Consumer Product Safety Standards: How NEISS Injury Data Becomes a Rule or a Recall

Each year the U.S. Consumer Product Safety Commission collects almost 400,000 product-related emergency department cases through the National Electronic Injury Surveillance System. Very few of them ever change anything. The difference is not luck: it is a legal pathway that turns injury counts into consumer product safety standards, into recalls, or into nothing. This article maps that pathway: which statutes create a mandatory rule, why Congress told the Commission to defer to a voluntary standard where one would do the job, what a firm must report, how a recall is classified, and whether standards demonstrably reduce injuries.

What NEISS Is and Why Consumer Product Safety Standards Depend On It

NEISS is a stratified probability sample of 100 hospital emergency departments drawn from a 2020 frame of 4,723 eligible U.S. EDs. Coders abstract every ED visit associated with a consumer product, apply CPSC’s own product, body-part, diagnosis, disposition and location codes, and write a narrative capped at 400 characters. CPSC’s FY2023 report to Congress calls NEISS “the foundation for many of the CPSC staff’s activities.” The reference post on how an ER visit becomes a NEISS record covers the mechanics in full.

Two features matter for regulation. Speed: CPSC notes the data “are available within 24 hours after an incident” in some cases. And follow-back: CPSC’s supporting statement gives a planning figure of roughly 470,000 product-related cases a year, of which about 2,375 are selected annually for in-depth investigation. Those investigations produce the model number, failure mode and event sequence a counting system can never yield.

One boundary matters at the outset. NEISS is a consumer product system: device adverse events sit with the FDA in MAUDE rather than NEISS, under the pathway described in the post on what FDA clearance does and does not guarantee. CPSC’s Recall Handbook excludes devices, drugs, food, vehicles and firearms.

The Statutes That Create Consumer Product Safety Standards

The Consumer Product Safety Act of 1972 created the Commission and is the parent statute for the authority to issue a safety standard, ban a product, require firms to report defects, and compel corrective action once a product is in consumers’ hands. CPSC administers subject-specific statutes alongside it, including the Federal Hazardous Substances Act, the Safe Sleep for Babies Act of 2021, Reese’s Law and the STURDY Act.

The statutory preference for voluntary standards

This is the most misunderstood feature of the system, and it explains why strong injury evidence so often produces an ASTM revision rather than a federal rule. Section 9, at 15 U.S.C. 2058, requires the Commission, before issuing a rule addressing a risk for which industry “adopted and implemented a voluntary consumer product safety standard,” to find either that “compliance with such voluntary consumer product safety standard is not likely to result in the elimination or adequate reduction of such risk of injury” or that “it is unlikely that there will be substantial compliance” with it.

An adequate, widely followed voluntary standard is therefore a legal barrier to a mandatory one. The agency’s burden is not “this product injures people” but “the voluntary standard does not adequately reduce this risk.” Injury surveillance is how that harder proposition gets tested, and it is why CPSC staff sit on ASTM, UL and ANSI committees.

The Consumer Product Safety Improvement Act

CPSIA, passed in 2008 and not 2022 as an earlier version of this article claimed, rebuilt the children’s product regime. Section 102 requires children’s products to be tested by a CPSC-accepted third-party laboratory and covered by a Children’s Product Certificate; general-use products carry a General Certificate of Conformity from the manufacturer’s own testing. Section 104 directed the Commission to review and where necessary make mandatory the voluntary standards for durable infant and toddler products, which is the hook under which the infant walker rule at 16 CFR part 1216 was issued, incorporating ASTM F977 by reference.

Four Routes From an Injury Report to a Rule

A fifth, narrower mechanism sits alongside these. Under 16 CFR part 1120 the Commission has deemed certain product characteristics substantial product hazards by rule, so a non-conforming product is automatically reportable and refused entry at the border: among them hair dryers without immersion protection, extension cords failing UL 817, and window coverings failing ANSI/WCMA A100.1-2018.

Four Documented Cases Where Injury Data Moved Consumer Product Safety Standards

Furniture and television tip-over

Lu and colleagues estimated 560,203 children under 18 treated for furniture or television tip-over injuries between 1990 and 2019 [1]. Clothing storage units were involved in 17.2 percent of tip-overs, and children under six accounted for 82.5 percent of those injuries. The rate rose 53.8 percent from 1990 to 2010, then fell 56.8 percent to 2019, both significant at p less than 0.0001. The authors concluded that education, warning labels and tip restraints “are not a substitute for strengthening and enforcing the stability requirements.”

