of all of these visual programming applications, the manipulation of graphic icons on the screen creates a significant amount of traditional code in an underlying application. Thus, the non-programming user is in fact doing highly sophisticated computer programming without knowing it.


Many readers of this column already do regular programming in their daily lives. They use languages that are close to "scripting" languages--highly specialized for their specific tasks. Those who are involved with statistics are typically quite expert in the SAS programming language, and many in S-Plus or the closely related open source R programming environment. Others are more involved with databases and are very facile with programming in the natural language-like SQL (structured query language) programming. SQL queries are so straightforward and easy to learn that many forget that they represent a form of programming. So, that takes care of 10%--25% of readers. What about the rest?

Another form of simple programming is the macro, found in many programs from Microsoft Word to Adobe Photoshop. In its simplest implementation, a macro can watch what a user does and repeat it on demand, when selected from a menu or sequence of key strokes. For example, if you wanted to sequentially do "search and replace" for 10 different character sequences in Word, you might begin the recording of a macro and then perform the searches. When finished, you would stop the recording of the macro and name it. Then, to perform the searches in the future you would invoke the macro, sit back, and watch. 

Behind the scenes, the recording of a macro creates a sophisticated set of Visual Basic for Applications (VBA) instructions. If you are a "power user" you can edit the VBA and improve the macro without redoing it. Unfortunately, many viruses use VBA in Word documents; thus, macros are disabled by default in most Word installations. 

In Adobe Photoshop, macros are called "actions." An action allows a user to record a large series of manipulations to photographs that can then be repeated over and over in batch on a series of photographs.  Macros and actions are examples of programming that can easily be performed by a nontechnical user.

Another form of macro is one that you may already be using. In Outlook, it is very easy to configure a filter for incoming messages that checks for certain conditions (sender, recipient, text in the subject or body) and takes an action such as forwarding or filing in a particular folder. Many people sort all their mail automatically in Outlook and review folders rather than the InBox. This macro also has VBA running underneath, but the software instructions behind the design and implementation are absolutely hidden from the end user. If you haven't created macros in Word or filters in Outlook, go ahead and give it a try. You may be amazed at the usefulness of these functions.

More and more people are becoming comfortable with the use of HTML, which is the very straightforward language by which simple Web pages are "programmed." Without adding scripted elements (e.g., Javascript, PHP, and others) most HTML has little to no logic built in, and an HTML Web page may not appear to meet the minimal requirements of computer programming. However, many who are using HTML for building Web pages are often using a Web-page building program (or online environment). That program can insert scripting, and even active, server-based elements into a Web page without the user writing a single line of code. By dragging and dropping an active element onto a WYSIWYG (what you see is what you get) Web page designer, the user is actually doing active computer programming. 

You might protest that the act of computer programming involves the writing of computer code--not true. Even sophisticated Java and C# programming is usually done in a visual, WYSIWYG environment in which large blocks of code are inserted into a program automatically.

SQL is used behind the scenes in a variety of database settings. For example, consider a search of a database on a Web site (think Amazon or eBay) or use of an ATM machine. You are actually using buttons to create custom SQL queries that are run against a database to either display your results or take actions like withdrawing money. The original developers of SQL at IBM thought that they were creating a language that could be used by everyone. Unfortunately, the rigidity of syntax was too frustrating for nontechnical users. However, when this can be automated, SQL is a very powerful language for everyone, on a daily basis.


While the mainstream of computer languages progressed from FORTRAN to Java, another very important side story developed. It may provide the bridge between computer programming and the casual programmer--that of scripting languages. We have already mentioned scripting languages. Javascript and PHP are used in the development of Web pages and VBA is a form of scripting language for Word and Excel macros. 

Scripting languages are a hodgepodge of different languages that range from very simple to fairly sophisticated. Some scripting languages are used nowadays to write large computer programs: Python is one example. But they are generally intended to be used to connect one component (application or data source) with another. Most of the languages we have discussed so far are compiled languages--before being used, they are typically converted into an executable file (.exe) consisting of machine code which operates directly on the operating system and the memory. (Java and C# are a bit different, but that's another column.) 

Scripting languages are interpreted languages. These programs exist in plain text form and are run through an interpreter program, which reads and executes each line of code at the time a program is run. As such, a program written in a scripting language may be slower and take more memory to run than a compiled program. But scripted programs are much easier to write and maintain, because many of the mundane tasks of programming are either done automatically or are not needed. The result is a program that can be written quickly and efficiently. 

The most common use of scripting languages is to tie together diverse applications. For example, each hour start a program, take its output, put it in as input into another program, and save that output into a file. One very basic scripting language is the set of MS-DOS commands used in batch files, which are simple text files ending in .bat (anyfile.bat). Batch files used to be very important for PC users, automating a number of maintenance tasks now managed by Windows and utility programs. It is still perfectly useful to create a batch file to automate some repetitive task. For example, if your daily work required you to start three programs simultaneously, and then open two specific files, you could accomplish this easily with a three-line batch file that can be double-clicked. All three programs will open, along with the two selected files. 


One form of computer "programming" is often taken for granted--the spreadsheet. The spreadsheet (typically, Microsoft Excel) is a highly specialized environment for programming a series of numerical calculations. While many people use Excel as a simple database, or for adding a column of numbers, most of us have had the opportunity to see very complex spreadsheets. If you haven't already done so, you might want to take a look at http://www.exinfm.com/free_spreadsheets.html, where you can find many useful, free spreadsheets. Interestingly, spreadsheets are an example of a programming philosophy known as "programming by example"--show the program what type of answer you want, and it figures out how to get it for you.

