The denominator changes what the rate means
An HHS inspector-general review reports risk assessments in 80% of cases in which ORR’s National Call Center spoke with a child. The full appendix shows what that conditional measure leaves outside its denominator: staff spoke with a child in 125 of the 250 sampled safety cases and documented a risk assessment in 100. Across the entire sample, that is 40%, not 80%.
What’s the Scoop With Broach matched the contact table with the screening table in the September 22, 2026 evaluation to examine how far the helpline’s documented safety checks reached. The calculation does not establish that the remaining cases received no help, that each case required the same response, or that 60% of children were unsafe. It identifies the scope of a published performance measure.
| Documented action | Report’s count and denominator | Share of all 250 sampled cases |
|---|---|---|
| Spoke with child | 125 of 250 | 50.0% |
| Risk assessment | 100 of 125 children-contact cases | 40.0% |
| Trafficking screening | 62 of 125 children-contact cases | 24.8% |
| Personal safety plan | 38 of 125 children-contact cases | 15.2% |
| Referral to one or more outside entities | 205 of 250 | 82.0% |
| Follow-up call | 104 of 250 | 41.6% |
Why multiplying 70% by 61% gives the wrong whole-sample answer
The summary says the center attempted contact in 70% of cases and reached children in about 61% of those attempts. Appendix A specifies that the attempt measure excludes 33 cases in which children made the initial call. Of the remaining 217, staff attempted contact in 152 and reached the child in 92. The 92 successful outgoing contacts plus the 33 child-initiated calls reconcile to 125 cases with direct child contact.
Among the 65 cases without an attempted outgoing child contact, records attributed 40 to missing contact information; 25 contained no explanation. The report says staff were not required to document that explanation. Those entries therefore describe a documentation limit and barriers to contact, not an independently established policy violation in every case.
Referral and follow-up are different actions—not a safety verdict
The center referred 205 of the 250 cases to at least one outside entity, including law enforcement, child protective services and trafficking-related support. Such action can occur without a successful direct conversation with the child. Conversely, a referral does not prove that an outside investigation occurred or that the underlying concern was resolved; the report describes instances in which outside entities declined referrals because of jurisdiction or eligibility requirements.
The evaluation does not publish the intersections needed to determine exactly how many referred cases also received follow-up calls. Subtracting the 104 follow-ups from the 205 referrals would therefore manufacture a count of allegedly unresolved referrals. Follow-up also was not required under the policy operating during the reviewed period. The audit measures documented actions, not the appropriateness or outcome of each response.
These are 2024 cases, and the operating arrangements changed
The sample covers reports from January through June 2024 involving running away, abuse or neglect, or trafficking concerns. Those three categories represented 48% of the potential safety concerns reported to the center, not all children released to sponsors. The report separately counts 1,944 unique concern reports involving approximately 1,800 children; it does not verify every allegation or turn the concern-report count into a count of missing children.
The review concerns a former contractor, replaced on December 15, 2025. Its endnotes report that, as of August 2026, the then-current contractor was required to conduct follow-up when notified that a child missed immigration court and to establish follow-up procedures for callers with safety plans. Those changes are material counterevidence to any claim that the 2024 follow-up rules necessarily describe today’s operation. The evaluation does not measure the later contractor’s results.
Method and evidentiary limits
We used the appendix’s exact case counts rather than multiplying rounded percentages, recomputed each whole-sample share, and checked the study’s methods, confidence intervals and policy endnotes. HHS-OIG selected a random sample of 250 cases within the three identified concern categories; it reports a 43.8%–56.2% confidence interval for the 50% direct-contact estimate. Our 40% figure is the descriptive ratio of 100 assessments to the 250 sampled cases, not a newly calculated population confidence interval.
OIG did not evaluate whether each action was appropriate and cautions that undocumented actions may have been missed. The report supports a precise accountability finding about the reach and presentation of recorded checks. It does not establish a current nationwide child-safety failure rate or a count of confirmed trafficking victims.
Sources and further reading
HHS OIG official report publication record, issued September 22, 2026 ↗
Photograph source and reuse: Ben Schumin / Wikimedia Commons · CC BY-SA 3.0 ↗