Highlights from ExL Pharma's 5th Data Monitoring Committees

Philadelphia, PAFebruary 22-23, 2010Highlights from ExLPharma’s 5th Data Monitoring Committees/DSMB

The Role of DMCs/DSMBs in Adaptive Trials:The Impact of DMCs/DSMBs on Decisions for Dosage Changes or Changes in Sample Size2

DSMB/DMC3Data Safety Monitoring Board (DSMB) or Data Monitoring Committee (DMC) as defined in FDA Guidance “Establishment and Operation of Clinical Trial Data Monitoring Committees”:a group of individuals with pertinent expertise that reviews on a regular basis accumulating data from one or more ongoing clinical trials. DMC advises the sponsor regarding the continuing safety of trial subjects and those yet to be recruited to the trial, as well as the continuing validity and scientific merit of the trial.

DSMB/DMC4DSMB periodically inspects “interim looks” at the data (prior to all planned patients being treated and followed) and makes recommendations about trial design/conduct based on these “interim looks” including:Continue trial as initially planned.Safety:Potentially stop the trial early if reasonable evidence of harm of new product (or dose) or control product Often based on incidence of AEs or serious AEsDMC safety recommendations not necessarily based on formal statistical rules

DSMB/DMC5Example of DSMB Safety RecommendationDouble-blinded trial (randomized): Placebo versus Experimental treatment in high risk populationPatients scheduled to be treated over several monthsAfter the first 40 patients were enrolled (20 in each treatment group) and treated for 1 day to several weeks, DSMB inspected unblinded data:Experimental treatment had 4 life threatening serious adverse events (SAEs)Placebo group had noneP-value=0.1060DSMB Recommendation was to terminate study despite “non-significant” safety p-value

DSMB/DMC6Example of DSMB Safety Recommendation (cont’d)DMC Recommendation to stop trial was not easyPerhaps benefit of drug outweighed the risk? Tough to tell at this early stage, but it was thought perhaps it was best not to wait to find out.Sponsor does not necessarily need to follow DSMB recommendationE.g., Sponsor may decide:Continue trial as isTemporarily halt trial until further investigationDiscuss with DSMB/DMC & regulatory agency reason for not following recommendation

DSMB/DMC7Example of DSMB Safety Recommendation (cont’d)Sponsor does not necessarily need to follow DSMB recommendation (cont’d)Potential issue if trial continues:Sponsor was made aware of the 4 experimental group SAEs during DSMB deliberationsWill this “awareness” affect subsequent study conduct (even unintentionally), causing bias in remaining safety data?Perception of bias?Suppose SAE prevalence “evened out” at end of study?

DSMB/DMC8DSMB periodically inspects “interim looks” at the data (prior to all patients being treated and followed) and makes recommendations about trial conduct including:Efficacy (“Adapt the trial”) (cont’d):Change primary endpoint (not common -- in my experience)Remove non-efficacious dose(s) (for multi-dose trials);Requires formal statistical rules in order to: Not increase “Type I Error” rate or “significance level” (chance of a false positive trial)Minimize “Type II Error” rate (chance of false negative trial)DSMB/DMC unblinded or “partially” unblinded

DSMB/DMC Committee Composition9Discussed in Section 4.1 of FDA GuidanceSponsor/Steering Committee appoints DMCCRO also sometimes asked by Sponsor to appoint DMCMembers consist ofAt least 2 cliniciansAt least 1 statisticianMay also have other members; e.g., Ethicist for high risk/vulnerable populationsNon-scientist; e.g., consumer rep; patient (not in trial)

DSMB/DMC Committee Composition10Ideally, DSMB/DMC members should haveExperience in indication being studiedClinical trials experienceDSMB/DMC experienceEspecially important for statistician on DSMB/DMCAll members may not have all 3Opinion: Especially important for Statistician to have at least b and c; Clinician/physician to have at least a (and b);

DSMB/DMC Committee Composition11Statistician should also have:Experience in statistical methods for clinical trialsExperience in statistical rules for “adapting” designMembers should be independent from SponsorNot a sponsor employeeNot an investigatorNo involvement in the study conduct or analysis planningNo serious conflicts of interestsAll potential conflicts disclosed prior to each meeting