If you are really ambitious, try learning a fun and easy scripting language that is tailor made for the Internet. It is called Rebol (http://www.rebol.com) and the core system is available as a free download: http://www.rebol.com/view-platforms.html. Rebol is easy to learn, and with it you can quickly write very useful scripts that can read Web pages, extract specific content, bulk email, and much, much more. Using Rebol, it wouldn't be difficult to write the simple, but functional eBay searching program from the beginning of this column. It is a very different and easy place to start in understanding the power of scripting languages.

If you want to learn a great deal more about the birth and development of software, I recommend the book 

 by Steve Lohr. Many of the ideas and examples in this column came from this book.			

Technology Viewpoint

Writing computer code isn?t as difficult as you might think.

Do It Yourself Software

Speaking on clinical research fraud at the 

 Summit in Brussels last year, we were asked by someone why we were worried about it. "After all" he said, "surely it's a victimless crime?" 

Much has been written about research fraud and there have been many high-profile cases. But there has been little—if anything—written about its impact on the victims. There are obvious victims: sponsoring pharmaceutical companies. But of great concern to us are the direct victims of the fraud: patients given inappropriate treatment or denied full safety assessment while taking an experimental drug. Then there are the invisible victims, patients of a doctor removed from practice having been found guilty of research fraud. We'd like to share the stories of some of these victims.   

Many protocols require patients to be treatment-naïve or uncontrolled on present medication. When we interviewed patients in a depression study, we saw patient notes altered to make it appear that the patient fitted the inclusion criteria. The records showed that a husband and wife participated in the same study. Both agreed that they had been asked to take part in the study, both discussed participation, and both had signed consent forms. So what, the husband was asked, was his history of depression? "I've never been depressed in my life!," he responded. The doctor had explained that it would help his wife if he went into the study with her. "I did it all for her." On the face of it this seems almost comical, but he had had a miserable 12 weeks with every known side effect. His wife, however, complained that she had received no benefit whatsoever from being in the study. He had received the active medication, and she—who needed the drug—had received placebo. Both of them suffered; she was denied medication and he was treated for a condition he did not have. Furthermore, his notes now show that he suffered from depression, which may influence future choices of treatment, employment, and insurance.  

Sometimes, the impact on patients can be financial as well as a loss of trust. One elderly woman on income support was phoned by her general practitioner (GP) late one afternoon. The GP needed her to come to surgery to have blood taken for tests. He persisted, and she took a taxi at a cost equivalent to two days food. We later discovered that she was in three studies at the same time, and that the blood samples were needed to fulfill the needs of the protocols. She had no knowledge of any of the studies, and was so distressed and appalled by the abuse of the doctor–patient relationship that she travelled to London (several hundred miles) to give evidence from a wheelchair at his court hearing. She died less than two weeks later.  

She was not the only patient in that practice who adored her doctor. Many protested when he was charged with research fraud, and held a collection to fund his appeal when he was found guilty.

 When he failed to launch that appeal, the realization dawned that the charges were true. Thousands of patients were deprived of their doctor; despite his fraud, he was widely held in great respect and affection, and his patients found it hard to understand.  			

New drug development revolves around the need to show efficacy, safety, and quality. Unless all three are measured accurately, any conclusions will be false. A clinical research associate once returned from a monitoring visit worried, but unable to pinpoint what it was that made her uneasy. Her manager, a scientist with many years' monitoring experience, performed the next visit. He asked the study nurse how she measured and recorded heart rate, and was told that she counted the radial pulse for 15 seconds and multiplied by four. Yet very few of her patients' heart rates could be divided by four. At this point she broke down in angry tears. The unit had many studies with insufficient research staff and she couldn't do all the tests laid down in the various protocols on all the patients. So she made up some of the data. Her justification was that she "never made up important data. That's all real." This was a drug just starting its course through Phase III, with much still unknown about its safety and efficacy. She had no reason to believe that the heart rate would be an unimportant measurement.  

One might expect investigators working with drugs known to affect heart rate to be more careful about recording cardiac function. But sometimes fraudulent doctors seem to get careless. Years ago, electrocardiograph (ECG) recordings were done on long strips of heat-sensitive paper, which were cut up and mounted on a sheet of paper. Part of the safety assessment in one study comparing a calcium channel blocker with a beta-blocker was an ECG taken before and after 12 weeks of treatment. One would expect that patients on the calcium channel blocker would show a faster heart rate on treatment, and those on the beta-blocker a slower one. Amazingly, in this study no patient showed any change whatsoever. Looking closely at the ECG tracings we realized the benefit of long strip recordings. The paper manufacturer's name was repeated all along the lower edge. When the section for each lead's recording (I, II, III, etc.) was cut up for mounting, the cut inevitably passed through words. It was easy to match the cut-up words and see beyond doubt that the investigator had done one double-length recording at visit one, then cut each lead into two parts before mounting onto two separate sheets. He hadn't noticed that lead I on the pretreatment trace had "

" written along the bottom, while on the same lead on the alleged 12-week trace was "

" This breach of safety assessment was quite clearly crucial to the safety of patients.  

