A statistical odyssey through the DoLS - with pictures!

It's probably a remnant from my days as a researcher in psychology, but I do love a good graph.  And so, I present to you, a statistical Odyssey through the deprivation of liberty safeguards quarterly figures, in pictures.  I should say, all the figures are for England only, as I can't find equivalent data for Wales [It's here -  cheers CB for the link].



Unexpected underuse of the deprivation of liberty safeguards
I think most people are aware that the DoLS are used far less than was anticipated - but still I find this graph really striking.  It charts the difference between the initial estimates of the number of people who might need safeguards in hospitals or care homes and the number of applications received under the DoLS in the first year.  On the furthest left are the figures given in the original Bournewood consultation.  In the centre is the predictions of the government's impact assessment.  On the right are the results from the first year of operation.

I love Baroness Hale's comments on the impact assessment predictions (from the Journal of Mental Health Law, Winter 2009):

The Government thinks that there are roughly 500,000 people in England and Wales who have a mental disorder and lack capacity living in institutions (including over 190,000 with severe learning difficulties and about 230,000 older people with dementia). Department of Health analysts concluded that around 10% of these would require additional restrictions for their own protection which might involve deprivation of liberty. For some reason they then predicted that there will be 21,000 assessments in 2009/2010, of which only 25% would lead to authorisation. They also predicted that assessments will rapidly decline to 6,600 in 2015/16 but a similar proportion will result in authorisations.

Who knows what basis the DH had for estimating 21,000.  I requested the consultation responses to their impact assessment, and many highly respectable respondents thought it was rather a low estimate.  Anyway, as we all know, the take-up of the DoLS was underwhelming to say the least; there's a brilliant report by the Mental Health Alliance on why the take-up might have been so low, a rather less brilliant report on the same by the CQC and the NHS Information Centre has produced their own statistical report for the first year's data.

Slow but steady(ish) growth in the number of applications under DoLS

This stacked chart shows the total number of applications for standard authorisation received by local authorities, broken up into those they granted and those they didn't.  There seems to be a gradual growth in the overall number of applications, and the proportion granted authorisation seems to have increased since the very early days (overall it's 48%, but it was 31% in the first quarter, and it's 56% now).  I wonder if this means that care providers are getting better at selecting those cases that need authorisation, or whether local authorities are granting authorisation more freely?

And here's the same chart but for PCTs - note that the range is much lower for PCT's (see Y axis) as they had less applications overall (the breakdown is almost exactly 75% local authority, and 25% PCT, for applications and authorisations).

Incidentally, the government's impact assessment predicted that the number of applications would fall year after the first year, presumably because they thought that all the unauthorised DoLS would be scooped up in the first year, and subsequent years would just 'top up' with newcomers.  I'm not sure why they thought it would continue to fall after the second year, especially given demographic growth of the most affected populations.  The impact assessment just says they envisage 'progressively fewer [applications] in subsequent years as all parties become familiar with the safeguards'.

Very uneven application rate between local authorities, not predicted by population size

I matched the number of applications received by each local authority by September 2010 against the size of their population in mid-2009 (the most recent estimate available from the ONS).  Each point on the chart represents a particular local authority. The line imposed on the chart is called a 'regression line' - it is where you would expect all the points to lie if all local authorities received the same proportion of applications per capita of their local population.  The fact that they are spattered all over the place signifies that there are other significant factors that are contributing to huge levels of variation between local authorities in the amount of applications they are receiving, which are nothing to do with the size of their population.  Of course, there might be demographic or socio-economic factors, but it could also be to do with their levels of activity in encouraging and informing care homes to make applications.

The green arrow points to a local authority that has had a tremendously high number of applications from care homes; the yellow arrow points to a local authority that has had far fewer than you would expect on the basis of its size and national average application rate.  (I've decided to be nice and not name them!).  If you look at the red arrow,  even though it's only a little above the line, it comes from an authority with a population of just over half a million, but the amount of applications it's received is what you'd expect from a local authority with a population twice the size on the basis of the national average.  So the differences are pretty staggering.  (I'm afraid because my research is primarily concerned with community care, I haven't done the same for PCTs.  Also, it's much harder to match the population size of a region to a particular PCT.)

