Another view: New ACOs need data and insight

OK, we all know CCGs have failed. They were a good idea. Let local clinicians on the ground, the GPs, who know the system, run things.

They see most of the patients. They aren’t influenced by latest fads or getting a professorship. They know what works and what doesn’t. It all seemed a good idea at the time.

Problem is, no one then gave them the power to achieve their aims. Acute foundation trusts don’t have to listen to them, so they don’t. They report to Monitor or whatever it's now called. They do have to make a profit and hit their four-hour target and…well, that’s about it.

What they certainly don’t have to do is help the local health and care economy balance the books?  

CCGs on the way to oblivion

So, CCGs are history and we move on again. Only this time the Government have decided they can’t be seen to be doing another top down reorganisation of the NHS, because they promised not to, just as everyone decides that accountable care is the way forward.

So, what do they do? They create the Sustainability and Transformation Programme (STP) process where every area must write its own locally-owned answer to the problem, just as long as that answer is an Accountable Care Organisation (ACO).

ACOs equal managing a unified capitated budget

And just what is an ACO? Effectively it puts all the services in one area into one organisation, gives it a unified capitated budget and tells it: “You have to live within your means. No really, you really do this time. We won’t bail you out, so you’d better start behaving responsibly”.

Problem is one half of the DH thinks the only people who should run the beast are the hospital managers, who are the ones who least get running a service to a capitated budget – as they are all on Payment By Results (PBR).  The other half thinks GPs should be allowed to have another go, but they need to merge into one first.

Once we are there everything sounds lovely, especially to me as a primary care doctor. Speak to any of these ACOs and the story they tell is when they form the hospital, docs try to take over, they are clever and important and their department needs more money. After about three years of near bankruptcy they realise they need to invest in prevention, screening and primary health care and they start hiring more GPs.

What information system does an effective ACO need?

This all sounds good, the problem is how do we get there? And are our current IT and clinical systems fit for this journey?

Lot of people talk about using the same computer systems at the front door as in the community – locally there are some moves to this. We already have all GPs on Emis, all community staff on it, so putting acute onto it makes some kind of sense.

There is a big debate about whether we just need to be sharing data across systems, or whether we need to drop all the data into one big pot. The problem is for current clinical care, sharing data is probably enough, for delivering an ACO I’m not sure it is, we need a lot more intelligence.

ACOs need intelligence not just data sharing

The operational strategy of an ACO should be to make decisions and plans that are cost effective. They should invest in prevention, in upstream early diagnosis, and in treating people aggressively to prevent them deteriorating and requiring an operation or admission.

Rapid access to diagnosis and specialist advice is massively important. This is more than allowing patients to Skype. This is about having as near to instant advice available for the clinician dealing with the patient.

If you don’t know what this rash is, or what drug to give next to this hypertensive, referring them and then waiting months for an answer is a huge waste. What we really need is to get the advice and opinion there and then. Whether this be via chat rooms, video links, bleeps and on call rotas really doesn’t matter. What I need though are tools that enable me to make the right decision when I see someone.

Good health intelligence vital to evaluate system change

And perhaps even more important than helping individual clinician decisions, is evaluating system change. If you change the system what effect is it happening?

Locally, several areas have introduced community matrons and case managers, but it’s difficult to understand whether they have made any difference. What do you measure? How do you measure it?

I do a lot of clinical trials research, and here the metrics are clearly thought out beforehand. Does this drug make X lower? Does it reduce incidence of Y?

By comparison, answering the question of whether investment in community matrons has reduced admissions is far more difficult. The local hospital is still struggling. Admissions are up, would they have gone up more without them?

So are we collecting the data we need to inform decisions about investments and changing services? Are we collecting too much unnecessary data? which is just as big a burden. We have to avoid wasteful bean counting.

Bean counting alive and well

I recently came across a member of staff at the local community provider whose whole-time job was collating data for the commissioners for some contract meeting. I seriously doubt the commissioner ever even looks at the data being provided.  It made absolutely no sense to me and was completely pointless in understanding whether that service is doing any good.

Another example, is my practice recently did a bone health project, looking at patients at risk of osteoporosis. The model we used predicts that over the next three years we would save around £50,000 in costs in A&E attendances and operations.

But will it? How do we find that out? My list size is growing; the local rate of attendance appears to be going up – the population is getting more elderly – it’s highly likely that I just won’t be able to tease this out. But I suspect that we really need to.

Without knowing which pieces of work, work, how do you know what to do or how to prioritise it?

How can we avoid shooting blind?

In a trial, you’d use a control – in this case some practices haven’t done this project, is there any way of tracking us versus them? EMIS won’t do this? can I get this from HEs/SUS data?

I’m about to go on a training course for something called Aristotle – I’ve no idea if this will do it. Seems like we are shooting blind at moment. We do good things. We don’t know if they work. We don’t then prioritise the good ones and stop the bad ones.

This publication, Digital Health, has long championed the idea of CCIOs. I agree, we need more clinicians thinking IT, but they should be thinking not just about hardware and software but data and analytics. The key really is information and its analysis to gain insight into performance and outcomes.

I was recently in a meeting where we were talking iPads and VPN and RDCs but people were a bit taken aback when I asked what information did we need to collect and what was a waste of time. We need experienced clinicians working with the data people to work out what we need to count and why and for how long.

Information and insight needed to model changes

We urgently need information on the whole system and tools to model changes. There is always the law of unexpected consequences. The local health economy is like an ecosystem you can’t just introduce a new species and expect things to work out.