Not all fraud is so evident. Ethical review of clinical research is vital to protect the rights of patients. But because it can take time to obtain independent ethics committee (EC) or review board approval, the temptation to start a study without it can become very strong. When investigating cases of alleged research fraud, we approach ECs to ask for details of other studies they have approved for the investigator under suspicion. We have seen several studies where EC approval has never been sought, although sponsoring companies have been told the contrary. With scanners and modern copying equipment, it is relatively easy to produce a letter seeming to give EC approval that would pass casual inspection. But the EC chairman only has to compare it with the minutes of relevant EC meetings for that fraud to become glaringly obvious. If that check is not made, a whole study's worth of patients at that site have been denied the most basic level of protection, that of impartial protocol review.  			

Another real root of ethical research is informed consent. Yet one of the most common responses we hear from patients in a study where we suspect fraud has taken place is "Study? What study?" Some of these cases have involved patient safety, some "just" an abuse of their rights. But all of these patients are victims.  

One young man is lucky to be alive. He presented to his GP with suicidal thoughts. His GP told him he had some new treatment that would help him enormously, but that as it was not on the market in this country he must sign a receipt for it. This the young man did, but his failure to respond so concerned his parents that they took him to hospital and got him effective treatment. It was only when we were asked to investigate possible research fraud that the truth became known. What he had signed was actually a consent form, yet he had been given no information that he was a research subject taking unlicensed medication. The fact that his allocation was to the placebo group merely made matters worse.  

The drug in the desk drawer is a familiar story. We were asked to look into a hormone replacement therapy (HRT) study. Although informed consent forms had been signed, the signatures all looked very similar to each other and to the handwriting of one of the study staff. The study involved the women taking HRT over several months, with a biopsy of their uterine lining being done before and after treatment. All had been supplied with the drug "as a personal favor" by their GP. Many had been with their GP for years and had trusted him with their family's care. To find that he had used them for personal gain and had performed invasive biopsies not for their good but to fulfill the terms of the protocol was devastating to all of them, and when last heard at least one was considering civil action for assault over the biopsy.  			

Recourse to the civil courts is also technically open to pharmaceutical companies, though we have not heard of this being exercised. Yet research fraud is an expensive matter for the pharmaceutical industry. It is not often that an entire study must be scrapped because of fraud, but we have recently confirmed as fraudulent the seventh investigator in an 11-center study. Clearly the results of that research will never be used. 

Withdrawal of a drug is much more expensive. When done because of the results of fraudulent research, the price is too high. William McBride, an Australian obstetrician, was one of the first to highlight the dangers of thalidomide for pregnant women, suggesting that the drug caused limb deformities in babies.

 Twenty years later he published research on Debendox, a drug for morning sickness, claiming it caused birth defects in rabbits.

 The manufacturers took the drug off the market and lawsuits followed. Only when nobody could reproduce his work did it transpire that McBride had altered research data: in fact, Debendox had no ill effects. Ten years later he was found guilty of scientific fraud by a medical tribunal and removed from the Medical Register.

 The obvious victim was the drug company, but what of the pregnant women who had already taken Debendox during early pregnancy? There must have been many thousands who continued their pregnancies in fear and dread.  			

Another victim of this case was Phil Vardy, a scientist working with McBride who blew the whistle on him. When he confronted McBride he was sacked, and found it impossible to get another job until he moved away. He lost his marriage in the process.  

It is hard to feel sorry for fraudulent doctors who lose their license to practice medicine, as the choice to perpetrate fraud was their decision. But it is callous not to feel sympathy for their spouses and families, suffering a drop in standard of living and public shame and dishonor. At least two doctors have taken their own lives when sponsoring pharmaceutical companies were getting close to unravelling their long-term research fraud. Both were family men, so two women were widowed and several children grew up without their fathers. In both cases they were popular GPs, so many thousands of people were affected by their deaths.  

Like these two, in most cases of fraud we have investigated deceit has been practiced for many years—on average five to 10 years. In many cases, one or several pharmaceutical companies have been using the site on an almost continuous basis. Quite often, therefore, one CRA or monitor has been visiting the investigator for a significant length of time. Pharmaceutical company staff understandably get possessive of their sites and fond of the staff, and it is easy for them to be flattered when an eminent researcher seems to welcome visits and calls them by their first name. But we have seen the human response to finding out that a respected investigator is a fraud: the CRAs are devastated. It is infidelity of a sort, and the fact that it is professional rather than personal doesn't make it any less painful. The CRA has to face the fact that he or she has been duped by someone they respected, someone they thought valued them. 

Research fraud can never be "a victimless crime." There are victims on many levels, obvious and hidden, financial and personal. Money is money, but we can't replace the lost trust in the patient's doctor, and we can't make them feel less used and abused. Nobody can retrospectively run safety assessments on patients who have been studied without the right tests at the right time. Nobody can make it right to carry out research without informed consent, to give these patients back their most basic human rights. Nobody can give widows and orphans what they want most in the world.  

 Summit, we asked a large group of delegates whether they had ever seen what they believed to be evidence of clinical research fraud. About 75% of them raised their hands. In Munich in 2004 we posed the same question to another group and saw a similar response. This is particularly worrying, as most of the delegates were at the front line of research. How many prosecutions have ensued? We know of very few outside the United Kingdom, but one pharmaceutical company with a multinational Phase III study has closed seven centers in one European country for fraud. It is time the proactive UK approach to the problem extended into Europe—there is no acceptable reason to ignore and condone such outrageous behavior. 

