In biopharmaceutical manufacturing, the cost of an undetected failure is not just a batch loss — it's patient safety, regulatory action, and months of recovery. Failure Mode and Effect Analysis (FMEA) is the industry's most rigorous framework for identifying, quantifying, and controlling process risks before they materialise.

Failure Mode and Effect Analysis (FMEA) Flow

A systematic five-stage process that moves from identifying what can go wrong, through quantifying its likelihood and impact, to implementing controls that reduce residual risk to acceptable levels.

Understanding the Risk Priority Number (RPN)

The RPN is the mathematical backbone of FMEA. It converts three expert-scored dimensions into a single actionable risk metric — enabling consistent prioritisation across all failure modes in a bioprocess.

Why Traditional Risk Assessment Falls Short

Bioprocesses are complex, multi-variable systems. A single manufacturing run involves dozens of interdependent parameters — inoculation density, pH control, dissolved oxygen, temperature, feed timing, agitation — each of which can deviate from specification in ways that affect product quality attributes (CQAs). Traditional qualitative risk assessment, relying on expert opinion alone, struggles to capture the true probability and detectability of failure modes across all these variables.

The data-driven FMEA approach addresses this limitation directly. By grounding risk identification and assessment in historical bioprocess data, platform knowledge, and quantitative statistical analysis, it replaces gut-feel ranking with evidence-based prioritisation. The result is a risk register that reflects actual process behaviour — not just what engineers expect to happen.

Inside Each FMEA Stage: A Bioprocess Perspective

Stage 1 : Risk Identification

The first stage asks: what could go wrong, and what would be the consequence? In bioprocessing, this means systematically mapping each process parameter (PP) to each Critical Quality Attribute (CQA) it could affect. The data sources used at this stage include existing development data from earlier runs, platform process knowledge accumulated across similar molecules, and published literature on known failure modes in comparable systems.

The output is a structured list of potential failure modes — not vague concerns, but specific, testable hypotheses about parameter-quality relationships. For example: "If pH drops below 6.8 during the exponential growth phase, ammonium accumulation increases, reducing CQA X."

Stage 2 : Risk Assessment

Each identified failure mode is then assessed across three dimensions. Severity is scored based on the potential impact of failure on product quality — a failure that compromises a critical CQA like potency scores higher than one affecting a non-critical attribute. Occurrence is scored based on how frequently the failure mode is observed in historical data. Detectability reflects how reliably existing monitoring and control systems would catch failure before it causes a batch impact.

Critically, assessment evaluates each parameter at both its "High" and "Low" ends of the operating range — capturing directional asymmetry in risk. A temperature excursion above setpoint may have a different risk profile than one below setpoint.

Stage 3 : Risk Evaluation via RPN

The RPN (Risk Priority Number) is calculated as the product of Severity, Occurrence, and Detectability scores. This single number allows the risk team to rank all identified failure modes on a common scale and classify them as High, Medium, or Low risk. The classification directly determines what action is required: high-risk parameters must be addressed through process characterisation studies; medium-risk parameters require targeted investigation; low-risk parameters may be accepted with documented control measures.

Stage 4 : Risk Reduction Through Characterisation

For high- and medium-risk parameters, risk reduction is achieved through process characterisation studies — designed experiments (DoE) and multivariate analysis that empirically establish how the parameter affects the CQA within the operating range. The study outcome either confirms that the risk is controllable within the design space, or reveals that additional control measures (tighter specification, enhanced monitoring, or process changes) are needed.

Stage 5 : Risk Acceptance and Residual Risk Management

Not all risks can be eliminated. Where process characterisation confirms that a risk cannot be reduced further, residual risk must be formally accepted — documented with appropriate control measures in place and approved through a structured review process. This residual risk register forms part of the regulatory submission package and underpins the process validation strategy.

How Graphtal Supports Data-Driven FMEA

Graphtal brings together historical data integration, advanced statistical tools, and hands-on facilitation expertise to make FMEA more rigorous, more defensible, and faster to execute than traditional approaches. Here is what that looks like in practice: