Melbourne's Metropolitan Ambulance Service was in a "state of bedlam" this week when a botched upgrade disabled the service's Computer Aided Dispatch system for 24 hours.
In addition to turning up to jobs more than 10 minutes late, multiple ambulances were mistakingly turning up to the same emergency, while others doubled-back because the archaic manual radio fall-back system could not determine the closet available car.
A spokesperson for the Liquor, Hospitality and Miscellaneous Union (LHMU), which represents the Victorian Metropolitan Ambulance Service, said the system failed to restore after it was shutdown at 10pm on Tuesday (August 14, 2007).
On Wednesday, August 15, 2007, the system was completely disabled for 24 hours.
"They were taking so long to dispatch cars to jobs that 10 minutes had elapsed by the time they got around to the next code one [top priority] call-out ," he said.
"It was total bedlam out there; every crew would have gone home without a break.
"Ambulances were crossing over each other and were going where other units had just been."
The crash was the second in five weeks, after IT postponed the upgrade following an initial spate of maintenance problems that disabled the system overnight.
The downed dispatch system uses Automatic Vehicle Locators (AVLs) over GPS to pinpoint the location of the state's ambulances on a map, and automatically routes calls to the nearest available unit.
A digital radio network alerts ambulances of call-outs using selective calling, and information including the job time, case number, address of incident, and a brief job description is sent to pagers inside the vehicles.
A mobile data terminal, similar to the on-board computers used by the police and fire services, was introduced last year to stream updated information to the pagers and reduce radio congestion.
A spokesperson for the Victorian government's Emergency Services Telecommunications Authority (ESTA), which took control of the dispatch system from its original designer Intergraph, said the system is updated every few years.
While he could not explain the exact details of the problem, he said it originated in the networking interface between two of the three emergency call centres, and has since been repaired.
"It was isolated when we found [problems] in data transfer from one call centre to another; there were a number of variables, but a solution has been devised which brought the other two centres back online," he said.
The system's time-to-answer call monitoring came under fire back in 2000 when it was found the routing architecture reset waiting times.
Unanswered calls were pushed though three tiered lines of priority after each call had rung-out for 15 seconds.