The true victims of research fraud are very real. They are just like us, and members of our family. And while research fraud is allowed to continue, they will continue to suffer.  

Whistleblower. In Fraud and Misconduct in Biomedical Research

, 3rd ed., S. Lock, F. Wells, and M. Farthing, Eds. (London: BMJ Books, 2001).  			

2 W.G. McBride, "Thalidomide and Congenital Abnormalities," 

3 W.G. McBride, P.H. Vardy, J. French, "Effects of Scopolamine Hydrobromide on the Development of the Chick and Rabbit Embryo," 

4 G.F. Humphrey, "Scientific Fraud: The McBride Case—Judgment," 

MBBS, is medical adviser and director, and 

is co-founder and chief executive of MedicoLegal Investigations Ltd., Nup End Business Centre, Nup End Green, Knebworth, Herts, UK SG3 6QJ, telephone and fax +44 (0) 1438 820000, email: 

*To whom all correspondence should be addressed. 

Feature Article

Because it can take time to obtain independent EC or review board approval, the temptation to start a study without it can become strong.

Clinical Research Fraud: A Victimless Crime?


In response to the "From the Editor" column about placebo-controlled trials (

, August 2004), we think that placebo use is more a tradition than an enigma. Placebo trials were the gold standard for evaluating new treatments for many years, but nowadays we need to change that point of view. Medical advances, clinical trials design development, the evolution of human subjects' protection, and ethical principles force us to rethink the idea of placebo as the best control arm.

 is not uncommon. Who would choose the placebo over the untested drug? We feel that most of the patients enrolled in the clinical trial would have chosen the untested drug if they had been given the choice. Only those patients who decide to take part in a clinical trial due to altruistic principles would have another point of view. But, most likely even those subjects wouldn't have any predilection to only receive the placebo as treatment.

We feel that researchers should not design a clinical trial using a placebo because they are not giving their subjects the option to receive the best treatment. Regarding medical practice, patients and health professionals should look at the differences between active treatment options before they decide to choose a treatment or medical intervention. By using placebo arms as control gold standards we are comparing the treatment to an infrequent or unrealistic decision point in medical practice.

Placebo arms should be questioned in clinical trials even when there isn't an active treatment. For example, testing a new treatment with objective measurements as an endpoint when no placebo effect is expected for this variable. In these situations we should consider the best supportive care as the possible control arm.

Placebo effect can be positive by itself. Patients and physicians are looking for positive effects, even though it may be a placebo effect, to improve their health condition. The quantification and quality of the effect compared with an active treatment is more important than determining the placebo effect in the overall positive effect. 

We feel that some researchers do not realize that they are exposing the placebo arm subjects to additional risks. Placebo treatment is always an experimental intervention and sometimes is an uncomfortable intervention. Negative effects in patients receiving placebo treatment have been reported frequently.

 The effects could be due to the nontreatment of disease itself or to the anticipation by subjects of adverse events that have been described to them as likely to occur in the active treatment arm. Only in those medical situations with no active treatments and with some expectation of placebo effect should placebo be used as a control arm in clinical trials. We believe that it is important to improve public opinion about clinical trials and there should be restrictions associated with placebo-controlled trials. This could lessen the effect of the guinea pig opinion of clinical trials. But, traditions are very difficult to change. Yes dear Editor, we agree with you about the woman of the story from the 

: "Unfortunately she received the placebo."

This is our personal opinion about placebo and not necessarily the official position of our center.


1. G. Chevtzoff and I.F. Tannock, "Placebo Effects in Oncology," 

Alexander Montenegro Surs,
DrPH, MSc, Prof, Project Manager
Magda Elaine Monreal Agero,
DrPH, Clinical Research Associate
National Co-ordinating Centre of Clinical Trials, Havana, Cuba

In response to Toby Hindin's  "From the Editor" column (

, September 2004), the state of the conduct of clinical research in the United States must first be placed in perspective, and the good that has evolved from the clinical trials enterprise must not be forgotten amidst the violations that have received the primary attention of the media. While one violation is one too many, a great number of investigations are conducted conscientiously and in full compliance with all applicable guidelines. Regarding legal parameters, I do not believe that it is a question of further enactment of laws, but of more educated application and beneficial enforcement guided by constant reflection on the ethics of research involving human subjects.

As an example of such an education and enforcement approach, the National Center for Research Resources, as part of the National Institutes of Health, has provided funding for approximately one full-time Research Subject Advocate (RSA) at each one of its 80 General Clinical Research Centers located throughout the United States. While the primary function of RSAs is to ensure that studies are designed and conducted safely and ethically with the protection of human subjects given highest accord, the scope of their activities are varied from individual subject attention to systems and process approaches--all viable constructs. 

Moreover, this concept has spurred interest at several parent academic research institutions to employ similar approaches to the broader net of clinical investigations being conducted. The benefits truly are to educate and protect all parties involved in clinical research from subjects to coordinators to investigators, and in some instances even sponsors. It is anticipated that such enforcement will enhance the current environment without compromising the objectives of advancing medicine and improving quality of life for the citizens of this country and the world.Note: This letter does not necessarily represent the views of other advocates for research subjects nor the Society of Research Subject Advocates.