That is what happened, but only after Congress acted. The STURDY Act directed the Commission on the point; CPSC voted in October 2022 to approve a final rule and in April 2023 to determine that ASTM F2057-23 satisfied the statute. The result is 16 CFR part 1261, applying to units manufactured after 1 September 2023.

Window covering cords

This is the cleanest published example of surveillance data being used to argue that a voluntary standard was not enough. Onders and colleagues combined NEISS with CPSC’s In-Depth Investigation file and estimated 16,827 window blind-related injuries (95% CI 13,732 to 19,922) among children under six from 1990 to 2015 [2]. Entanglement accounted for 11.9 percent of cases, 98.9 percent of those involving blind cords. The investigation file told a darker story: of 231 cord entanglement incidents from 1996 to 2012, 67.1 percent ended in death. The authors concluded that despite existing voluntary standards “a mandatory safety standard that eliminates accessible window blind cords should be adopted.”

That standard now exists. 16 CFR part 1260 took effect on 30 May 2023 and requires custom window coverings to comply with section 4.3.1 or 4.3.2.5.2 of ANSI/WCMA A100.1-2018 instead of section 4.3.2, meaning cord shrouds, retracting cords, or restraints that prevent a hazardous loop. Note the asymmetry: the ED data said most injuries were minor, since 93.4 percent were treated and released, while the investigation data said the tail was lethal. A regulator reading only the counts would have drawn the wrong conclusion.

Button cell batteries and high-powered magnets

Chandler and colleagues estimated 70,322 battery-related ED visits (95% CI 51,275 to 89,369) among children under 18 from 2010 to 2019, with button batteries implicated in 84.7 percent of visits where type was described [3]. The rate rose significantly from 2010 to 2017 (P = .03), and was 24.5 per 100,000 among children aged five and under against 2.2 among those aged 6 to 17. Their conclusion was a direct call on the regulatory system: “Prevention efforts have not significantly reduced injury rates; therefore, regulatory efforts are needed.” Reese’s Law produced 16 CFR part 1263, incorporating ANSI/UL 4200A.

Magnets followed the same pattern on a messier timeline. Reeves and colleagues used NEISS to estimate 14,586 pediatric presentations for suspected magnet ingestion from 2010 to 2015 [4]. Visits peaked at 3,167 (95% CI 1,612 to 4,723) in 2012, the year of CPSC’s action, then fell to 1,907 (95% CI 1,062 to 2,752) by 2015, with a significant downward post-action trend (P = 0.03). A court decision later vacated that ban; the eventual answer was 16 CFR part 1262, requiring any loose magnet small enough to fit the small-parts cylinder to have “a flux index of less than 50 kG2 mm2.”

Nursery products and infant walkers

Gaw and colleagues estimated 1,391,844 nursery product-related injuries (95% CI 1,169,489 to 1,614,199) among children under three between 1991 and 2011 [5]. The annual rate fell 33.9 percent to 2003 then rose 23.7 percent to 2011, and the authors attributed the fall to declining baby walker injuries. Sims and colleagues then evaluated the walker standard directly, analysing 230,676 infant walker injuries among children under 15 months, 74.1 percent from falling down stairs [6]. In the four years after the 2010 federal mandatory standard the average annual number fell 22.7 percent (P = .019) against the four years before, though the authors noted the decrease “may, in part, be attributable to the standard as well as other factors, such as decreased infant walker use.”

Do Consumer Product Safety Standards Actually Reduce Injuries?

A small but rigorous evaluation literature answers this, much of it written by CPSC’s own Directorate for Epidemiology. Hanway and Rodgers evaluated the ASTM voluntary standard for liquid laundry packets using semiannual NEISS estimates from 2012 to 2018 and a negative binomial model controlling for the growth in packet use [7]. The standard was associated with a 49.4 to 61.6 percent reduction in the injury rate among children under five, and may have prevented 9,200 to 23,000 ED-treated injuries. A companion interrupted time series on poison centre exposures found a 28.6 percent reduction in reported exposures [8].

The largest effect comes from an older intervention. Rodgers and Garland, controlling for hair dryers in use and household ground fault circuit interrupter penetration, estimated that immersion protection requirements cut the hair dryer electrocution death rate by 96.6 percent (95% CI 90.8 to 98.8), preventing roughly 280 deaths between 1987 and 2007 [9]. That was a voluntary standard, not a federal rule: the statutory preference working as designed.

