Failure Recovery

Lab recovery begins with the actual observation, not the planned outcome. A failed assay, blocked batch, incomplete handoff, or lossy LIMS export must remain visible in durable history while the next action is assessed from scientific and operational evidence.

flowchart TD
    F[Failure or blocked handoff] --> O[Preserve plan, observations, and artifacts]
    O --> T[Classify failure and triage batch]
    T --> R{Responsible to rerun?}
    R -->|no| H[Hold, redesign, or retire]
    R -->|yes| P[Reassess controls, materials, capacity, and risk]
    P --> V{Handoff validation passes?}
    V -->|no| H
    V -->|yes| N[Create a new attributable plan]
    N --> C[Compare outcomes and feed learning]

Preserve the incident record

Retain the experiment plan, batch and assay identifiers, upstream recommendation and evidence references, schedule, material reservations, required controls, risk assessment, handoff validation, LIMS export and loss notes, observations, and outcome assessment. Record what was actually executed and observed. Do not edit the original plan to resemble the recovered attempt.

Classify failures by their owning boundary:

Condition	Response
Upstream recommendation is ambiguous, stale, or on hold	Reopen review; do not schedule from inference.
Required control, sample, reagent, staff, instrument, or capacity is unavailable	Keep the batch blocked and rebuild practicality and schedule reports.
Assay feasibility or predicted failure risk is unacceptable	Redesign or refuse the handoff.
Observation is missing or inconsistent with the planned assay	Reconcile identifiers and custody before scientific interpretation.
Execution completed but QC or outcome assessment failed	Triage the failure class and decide rerun, redesign, or retirement.
LIMS mapping loses required meaning	Correct the mapping and issue a new export bundle with loss reporting.

Authorize a new attempt

A rerun is a new attributable plan. Reassess decision readiness, assay risk, controls, material feasibility, instrument and staff capacity, schedule pressure, and blocking review gates. Preserve changes to protocol, batch design, controls, or acceptance criteria as explicit differences. A failed assay should feed future priority and planning models rather than disappear behind a successful batch summary.

Recovery is complete when the new handoff passes validation, its LIMS export is reviewable, observations reconcile to planned assays, the outcome assessment is recorded, and the learning record explains what changed. A scheduled slot or a successful serialization alone does not demonstrate a responsible recovery.