Carson R. Reider, PhD
Research Subject Advocate
General Clinical Research Center
The Ohio State University


Letters to the Editor

Readers respond to articles and commentaries.

In the biomedical research industry, and at Pfizer in particular, we are committed to what we call "The Three As"—making medicines available, accessible, and affordable.

Peter B. Corr, PhD, Senior Vice President, Science and Technology, Pfizer Inc.			

These "Three As" are closely interconnected. The more therapeutics we make available and the more accessible we make them, the more affordable they become. The whole system, however, rests on the foundation of availability. And availability is a matter of science, supported by business. 

Without a doubt, recent discoveries will yield many valuable therapeutics over the next decade or so. But we're at a critical juncture in biomedical science. In a sense, we are the victims of our own success. Science has become such a massive, complex body of knowledge that no person, discipline, institution, or industry can manage it alone. 

The technological demands and complexity of today's research methods are just mind-boggling—and expensive. At the same time, a growing array of regulations generates its own costly challenges. 

But I am optimistic. In fact, I believe we are on the cusp of a new paradigm of biomedical research, which centers on an astute assessment of medical need, careful assay of scientific possibility, and deliberate partnerships to maximize our use of resources. 

I believe the biomedical industry is the organizing force behind this new paradigm, and is the dominant factor in biomedical research today. Yet the complexities, risk, and expense of research demand that all members of the research community coordinate their efforts.  

We simply must develop rich networks of partnerships from all over the world—between public and private research organizations, basic scientists and engineers, academia, and the biomedical industry, including both not-for-profit and for-profit enterprises. 

We can accomplish much more with less cost and duplication of effort when a variety of institutions and organizations pool and coordinate their resources and bring their special strengths to bear. Clearly, we have begun to recognize that our work is complementary and that the practical, institutional, collegial, and legal structures of these new forms of partnering should begin to take shape. 

This new paradigm of biomedical research partnerships will have a huge impact on the Three As. In short, intelligently formed, interactive, cooperative ventures will turn more science into therapeutic discoveries that will change the course of devastating human diseases. 

Impediments to access today suggest that we must work as part of a team with national governments, nongovernmental organizations, multinational agencies, medical professionals, patient organizations, academic institutions, and an informed media to build sustainable infrastructures that will improve health conditions around the world. This will provide access and infrastructure to deliver these therapies in regions of the world where this is not yet possible with consistency. 

By sharing resources and infrastructures, such partnerships also limit the risk to all partners. With more institutions and organizations "owning" some part of the process, we'll have more avenues to distribute medicine and more people with a stake in its success. 

What's more, better coordination between regulators, government agencies, and industry will remove barriers and facilitate further development of needed infrastructures. 

Now, what about the third A—affordability? Today, the biomedical industry routinely invests billions of dollars in programs to improve access and affordability. 

But let's be clear. The discovery, development, and delivery of biomedical innovation is risky, complex, time-consuming, and very expensive. So, we expect those who can afford to pay for medicines to pay. And for those who can't afford to pay, we will do all that we can to see that they have medicines they so desperately need. For instance, Pfizer Helpful Answers (

) and other programs such as Together RX Access help those without prescription drug coverage every day with access to prescription products. 

Sharing responsibilities and risks among a broader group of public health stakeholders enables us to harness our unique skills, capabilities, and capacities to expedite delivery and needed treatment. 

With sure hands and wise hearts, we can, together, provide an attainable level of good health care as fairly as possible for everyone on the planet, and bring the light of hope and health to those still groping in the dark.

 PhD, Senior Vice President, Science and Technology, Pfizer Inc.

A Closing Thought

In the biomedical research industry, and at Pfizer in particular, we are committed to what we call "The Three As"-making medicines available, accessible, and affordable.

A New Paradigm for Biomedical Research

Industry's proposal for a shared database has merit, but is it "too little too late"?


The rest of the world caught up with Europe in January, when an agreement was finally reached on an international joint position statement on disclosure about clinical trials. At least that's what Europe is saying. 

The European Federation of Pharmaceutical Industries and Associations (EFPIA) claims that it was ready to come out with a commitment on public access to trial data way back in 2004. But it deferred its own publication, it says, so that a common announcement could be made, in concert with other major drug industry associations around the globe.

In the end, when the 6 January "Joint Position on the Disclosure of Clinical Trial Information via Clinical Trial Registries and Databases"

 was released by EFPIA, the International Federation of Pharmaceutical Manufacturers and Associations (IFPMA), the Japanese Pharmaceutical Manufactures Association, and the Pharmaceutical Research and Manufacturers of America, it appeared to be more of a damp squib than a firework display of public accountability.

The public image had deteriorated significantly for clinical research, in the wake of the late-2004 controversies over the safety of many major products--notably the COX-2 inhibitors. So the commitments in the joint statement were not greeted with universal acclaim. "Too little, too late," was typical of some of the reactions to the declaration.

The unfortunate timing should not detract from the intrinsic merit of the new arrangements. They amount to an important engagement to increase the transparency of clinical trials sponsored by companies in membership of the signatory associations--which represent the vast majority of the world's innovative drug firms.


Under the new arrangement, details of clinical trials initiated on or after 1 July 2005 will be submitted for listing in a free, publicly accessible clinical trial registry within 21 days of the initiation of patient enrollment. Any one of a number of Internet-based registries may be used--such as the clinical trial registry maintained by the U.S.'s National Institutes of Health at www.clinicaltrials.gov.

