GOOD STORIES
A leading medical device manufacturer partnered with Emendo R&D to bring statistical consulting for medical devices into the development of new product formulations. The goal was to build a rigorous foundation for testing and validation that would inspire confidence in product performance and support regulatory compliance from day one.
THE CHALLENGE
Introducing new formulations required demonstrating that performance remained consistent with existing products. However, the comparison between formulations could be affected by:
To ensure sound decisions and safeguard product quality, a robust framework for analyzing and managing data was essential – one that applied the right statistical analysis methods in medical product formulation development to capture variability and provide confidence in the results.
OUR APPROACH
Before we could define a systematic approach, it was essential to recognize the complexity of the task. Introducing new formulations required not only a comparison against legacy benchmarks but also a deep understanding of variability in testing, production scale, and manufacturing batch conditions.
Without a structured framework, results risked being inconclusive, slowing development and undermining confidence in decision-making.
We started by systematically collecting data using selected test methods and benchmarking the performance of legacy devices. We produced small batches at laboratory scale for agility and collected data on manufacturing scale for validation.
After quantifying variability due to test methods, batches, and production scale, we introduced changes in formulation and looked for performance shifts at laboratory level.
To ensure accurate and repeatable evaluations, we developed bespoke scripts for data analysis. These scripts automated calculations and visualizations, reducing manual effort and enabling consistent results.
Depending on the test method and level of variability, we selected different statistical analysis methods to test for equivalence. Typically, the data structure was hierarchical, and linear mixed models were most suitable to consider batch variability and properly assess the effect of formulation changes.
THE RESULTS
The project delivered several core outcomes that directly strengthened the evaluation of new formulations and the reliability of decision-making. These outcomes provided both immediate value and a solid foundation for future development:
Data-Driven Formulation Evaluation: New formulations were systematically assessed, providing robust insights into their performance relative to established benchmarks.
Comprehensive Database: A centralized database was established as a valuable resource for historical data comparisons, supporting ongoing research and development while reducing errors caused by fragmented data handling.
Efficient Analysis Pipeline: Bespoke scripts streamlined the analysis process by automating calculations and visualizations, which reduced turnaround times and ensured consistent accuracy.
In addition to the core outcomes, the project delivered a range of further advantages that strengthened both process robustness and data reliability. These benefits ensured that improvements were not only immediate but also sustainable over time:
Enhanced Process Control: By collecting data at large manufacturing scale, we established control charts that identified trends and potential issues early. This proactive monitoring allowed the team to take timely corrective actions, improving stability and safeguarding product quality.
Improved Method Reliability: By analyzing operator and date effects with gauge R&R, we were able to identify and mitigate sources of variability. This strengthened the reliability of test methods and ensured that results were both reproducible and trustworthy across different conditions.
Statistical Study Design: Beyond analysis, statistics played a critical role in shaping the test studies themselves. Carefully structured designs ensured that every study generated meaningful, actionable insights, maximizing the value of both time and resources invested.
Explore how we helped a MedTech client navigate the EU MDR transition by leading a science-driven material reformulation. The result: a fully compliant device that preserved critical product claims, improved performance, and reduced cost of goods – turning regulatory risk into competitive advantage.
Discover how we achived cost savings while ensuring compliance with regulatory standards when we supported a client in the development of a user-friendly injection device.
When a Danish biotech faced a two-year-delayed drug product tech transfer with fragmented communication, lost knowledge, and regulatory risk, Emendo stepped in. Through expert CMO selection, structured governance, and hands-on execution, we transformed the project into a compliant, collaborative, and submission-ready success — securing supply continuity and commercial readiness.
HAVE A QUESTIONS
Every project comes with unique challenges. Tanguy Floch, Senior Scientific Consultant, helps life science companies turn complex data into clear, reliable insights. Reach out to Tanguy to explore how we can support your needs.
Tanguy Floch
Senior Scientific Consultant
Linear mixed models are advanced statistical methods used when data is influenced by both systematic factors (fixed effects, such as formulation type or process conditions) and random factors (such as operator or batch differences). By accounting for both, these models provide a more accurate picture of how changes in formulation impact performance while properly handling variability.
Control charts are monitoring tools that track process behavior over time against statistically defined limits. They help distinguish between normal variation and unusual shifts, making it possible to detect potential issues early and take corrective actions. This ensures stable processes and consistent product quality.
Gauge R&R (Repeatability and Reproducibility) is a type of study used to assess the precision of a measurement method. Multiple operators measure the same set of parts several times, and the results are analyzed to separate repeatability (variation due to equipment) from reproducibility (variation between operators). This helps determine whether a measurement system is reliable enough for decision-making.