DSMB/DMC Committee Composition12Each DSMB/DMC has a chair.Chosen by sponsor and other DSMB/DMC membersFDA Guidance: chair capable of “facilitating discussion/integrating different points of view”.Chair provides DSMB/DMCs recommendation of future study conduct to sponsor or steering committeeAfter inspection of interim data, DSMB/DMC members “vote” on a recommendation for future conduct of studyWorks best if “odd” number of DSMB/DMC voting members

Statisticians in DSMBsOften 3 statisticians involved in the DSMB/DMC processDSMB/DMC StatisticianVoting member of DSMB/DMC; helps interpret statistical aspect of results to other voting members, especially adaptive results and statistical repercussions of adapting the design.Analysis Statistician Generates interim results (often not a Sponsor employee)Unblinded to treatment group for each patient“Independent Statistician”Presents results to DSMB/DMC; answers DSMB/DMC statistical questions Non-voting“Go-between” between DSMB and Analysis StatisticianAll statisticians are not to be involved in analysis of final data or analysis planning

“Fixed” Design“Fixed” clinical trial design (2 treatments)Compare “cure” rate across two anti-infectives:Specify an appropriate (superiority) null and alternative hypothesis:H0: πE = πCH1: πE ≠ πCwhere πE and πCare the true sensitivities of the experimental and control diagnostic, respectively.To calculate sample size, make assumption of the true πE, πCDetermine sample size required to yield adequate power to reject H0 in favor of H1 at a given significance level14

“Fixed” Design“Fixed” clinical trial design example (cont’d)For determining sample size, assume πE =0.80 and πC =0.70Two-sided 0.05 level of significance (false positive rate)Desire 80% Power1:1 Treatment AllocationSample size per group (using two-sample z-test for proportions): 294 per group15

Adaptive Trial Designs16Adaptive designs allow for re-designing the trial at some point during the trial, usually based on interim look(s) at aggregate efficacy dataStop trial for “futility” of experimental diagnosticStop trial for overwhelming efficacyIncrease sample size to ensure adequate (conditional) power by end of studyDecrease sample size if (conditional) power >>original powerRemove non-efficacious dose in dose-finding studyChange primary endpoint

Adaptive Trial Designs17The interim “look” at the data can be blinded or unblinded.Adapting the design of the study often has statistical repercussions, especially if re-design is based on an unblinded analysisHere, we will focus on adapting trial based on an unblinded interim analysis.Also may have operational repercussions

Issues to Consider18Once DSMB gives recommendations arrive, could bias in final results occur?If DSMB says “increase the sample size by xx”:if Sponsor follows recommendation, study personnel may realize the product may be efficacious, and just needs a little more sample size to prove its efficacy with good confidence.Cause bias in future study conduct? If Sponsor does not follow recommendation, does Sponsor feel product then has small chance of being efficacious and bias remainder of study? E.g., may move resources off study, slowing study down or reducing quality?

Issues to Consider19Once DSMB gives recommendations arrive, bias could result If DSMB/DMC recommends “continue the study as is”, Study personnel who may not understand the requirement for overwhelming efficacy, may think the product is not efficacious, otherwise study would have stopped for efficacy?Causes bias on remaining study personnel?

Dosage Change20Most “pivotal” (Phase III) drug/biologic trials contain one dose of the experimental treatment and a control groupSome Phase III may contain more than one dose of experimental however (especially those that may “skip” Phase II to get product to market faster in case of serious illness, e.g., ALS)Many Phase II’s have multiple doses of experimental to determine the dose of study in future pivotal Phase III

Dosage Change21Suppose we have a trial with multiple doses of experimental treatment, and one control groupFormal efficacy rules can be built in to “drop” inefficacious dose after interim analysis;E.g., drop dose(s) with CP<20% when compared to controlMaybe use a more liberal (higher) threshold, especially in Phase II, if other doses showing more promise.Increase enrollment in doses with CP<80% but showing promise?In Phase II, control of false positive rate may not be as much of concernIn Phase II, may not necessarily have to show treatment effect at 0.05 level of significancePerhaps calculate CP at 0.10 level of significance.

