Summary

  • Hanford contractor CPCCo identified 2,734 issues in fiscal 2024; its federal field office said it tracked 346 CPCCo issues. These are separate tracking scopes, while GAO found headquarters lacked a systematic way to use field-office and contractor issue data for cross-site oversight.
  • Headquarters' own tracker held 64 issues identified across the Environmental Management complex in 2024–25. Officials could not explain the selection rules or show how tracked issues shaped risk-based assessment schedules, which headquarters had not implemented since 2020.
  • Headquarters officials were unaware of a recurring series of 15 radiological events at Idaho from June 2024 to February 2025, though the contractor said it reported them and field-office officials said they sent a quarterly report. The useful control is a traceable path from site warning to federal decision.

Three ledgers, one cleanup mission

At Hanford, the Central Plateau Cleanup Company logged 2,734 issues in fiscal 2024. The federal field office said it tracked 346 issues for that contractor. A year later the figures were 1,908 and 320. Those numbers invite a dramatic subtraction. Resist it. The contractor and field office do not necessarily use the same inclusion rules; the difference is not a count of problems that vanished. It is a demand for the mapping that tells an overseer which problems crossed from one record to another and why.

The Government Accountability Office's August 2026 report examined three contracts in the Department of Energy's Office of Environmental Management: Hanford, Idaho and Los Alamos. The sample was selected to probe oversight, not to estimate a system-wide defect rate. The mission is enormous. EM is responsible for cleanup at 15 sites contaminated by nuclear weapons production and research. As of April 2026, it expected the remaining work to run as late as 2100 and cost $641 billion to $840 billion. Roughly 95% of its $8 billion annual budget goes through contracts. More than 25,000 contractor employees work in the program.

This is why the issue ledger matters. DOE relies on contractor assurance systems: arrangements in which contractors check their own work, identify and report trouble, and correct it. Federal field offices oversee those systems; headquarters is supposed to use their information to direct scarce oversight toward higher risk. A private company's internal finding becomes part of a public safety decision only if the reporting chain preserves its meaning.

“Effective” without a ruler

For the latest available 2021–22 declarations, the Hanford and Idaho field offices and their contractors described their assurance systems as generally effective. Los Alamos declined to make an effectiveness determination. GAO could not establish how the positive conclusions were reached. The declarations did not define effectiveness or supply measurable criteria, and they sat beside reports of recurring performance trouble.

The Idaho record shows the weakness of an unmeasured adjective. Contractor IEC documented 15 radiological events over eight months, from June 2024 to February 2025, that put personnel at risk of overexposure and could have spread contamination. GAO says headquarters officials were unaware of the series. IEC said it reported the events to a noncompliance system; field-office officials said they included them in a quarterly report sent to headquarters. The question is where that warning stopped being a signal for headquarters' oversight agenda.

Headquarters' own issue system tracked 64 items identified in 2024 and 2025 across the whole EM complex: 24 at Hanford, 17 at Los Alamos, none at Idaho and 23 elsewhere. This is a separate scope and period from the Hanford contractor's 2024 ledger. It cannot be divided into 2,734 to yield a meaningful capture rate. GAO's sharper finding is that headquarters officials could neither explain how they selected entries from different sources nor demonstrate how those entries informed the assessment schedule.

A map that changes scale at every desk

Even the severity labels differ. At Idaho, the contractor classified issues into five significance levels and the field office used three. At Los Alamos, the contractor used three and the field office four. An issue marked serious at one desk cannot simply be pooled with another site's count. Hanford's field office introduced common sitewide terms and metrics in 2022 and said this helped it compare contractors at that site. Headquarters has no equivalent, documented cross-site map in GAO's account.

The office also failed to run its scheduled independent checks. EM headquarters had not implemented a field-office oversight assessment schedule since 2020. It drafted annual schedules for 2024 and 2025, including reviews of assurance-system declarations, but did not carry them out. A headquarters that lacks a shared issue vocabulary and does not execute its review calendar has little way to test whether “generally effective” describes a working control or a reassuring line in a form.

GAO recommended four linked changes: define assurance effectiveness and measurable criteria, specify the data and terms headquarters needs for risk-based oversight, run the field-office assessment schedule, and put clear assurance expectations and consequences in contracts. DOE agreed with the recommendations. The next proof lies in executed processes, not in agreement with a report.

Sources