Handover concentrates risk because information and authority change hands at the same moment
Aliases: Risk concentration at shift handover · industrial human factors
What it is
Handover concentrates risk because information, responsibility, and control authority change hands at the same time. An event arising during the transfer can fall into the gap between two teams' awareness and authority — the outgoing operator feels the shift is ending and treats it as the next team's problem, while the incoming operator has not yet settled in and assumes the outgoing team is still watching. This handover vulnerability is not a matter of individual carelessness; it is a window the transfer act itself creates.
Why it happens
Before handover, the outgoing operator's attention has already shifted from monitoring the process to organizing a summary, while the incoming operator has just arrived and has not built a full picture yet. The three channels carrying information — spoken word, written record, and live display — can drift out of sync: something said aloud never makes it into the record, or a record is updated with no verbal confirmation to match. Without one unambiguous moment when control authority changes hands, both parties may default to assuming the other is monitoring — a pattern close to diffusion of responsibility, where two people present in the room each believe the current moment belongs to the other, and neither actually owns it.
Under what condition does this stop happening? An unambiguous transfer act — an explicit verbal handoff, a system role switch, a step that only takes effect once the incoming operator reads it back and confirms — narrows this diffusion window sharply, even without lengthening the handover itself. Shrinking the high-risk window is less about extending overlap time than about making sure "who is responsible right now" has exactly one answer at every second, with no gap.
Studying it
Event timing can be analyzed relative to the handover window, but shift timing, fatigue, and operational load must be controlled as confounds — handovers cluster at particular times (commonly around night-shift start), which also coincide with circadian troughs or load changes, so a raw count of events inside the handover window risks conflating a handover effect with a circadian effect; the two need to be separated before drawing conclusions. Simulation studies can insert changes before, during, and after transfer and measure who notices first and who claims ownership of the response.
Where it stops holding
"Handover is a high-risk period" is not a universal rate that transfers unchanged across industries; differences in reporting practice, overlap duration, and shift structure change the statistics substantially, so borrowing a conclusion from another organization or sector is unreliable — local event-timing data should be checked directly. Overlap also has a second effect: a brief period with two people present functions as temporary double-checking and can raise detection of existing anomalies. Whether the net effect raises or lowers risk depends on whether the organization has defined clear ownership rules for the overlap period, not on the existence of overlap itself.
Distributed crews that are not co-located are another boundary case: when the two sides hand over by phone or message rather than in person, "an unambiguous transfer act" cannot be just a spoken confirmation — it needs a system-recorded trace as well, or the diffusion-of-responsibility window ends up wider than it would be for a co-located handover.
Asking the outgoing operator to stay on through a developing event, specifically to avoid the responsibility gap at handover, does remove the transfer window itself — but trades it for a different risk: judgment degrades as time on shift keeps extending. The two risks cannot both be eliminated at once; the choice is a trade-off, not a default answer of simply skipping the handover.
Applying it
Provide a handover window with genuine overlap and restrict nonessential operations during it. Mark the exact moment control authority transfers with an explicit verbal statement or a system state change, and require that transfer to take effect only after the incoming operator actively reads it back — not merely announced unilaterally by the outgoing operator. Events arising inside the overlap window should enter a queue owned jointly by both operators until the incoming operator formally accepts it, so no gap is left unowned. When reviewing an incident, pull out entries that fall inside the overlap window specifically and check the responsibility state, detection time, and response delay at that moment, using it as evidence for whether the handover procedure itself needs adjustment.