For all the sophisticated technology employed in bridge construction, the maintenance of bridges still depends largely on visual inspection of the structures. The visual inspection method, in fact, is the predominant technique used for bridge inspections in the US

The weakness of this type of inspection system, which is less rigourous than in the UK, has been common knowledge amongst engineers for some time.

Indeed,between 1998 and 2001, the US Federal Highway Administration's (FHWA) Nondestructive Evaluation Validation Center undertook a major study to examine the reliability of this method for highway bridges.

The study was designed to measure the accuracy and reliability of both routine and in-depth inspections, study the influence of key factors that might affect bridge inspector performance, and examine differences in State inspection procedures and reporting styles. The results have shown that the methods used and data collected in routine inspections can vary considerably from State to State.

In short it found that the methods were not sustainable. But little appears to have been done to act on the report. Recommendations made by the FHWA Center included:

* Revising the condition rating system to increase accuracy and reliability;

* Increasing the training of inspectors with regard to frequently reoccurring types of defects that they are likely to encounter and methods that would allow them to more easily identify these faults;

* Performing more research to determine if setting minimum vision standards for inspectors would benefit the inspection process.

It is unclear if all or indeed any of these recommendations were implemented nationwide.

Although there is no suggestion that better methodology would have prevented the Minnesota bridge collapse - it was inspected annually - there is some embarrassment that obvious weaknesses in bridge inspection procedures were apparently allowed to go unchecked.

In the UK where many of the bridges are far older, visual inspections are usually allied with computer modeling. The UK Highways Agency has already checked its computer models and is confident that none of the UK's 17,000 bridges are of the same construction as the Minnesota bridge.