FDA Addresses Industry Concerns with Quality Metrics

Agency officials are emphasizing that new quality metrics data will be just one factor that FDA uses to assess a facility and drug products, seeking to address industry concerns that providing FDA with specific information on production shortcomings could lead to public criticism and enforcement action. Russell Wesdyk, acting director of the Office of Surveillance in the Office of Pharmaceutical Quality (OPQ), Center for Drug Evaluation and Research (CDER), talked about a “phased approach” for requiring submission of this information at the recent Quality Metrics Conference sponsored by the Parenteral Drug Association (PDA), and emphasized that the data would be used for surveillance, as opposed to enforcement. Over time, he added, the additional information may help the agency identify a “dean’s list” of quality operations that warrant fewer inspections and less oversight of postapproval changes.

FDA staffers involved with defining a “consensus set” of metrics explained that FDA needs to find a better way to assess the quality and risks at thousands of production sites around the world, when the agency can conduct only about 1500 inspections each year to ensure compliance with good manufacturing practices (GMPs). Wesdyk recalled extensive discussions about metrics between industry and FDA over the past two years that supported the initiative and led to FDA’s draft guidance on the program last July.

Yet industry concerns about the draft policy, as voiced at a public meeting on the topic in August, reflected confusion about definitions and proposed procedures. Tara Gooen Bizjak, senior science policy advisor in OPQ’s Office of Policy, acknowledged that existing metrics often are different even within the same company. Manufacturers are unclear about the definition of official “reporting establishment” and if metrics will have to be submitted for each active ingredient and product from a facility. 

Bizjak anticipated that quality metrics will provide a more objective measure of quality at a site and further facilitate risk-based inspection scheduling. She and others emphasized that poor quality data would not necessarily lead to enforcement action, but should help identify problems that would trigger consultation with OPQ on how best to address the issue. Failure to submit requested quality metrics data, however, would be considered a violation.

One fear of manufacturers is that FDA investigators will raise questions during site inspections about metrics calculated by FDA. Wesdyk said that he expects at least a year-long learning period following submission of initial quality metrics data before all parties feel comfortable with how FDA analyzes the information. 

FDA efforts to communicate its broader goals for the metrics initiative seem to be building support among pharma manufacturers. Jeffrey Coates, director of risk modeling for global manufacturing and supply at GlaxoSmithKline, explained how his company is using metrics data internally in developing site risk model profiles that can generate early warning signals before negative events occur. And while there’s always some risk of unintended consequences, Steven Lynn, global head of compliance and audit group at Novartis, acknowledged that a lot of thought has gone into determining what metrics should be and how they will be used by FDA and by industry. Quality metrics need to be relevant and accessible at the operational level, Lynn said, and they should be risk based and generate actionable signals and knowledge that management values.

To move forward with implementation, OPQ staffers are preparing a technical document that defines terms and provides specifications for how manufacturers should submit metrics data to the agency. The plan is for manufacturers to provide information on specific production measures and FDA will use that data to calculate metrics, such as lot acceptance rate, quality complaint rate, invalidated out-of-specification rate, and on-time rate for annual product review or product quality review. This information should help the agency develop objective measures for the quality of a drug product and site.

At the same time, the agency is building an IT system able to accept this data for some 5000 manufacturing sites that produce 60,000 medicines, often with five or more facilities involved in a product, explained Karthik Iyer, acting branch chief in the Office of Quality Business Informatics of the Office of Surveillance. The system has to be able to analyze the data submitted in order to produce useful reports on variation in product quality. Ultimately, FDA would like to be able to determine if metrics correlate to product type and dosage form; if they can signal problems likely to lead to recalls or post-approval quality defects; and if they have predictive power or change over time.

The broader aim of the metrics program is to encourage manufacturers to implement new technology and to reward firms that achieve high grades with less frequent inspections and “appropriate” reporting of post-approval changes, Bizjak noted. This information also should help manufacturers identify situations where there may be a greater risk of drug supply disruptions, she added.

The metrics program “won’t be perfect from the start,” commented Amgen senior vice president Martin Van Trieste. But he believes the initiative will be “transforming,” as it will provide consumers and payers with important information about the level of quality of prescription drugs. He also predicted at the PDA conference that manufacturers will see costs go down from adopting modern quality systems that “get things right the first time.”

Quality metrics should promote responsible practices by manufacturers and a quality driven corporate culture, said Wesdyk. The public also should gain from fewer recalls and quality related drug shortages. And all parties should benefit from FDA’s ability to reward companies for quality operations, without extensive oversight.