Scope: examine every item first closed in one outsourced work lane during a fixed four-week period. Include items later reopened, corrected without reopening, cancelled, or still disputed. Define closure, reopening, and the unit of work before extraction.
Methodology: join the queue history, instruction version, reviewer note, change log, and final disposition by stable identifier. Code each reopen as missing required work, changed request, new information, reviewer disagreement, system event, duplicate, or unresolved. Double-code a sample and report disagreements, exclusions, missing histories, counts, and rates.
Inference boundaries: GAO, NIST, and ILO provide principles for records, governance, monitoring, and remote work. They do not set an acceptable reopen rate or prove that a worker, reviewer, vendor, or location caused a reopening. The study is descriptive and supports a narrower process test.
Limitations: deleted events, informal approvals, changing queue definitions, unequal case difficulty, and repeated items can bias results. A low rate can also reflect failures to reopen. Report the denominator and taxonomy with the result. References are listed below so readers can inspect the primary guidance used.