Effective clinical study design can minimize FDA and IRB involvement: Temple.

WASHINGTON -- Quality in a clinical trial typically focuses on accuracy, compliance with recording what was expected and consistency of case report forms (CRFs).

Often what is ignored, noted a senior Center for Drugs official here, is the design of the study itself.

Robert Temple, M.D., associate director for medical policy at CDER told a May 9 FDA/Drug Information Assn. (DIA) meeting: "The study design should assure patient safety and unbiased observations through blinding and other means."

Temple added that "clinical investigators should have assurance of the completeness and accuracy of observations on the site, central monitoring and auditing data."

He noted that a well-designed study can be achieved in many ways, but in the end, it benefits the clinical researcher because "an already good study design won't need help from an IRB or FDA." Well-defined endpoints may reduce the need for auditing. Also, compulsive investigations can make auditing less critical. And, "past knowledge of a drug can reduce the need for safety monitoring."

The FDA official described regulations that cover the obligations of investigators and study sponsors as "broad."

"Regulations state that you must select good investigators and have a Form 1572 signed, which means the investigator agrees to personally conduct or supervise the trial." Temple stressed that many researchers are disqualified on this point.

He said that clinical investigations must be monitored in two senses: to make sure that investigators are doing what they agreed to do, including selecting qualified monitors and stopping shipments to non-compliant investigators.

They also must watch accumulating evidence related to safety. He added that while it is "highly relevant" what exact safety measure should be in clinical trials, "the regulations are silent on this."

Sponsors do have ICH [International Conference on Harmonization] E6 for guidance on monitoring, but it leaves room for variability. He noted that in some cases, there could be no monitoring required but the regulation does not say why or when that would be.

As far as GCP guidance, there are few specifics about safety monitoring, such as, specific tests, frequency of testing, analyses needed of ongoing safety and concomitant therapy.

Other subjects not addressed in FDA regulations regarding clinical trials, Temple said, include:

1. Are people paying attention to patient problems?

2. Are adverse events reviewed promptly by a knowledgeable person?

3. Are abnormal data seen right away?

4. Is the sponsor looking at all data and putting it together?

5. Are all the data called for being recorded and recorded accurately?

6. Can we be sure that blinding was maintained?

Temple said that in early trials, "there is major interest in showing patient safety, but getting the right answer is also critical to response. Data will play a minor role in marketing decisions by FDA."

He warned attendees to be on the look out for intellectual bias in academic studies. The FDAer also asked, "To what extent can central monitoring of completed data and CRFs substitute in later trials for on site monitoring? Can approaches used in large phase IV trials and used by FDA inspectors be utilized?"

Temple also stated: "We want to continue to be sure that serious adverse events are still being sought and reported. How do we do that?"

Robert Califf, M.D., vice president for clinical research at Duke University, added in a subsequent presentation: "Clinical trials in the U.S. are going the way of the textile industry and that is bad for patients" in the United States.

He also provided a top-10 "to-do" list for clinical trials in the United States:

1. Increase the national clinical research IQ--it is not enough to just have a narrow focus on your job;

2. Fund clinical trials quality research--develop clinical research ontology and pragmatic assessment of quality measures for clinical trials;

3. Register all trials at clinicaltrials.gov--truth will come out with forced compliance;

4. Require assessment of blinding;

5. Have more than one pair of eyes on analysis;

6. Accelerate data standards and controlled vocabulary development;

7. Develop a system to certify centers rather than evaluate each individual trial;

8. Public vetting of trial designs before they start;

9. Require quality assessment plan as part of the protocol; and

10. Provide public funding based on meeting these quality issues.

By Joseph Pickett Managing Editor