The entry in the registry about each trial will include a brief title and trial description in lay terminology; phase; type (e.g., interventional); and status and purpose of the trial--whether it is for treatment, diagnosis or prevention. It will indicate the intervention type (e.g., drug or vaccine); the condition or disease; the key eligibility criteria, including gender and age; the location of the trial; and contact information.

Each trial listed in the registry will be given a unique identifier, which will permit registry users to track the trial through multiple databases, including clinical trial results databases, to ensure transparency.

The declaration also indicates that industry "is also prepared to explore the concept" of a secure, nonpublic third party electronic repository for subsequent disclosure to medical journals. In fact, Europe has already set up the EudraCT database of interventional clinical trials initiated in the European Union since 1 May 2004. This is accessible to European regulatory authorities from the time of trial initiation. Some data fields will subsequently be made publicly accessible once the product is approved.

There are some qualifications to the commitments. They cover only "hypothesis-testing" or "confirmatory" clinical trials using statistically valid plans for data analysis, and aim at providing firm evidence of safety and/or efficacy to support product claims. They do not apply to "exploratory" trials aimed at setting direction or generating hypotheses for possible future studies.

Further, they apply only where there are no alternative national requirements. And the disclosure must maintain protections for individual privacy, intellectual property and contract rights, as well as conform to the regulations in relevant countries.


For completed clinical trials, the commitments go even further. From the beginning of 2005 onward, the results of all clinical trials (other than exploratory trials) conducted on an approved, commercially available drug should appear by 13 September 2005 on a free, publicly accessible, clinical trial results database, regardless of outcome. 

And where trial results from exploratory trials could have significant medical importance and may have an impact on a marketed product's labeling, they should also be publicly disclosed. Similarly, where trial results for an investigational product that has failed in development have significant medical importance, study sponsors are also encouraged to post the results.

If trial results are published in a peer-reviewed medical journal, the database should include a citation to or link to the journal article and/or a summary of the results. This will be in a standard, non-promotional format covering trial design and methodology, results of the primary and secondary outcome measures, and safety results. If trials results are not published in a journal, the summary results should be posted on the database.

The results should include the unique identifier used to register the trial at inception, and should be posted within one year after either the drug is available or postmarketing trials are completed.

Now companies are being invited to subscribe to the joint position. Then it will be up to them to establish a process of verification for both the clinical trial registry and the clinical trial database, and to make public how they will adhere to these standards. And all trial sponsors are encouraged to commit to keeping registries accurate and up to date.


Meanwhile, the European Medicines Agency has decided to accelerate its review of COX-2 inhibitors, after Vioxx raised cardiovascular concerns.

In December, the agency reviewed summaries of two clinical trials relating to celecoxib and found that while they were not consistent or conclusive, the data indicated that an increased risk of serious cardiovascular events seen in one trial--the Adenoma Prevention with Celecoxib trial--"may be related to the dose and duration of treatment."

The similar Prevention of Spontaneous Adenoma Polyps trial did not show any evidence of an increased risk of serious cardiovascular events, overall or in subgroups. The reason for the different results "is not apparent," the agency remarked, and will continue analysis.

Based on the preliminary assessment, the agency will not be recommending any immediate regulatory action to restrict the use of celecoxib and other COX-2 inhibitors, but it undertook to review on an expedited basis the full clinical trial data and further information on other COX-2 inhibitors in January 2005. The companies concerned were requested to submit all-new data and analyses in early January for review by the Committee on Medicinal Products for Human Use.

Meanwhile, the agency has limited its advice to encouraging prescribers to carefully follow the latest version of the summary of product characteristics for COX-2 inhibitors, especially regarding the warnings and precautions in patients with a history of cardiovascular disease. And it has reminded patients that they should be aware that all COX-2 medicines available in Europe already contain warnings regarding heart problems. 

There is a lot riding on the outcome of these continuing reviews of COX-2 medicines. The credibility of the products are at stake. So are the regulatory processes themselves, in Europe and around the world. And the joint declaration from the industry on public access to registries and databases will not be enough to satisfy widespread concerns.

1. http://www.efpia.org/4_pos/sci_regu/Clinicaltrials2005.pdf			

View from Brussels

Industry?s proposal for a shared database has merit, but is it ?too little too late??

Tardy Europe Finally Signs Up on Data Transparency

It is axiomatic, because we work in clinical research, that we think we know what clinical research is. It's too fundamental an issue, too basic a concept, to give it much thought. However, there are compounds administered to patients (not subjects), as well as procedures performed on patients that are unapproved, untested, and could in no way be considered the generally accepted standard of care. These therapies, albeit well intentioned, are largely unreviewed for scientific merit. For example, a physician hears from a colleague that off-label use of a drug might be helpful in treating a patient's clinical condition and so prescribes the drug for this off-label indication. There are no statutory prohibitions on this approach, and the physician feels that he or she is in the best position to make this determination. Regulatory agencies are loath to interfere with these kinds of clinical approaches, lest innovative care be stifled.  