The proportion of applications that are granted authorisation

This graph plots the rate of authorisations (ie. the proportion of authorisations granted for each application received) against the total number of applications received.  I know it sounds odd, but it does tell us some quite interesting things.  Firstly, the regression line on this graph gently slopes up; that tells us that the more applications a local authority receives, the more likely it is to authorise them.  This ties in with overall trend that the authorisation rate has risen over time, as supervisory bodies have received more applications.  Now, this may be because teams that receive more applications for some reason end up authorising more; perhaps experience leads to assessors being more inclined to authorise.  Another possibility is that those local authorities that have had lots of applications are the same ones that are ensuring care providers have lots of information about DoLS, and so they are tending to receive less inappropriate applications to authorise.  What do you think?

Notice also that some local authorities have extremely high (100%) or extremely low (0%) rates of authorisation.  These are clustered to the left of the chart - signifying that these 'extreme' authorisation rates are local authorities that haven't received many applications.  Statistically speaking, 'spurious' outliers are more likely to occur where the sample size is much lower ('sample size' being the overall number of applications received).  But it could possibly signify some rather odd biases in teams that are receiving very few applications, it's hard to tell from raw data alone.  Hope nobody's dropped off yet, zzzzzzz.... (seriously though, look at that lovely bell-curve distribution, what's not to love?)

High proportion of authorisations concern people with dementia

Er, this graph does exactly what it says on the tin.  Here's the local authority figures:

And here's the PCT figures:

The DH only started collected data on learning disabilities (the other area my research concentrates on) a bit later on; I'll post that graph when I get around to making it.  If anyone puts in a request for age groups or other disabilities, I'll see if I can drum up the enthusiasm to put those in too.

When deprivation of liberty is said to be occurring, but is not in a person's best interests...

A quick non-statistical deviation.  These can be really messy situations.  Essentially, the best interests assessor (BIA) has found that a person is currently deprived of their liberty (de facto detained), but they cannot authorise the detention because it is not in their best interests.  Due to a staggering oversight by the drafters of the deprivation of liberty safeguards, there's no clear remedy for these situations.  The problem is that unless the BIA can authorise it, the detained person can't have access to the safeguards.  Which is (in my mind anyway) a massive shame - because surely the times you would want to have specialist advocacy services, access to free legal advice and representation, and a 'fast track' to bring your case before the Court of Protection would be the occasions when someone is deprived of their liberty and it's not in their best interests.  As I understand it, lots of local authorities refer these cases to safeguarding, but the impression I get from talking to BIA's is that lots of safeguarding teams don't really get the DoLS (and not all, I'm sure some are very hot of DoLS).  These situations can be really hard to resolve, and often revolve around disputes over funding for less restrictive alternatives.  Sometimes this might relate to the complex interaction between community care law and the DoLS (that's a juicy blog post waiting in the wings).  Other times, perhaps even more stubbornly difficult, detainees themselves (if they have financial capacity and are self-funding) or their deputies/attorneys may be unwilling to release the resources for less restrictive care when a placement is self-funded.

In any case, however, these cases form a very low proportion of the overall number of applications:

You can't really see it on the chart above, but what's really interesting is that despite the overall rise in the number of applications, the number of cases where a BIA finds that deprivation of liberty is occurring but it is not in a person's best interests is gradually declining, with a really dramatic drop-off for the final quarter of 2011:

Now, this could be because care practices are getting much better, and care providers are much more attuned to only using restrictive practices when absolutely necessary.  Another possibility arises from the way the BIA's themselves are responding to these situations.  It is my understanding that the Department of Health has informally advised BIA's to give preference to authorising detention where they find it occurring, even if the placement itself is not in a person's best interests.  At first, I found this approach really hard to understand, as it seemed to condone and legitimise unlawful deprivation of liberty.  However, as time has passed I've recognised that it does provide the detainee themselves with far better resources to challenge their detention than would otherwise be available.

