Evidence Brief · DA-EB-ENG-001

What a Poor Bridge Rating Can Establish

A National Bridge Inventory Poor rating records component deterioration; it is not by itself a finding that a bridge is unsafe.

Evidence, not advice. This brief does not determine the condition, capacity, safety or legal status of any structure, and it is not an engineering determination. A qualified reviewer would be a licensed professional engineer with bridge inspection and load rating experience, ideally with National Bridge Inspection Standards program management experience. This brief has not had that review.

Question

What can a National Bridge Inventory condition rating establish about a bridge, and what can it not establish?

Who can use this

  • Analysts and researchers interpreting published National Bridge Inventory tables.
  • Public agency staff and policy researchers explaining the difference between condition, load capacity and operating status.
  • Editors and communicators checking whether a statement about a Poor classification goes beyond what the data can support.

Chronology

  1. 1968 Congress first requires the Secretary of Transportation to establish bridge inspection standards.
  2. 1971 The original National Bridge Inspection Standards are published.
  3. 1995-12 FHWA issues the Recording and Coding Guide containing the earlier general condition rating descriptions.
  4. 2007-08-01 Interstate 35W highway bridge collapse in Minneapolis, later investigated by the NTSB.
  5. 2017-01 The Pavement and Bridge Condition Performance Measures final rule is published, the basis for Good, Fair and Poor.
  6. 2018 Structurally deficient is eliminated as a program term, effective with the 2018 data archive.
  7. 2022-01-28 Fern Hollow Bridge collapse in Pittsburgh, later investigated by the NTSB.
  8. 2022-03 FHWA publishes the Specifications for the National Bridge Inventory with revised rating descriptions.
  9. 2022-05-06 The National Bridge Inspection Standards final rule is published.
  10. 2025 Last year inventory data was submitted in the Coding Guide's ASCII format.
  11. 2025-06-15 As-of date of the latest bridge condition by posting status table located.
  12. 2026-08-04 Date stamped on the 2026 bridge condition by highway system table.

Claims

Observed

C-01 FHWA assigns the Good, Fair or Poor summary from the lowest rating among the deck, superstructure, substructure or culvert items, with 4 or below classified as Poor. Federal Highway Administration
Source text: “If the lowest rating is greater than or equal to 7, the bridge is classified as Good; if it is less than or equal to 4, the classification is Poor.”

C-02 The department's report to Congress states that a poor classification does not by itself imply that a bridge is unsafe. U.S. Department of Transportation
Source text: “The classification of a bridge as poor does not imply that the bridge is unsafe.”

C-03 The same report describes the classification as a measure of deterioration relative to the bridge's original built condition. U.S. Department of Transportation
Source text: “the classification indicates the extent to which a bridge has deteriorated from its original condition when first built”

C-04 The report also states that a poor bridge may need more inspection, monitoring or weight restrictions in order to remain safe for public travel. U.S. Department of Transportation
Source text: “A bridge in poor condition may require more frequent inspections, closer monitoring, or weight restrictions to ensure it remains safe for public travel.”

C-05 Under the current specification, code 4 denotes widespread moderate or isolated major defects with the strength or performance of the component affected, whereas code 5 is described as not affecting them. Federal Highway Administration
Source text: “Widespread moderate or isolated major defects; strength and/or performance of the component is affected.”

C-06 The specification treats load posting and condition rating as different things: posting alone, on a bridge designed for less than current legal loads, does not change the rating code. Federal Highway Administration
Source text: “Load posting alone, for an existing bridge designed for less than current legal loads, is not considered a defect and does not affect the condition rating code.”

C-07 Federal regulation defines load rating as a separate analysis of what live load a bridge can safely carry. U.S. Government Publishing Office, Code of Federal Regulations
Source text: “The analysis to determine the safe vehicular live load carrying capacity of a bridge using bridge plans”

C-08 Closure rests on documented criteria that consider both condition and load carrying capacity, not on a summary classification alone. U.S. Government Publishing Office, Code of Federal Regulations
Source text: “Develop and document criteria for closing a bridge which considers condition and load carrying capacity for each legal vehicle.”

C-09 Regulation requires intervals not exceeding 12 months, under the simplified method, for bridges with a deck, superstructure, substructure or culvert rated serious or worse, coded 3 or less. U.S. Government Publishing Office, Code of Federal Regulations
Source text: “must be inspected at intervals not to exceed 12 months”

C-10 Procedures for critical findings must, at a minimum, include a component rated critical or worse, which the regulation ties to a code of 2 or less. U.S. Government Publishing Office, Code of Federal Regulations
Source text: “A deck, superstructure, substructure, or culvert component to be rated in critical or worse condition”

C-11 FHWA states that closure criteria are expected to be general because not every factor requiring closure can be anticipated. Federal Highway Administration
Source text: “All factors requiring bridge closure cannot be anticipated; therefore, criteria are expected to be general in nature”

C-12 In one investigation, the NTSB concluded that a recorded rating of 4 overstated the condition, so a recorded rating can itself be inaccurate. National Transportation Safety Board
Source text: “the overall bridge condition rating was overestimated when evaluated at a 4 (Poor) rating”

C-14 The term structurally deficient was retired from the national program with the 2018 data archive and replaced by Poor. Federal Highway Administration
Source text: “Effective with the 2018 data archive, this term was eliminated from use in the National Bridge Inspection Program.”