Not every product follows suit. Zakeri and colleagues examined hoverboard injuries in NEISS from 2015 to 2022, capturing 4,718 records for a national estimate of 154,121 [10]. Injuries peaked in 2018 and 2020 then declined, but the authors would not attribute that to regulation, concluding only that “continued efforts toward improving safety measures regarding hoverboard use are warranted.” That restraint is the difference between an evaluation and an assertion.

What Firms Owe the System: Reporting, Testing and Recalls

Section 15(b) requires manufacturers, importers, distributors and retailers to notify the Commission immediately on obtaining information reasonably supporting the conclusion that a product fails to comply with an applicable safety rule, or with a voluntary standard the Commission has relied on under section 9, or contains a defect that could create a substantial product hazard, or creates an unreasonable risk of serious injury or death. CPSC’s reporting guidance sets the deadline at 24 hours, while allowing up to 10 working days to investigate whether information in hand is reportable. Note the second clause: failing a voluntary standard the Commission has relied on is itself reportable, which gives an ASTM or UL clause real legal weight without any federal rule. An earlier version of this article cited 16 CFR 1303 as a testing mandate. It is not: part 1303 bans lead-containing paint.

Once a report lands, CPSC Compliance staff assess whether a defect exists and how substantial the risk is, using the section 15 criteria of pattern of defect, units distributed, severity of risk and likelihood of injury, then apply a hazard priority classification. A Class A hazard “exists when a risk of death or grievous injury or illness is likely or very likely”; Class B and Class C step down from there. All three, in CPSC’s words, “create a substantial product hazard that requires corrective action,” and the remedy can be “any repair, replacement, refund, or notice/warning program.” The Fast Track Product Recall Program, introduced in 1995, lets a firm avoid a staff preliminary determination entirely if it implements an acceptable recall within 20 working days of its report. CPSC reported that in 2023, 99.19 percent of Fast Track recalls were initiated within 20 days.

What NEISS Evidence Can and Cannot Support

The original title of this article promised to identify the standards that “could have prevented” specific injuries. No honest analysis of NEISS can deliver that, and the claim has been removed. CPSC’s own wording is that NEISS captures injuries “associated with (but not necessarily caused by) specific consumer products.” A record establishes that a product was present at an ED visit, not that it failed. Counterfactual claims need an evaluation design of the kind used in references 7 to 9.

Four further limits shape what a regulatory-facing NEISS paper can claim. Coverage: NEISS sees ED-treated injuries only. Granularity: it yields national estimates and cannot support state or county surveillance, the most frequently reported limitation in a scoping review of 167 NEISS All Injury Program articles, cited by 38 of them [11]. Case validity: a record-linkage study found 13.32 percent of sampled NEISS cases were non-injuries by the ICD-9-CM diagnosis in the same chart, and concluded that best practice “should include manual review of case narratives” [12]. Precision: an FDA analysis of e-cigarette battery explosions derived an estimate of 2,043 ED visits from 52 unweighted records [13], so read the case count before the headline number. The companion posts on the biases baked into NEISS sampling and whether NEISS injury statistics are reliable go deeper. One further caveat: CPSC is replacing the legacy manual-abstraction design with standardised electronic data exchange covering all 50 states, expected to be fully effective by the beginning of 2027.

Designing a Study That Can Inform Consumer Product Safety Standards

Report both numbers. Shields and colleagues state plainly that 759 actual pressure cooker records produced a national estimate of 28,337 ED-treated injuries (95% CI 24,588 to 32,086) over 2003 to 2019 [14]. The weighted estimate alone hides how thin the evidence is; the raw count alone understates it.

Use the design variables. NEISS is a stratified cluster sample, so an estimate computed without declaring Stratum, PSU and Weight carries standard errors that are wrong, usually too small. We showed the same effect in the piece on running the same NHANES analysis three ways, and our free incidence rate calculator keeps a rate denominator honest.

Read the narratives. Every case above mapped to a standard because someone classified the mechanism, entanglement versus struck-by, ingestion versus nasal insertion, and none of those distinctions exist in the coded fields. The post on the free-text narrative field covers what they carry.

