Guest Column | August 2, 2016

Can You Support Quality Metrics With Lobotomized Data?

daniel-matlis_pharm

By Daniel R. Matlis, President, Axendia

Life science companies are obsessed with collecting dumb data. We assemble it, retain it, and hoard it to support regulatory and legal requirements.

But what happens to these vast amounts of collected data?  Unfortunately, most of the data is “lobotomized” as soon as it hits the paper it was printed on or the electronic document it was saved to — vast amounts of product and process intelligence that could be used to improve control over manufacturing and product quality go unused.

FDA intends to drive life science companies to harness some of that intelligence. The Food and Drug Administration Safety and Innovation Act of 2012 (FDASIA) provides FDA with the ability to conduct electronic inspections (e-inspections) and target onsite inspection based on quality metrics.

By shifting to a metrics-based approach, the agency wants to “encourage” industry to implement state-of-the-art, innovative quality management systems (QMS) that drive a focus on improving product quality, rather than simply ensuring compliance to regulatory requirements.

“We don't have the capacity to be in every firm all the time,” Howard Sklamberg, FDA deputy commissioner for global regulatory operations and policy, said during the FDA’s public meeting on the drug supply chain provisions of Title VII of FDASIA in July 2013. “We want to encourage firms to prioritize quality — and encourage their boardrooms and their pocketbooks — because quality costs money.”

To support this effort, FDA released for comment its draft guidance document Request for Quality Metrics in July 2015. While the biopharmaceutical industry, in general, supports FDA’s effort to implement a quality metrics program, it also has some lingering concerns.

Grading Industry On A Curve

FDA’s Center for Drug Evaluation and Research (CDER) and Center for Biologics Evaluation and Research (CBER) intend to request quality metrics data reports from industry for a one-year period in electronic format. Following are the types of data FDA proposes industry report, both by product and by establishment.

  • The number of lots attempted of the product
  • The number of out of specification (OOS) lots of the product rejected during or after manufacturing
  • The number of attempted lots pending disposition for more than 30 days
  • The total number of OOS results for the product, including stability testing
  • The number of lot release and stability tests conducted for the product
  • The number of OOS results for lot release and stability tests for the product that were invalidated due to lab error
  • The number of lots attempted that were released for distribution or for the next stage of manufacturing the product
  • The number of product quality complaints received for the product
  • Whether the associated annual product reviews (APRs) or product quality reviews (PQRs) were completed within 30 days of annual due date for the product
  • The total number of APRs or PQRs required for the product

Analysis of this data would be used to grade each company fit them along a bell curve.  According to Mr. Sklamberg, FDA would direct resources to those companies that are on the lower end of the curve.

Based on its analysis of quality metrics across the industry, FDA plans to develop a risk-based inspection schedule.  Establishments that demonstrate their commitment to quality and have highly controlled manufacturing processes can be inspected less often (lower priority for inspection occurrence) than those that do not (higher priority for inspection frequency).

In the event an establishment does not report required quality metrics, FDA may deem products manufactured or processed at that establishment as adulterated and subject to enforcement action.

Industry Response

On June 6, 2016, the International Society for Pharmaceutical Engineering (ISPE) released findings from Wave 2 of its Quality Metrics Pilot Program.  While ISPE reiterated the industry’s support of FDA’s effort to implement a quality metrics program, its analysis of the Wave 2 data showed that the effort to collect metrics consistent with FDA’s draft guidance was approximately three times that estimated by the agency in the Federal Register Notice.

Another key finding was the need to standardize quality metrics across the industry.  For example, while process capability/performance measures are extensively used by companies to help control processes and identify continual improvement opportunities, their definition and use of varies between companies.

Based on its Wave 2 findings, ISPE recommendations to FDA include:

  • A small, targeted launch to minimize the industry burden
    • This would also allow FDA and industry to learn, for example, about topics like implementation of standardized definitions and the collection, submission, and analysis (e.g., statistical analysis) of data.
  • A phased introduction, as an alternative to “starting small”
  • Initially focusing on three of the proposed metrics while simultaneously considering varied definitions of these metrics
    • This would defer some metrics and data points.
  • Greater transparency regarding the manner in which data will be assessed, and how outcomes and conclusions will be determined and communicated

Conclusions

FDA’s e-inspections program based on quality metrics will be a key step in its plan to encourage industry implementation of state-of-the-art, innovative quality management systems that prioritize product quality improvement above mere regulatory compliance.

While lobotomized data is still prevalent in our industry, the requirement to submit metrics annually in electronic format will make the use of dumb-data unmanageable.

To support e-inspections, most companies will be forced to discontinue the use of paper records in favor of electronic batch records systems and manufacturing execution software (MES).  Integration with complementary systems like QMS, product lifecycle management (PLM) systems, enterprise resource planning (ERP) and laboratory information management systems (LIMS) will play a key role in standardizing definitions and achieving a single source of truth to support quality metrics reporting. 

Stop lobotomizing your data!

The time to begin preparations is now.  Is your company, and are your systems, ready to support e-inspections and be graded on a curve?

About The Author

Daniel R. Matlis is founder and president of Axendia, an analyst firm providing trusted advice to life science and healthcare executives on business, technology, and regulatory matters. He has over 25 years of experience in the industry, spanning the entire value chain. He is also an active member in FDA’s Case for Quality Initiative and has presented Axendia’s research findings to industry executives and the FDA.

Dan began his career at Johnson & Johnson (Ethicon), where he provided leadership in the areas of technology, regulatory compliance, and business. Most recently, he was a partner, VP, and GM at a leading life science consultancy firm.

Dan holds a BS in electrical engineering from Polytechnic University (now NYU Polytechnic School of Engineering) in New York, and an MS in management from the New Jersey Institute of Technology.

About Axendia

Axendia, Inc. is a leading trusted advisor to the life science and healthcare industries. The firm provides trusted counsel to industry stakeholders on business, regulatory, and technology issues. For more information, visit www.axendia.com  (where you can read the company’s Life-Science Panorama blog) or email info@axendia.com. You can also follow Axendia on Twitter and LinkedIn.

The opinions and analysis expressed in this research reflect the judgment of Axendia at the time of publication and are subject to change without notice. Information contained in this document is current as of publication date.  Information cited is not warranted by Axendia but has been obtained through a valid research methodology.  This document is not intended to endorse any company or product and should not be attributed as such.