Cause and Effect: Why Thanksgiving Sides Fail

Collage of Thanksgiving cooks in home kitchens with the title Cause and Effect: Why Thanksgiving Sides Fail and the label Fishbone Diagram

Thanksgiving season is a good moment to look at a tool that appears in nearly every investigation: the fishbone diagram, also called an Ishikawa diagram or a cause and effect diagram. It organizes the causes of a problem into categories, with branches leading back to the problem at the head of the fish. In our video, six cooks and six defective side dishes serve as an entry point. This post takes each failure into the life sciences settings where the same cause appears, whether the product is a drug, a biologic, or a medical device, and into the evidence a reviewer expects to see when a deviation or nonconformance investigation leads to a corrective and preventive action (CAPA).

Most manufacturing teams know the 6Ms: Materials, Machine, Method, Measurement, Mother Nature (Environment), and Manpower (People). We use the labels Materials, People, Environment, Machine, Measurement, and Method. Pairing each branch with a function that owns it is not an industry standard; it comes from the stakeholder roles in our CAGES® framework, where Operations, Projects, Quality, and Regulatory each hold part of the picture. In the video, Operations holds Materials and People, Projects holds Environment and Machine, and Quality holds Measurement and Method. A branch with no owner is a branch no one contains.

Materials

The bag said yams, and that was all it said. In medical device manufacturing, the counterpart is a polymer resin that meets its certificate of analysis while varying from lot to lot in a property the specification never named, such as moisture content before molding, melt flow, or the level of a processing additive. The parts can pass dimensional checks and still crack in use or degrade after sterilization. Containment starts with specifications built on the attributes that matter to how the part is used, continues through supplier qualification and change notification, and includes incoming testing that checks more than identity. The same pattern appears in drug manufacturing, where an excipient can differ in particle size or moisture and change how a blend compresses, and in biologics, where a media component or chromatography resin lot can shift cell growth or purification performance. An investigation that ends at "the lot was within specification" has not yet asked whether the specification was the right one.

People

Nobody showed the cook what "until the berries pop" looks like. Biologic manufacturing depends on manual technique in places where a written procedure cannot carry the judgment: connecting single-use assemblies, sampling from a bioreactor, seeding a culture. An operator who has read and been tested on the procedure may not yet handle a connection the way a demonstrated operator does, and the difference can appear as contamination, variable cell growth, or a lot that took weeks to produce and cannot be released. Qualification through observed performance, and through aseptic process simulation where the process calls for it, turns training into evidence. Drug manufacturing sees the same gap in aseptic gowning and technique, and device manufacturing sees it in manual assembly, bonding, welding, and visual inspection, where two trained people can reach different results on the same part. Demonstrated competence belongs before unsupervised work.

Environment

Four burners and a closet-sized kitchen turned dough sticky and rolls dense. In drug manufacturing, humidity and temperature change how powders behave: hygroscopic materials cake and flow poorly, weighing and blending drift, and coating solutions spray and dry differently as conditions shift. Storage areas and warehouses add their own swings across seasons and locations within a building. Containment depends on defined limits (alert and action levels where they fit), a defined response when a reading drifts, and a documented assessment of the product made or stored during an excursion. Device assembly has direct counterparts in adhesive bonding and UV or heat cure steps, where a seasonal humidity change can alter cure and bond strength with no change to the written procedure, and in the cleanroom particulate and bioburden limits for sterile or implantable products. Biologic processes add incubator temperature and carbon dioxide levels, and differential pressures in classified rooms, where small shifts affect cell growth and contamination risk.

Machine

The oven dial read 350 degrees, and the corners burned. In drug manufacturing, a lyophilizer is the clearest example: shelf temperature can differ across a shelf and from shelf to shelf, so vials in one position finish with different cake appearance and residual moisture from vials in another, even though the controller shows one setpoint. Depyrogenation ovens and tunnels show the same hot and cold spots. Qualification addresses this with thermal mapping from calibrated sensors placed in worst-case locations and load patterns, and requalification follows because equipment ages. Device sterilization chambers, whether steam or ethylene oxide, have cold spots that depend on load configuration and packaging, which is why a sterilization cycle is qualified for the loads it will actually see. Biologic manufacturing relies on freezers, controlled thaw equipment, and storage where a unit's position affects its temperature history. A setpoint records what the equipment was asked to do; a map records what it did.

Measurement

The scale read low, and the cook never questioned it, because it had never given a reason to. Precision is not accuracy, and a consistent bias is hard to see. In device manufacturing, torque drivers and force gauges decide whether a screw is tight enough or a catheter bond is strong enough, and a gauge that has drifted can pass weak product with every reading looking steady. Calibration against traceable standards at suitable intervals, plus verification before use where a drift would matter, is how the bias is caught. When an instrument is found out of tolerance, the investigation looks back at every result or lot measured since the last good check, which is why the calibration interval matters. Drug manufacturing sees the same issue with balances at dispensing and pH meters, and biologic manufacturing with pipettes, cell counters, and conductivity and pH probes in buffer preparation. For inspection methods, measurement system analysis shows whether the measuring system can tell good product from bad.

Method

The cook followed the recipe exactly, and the result was glue. A purification step in a biologic process shows how it happens in practice: a chromatography column run to a fixed load or time performs well at nominal conditions, but when the feed's concentration or impurity profile shifts, the same instruction produces a different result. A drug process can fail the same way, for example a wet granulation run for a set mixing time when incoming moisture changes, and a device process can too, for example a heat seal on a sterile barrier pouch run at a set temperature and dwell that leaves no margin when film thickness shifts. Process validation answers this by incorporating the process design, qualifying it with ranges that are challenged rather than assumed, and then continuing to verify it in routine use. The test of a method is that it delivers an acceptable result within its parameters, not that it was carried out as written.

What this means for validation.

A fishbone diagram names the places where a defect could have originated; it does not show that any cause is contained. We use containment as the overall term for a cause that is identified, controlled, and validated. Regulatory expects that chain, and a diagram alone covers only the first link. Validation supplies the last link: evidence, gathered under the conditions the process will actually see, that the control on each cause holds. Three questions make the link concrete for any branch. Who owns the cause? What control is in place? What evidence shows that the control works?

In the video, the turkey came out perfectly, along with a wishbone. But Regulatory does not rely on wishes, and neither should an investigation.

Watch the full video here: https://youtu.be/_wYvx56jyfM

If your organization wants to learn how to prove its defect causes are contained, Brayearst Validation Consulting is ready to help. A validation spot check of an existing investigation is one practical place to start. Learn more at www.Brayearst.com.

© 2026 Brayearst LLC. All rights reserved.

Stephanie Brandford

Brayearst Validation Consulting specializes in high-impact workforce transformation through its proprietary CAGES® validation framework. Built on a foundation of Six Sigma methodology and decades of Life Sciences expertise, Brayearst bridges the critical gap between complex engineering requirements and stringent regulatory compliance.

We empower Life Sciences organizations to move beyond mere documentation toward Validation Excellence . Our core focus is the delivery of specialized training programs that standardize how technical teams define, execute, and defend their manufacturing processes. By leveraging the CAGES® methodology, our clients achieve a higher state of audit readiness, ensuring that every claim made in a regulatory dossier is backed by scientifically sound, audit-proof evidence.

https://www.brayearst.com
Next
Next

Cleaned Is Not Clean Enough: What a Shared Equipment Finding Teaches About Cleaning Validation