Stakeholders want more opportunities to discuss MIDD with FDA

Stakeholders want more discussions on the kinds of data that can be used when stakeholders consider using model-informed drug development (MIDD) approaches. They note that there is uncertainty about what kinds of data the US Food and Drug Administration (FDA) will accept when such models are employed.

In August, the Center for Drug Evaluation and Research (CDER) and the Center for Biologics Evaluation and Research (CBER) asked stakeholders for feedback on how to advance model-informed drug development (MIDD). More specifically, the centers want to know how to increase the use of MIDD approaches in regulatory decision-making and how drug development efforts are incorporating MIDD.

The American Society for Clinical Pharmacology & Therapeutics (ASCPT) responded to the request and provided three recommendations for increasing the use of MIDD in regulatory decisions. First, the group recommended reducing barriers to adopting MIDD. It said sponsors often consider MIDD approaches too risky because there is uncertainty about how FDA plans to allow their use.

"Increased transparency and communication by FDA in the suggested areas would enable greater MIDD utilization across the field, to supplement what has already been done by FDA in rolling out the Quantitative Medicine Center of Excellence (QMCOE), and providing an introduction to MIDD for strategic partners," said ASCPT. The group recommended the agency develop guidance and best practices related to the topic, document what is considered acceptable early access to pharmacokinetic (PK) and pharmacodynamic (PD) data, and other recommendations.

ASCPT also recommended that FDA work to establish shared virtual population databases and model repositories that researchers could use when considering MIDD approaches. That could include establishing a publicly accessible centralized MIDD knowledge repository and developing a mechanism to encourage non-competitive sharing of validated disease models.

"The application of MIDD approaches by sponsors would be supported through efforts to increase the transparency of how models are evaluated and expectations of documentation and qualification," ASCPT added. "Enhanced regulatory decision-making by MIDD approaches would be facilitated by providing greater transparency to sponsors regarding the agency’s expectations in model evaluation and providing a better understanding of what good looks like, particularly in emerging or developing areas such as use of [artificial intelligence/ machine learning (AI/ML)] or mechanistic models such as [physiologically based pharmacokinetic (PBPK)] and quantitative systems pharmacology (QSP)]."

The Critical Path Institute (C-Path) also wrote to FDA with recommendations centered on using real-world evidence (RWE) and real-world data (RWD) to promote the use of MIDD approaches. The group said that FDA’s RWD draft guidance on using electronic health records (EHR) details the two different kinds of missing data, and it is an issue that needs more discussion because it may be useful when considering MIDD models.

“RWD offers model-informed drug development (MIDD) programs strength in its abundance, especially in rare and orphan diseases,” said C-Path. “But a larger sample size often comes with the caveat of higher inter- and intra-source heterogeneity.”

“There are currently unaccounted for factors that pose a challenge to determining if data is missing in a traditional sense, and if so, how it should be interpreted,” the group added. “Additionally, if data is indeed traditionally missing, extensive resources may be required to validate why the data is missing.”

C-Path recommended that the QMCOE should consider informal communication mechanisms such as scientific advice meetings and Critical Path Innovation Meetings, to discuss MIDD-related activities.

The Metrum Research Group (MRG) wrote to FDA about having further discussions about MIDD. It said stakeholders outside of pharmacometrics could have more engagement on the topic if the agency broadened the meaning of MIDD. More specifically, the group argued that limiting the definition of MIDD approaches to exposure-based biological and statistical models is too narrow and should instead use MIDD approaches that leverage comprehensive information.

"Whenever key questions need to be answered without the benefit of adequately powered randomized controlled trials, there will be a need to draw causal inferences from non-randomized comparisons (again, this is true whether or not the relevant data sources include 'exposure')," MRG added. "Accordingly, we propose that basic concepts from the causal inference literature should play a central role in MIDD guidance."

Comments

Stakeholders want more opportunities to discuss MIDD with FDA

Related topics

Recommended content

Welcome to the new RAPS Digital Experience