Frame the comparison against a policy date and control for exposure. Without a denominator for how much product was in use, a fall in injuries after a standard is uninterpretable: the product may simply have sold less. Then write up to the STROBE checklist for observational studies. Where a causal claim is intended, the design must earn it, which is where instrumental variable approaches enter.

Common Mistakes to Avoid

• Treating a NEISS count as proof of product failure. A bunk bed study estimating 23,000 children aged 0 to 9 treated annually for bunk bed falls tells you where to look, not what failed [15].
• Assuming a mandatory rule is the default outcome. Under 15 U.S.C. 2058 it is the exception. Check whether a voluntary standard already covers the risk before claiming a regulatory gap.
• Citing a regulation you have not opened. 16 CFR 1303 is a lead paint ban, not a testing mandate, and CPSIA is 2008.
• Ignoring the unweighted case count. CPSC will not report an estimate below 1,200, based on fewer than 20 cases, or with a coefficient of variation above 33 percent.
• Summing CPSC’s product categories. Up to three products can be coded per case, so adding them double counts people.

Tools and Resources

Every primary source used here is free: the CPSC NEISS programme page, the eCFR text of each standard cited, and CPSC’s FY2023 Report to Congress. Still choosing a data source? The guide to which health database your study should use compares them, and how epidemiologists use NEISS covers the detection end.

Professional Services for Injury Surveillance Studies

A regulatory-facing injury study fails in predictable places: the question is already published, the variable you need is missing in the years you want, or the analytic sample collapses once a real case definition is applied.

Everything in this category is listed under Registry and Real-World Data Studies. There is no NEISS-specific package: a NEISS project almost always lands in one of the three above.

Frequently Asked Questions

Does NEISS data prove that a product caused an injury?
No. CPSC describes NEISS as capturing injuries associated with, but not necessarily caused by, specific consumer products. Establishing that a different standard would have prevented an injury requires an evaluation design with an exposure denominator and a defined intervention date.

What is the difference between a mandatory and a voluntary consumer product safety standard?
A mandatory standard is federal law, codified in Title 16 of the Code of Federal Regulations and enforceable by CPSC. A voluntary standard is written by a body such as ASTM, UL or ANSI and is not law by itself. Under 15 U.S.C. 2058 the Commission generally cannot issue a mandatory rule where an adequate voluntary standard exists and is complied with.

How long does a company have to report a defective product to CPSC?
CPSC guidance sets the Section 15(b) reporting deadline at 24 hours from obtaining reportable information, with up to 10 working days permitted to investigate whether information in hand is in fact reportable.

Which CPSC standards were driven by injury surveillance evidence?
Documented examples include 16 CFR 1261 for clothing storage unit tip-over, 16 CFR 1260 for window covering cords, 16 CFR 1263 for button cell batteries, 16 CFR 1262 for magnets, and 16 CFR 1216 for infant walkers. Each is supported by a peer-reviewed analysis of emergency department injury data, cited in the references.

Do published studies show that consumer product safety standards reduce injuries?
Yes, for several specific standards. The voluntary standard for liquid laundry packets was associated with a 49.4 to 61.6 percent reduction in the child injury rate; immersion protection for hair dryers with a 96.6 percent reduction in electrocution deaths; and infant walker injuries fell 22.7 percent in the four years after the 2010 mandatory standard. Effects are standard-specific and should never be generalised across product categories.

Can I use NEISS data to evaluate a standard in my own study?
Yes. The public files cover treatment years 2006 to 2025 and are free. You will need the design variables Stratum, PSU and Weight, an exposure denominator, a defined intervention date, and a plan for classifying mechanism from the narrative field. Report the unweighted case count alongside every national estimate.

Conclusion

Injury surveillance does not prevent injuries. It supplies the evidence that lets someone else act, and statute determines who that someone is. Most of the time the first mover is an ASTM or UL committee, because Congress wrote a preference for voluntary standards into the Consumer Product Safety Act. Sometimes the evidence shows that route has failed and a rule follows. Sometimes the answer is a recall.

The papers that changed something all did the same unglamorous things: they reported the raw case count next to the weighted estimate, classified the mechanism from the narrative, controlled for how much of the product was in use, and stopped short of causal language the design could not support. That is why a handful of NEISS analyses end up in a rulemaking record while hundreds do not.

