Injury data are changing

CPSC’s July 22, 2026 plan would extend its injury-surveillance network across all states using electronic health records, with full effect expected in early 2027. It promises faster detection and data minimization, including support for de-identification before records reach CPSC. These are implementation commitments, not demonstrated reductions in injuries. ^[1]

The existing NEISS system estimates national product-related emergency visits from a probability sample of hospitals. It is a way to detect patterns beyond the patients actually observed, rather than a census of every injury or every hospital. ^[2]

Sampling requires restraint. CPSC flags estimates below 1,200, samples with fewer than 20 records, or coefficients of variation above 33 percent as unstable. Its raw case counts have different weights and should not be used as comparable national totals. ^[3]

Even a trend can reflect a coding change. CPSC’s query instructions explain that baby-slings coding changed in 2019; searching only the newer code makes earlier cases disappear. Readers comparing years should check the comparability table and uncertainty measures before treating a rising or falling number as a safety outcome. ^[4]

About this item: Retrospective report. No first-hand reporting or independent testing is claimed. Corrections policy.