Any proposal to expand the ClinicalTrials.gov registry that does not require submission of trial protocols "does not do the whole job," FDA Office of Medical Policy Director Robert Temple said at a June 19 session of the Drug Information Association in Atlanta.

If the goal of legislators is to provide public access to information from clinical studies, the only way to ensure that is to require that all the data is provided, he said.

"The [proposal] that is facing potential legislation calls for a fair amount of description for each study, but it doesn't actually ask for the protocol," Temple said. "You aren't smart enough to put everything into categories - nobody can do that. You really should see the protocol."

Part of the FDA Modernization Act of 1997, the ClinicalTrials.gov registry was launched in 2000 with a mandate that sponsors post all Phase II-IV studies for products to treat serious or life-threatening diseases (¹ [A#00620100018], p. 24). Legislation to renew and expand the public online clinical trial registry is currently pending, and bills cleared by both the House and Senate would require, in addition to trial registration, both the posting of results and enforcement mechanisms to ensure compliance.

Bills Would Demand "Results"

A public clinical trial results database is uncharted territory for all involved - industry, medical journals, the public, and legislators - and getting it right presents what Temple called a "thorny problem." The House and Senate bills approach results registration from prescriptive and descriptive angles, respectively.

Beyond linking to existing clinical trial results information, the Senate proposal, contained in the drug safety legislation authored by Sens. Edward Kennedy, D-Mass., and Michael Enzi, R-Wyo., leaves the details of the results database to NIH.

The proposal calls for the director of NIH to conduct a "feasibility study" to determine the "best, validated methods of making results of clinical trials publicly available," and to then provide the Secretary of HHS with a recommendation for a proposed rule. A negotiated rulemaking committee, consisting of members from FDA, NIH, PhRMA and others, would then decide which clinical trial results information should be included; a format and standard procedure for results submission, including timing of submission and posting; a standard procedure for results information verification; and an implementation plan.

In contrast, the provisions in the House bill, based mainly on language originally offered by Reps. Henry Waxman, D-Calif., and Edward Markey, D-Mass., detail specific requirements for the results registry, the most controversial of which being the posting of trial results summaries - a technical summary discussing endpoints, relevant statistics, and safety information; and a lay summary of the trial and results written for patients, with descriptions of and reasons for any changes in trial design, as well as significant safety information.

Non-Technical, Or Technically Promotional?

The prospect of lay results summaries has inspired industry anxiety, in that the patient-friendly summary language might veer dangerously close to promotional materials in the eyes of FDA.

PhRMA Senior Assistant General Counsel Scott Lassman told "The Pink Sheet" that "particularly because there is a provision in there saying that it is a criminal penalty to post information that is promotional, we think that this creates a huge risk for companies when they are forced to distill this very complex scientific information down into lay language."

At DIA, Temple warned that the risk of being cited "comes with the territory. ... You should restrain your language. ... I think there's going to be a requirement to post those. ... This will be a cautious, sober, careful mention given that everybody's watching."

The House proposal also requires inclusion of the percentage of subjects who dropped out of the trial, the trial's completion date and links to any adverse regulatory actions taken by FDA based on the substance or misrepresentation of the design or outcome of the trial.

Biased clinical trial results have long been a concern of patient advocacy groups. The results database will help to quell some, but not all, of that anxiety by ensuring that all results - rather than the disproportionate amount of positive results suspected by some - are made publicly available, but concern remains that bias could remain in study designs or analyses.

"The results part makes everybody nervous because they're worried about biased results, and I share that worry. But if you want to have results, you have to take that risk. I don't know what the alternative is," Temple said.

Focusing On The Goal Of The Registry

Temple said that the best way to avoid bias in results presentations would be for registry proposals to be modeled after the data FDA receives.

"Just to remind everybody, 100 percent of what we get comes from people who have preferred outcomes - it's not evil to have a preferred outcome. ... [The results] come from people with a bias - we know that. But we don't have much of a problem because we get all the studies under law - we get all the protocols, amendments, statistical analysis plans, we get all the data from all the studies, we get [case report forms] of deaths and any drop outs, we can check on analysis, we can do anything," he said.

"So we're not worried about bias because we can do our own analysis. That really seems to me to be what a registry should be at least seeking toward - you don't have to worry about this and that if you have all the data," Temple said.

"If you want to examine the study closely, anybody -whether it's FDA, a peer reviewer for a journal, a physician, a patient - if they really want to know, they need to look at the data, design and outcome just as FDA does," Temple said.

"You need to know about changing hypotheses, changing influence, changing statistical analysis, whatever the plans were to leave some patients out, multiple endpoint assessment findings, all subgroups and all the time working on it, and you need to know how the analyses were done, whether adverse events were grouped in this or that way - you need to know all that stuff. Then of course you have to know about all the studies," Temple said.

"My plea is that every registry proposal should be looked at for its ability to solve or at least contribute to a solution to the problem that led to desire for a registry," he said.

Both House and Senate bills contain industry compliance enforcement mechanisms. The House bill allows a sponsor 30 days to correct the finding of any neglected or misrepresentative information, after which the noncompliance would be reported to the federal research agencies and the Office of Human Research Protections; noncompliance notification would be made public, and civil penalties would be applied.

Under the Senate bill, federal grant recipients would also have 30 days to correct any noncompliance reported by the applicable federal agency, and sponsors would have to submit with NDA or BLA submissions certification, such as National Clinical Trial control numbers, showing that all the required information had been added to the registry.

The enforcement mechanisms follow a 2005 FDA evaluation that found industry compliance with the clinical trial registry mandate to be "mixed" and to vary by FDA review division (² Also see "ClinicalTrials.Gov Compliance Varies By FDA Review Division, Report Finds" - Pink Sheet, 22 August, 2005., p. 3).

A policy adopted in 2005 by the International Committee of Medical Journal Editors expanded the scope of clinical trial registration by compelling registration of Phase II and III trial data as a condition of publication. As a result, the number of trials registered on ClinicalTrials.gov jumped from 13,153 to 22,714 trials one month after the policy went into effect. Today, 41,000 trials are registered.

A recent expansion of the ICMJE policy to require registration of Phase I trial protocols for all trials that begin enrollment on or after July 1, 2008 - a data submission requirement absent from current versions of the bills - could subvert congressional efforts to establish a single, national set of trial registration requirements (³ Also see "Phase I Trial Registration Required For Publication, Medical Journals Say" - Pink Sheet, 11 June, 2007., p. 16).

Identical preemption clauses in both versions of the pending legislation state that "No state or political subdivision of a State may establish or continue in effect any requirement for the registration of clinical trials or any requirement for the inclusion of information relating to the results of clinical trials in a database."

- Jamie Hammon (j.hammon@elsevier.com)

Trial Registry Bill Should Request Protocol With Results, FDA’s Temple Says

More from Archive

More from Pink Sheet