References

1. Lu C, Badeti J, Mehan TJ, Zhu M, Smith GA. Furniture and television tip-over injuries to children treated in United States emergency departments. Inj Epidemiol. 2021;8(1):53. doi:10.1186/s40621-021-00346-6. PMID 34446115.
2. Onders B, Kim EH, Chounthirath T, Hodges NL, Smith GA. Pediatric injuries related to window blinds, shades, and cords. Pediatrics. 2018;141(1):e20172359. doi:10.1542/peds.2017-2359. PMID 29229682.
3. Chandler MD, Ilyas K, Jatana KR, Smith GA, McKenzie LB, MacKay JM. Pediatric battery-related emergency department visits in the United States: 2010-2019. Pediatrics. 2022;150(3):e2022056709. doi:10.1542/peds.2022-056709. PMID 36032018.
4. Reeves PT, Nylund CM, Krishnamurthy J, Noel RA, Abbas MI. Trends of magnet ingestion in children, an ironic attraction. J Pediatr Gastroenterol Nutr. 2018;66(5):e116-e121. doi:10.1097/MPG.0000000000001830. PMID 29135818.
5. Gaw CE, Chounthirath T, Smith GA. Nursery product-related injuries treated in United States emergency departments. Pediatrics. 2017;139(4):e20162503. doi:10.1542/peds.2016-2503. PMID 28289137.
6. Sims A, Chounthirath T, Yang J, Hodges NL, Smith GA. Infant walker-related injuries in the United States. Pediatrics. 2018;142(4):e20174332. doi:10.1542/peds.2017-4332. PMID 30224365.
7. Hanway SJ, Rodgers GB. Impact of the voluntary safety standard for liquid laundry packets on child injuries treated in US hospital emergency departments, 2012-2018. Am J Public Health. 2020;110(8):1242-1247. doi:10.2105/AJPH.2020.305650. PMID 32552030.
8. Rodgers GB. Evaluation of the impact of the voluntary safety standard for liquid laundry packets on the rate of child exposures reported to Poison Control Centers in the US. Inj Prev. 2022;28(1):43-48. doi:10.1136/injuryprev-2020-044115. PMID 33833071.
9. Rodgers GB, Garland S. The impact of immersion protection requirements on hair dryer electrocutions in the USA. Inj Prev. 2012;18(6):371-376. doi:10.1136/injuryprev-2011-040215. PMID 22589363.
10. Zakeri BS, Fox-Good CG, Nair M, Jaworski HM, Froehle AW. Orthopedic injuries due to hoverboard use, reported in the NEISS database, 2015-2022. Injury. 2024;55(11):111769. doi:10.1016/j.injury.2024.111769. PMID 39137476.
11. Navon L, Chen LH, Cowhig M, Wolkin AF. Two decades of nonfatal injury data: a scoping review of the National Electronic Injury Surveillance System-All Injury Program, 2001-2021. Inj Epidemiol. 2023;10(1):44. doi:10.1186/s40621-023-00455-4. PMID 37679835.
12. Thompson MC, Wheeler KK, Shi J, Smith GA, Xiang H. An evaluation of comparability between NEISS and ICD-9-CM injury coding. PLoS One. 2014;9(3):e92052. doi:10.1371/journal.pone.0092052. PMID 24658100.
13. Blakney AM, Griggs E, Farrell S, Lindsey K, Chang JT. Electronic nicotine delivery system (ENDS) battery-related explosions and burns presenting to United States emergency departments, 2020-2024. Prev Med. 2026;207:108563. doi:10.1016/j.ypmed.2026.108563. PMID 41875938.
14. Shields W, Dong Y, Jager L, Shiang E, Frattaroli S, Omaki E. Using the NEISS database to understand pressure cooker related injuries in the USA. Inj Prev. 2023;29(6):506-510. doi:10.1136/ip-2023-044910. PMID 37666516.
15. Mack KA, Gilchrist J, Ballesteros MF. Bunk bed-related injuries sustained by young children treated in emergency departments in the United States, 2001-2004, National Electronic Injury Surveillance System – All Injury Program. Inj Prev. 2007;13(2):137-140. doi:10.1136/ip.2006.013193. PMID 17446257.

Primary regulatory sources: 15 U.S.C. 2058; 16 CFR parts 1120, 1216, 1260, 1261, 1262 and 1263 (eCFR, current edition); U.S. Consumer Product Safety Commission, Recall Handbook; U.S. CPSC, 2023 Report to the President and Congress; U.S. CPSC, Duty to Report guidance. All accessed 9 September 2026.