Friday, 9 September 2016

Through the round window

This week, for the first time the MyNHS website published some information on services for people with learning disabilities in England, broken down to the level of individual Clinical Commissioning Groups (CCGs). The relevant webpage is here – although I found it quite slow to load, it did work for me in the end.

Although knowledge is not necessarily power, any steps towards transparency are to be welcomed, and although I don’t know this I guess the hand of Dominic Slowie behind it. Dominic will be much missed as National Clinical Director for Learning Disabilities. In this blogpost I simply want to describe what information has been published online to help people understand what is (and isn’t) there when looking up their local CCG area.

There are two types of information published online here (a much wider set of information about people with learning disabilities broken down by local authority area in England is available here).

The first is the number of people with learning disabilities in specialist inpatient services commissioned by the CCG. This is written as the number of people in inpatient services per million total adult population registered with GPs in the local area, so that areas with very different population sizes can be compared. Overall, the median number of people in inpatient services per million adult population (if you ordered the 209 CCGs from lowest to highest, the median is the number reported by the CCG in the middle) was 62 people in inpatient services per million adult population. As the graph below hopefully shows, the variation is huge. Apparently, if you’re a person with learning disabilities in Swindon or Wiltshire (where 16 people with learning disabilities per million adult population are in specialist inpatient services) you are nearly six times less likely to be put into an inpatient service than if you live across Cumbria and the North East (92 people per million).

Before marching to your local CCG offices with pitchforks (if you can find them - they’re quite often hidden in business parks nowadays) there are a couple of caveats to bear in mind about this information. The first is that more than half of the people with learning disabilities in inpatient services (and attributed in the online information to individual CCGs) are not actually commissioned by CCGs. The latest NHS Digital Assuring Transformationdata for July 2016 reports that 1,175 people with learning disabilities (47%) were in inpatient services commissioned by CCGs and 1,345 people (53%) were in inpatient services commissioned by NHS England ‘specialist hubs’ (these are regional ‘specialist’ commissioners who tend to commission more secure and forensic services). So while I’m sure there are close working relationships between CCGs and these specialist hubs, it’s important that NHS England share in this transparency. This level of detail is available (for each CCG area, how many people with learning disabilities are in inpatient services commissioned by CCGs and by these hubs) (Table 10 here, if you’re interested) but is not in the online information.

A second caveat is that information is set against individual CCGs when it has been provided collectively. For example, all the CCGs across Cumbria and the North East report the same number of people per million (92, the highest numbers in the dataset) – presumably some individual CCGs have lower numbers and some have even higher numbers. Finally, given the pressure CCGs are under from NHS England to cut inpatient service numbers, it’s important to bear in mind the ‘Daisy Unit’ phenomenon. The Daisy Unit is a newly built 9-bed unit in the grounds of a psychiatric hospital, for people with ‘complex’ learning disabilities to ‘receive specialised services’. But apparently it’s not an inpatient unit, it’s a ‘specialist residential home’. Coincidentally it’s been commissioned by Wiltshire CCG, an area with the joint lowest inpatient numbers in the country according to this online information.

The second type of information reported online is the percentage of people with learning disabilities identified in GP registers who had an annual health check. The median figure for this was 46.5%, so just less than half of people identified as a person with learning disabilities with their GP got an annual health check in the ‘average’ CCG. Again, the graph below shows huge variation between CCGs. Apparently, if you live in the area of Coastal West Sussex CCG (13%), you are 6.5 times less likely to have had an annual health check than if you live in the area of Eastern Cheshire (85%).

Again, there are a few things to bear in mind when trying to make sense of your local area. First, in fully 19 CCGs (9.1% of the total number of CCGs), their information was so bad that a percentage could not be calculated. Second, in the year reported, the age range of people of learning disabilities eligible for an annual health check changed from 18+ years to 14+ years, so some areas may not have been very well prepared for this. Third, if a local area is better at actively finding and registering people with learning disabilities on GP registers their percentage is likely to go down (as the number ‘under the line’ gets bigger). The online information does report the total number of people with learning disabilities registered with GPs so it’s worth looking at that as well as the overall percentage.

One thing I haven’t mentioned is the first column in the webpage, a ‘panel assessment’ of CCGs when it comes to services for people with learning disabilities. The headline figures are that 192 CCGs (91.9%) were rated as ‘needs improvement’, 16 CCGs (7.7%) were rated as ‘performing well’, and one CCG (0.5%) was rated as ‘greatest need for improvement’. Because the panel only had two bits of information to go on, their rating system seems quite straightforward – if a CCG was reported as having less than 20 people with learning disabilities per million in inpatient services (3 CCGs) OR if a CCG was reported as hitting 70% of more of people with learning disabilities getting an annual health check (13 CCGs) they were rated as ‘performing well’. As long as you’re doing well on one of these indicators, you can be doing badly on the other one and still be ‘performing well’. With only two bits of information, I’m not sure how useful this panel rating is, to be honest.

Hopefully this blogpost is useful if you’re interested in looking up your local CCG. Whether these first steps towards transparency are really designed to improve the accountability of CCGs to their local populations, well that’s another question.


  1. Thank you for the summary. I must say I agree about the (not) usefulness of the panel rating with only two data items. Numbers are clearly adjusted to give per million figures. Kernow CCG for example serves a population of 532,000. This means that not 29 but around 15 people with learning disabilities from Cornwall are said to be in specialist inpatient facilities. I must say I would also be surprised if this was true since there are none in county and the policy is to use generic psychiatric in-patient beds always in the first instance and as far as possible for longer admissions. They had brought all but 6 back from out of county hospital 2 years ago. It may have drifted back up - but that would be quite some drift. So I think it might be worth looking at the definition of "Specialist in-patient unit". Perhaps they included generic psychiatric beds or specialist out of county social care. Another problem is that being one of 90% who need improvement is not exactly a spur to action in CCG-land where they are trying to manage deficits, Sustainability & Transformation and acute sector overspends. PS why do they keep pitchforks in business parks these days? :-)

  2. I do not know that this information is accurate for other CCGs. No doubt all could improve their services for people with ID but cannot see the point of publishing inaccurate data. At least in their description of services commissioned by our KCCG the figures seem to be plain wrong. We have 7 in Out of County Hospitals, 4 of those people have Learning Disabilities & 3 High functioning autism. The population of Cornwall is 532,000 so that is more like 15 per million than 29. NHS England either have an unusual definition of Specialist Services or have misunderstood or wilfully misrepresented the situation. Do figure perhaps include people in generic psychiatric in-patient beds? Social Care out of county? Forensic patients who have neither LD not autism? Who knows. The Commissioners certainly do not, so it is hard to see, though they may need improvement, how Kernow CCG are supposed to achieve it. And I would be interested to know if these or similar errors are repeated up and down the country. Bad data can not inform good decision making. What, do you suppose, is the reason NHS England are a collecting & b publishing this data? In Cornwall the commissioners health & social care providers have worked hard, and the Commissioners have spent a lot of money, since 2006 to avoid the use of Specialist Units. Hard to show that time & money were well spent with this degree of error in national, public reporting.