The 7 cognitive biases that sabotage our decisions in pharmaceutical Quality (and 3 ways to defuse them)

Every day, a Quality manager makes dozens of micro-decisions: release a batch or not, start a production run, accept a deviation or challenge it. All these decisions go through our brain, and our brain loves shortcuts.

These shortcuts are called cognitive biases. They let us decide quickly, which is valuable. But in a regulated environment, where an error can have health consequences, they can also distort the perception of a risk, make us miss a weak signal or freeze an organisation. Knowing them doesn’t make you immune. It does, however, let you spot them, and put safeguards in place.

Individual biases: we see what we want to see

1. Confirmation bias

A deviation comes in. The problem is known, it’s not the first time, the investigation will be quick: we already know what we’re going to check. If the investigation is limited to looking for what confirms the initial hypothesis, that’s confirmation bias. Contradictory evidence falls through the cracks, and if the problem is recurring, it may be precisely because something was missed the previous times.

The reflex: actively look for data that contradicts the hypothesis, and ask for the opinion of someone who wasn’t involved in the analysis.

2. Availability bias

“That happens all the time.” Then you look at the figures: two occurrences in six months, for an activity performed every week. Our memory retains striking, negative events and gives them a frequency they don’t have. The same mechanism leads people to re-check a data item at every batch record review, even though no error has been found there in a hundred records.

The reflex: rely on an objective history rather than on memory, and regularly put the real frequency of problems into perspective.

3. Probability weighting bias

An exceptional, highly visible defect is found during an audit. The team invests massively so that it never happens again, while less spectacular but far more frequent deviations wait. The brain overweights the rare, striking risk and underweights the commonplace one.

The reflex: use an objective rating method that combines severity, occurrence and detectability, to put each risk back in its rightful place.

Contextual biases: “we’ve always done it this way”

4. Authority bias

In an investigation meeting, a recognised expert very quickly gives their opinion on the cause. Impressed by their experience or status, the team follows that lead without exploring the others. The same bias keeps alive a check added ten years ago after an inspection remark, even though the organisation has completely changed since.

The reflex: introduce collegial decisions, where people always ask whether the data confirms the expert’s opinion.

5. Consistency and status quo bias

A human error? Add a check, “as always”. A simpler method has been identified? Nothing moves when it comes to updating the procedure, because the new seems risky. This need to stay consistent with what has already been done produces an inflation of checks and holds back continuous improvement.

The reflex: add a simple question to your reviews: “Is this still relevant?”

Cultural biases: us versus them

6. Attribution bias

A field isn’t filled in on a check sheet. The conclusion is negligence, when in fact the screen is confusing and the procedure complex. Attributing the error to the person rather than the system gives an illusion of control, but prevents learning. The same errors come back, and teams lose motivation.

The reflex: move from “who is to blame?” to “what do we learn?”, and analyse root causes with the people concerned.

7. Tribalism bias

“They don’t know the process, let us handle it.” Production on one side, Quality on the other, and each tribe distrusts the other. Information stops flowing, investigations get bogged down, and decisions become less reliable.

The reflex: analyse deviations in Quality–production pairs, and organise cross-immersions.

Three ways to defuse biases

• Decide on data, not memories. Deviation history, objective risk rating, indicators: figures put impressions in their place.
• Decide together. A multidisciplinary committee, an outside view, the right to question: diversity of viewpoints is the best antidote to individual biases.
• Record and revisit your decisions. Document the reasoning, then regularly go back over the checks and practices in place, so that habit doesn’t become a rule.

Biases never act alone: authority reinforces consistency, fear feeds the status quo. Recognising them gives teams the chance to decide better, faster and more calmly.