Control Dependency: Definition, Signs and Fixes
What is control dependency?
Control dependency is the degree to which problems in an operation get detected, escalated and closed only because a person notices them: errors caught by re-checking, deviations that surface when someone happens to look, incidents that close slowly because nobody tracks them. It is one of the three layers of the HDI, with coordination dependency and cognitive dependency. Required reviews, such as QA review of an executed batch record, do not count as control dependency in themselves; HDI asks what else rests on people noticing.
Why does control dependency matter?
People's attention varies with workload, fatigue, holidays and staff turnover, so an operation that relies on it catches problems unevenly and late. In regulated life sciences the consequences are direct: a deviation found late is harder to investigate, a batch-record error found at review delays release, and an untracked error repeats. In autologous cell therapy, where each batch is made for one patient, a problem caught late can reach that patient.
What are the signs of high control dependency?
Quality relies on a second person re-checking the first beyond what procedures require. The people doing the work find the errors, not the system. Deviations get logged when someone remembers. Closure regularly runs past the procedure's target, and nobody can say how many errors occurred last month. The HDI screening asks about these behaviours, and they show up in the Control layer score.
How do you reduce it?
Move detection from people to systems: validation at data entry, monitoring and alerts on key parameters, tracked deviations with clear escalation rules, and dashboards that show open issues without anyone assembling them. People keep the judgement; systems do the watching. The Phase 2 audit shows where detection depends most on people.
Related terms
- Operational leak: manual re-checking is one of its components.
- Perception gap: leadership often misjudges this layer.