'We will not achieve the 18-week target by carving out resources for specific clinical specialties, as we've done for cancer, because 18 weeks applies to all conditions. We need something radically different.'

Are you doing anything about 18 weeks? No? You're not the only one, but when you do, where are you going to start?

There seems to be a range of approaches: following a sample of patients through their journey, from GP until discharge; examining diagnostic durations and pathways; looking at outpatient clinics and determining the proportion of first-time appointments, second and so on.

Given that there are gross inefficiencies in the current system, choices are two-fold: invest huge additional resources, or change the system. My money is on the latter.

We will not achieve the 18-week target by carving out resources for specific clinical specialties, as we've done for cancer, because 18 weeks applies to all conditions. We need something radically different. In the last decade spending has doubled but capacity has not, so it is unlikely large increases in resources are going to be made available to achieve the target.

A different approach could be borrowed from the Italian economist Vilfredo Pareto. This is to identify and manage the runners, repeaters and strangers within the customer base.

Runners are groups of products in sufficiently high volume to dedicate facilities. For example, acute hospitals in south London might determine there is sufficient volume for a dedicated primary orthopaedics facility.

Repeaters are groups of products with intermediate volume not sufficient to dedicate facilities. These need to be regular and are the foundation of scheduling, be it daily, weekly, monthly, morning or afternoon. The key is for them to be completed at the same regular interval. For example, a specific day surgery procedure might take place every day from 9am-10.30am.

Strangers are groups of products with low volume and should be scheduled around the repeaters and runners. These are likely to be the cases that break 18 weeks most frequently.

We need to understand the real demand. For example, how many orthopaedics patients require primary knee and hip surgery? Which can we readily identify as runners? One way of quickly identifying a cohort of patients that meet these requirements is to 'Pareto' all the patients and find which 20 per cent of patient types account for approximately 80 per cent of the volume.

The criteria for identifying the patients have to be determined and there are likely to be multiple factors that make up the patient type. Within orthopaedics some of the defining characteristics of the runners might be a combination of age, American Society of Anesthesiologists grade, support at home and fitness.

The repeaters are the more complex but regular volume, so these might be 80 per cent of the remaining volume, accounted for by 20 per cent of the remaining patient types. The strangers will be the balance.

I don't think we will hit the 18-week target through managing demand and restricting access. It will be nigh on impossible for primary care trusts and acute trusts to limit access enough to process elective care within 18 weeks simply by managing demand. As we make progress towards the target we may see an increase in demand for services as 'marginal condition' patients decide it is worth seeking treatment.

Can you identify your runners and set up dedicated facilities? Examples might be identifying them through annual check-up outpatient appointments, orthopaedics and one-stop clinics.

Whatever the methodology, the only thing guaranteed to fail is to do nothing.

Andrew Castle is service improvement consultant at the NHS-funded South West London Improvement Academy