Dosage Increase22Many Phase I trials are non-randomized, dose-escalating Enroll first k patients into lowest doseFollow for safetyDSMB inspects safety results and gives the “OK” to move on to next doseNext set of k patients enrolled to next (higher) doseContinue through pre-planned doses

Primary Endpoint Change23Medical device clinical trial example: Efficacy endpoint is survival at 30 days.DSMB/DMC performed formal pre-planned interim efficacy/futility analysis on endpointAssessed if trial could be stopped early for overwhelming efficacy, futility, and sample size recalculation as necessary.DSMB/DMC also decided to inspect CP for “3-month survival” outcome at interim stageNot in DSMB/DMC charterNot Necessarily a problem, depending on what DSMB/DMC plans to do with this infoLess patients available than at 30 days

Primary Endpoint Change24Medical device clinical trial example (cont’d): Reason for request unclear:DSMB/DMC considering recommending change in primary endpoint?Data-driven decision?May increase chance of false positive study Sponsor requested to see interim 3 -month dataStrongly recommend against thatData-driven decision and potential introduction of bias into trialPossible approach with no real repercussions:Use 3-month data to plan another study with 3-month outcome as primary endpoint.

Conclusion25DSMBs/DMCs consist of experienced clinicians and at least one experienced statisticianDSMBs/DMCs responsible for subject safety and scientific merit of studyDSMBs/DMCs make recommendations on study design and conduct issues while the trial is ongoingSuch recommendations may have repercussions and may introduce bias in remainder of studyImportant for DSMBs/DMCs to consider such issuesImportant to minimize such bias

Acting on a DMC/DSMB Recommendation:How to prepare for and respond to DMC/DSMB recommendations

1. Identifying the need for DMCs/DSMBs - 11998 - NIH Policy For Data and Safety Monitoring2001 - Draft FDA Guidance for Clinical Trial Sponsors: Establishment and operation of clinical trial data monitoring committees2003 - Draft CHMP Guideline on Data Monitoring Committees2004 - NEJM Data Safety and Monitoring Boards2004 – Small group processes relevant to data monitoring committees in controlled clinical trials: an overview of reviews2005 - Issues in data monitoring and interim analysis of trials - DAMOCLES Study Group2006 - Final CHMP Guideline on Data Monitoring Committees2006 - Final FDA Guidance for Clinical Trial Sponsors: Establishment and operation of clinical trial data monitoring committees2008 - JAMA commentary: When should data and safety monitoring committees share interim results in cardiovascular trials?Clear expectationsIncreased focus on the DMCEvolving role of the DMC

1. Identifying the need for DMCs/DSMBs - 2Changing Characteristics of Development Portfolios Serious diseases

Greater HA scrutinyIncreasing Complexity of Development Programs Earlier in development phase

Large multi-site, multi-national clinical trials

Multiple simultaneous indicationsPotential for Variation in DMC Interactions Scope of DMC

Interpretation of roles and responsibilities

Alliance partnershipsGreater need for DMCsImportance of clear DMC decisionsOpportunity for improvement

1. Identifying the need for DMCs/DSMBs - 32006 Guidance for Clinical Trial Sponsors: Establishment and Operation of Clinical Trial Data Monitoring Committees2. Determining Need for a DMCAll clinical trials require safety monitoring, but not all trials require monitoring by a formal committee that may be external to the trial organizers, sponsors and investigators.2.1 What is the Risk to Trial Participants?A fundamental reason to establish a DMC is to enhance the safety of trial participants in situations in which safety concerns may be unusually high, in order that regular interim analyses of the accumulating data are performed. We recommend that sponsors consider using a DMC when:The study endpoint is such that a highly favorable or unfavorable result, or even a finding of futility, at an interim analysis might ethically require termination of the study before its planned completion;

There are a priori reasons for a particular safety concern, as, for example, if the procedure for administering the treatment is particularly invasive;

There is prior information suggesting the possibility of serious toxicity with the study treatment;

The study is being performed in a potentially fragile population such as children, pregnant women or the very elderly, or other vulnerable populations, such as those who are terminally ill or of diminished mental capacity;

The study is being performed in a population at elevated risk of death or other serious outcomes, even when the study objective addresses a lesser point;

Highlights from ExL Pharma's 5th Data Monitoring Committees

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Highlights from ExL Pharma's 5th Data Monitoring Committees (20)

More from ExL Pharma (20)

Highlights from ExL Pharma's 5th Data Monitoring Committees