The wrong call for the Ambulance Service

The National Audit Office produced another report this month.  This time it was on the Ambulance Service.  And its headline finding was that the focus on response times had caused inefficiency in the system.  This is such a muddled finding, that Militant Manager does not even know where to start.

But let me try.  I think you can justify response times in three ways.  Firstly, it is a common sense target.  Secondly, its robust and meets all the criteria for targets.  Thirdly, the "inefficiencies" that the NAO finds are due to an incomplete analysis.

First, let's try the common sense approach.  Response times is what Joe Public expects from his ambulance service.

I can demonstrate this through a thought experiment.  We are going to ask somebody what they feel is an appropriate target for the ambulance service in this experiment.  Imagine we are in Sheffield city centre on a Saturday morning - shopping time - bound to find lots of people.  But wait - those idiotic 1980s town planners thought it would be a good idea to open a huge shopping centre near the M1 - Meadowhall.  So we cannot find anybody.

Luckily those town planners tried to cover up their idiocy by commissioning the Winter Gardens, and that is where we find Aron Ralston (not Joe Public unfortunately).  We ask Aron the simple question: "If you had to think of a target for an ambulance service - a target that they would then pull all stops to meet - what do you think it would be?"  Aron is a bit strange, saying "The key target should not be response time - I do not mind if it takes 127 hours; what I want as the target is the handover time and information sharing at the hospital?"

The difficulty of imaging proves my point that response times is the common sense target, and you need to go to the Winter Gardens in Sheffield to find anybody who does not think so.
Secondly, the target meets all the criteria that you would want a target to have.  Better people than me have worked this out as SMART targets.  Specific, Measurable, Attainable, Relevant, Timely.  Response times fits all those criteria to a "T."

The third issue is the key one.  The NAO analysis is incomplete.  It focuses purely on the unintended consequences of the policy - more than one ambulance scrambled for Category A calls, and then stood down; but not on the other aspects - the intended consequences; and what would have happened under an alternate policy environment.

There are many aspects to every policy and its effects.  Militant Manager has created a handy two-by-two matrix to illustrate all the aspects (MM should have been a management consultant).  You can find this matrix here.  Causes are split into "Policy" and "Other factors" and Effects are split into "Intended" and "Unintended."

All policies will have unintended consequences.  And the NAO has focused on the unintended consequences of the policy, quantified them and published them as avoidable inefficiencies.  It is a bit like looking at a football match, identifying all the shots that missed the goal, and then figuring out the "points lost" as a result of those missed shots.  Yes, that is interesting and difficult analysis, but in itself it tells you little: what you can control is the shots; not whether it misses or not.

The true answer lies in looking at both the unintended and intended consequences, quantifying both; and then comparing all of this to an alternate policy environment.

One would well argue that the dual scrambling of ambulances is caused by this response time target, but so is better response times (the intended consequence) - which would have resulted in greater public satisfaction benefits, as well as clinical benefits.  You need to consider them all in the round.

And then you need to consider them against at alternate policy.

The NAO's alternate policy is a bit weak.  Some amorphous set of indicators, which may not comply with public expectations, be SMART or result in pure good effects.

********

My main quibble is what the NAO has decided to lead with in its public announcements: the response time target - which is robust, SMART, well accepted, and, on balance, desirable.  The actual report has some other very interesting findings which do point to the inefficiency of ambulance services: 5-8% sickness rates, increasing costs per trip, reducing labour productivity . . .  None of these are caused by or cause the dual scrambling.