The government's latest plan for improving our ambulance service says the 12-minute target for responding to calls will be met in not less than 92.5 per cent of cases. That is a respectable performance when you consider traffic congestion, breakdowns and problems with staffing levels, which can sometimes occur. Unfortunately, though, there will still be no system for discriminating between life-threatening, less critical and non-acute cases - or even frivolous ones that do not require ambulances at all. The delay in implementing such arrangements does nothing to enhance service quality. It is disappointing that we will now not get a final decision - let alone any action - on a new dispatch system aimed at ensuring that ambulances are sent to the most urgent cases first until 2014 at the earliest. The Fire Services Department had already put back the target date from 2012 to 2013. The Security Bureau says the latest delay is to address public concerns. A senior ambulance officer who is familiar with the plan says it would take time for the public to understand the new system. The government is understandably cautious about 'rushing' the new scheme's introduction because of fears expressed by staff unions and patients' rights groups that it could lower the level of service, since not every patient - or person calling for an ambulance - could be precise about his or her condition over the telephone. Ambulance calls would be categorised, on the basis of callers' answers to dispatch operators' questions, as life-threatening, serious or non-acute, with response targets of nine, 12 and 20 minutes, respectively. At present, emergency receptionists need only note addresses and contacts to direct ambulance crews to the right location, although they may seek more details of a patient's condition. Prioritising would involve asking clear, structured questions about a patient's condition and making potentially life-and-death decisions. There would be little room for error, and mistakes would be inevitable. But a system that does not discriminate in life-and-death matters is fundamentally flawed, as it has built-in errors. There is a need for a more sophisticated approach, making the best use of available resources. Safeguards could be put in place to minimise the risk of inaccurate prioritising of calls. When in doubt, for example, calls should be treated as an emergency. That is how other advanced economies have filtered out non-emergency calls. But this is shaping up as another policy on which there is inaction by an unelected government that craves consensus before it will act. Prioritising is no more risky than a system with a one-size-fits-all benchmark of accountability for service performance. The government should not delay introducing the scheme any longer. The public can be educated to accept it. The most effective education would be experience. Talk of trying out a priority system in selected districts before 2014 should be translated into action, at least. That said, the latest plan includes some worthwhile initiatives in the meantime, such as the introduction from next year of simple first aid advice by dispatch operators for easily identified injuries such as burns, bleeding and bone fractures pending the arrival of an ambulance, and a full protocol for questioning a patient or emergency caller by 2013. Tiered ambulance response targets are a logical extension. The government should strive to match best practice in comparable societies around the world.