Derived

C-15 In the 2026 table, 41,148 of 624,939 bridges are classified Poor, about 6.6 percent by count and about 4.8 percent by deck area. Federal Highway Administration
Basis: 41,148 divided by 624,939, and 217,000,522 divided by 4,548,152,281 square feet, using the TOTALS row for all bridges in SRC-02.

C-16 As of 06/15/2025, of 41,685 bridges classified Poor, 19,602 (47.0 percent) were open with no restriction, 17,571 (42.2 percent) were posted for load and 2,820 (6.8 percent) were closed. Federal Highway Administration
Basis: Division of each status count in the Poor column by the Poor total in the national block of the 2025 sheet in SRC-03; the status rows sum to 41,685.

C-17 Restriction and closure are not confined to Poor bridges: in the 2025 table, 767 of 3,587 closed bridges (21.4 percent) and 44,870 of 62,441 bridges posted for load (71.9 percent) were classified Good or Fair. Federal Highway Administration
Basis: Good plus Fair counts divided by the row totals for Closed and Posted for Load in the national block of SRC-03.

Inferred

C-18 A Poor classification alone supports neither the conclusion that a bridge is unsafe nor the conclusion that it is safe; it is a summary of component deterioration, and operating decisions draw on additional inputs. Federal Highway Administration, U.S. Government Publishing Office, Code of Federal Regulations, Federal Highway Administration, U.S. Department of Transportation, Federal Highway Administration
Basis: Follows from C-01 through C-04 on the meaning of the classification and C-07, C-08 and C-11 on load rating and closure criteria being separate inputs, and is consistent with the mix of statuses in C-16 and C-17.

C-19 The Poor category groups ratings with materially different regulatory consequences, so the summary term conceals information that the underlying code retains. Federal Highway Administration, U.S. Government Publishing Office, Code of Federal Regulations
Basis: C-01 places codes 0 through 4 in one category, while C-09 attaches a 12-month limit at code 3 or less and C-10 attaches critical finding procedures at code 2 or less.

Limits and unknowns

  • The sources carry a tension that this brief does not resolve: SRC-09 says poor does not imply unsafe, while SRC-07 says code 4 means strength and/or performance of the component is affected, and SRC-09 notes that restrictions or monitoring may be needed for a bridge to remain safe. Not unsafe by implication is not the same as a finding of safe.
  • The derived shares use tables from two different years: condition counts from 2026 and posting status from 2025. A 2026 posting table was not found on the FHWA path checked, so the two sets should not be combined.
  • SRC-03 does not state why a bridge is closed or posted. The counts show co-occurrence only and do not show that a rating caused a status.
  • The rating descriptions differ between SRC-08 and SRC-07. Comparison across the standards transition carries uncertainty that this brief did not measure, and the crosswalk files were not analyzed.
  • SRC-12 documents one case. No source reviewed gives a national rate at which recorded ratings differ from what a closer review would assign.
  • The 25th edition of SRC-09 describes 2008 to 2018 and its transmittal letter anticipated a 26th edition; no 26th edition was located, so a later statement of the same point may exist.
  • eCFR and the Federal Register returned an access request page, and Congress.gov, the Bureau of Transportation Statistics and GAO returned HTTP 403 to automated access, so they were not used. The regulation was read through the govinfo 2025-04-01 edition, and the November 2025 performance measures rule mentioned in SRC-01 was not read directly.
  • The 2025 posting page in HTML, as retrieved, ended at the state of Florida, so national totals were taken from the FHWA workbook. The 2026 table's date stamp of 8/4/2026 also differs from the June 15 archive date stated in SRC-01; the reason is not stated.
  • Coverage is limited to structures defined as highway bridges in SRC-01 and SRC-05. Ratings reflect the most recent inspection, so a rating can lag later changes in condition. Individual bridge records were not reviewed.

Prohibited uses

  • Stating or implying that any named structure is safe or unsafe on the basis of this brief or a National Bridge Inventory rating.
  • Using this brief as an engineering determination, inspection finding, load rating or substitute for the responsible bridge owner's records.
  • Using this brief to assign, allege or disclaim liability or fault for any party or event.
  • Presenting NTSB probable cause findings summarized here as findings about any structure other than the one investigated.
  • Ranking places or owners as more or less safe from counts of Poor bridges.
  • Using this brief for travel, purchasing, insurance, investment or legal decisions about a structure.

Safe outputs

  • An explanation of how Good, Fair and Poor are assigned from component ratings.
  • A statement, attributed to the source, that the federal report says a poor classification does not imply a bridge is unsafe.
  • A statement that condition rating, load rating, load posting and closure are distinct records and processes.
  • Aggregate counts and shares from the published FHWA tables, labeled with table year and as-of date.
  • A plain description of what the published data cannot show about a particular bridge.

How to cite this page. Delta Arc, LLC. “What a Poor Bridge Rating Can Establish.” DA-EB-ENG-001, version 1, as of 2026-11-02. https://thedeltaarc.com/research/evidence-briefs/nbi-poor-rating-limits/

Published by Delta Arc, LLC · Editor: Michael Davis · Version 1 · As of 2026-11-02 · Engineering and Reliability · Report a correction