 Photography: Eyewire  Artwork: Paul Belci			

Innovation and experimentation are part of the daily life of some physicians and dentists. They may inadvertently cross the line from innovative care to clinical research. Both medical and dental activities are regulated by state law regarding their practice, but clinical research has additional GCP requirements, including FDA and IRB oversight. Individual patients may benefit from innovative care, but society also benefits from innovative clinical research; individual patients must be protected from substandard care and research subjects must additionally be protected from substandard research. So there are significant differences in processes and procedures applied to specific clinical situations depending upon a regulatory agency, sponsor, or IRB determination of whether it is innovative care or clinical research. 

This short paper explores current thinking/best practices regarding distinctions between routine care, innovative care, and clinical research including research that may represent innovative care. It begins with three scenarios that approach, and possibly cross, an unmarked, "fuzzy" border into clinical research, describes how regulatory agencies and IRBs look at these scenarios, and finally adds dimension and perspective to the discussion. 

 A pulmonologist decides he would like to experiment with a compound that has been previously studied 30 years ago in the treatment of asthma. He obtains IRB approval to try this compound on a few subjects, but like the surgeon below, he "innovates" and makes changes as his series of patients expands; he varies the dosing conditions and he changes the premedication treatment. 			

 For a certain kind of benign tumor that's usually found in the femur, the standard treatment is to remove the tumor, curette the area, then fill the hole in the bone with cement, or cement plus some hardware, to strengthen the bone and hopefully avert a later fracture. 

An orthopedic surgeon thinks this last step might be accomplished better with cement and a slight modification of the hardware that is usually applied, or by adding hardware specially designed by the surgeon. He has a patient he thinks might benefit from this innovative approach and he will decide which hardware to install once his patient's bone is exposed. He expects he'll probably use this surgical approach for more than just this one patient, and if he does, he'll probably do a retrospective chart review and publish his results. However, his description of this activity is that he's just going to do what he thinks is best for each individual patient, and he is being an innovator in his surgical practice. 

 An infectious disease consultant has just read a paper describing the use of an old generic antibiotic for staphylococcal skin abscesses. Coincidently, she is called to the ER that weekend to see such an abscess. With the results of the study fresh in her mind, she uses the old and cheaper, but non-standard of care—antibiotic—on this patient. 

None of these physicians seeks FDA approval of their activities, and neither consents patients with a consent that describes the treatment as experimental. Are these approaches research that fall under GCP rules and regulations or are they simply innovative care? Can treatment of a single patient constitute "research?" 

An examination of this question begins with another question: what is the FDA's expectation regarding these clinical activities? A March 31, 2003 FDA warning letter issued to Johns Hopkins asthma specialist Albis Togias is informative (Reference: 02-HFD-0303) regarding Scenario #1. First, Togias is specifically cited for violation of the Federal Food, Drug, and Cosmetic Act (USC 331(d)) by engaging in a prohibited act by causing the introduction or delivery of an unapproved new drug in interstate commerce. Second, Dr. Togias failed to submit an IND for the conduct of a clinical investigation with an investigational new drug as required by 21 CFR 312.20(a). The warning letter states that a clinical investigation is defined as "any experiment in which a drug is administered or dispensed to, or used involving, one or more human subjects — except for the use of a marketed drug in the course of medical practice" (21 CFR 312.3). These statements make it clear that participation of even one subject is a regulated activity. Dr. Togias should have submitted an investigator IND to the FDA. On the other hand, had he been using a marketed product, the case would have been less clear cut. The case is also clouded by the fact that this study was similar to a Phase I study because the subjects were healthy volunteers and were not being "treated" for a specific disease.  			

Regarding the surgeon's activities in Scenario #2, FDA regulations (21 CFR Part 812) define two types of device studies—"significant risk" and "nonsignificant risk" studies. A significant risk device study is a study that presents potential for serious risk to the health, safety, or welfare of a subject and 1) is intended as an implant; 2) is used in supporting or sustaining human life; 3) is of substantial importance in diagnosing, curing, mitigating, or treating disease, or otherwise prevents impairment of human health; or 4) otherwise presents a potential for serious risk to the health, safety, or welfare of a subject. A nonsignificant risk device study is one that does not meet the definition for a significant risk study. 

If the activity is a significant risk device study, then the surgeon must obtain approval from both the FDA and the IRB before the surgeries are initiated. The surgical procedures described in Scenario #2 above would fall under the significant risk category, and since it would appear as though this procedure has elements of research, such as modified and/or unapproved hardware, it requires FDA and IRB oversight. 

So in the first two cases, unauthorized research has been undertaken. In the third case, despite the fact that a nonstandard therapeutic intervention has been initiated, this is not research because it involves a drug marketed for an approved indication. 

The Office of Human Research Protection (OHRP) has provided yet another way to address the care versus research question. In February 1998 it published a chart that describes when a patient becomes a subject. The chart begins by asking, "Is there an intervention or an interaction with a living person that would not be occurring or would be occurring in some other fashion, but for this research?" A "yes" answer means that human subjects are involved and 45 CFR 46 applies. In other words, if the care provider's intervention is proceeding differently from normal practice, then the patient is a subject and the activity is considered research. The OHRP definition calls into question the conclusion that Scenario #3 is not research. 

Secondarily, from a less regulatory perspective, the following questions might be asked about these clinical procedures: 

Was there a clear intent before treating the patient to use systematically collected data on a series of patients receiving similar treatments?

Was the hardware, drug, or treatment being used for an unapproved indication?

Are separate data sheets for research files being completed for the purpose of reviewing patient outcomes? Is there organized data gathering?