One peculiar conversation I've had with various people though, seems to indicate that they think that authorising will protect care providers and local authorities from litigation.  I'm not so sure it does (although, I'm not sure how much protection from litigation they honestly need, given the massive underuse of the DoLS as a whole and s21A in particular).  The thing is, just authorising a placement doesn't actually make it any more legal if it's truly not in their best interests in the first place.  It's still within the scope of the courts to find that any authorised detention was not in a person's best interests - (the court doesn't act as a 'rubber stamp' for authorisations).

I've also heard arguments that whilst the deprivation of liberty itself is not in someone's best interests, the act of authorising it is because it's in the detainee's best interests to give them the safeguards.  It's not that I disagree with the sentiment here - undoubtedly people will usually be better off with safeguards than without - it's just that this seems to misconstrue the law.  The decision of the BIA is not whether authorising is in someone's best interests, it's whether deprivation of liberty is.  One approach some supervisory bodies may take is to say, well, if the alternative to the deprivation of liberty is to stop restricting them altogether (ie. to let them wander off unassisted, to return them home without a package of support etc), then  it's not in their best interests to do that, but we should authorise it while we work towards finding a solution to bring this crappy situation to an end.  So deprivation of liberty, whilst in itself an overly heavy-handed approach, can be authorised for a short duration whilst less restrictive alternatives are sought.  Some supervisory bodies may impose conditions to help bring a deprivation of liberty to an end.  The problem is that the DoLS don't grant them the powers to enforce those conditions.  Suppose the condition is "the care home must find more staff so they don't have to leave residents in low seating all day to stop them wandering, and take them out more"?  Well - what if the care home hasn't got the resource to do that?  What if the condition is "this person should be supported to live in their own home", but the public authority won't pay for the necessary package of care?The DoLS don't create powers to force public authorities or families funding a placement to cough up to implement these conditions.  The best a supervisory body can do then is refer the whole mess to the Court of Protection (and the s21A referral mechanism won't be possible, because the DoLS don't apply...) - but even the Court of Protection itself may not have the powers to pry more resources out of people (see judgment 1 in this case, paragraphs 22 onwards, and future blog post!).

Anyway, I've digressed rather - but the basis for the decline might just be that it's a messy situation and BIA's are doing their best to give people the means to help bring unnecessary or inappropriate detention to an end.  Of course, this does rather make  you wonder what's happening in the cases where authorisation isn't granted...  Interestingly enough, we see a similar pattern of decline over time for authorisations refused because a person is found to have mental capacity, but only for local authorities and not for PCTs:

Very low, and declining, numbers of third party referrals

This 'donut' chart shows the overall proportion of third party requests making up the total number of applications received under the DoLS (correct to December 2010):

It's very small isn't it?  The skinny green slice that you probably can hardly see are the times when a supervisory body receives a third party request to assess whether a situation is a deprivation of liberty or not, but they do not conduct an assessment.  The supervisory body can refuse on the basis that the request is 'frivolous or vexatious', or an assessment has already been conducted and the situation hasn't changed (Schedule A1 s69).  A third party doesn't have a right to request an assessment by the supervisory body if they haven't already made a request of the managing authority (s68); it's not clear, but I assume these cases aren't included in the figures.

Anyway, the point is that there is a concern (expressed by the CQC and the Mental Health Alliance) that there may be large numbers of people who are deprived of their liberty but for whom authorisation has not been sought by the managing authority.  What these charts tell us is that the 'third party referral mechanism' isn't making up for this shortfall.  This may well be because family members or others who are in contact with de facto detainees are either not aware of the safeguards, or they don't want to rock the boat, or, they just don't see the point of them as they view the restrictions as being in someone's best interests.  Interestingly, the overall number of third party referrals is falling over time, albeit with a slight recent bounce for PCTs:

I suspect this may be because managing authorities are increasingly likely to apply for authorisation when a third party raises it with them, so the concerned third party never has to make a direct request of the supervisory body.

Phew, that's all folks.  All this data, incidentally, is derived from combining the DoLS quarterly statistics available from the NHS Information Centre, with a little help from the Office of National Statistics.  I hope you didn't find it too boring, I realise not everyone loves graphs as much as me.  Please let me know in the comments if you have any theories to explain the figures, any requests for future charts, or spot any mistakes...