Changing the thresholds - shaking it all about

The third thing that the QOF Advisory Committee dealt with was possible changes to the thresholds for existing indicators. These are the triggers for payment in the clinical indicators. Briefly practices start getting paid when their achievement in a given indicator reaches the lower indicator (currently 40% for everything except MH 6 and DEM 2 where it is 25%). They gain points smoothly until the upper threshold at which they gain the maximum points. Maximum thresholds current vary from 50% to 90%.

The committee does not seem to have approached this task with any great enthusiasm. However the the potential changes in thresholds are the most significant aspect of the whole of its recommendations so far. Practices have been excellent at hitting the targets already set. The committee noted that on the only occasion that thresholds have been increased, a couple of years ago, there was an overall increase in exception reporting as practices became more diligent at seeking these out.

There area a lot of caveats and the committee recommend that most of the details need to be worked out in negotiations - it regards the overall effect on practices at beyond its brief. When moving the thresholds the committee has suggested this should be on a historic basis, picking a threshold that would have previously been missed by half or three quarters of practices. This has the strange effect that success will be met by things becoming harder and failure in an indicator will result in it being easier. There will be a disincentive to achieve anything above the upper threshold although I would expect this effect to be mild.

There is also a proposal to change the bottom threshold of indicators (currently 40% for most inductors) to a level that 95-99% of practices have achieved although it is acknowledged that there is little evidence for setting at any particular level.

The practical upshot of all of this for practices is more work to get the same money, or the same work to get less money. I have looked where the centiles (1st,5th,50th and 75th) in the 2008 data are. You can see a full list of where these new thresholds would lie(Excel or Opendocument spreadsheets) . This covers one year only and the committee has suggested looking back over two but I would not expect a huge difference. No indicator gets easier to achieve and many - even at the 50th centile - get much harder. I have not yet modeled the changes at practice level, things are still a little uncertain, but this could be similar to the square root loss in terms of its effects. Even the Advisory Committee notes that the losses are more likely to be concentrated in more deprived areas and may increase health inequalities.

The effects of these changes will not only make many indicators harder but also make the framework more complicated. All of this now passes to the negotiators and we wait to see what is agreed. Personally I think it unlikely we will see much change in the lower threshold. The upper threshold is likely to be much more fluid.

Going out?

The QOF advisory committee also looked at indicators that might be removed. This makes sense as there is no point in suggesting new areas without trying to free up some cash to pay for them. You can read the full details in the minutes but I will give a summary here.

This was split into two areas. As nobody really knows what happens to service delivery when an indicator is removed they were classified as low risk and higher risk. The low risk areas were:

• CHD 5, Stroke 5 and DM 11 - The measurement of blood pressure in coronary heart disease, stroke/TIA and diabetes.
• DM 5 - The recording of HbA1c in diabetes
• DM 16 - The recording of cholesterol in diabetes

You may have spotted a theme here. These are all about the recording of measurements, or to use the NICE jargon, process indicators. They do not represent outcome. They all have matched outcome indicators. The outcome indicators are not quite a perfect match, of course. Outcome indicators have a much longer exception time after diagnosis or registration (9 months) than the process indicators (three months). They are also much more likely to be affected by other exception reasons such as patient preference for medications and patients having maximum tolerated therapy.

The QOF committee does not currently deal with either the points value of indicators or the thresholds, which is a pity. The effective indicator only works as a combination of criteria, thresholds and points and any division into parts is likely to miss some aspects of the whole. What we can say is that simple removal of these areas would represent a disinvestment in these areas. The degree varies by area. In diabetes the process indicators tend to be about a quarter of the value of the outcomes. The loss of these process indicators is in total about £5 per patient with diabetes. The figures for CHD and stroke are £4.30 and £2.64 respectively (these figures are for the whole register- the amount you gain or lose by performing a blood pressure on a patient already on the register, within the thresholds would be double these figures). To put it another way these indicators are worth £2,282 for an average practice in England.

It is possible to overstate the effects of these losses but there is now no incentive payment for measuring blood pressure in patients unlikely to hit the outcome targets - the patients most at risk.

The committee also suggest three indicators for removal which they state have a higher level of risk.

• DM 22 - eGFR or creatinine testing in patients with diabetes.
• MH 4 - Creatinine and TSH testing in patients on lithium
• Thyroid 2 - Checking TSH in patients on thyroxine - the committee also noted this would leave Thyroid 1 (the register) on its own with no other indicators. They did not actually write "so what would be the point?" but I'm sure I heard them in my head.

Diabetes would still require enough blood tests that the renal function box is virtually bound to be ticked at some point. It was, however, at least a nod to the importance of renal function in diabetes, especially in patients taking metformin.

I am no psychiatrist but lithium certainly seems to be yesterday's drug these days. Around 5% of practices have no patients at all taking lithium. I personally would drop all the lithium indicators.

There really does seem to be little point in keeping the thyroid register if the indicator goes. It is automatically generated. Could we see the first clinical area to be dropped from QOF since its introduction?

Next time - the possibility of changing the thresholds.

Coming in?

The QOF Advisory Committee at NICE has published the minutes of its first couple of meetings which took place in June and July. I must admit that I have quite high hopes of this committee. In fact I have to declare that I applied to sit on it myself although was not selected. The very fact that its minutes have been published is a sign of some improvement in the whole process. We can all see the indicators being proposed and developed. In this posting I will look at the new indicators that they have put on what they describe as the "menu".

On a really very busy day in June they looked at a lot of proposed indicators, and found most of them wanting. The majority were sent back for more work. This has had the unfortunate result of leaving their menu looking somewhat bare. Additionally the committee has not suggested the number of points or ranges that should be attached to each indicator. In fact there is so little detail that these seem more like a list of raw ingredients than fully prepared dishes. These proposed indicators will go on to the negotiators and may possibly become part of QOF.

First the good news. A proposal for staged indicators for blood pressure in diabetes at 150/90 and 140/80. It has long been a perverse part of the QOF that it was much more lucrative to deal with mild disease than more significant disease and this would both reward tighter control and extend the reach of the QOF. Much will depend on the points awarded to each area - no details are given.

Much lower on the practicality scale is the suggested annual thyroid blood tests for patients with Down's syndrome who are over 18. The clinical and financial case for this in the briefing papers is somewhat thin at best. Even worse is the practicality of this indicator. A rough calculation (based on a birth prevalence of 1 in 1000 and a life expectancy of 50) suggests that a typical practice would have around three patients to check annually. Small practices would have fewer and, as we are dealing with such small numbers, many are likely to have none at all. These practices would not be eligible for the points at all in a similar way to that seen already with the lithium indicators. In 2007-8 579 practices (out of ten thousand or so) gained no points for these as they had no patients taking lithium.

The final indicator on the menu proposes the mentioning of contraception and conception advice to women of reproductive age currently treated for epilepsy. It seems to assume that they are quite forgetful and suggests they should be reminded annually. It is difficult to argue that this is a bad thing to do but there is little evidence for any positive effect either. Even the economic experts say that there is so little evidence that assessment is impossible.

It is a rather disappointing menu with little that is new and exciting and little that seems likely to be of significant clinical benefit. The committee sent many other potential indicators back as "must do better" and we wait to see if they can. Next time - what might be going out.