Are extra tests being performed that are not the standardof care?

Is the innovative care being delivered consistently across a series of patients according to an "unwritten" protocol in order to keep processes and procedures uniform?

Answers of "yes" to some or all of these questions would suggest that the innovative approach is actually clinical research, which must undergo IRB scrutiny and be conducted under an IND. A more subtle question might be: Was the decision to use the hardware, and or other treatment approaches, made on a case-by-case basis, after diagnostic "discovery" began? 

In other words, is every single case treated differently, or has the potential for being treated differently? An answer of "yes" to this question would tend to suggest that the approach is less likely to be research. However, one would still expect that this decision-making process was explained to the patient before the surgery or treatment started. 

A third consideration is the locus of authority within an institution to make decisions about innovative procedures. Is there a hospital oversight committee, such as an innovative care committee, in place to review such approaches, and does the committee work closely with the hospital IRB? An oversight committee of this type protects the hospital from legal and regulatory risks but also encourages innovation in treatment modalities. 

Lastly, there may be a medico-legal duty to create research records and perform contemporaneous chart review on early patients receiving innovative care as a way to adequately protect their safety. To do less might be considered, for a procedure that is sufficiently different from the standard of care, malpractice. 

While the innovative physician's intentions regarding use of his or her findings may incrementally change over time, rather than change for one treatment, it would probably be best to switch to a research project and IRB approval earlier rather than later for a series of similar treatments. 

There have been previous attempts at drawing a distinction between research and practice. The Belmont Report, in its introduction, addresses the boundaries between practice and research: 

"It is important to distinguish between biomedical and behavioral research, on the one hand, and the practice of accepted therapy on the other, in order to know what activities ought to undergo review for the protection of human subjects of research. The distinction between research and practice is blurred partly because both often occur together (as in research designed to evaluate a therapy) and partly because notable departures from standard practice are often called 'experimental' when the terms 'experimental' and 'research' are not carefully defined.  

"For the most part, the term 'practice' refers to interventions that are designed solely to enhance the well being of an individual patient or client and that have a reasonable expectation of success. The purpose of medical or behavioral practice is to provide diagnosis, preventive treatment or therapy to particular individuals. By contrast, the term 'research' designates an activity designed to test an hypothesis, permit conclusions to be drawn, and thereby to develop or contribute to generalizable knowledge (expressed, for example, in theories, principles, and statements of relationships). Research is usually described in a formal protocol that sets forth an objective and a set of procedures designed to reach that objective.       "When a clinician departs in a significant way from standard or accepted practice, the innovation does not, in and of itself, constitute research. The fact that a procedure is 'experimental,' in the sense of new, untested or different, does not automatically place it in the category of research. Radically new procedures of this description should, however, be made the object of formal research at an early stage in order to determine whether they are safe and effective. Thus, it is the responsibility of medical practice committees, for example, to insist that a major innovation be incorporated into a formal research project. 

"Research and practice may be carried on together when research is designed to evaluate the safety and efficacy of a therapy. This need not cause any confusion regarding whether or not the activity requires review; the general rule is that if there is any element of research in an activity, that activity should undergo review for the protection of human subjects." 

The Belmont Report makes it clear that if there is an "element of research" in the treatment, then the GCP rules of research apply: the FDA should be notified, IRB approval is required, and the consent should have all the mandated research sections and explain that the treatment is experimental. 

One classic clinical trial text is Curtis Meinert's 

Clinical Trials: Design, Conduct, and Analyses

"A research activity that involves administration of a test treatment (e.g., a drug, surgical procedure, diagnostic test, or medical device) to some experimental unit in order to evaluate the treatment. The term is subject to wide variation in usage. In some cases it may refer to the first use of a new treatment in man without any control treatment. In other cases it may refer to a rigorously designed and executed experiment involving a test and control treatment and randomization. The experimental unit in most cases is man (or a larger unit involving man, such as a hospital ward), but can be some other experimental animal." 

Meinert goes on to state that his use of the term involves human beings. He also appears to support the notion that "first use" constitutes research. His caution that the term "clinical trial" is subject to wide variation in usage is a good one. In fact, the terms "clinical trial," "clinical study," "clinical investigation," and "clinical research" are generally used in an interchangeable manner, but there is no universal or regulatory agreement marking a sharp distinction between innovative treatment and these terms. 

Reading through a recently published text entitled 

Abuse of Man: An Illustrative History of Dubious Medical Experimentation

 one is struck by the long-term historical trends of inhumane and callous treatment of patients under the guise of improving the quality of care. These kinds of unregulated experiments continued well past World War II and may still be ongoing, only in less obvious forms and more secretive venues. Their legacy is one of public and professional outrage only after these treatments cause harm and are disclosed to a wider audience. 			

Better to err on the side of caution, particularly in light of recent clinical trials litigation,

 and fully protect patients by seeking FDA or EU approvals, IRB review, and administering appropriate consents. Until the FDA, which is reportedly studying this issue and other regulatory agencies, weighs in with more definitive guidance, care providers should consider submitting investigator INDs and abiding by GCP regulations, guidelines, and standards. 

Clinical Trials: Design, Conduct, and Analysis

 (New York: Oxford University Press, 1986). 			

If the activity is a significant risk device study, surgeons must first obtain FDA and IRB approval.

Applied Clinical Trials-02-01-2005