This site is intended for health professionals only

Taking exception


3 August 2011

Share this story:
Twitter
LinkedIn

DR GAVIN JAMIE
Swindon GP

Webmaster of the QOF Database

 

DR GAVIN JAMIE
Swindon GP

Webmaster of the QOF Database

 

DR GAVIN JAMIE
Swindon GP

Webmaster of the QOF Database

 

DR GAVIN JAMIE
Swindon GP

Webmaster of the QOF Database

 

Exception reporting is probably the most controversial and misunderstood aspect of the Quality and Outcomes Framework (QOF). It is certainly the area in which practices are most likely to face challenge from primary care trusts (PCTs) or health boards.

Some, particularly in NHS management, believe exception reporting is a negative feature of the QOF. Considerable pressure is applied to PCTs to reduce or review exception reporting; a pressure often passed down to practices through challenges to figures.

Exception reporting has also been the subject of extensive research and has on occasion made the headlines (often with naive conclusions). Unfortunately PCTs, practices and researchers all have different views of what exception reporting actually is and how it is used. In particular, the statistics have different meanings for each of these groups.

Evidence suggests practices are generally reluctant to exception-report unless necessary to hit the targets. This has been interpreted as evidence of cheating by some academics. When looking at these figures it is important to understand what exception reporting is, how it came to be and how it is used.

Fairness and discretion
At the time the QOF was being designed, the anti-MMR movement was in full swing and immunisation targets were proving hard to reach for some practices. In a number of well-publicised cases, patients putting targets at risk were asked to leave practice lists.

Similarly, it was quite possible for practices to lose money for failing to do something dangerous to a patient. QOF targets were much more complex than immunisation targets and could not hope to include all possible cases within their specification. Exception reporting was included to ensure fairness and allow discretion.

For the first couple of years, no detailed information was gathered on exception reporting. When a reporting system was produced it was as a 'bolt on' to existing systems. A patient is only counted once, even if they fall into more than one category of exception reporting. There is no official guide to which reason will be reported where a patient falls into several categories, and this may vary between practice systems.

You may also hear about 'exemptions'. This is where an indicator does not apply to a patient by design. The smoking cessation indicator does not apply to non-smokers – this is an exemption rather than an exception. Unfortunately the terms sound similar and look even more similar written down – I will not refer to exemptions again.

The basic function of exception reporting is to avoid penalising practices where an action would be unwise, unfeasible or against the wishes of the patient. Additionally, in outcome indicators, there was a further reason where the patient was on the maximum appropriate therapy. So we have four general areas to consider.

Exception reporting a patient from an area does not remove that patient from the disease register; so prevalence adjustments are unchanged. Patients are often excepted from the whole of a disease area but in instances where therapy is contraindicated or at maximum levels it may be for a single indicator only. If a patient is exception reported but also meets criteria of an indicator then that exception report is ignored. So a patient may be excluded from the diabetes indicators due to recent diagnosis but a cholesterol level of less than five will still be counted as a success.

What does the practice need to do?
Most patients are exception reported without any direct action by the practice. For process indicators (measurement and history taking) patients are automatically excepted for three months following new diagnosis or registration with the practice – whichever comes later. For outcome indicators (eg, cholesterol or blood pressure levels to target) this period is extended to nine months. These exceptions are applied automatically by the data extraction software.

For example, a practice with a 10% annual list turnover could have an exception rate of up to 2.5% for most indicators or 7.5% for outcome indicators. Practices with higher turnover would have a larger proportion of exceptions.

Similarly, turnover within a disease population will affect results. Lots of new diagnoses, such as may arise from the new diabetes diagnosis criteria, will push up exception reporting rates. Conditions with a high mortality rate will also tend to have higher automatic exception reporting due to turnover within the disease area. A greater proportion of patients will have the diagnosis made close to the end of March.

The exception-reporting rates tend not to be as high as they could be, as where indicators have been achieved the patients are not exception reported. However, these automatic exceptions remain probably the largest group, a factor not always recognised by organisations comparing crude rates of reporting.

Other areas of exception reporting require direct action by the practice. This is relatively simple where a treatment is simply not appropriate. At its simplest this may be because of a drug that is contraindicated.

Expiring and non-expiring
Exception reporting codes can be either 'expiring' or 'non-expiring'.

Non-expiring codes apply where there is no prospect of the treatment ever being appropriate and generally apply to allergies and other adverse reactions to drugs. If a patient is allergic to aspirin there is little sense in having to enter this information afresh every year. Simply entering these codes once will apply to all future years.

Much more common are the expiring codes. These apply only to the year in which they are entered although they can be used up to 15 months before the year end; with careful timing a single code may apply to two years. These codes will apply where the medication is not tolerated, not indicated or contraindicated. The last reason is a particular point of confusion. For example, if a patient has asthma and coronary heart disease then beta-blockers must be recorded as contraindicated on an annual basis.

In some cases more than one drug may fulfill the criteria. Here, each of the drugs must be individually exception reported. In CHD 9, for instance, aspirin, clopidogrel or warfarin may be used. A patient will only be exception reported if separate codes are entered for each of these three drugs. Where there are contraindications (such as gastrointestinal bleeding) rather than specific allergies this can easily be overlooked.

When (and when not) to exception report
Whole clinical areas may be inappropriate for some patients. A patient in terminal care is unlikely to benefit from active blood pressure management and could be excepted as not suitable.

Less clear-cut are cases where a patient is taking the maximum tolerated dose of a medication. This applies to outcome indicators such as blood pressure, cholesterol levels, epileptic fits or diabetes control. There are always going to be patients whose treatment is limited by unwanted effects that simply become intolerable at higher doses.

The definition of tolerance is becoming increasingly blurred. The traditional medical approach has been to increase the dose until either the therapeutic effect is achieved or the patient has too many side effects to continue. More recently, guidance from the National Institute for Health and Clinical Excellence (NICE) and PCTs has introduced the concept of the maximum dose that the commissioner will tolerate.

Guidelines around cholesterol therapy, for instance, do not advocate the routine use of more expensive drugs above 80mg of simvastatin on cost-benefit grounds. Agreeing a policy with your PCT during the year for what exactly is considered maximal appropriate therapy should make things simpler if exception numbers are challenged after the year end.

The final reason for exception reporting is informed dissent on the part of the patient. There are two ways this can be expressed. Most simply is a statement of dissent from the patient, which can then be coded. This may be declining a particular drug, such as insulin in diabetes. Alternatively the patient may explicitly decline to attend for any treatment.

More commonly, however, patients just don't turn up for review. Although this is technically considered separate from informed dissent, the codes used are the same. The guidelines state that they must be given three personal invitations to attend before the exception code can be entered. These are normally in writing, although they could also be by telephone or a combination of the two.

The interval between invitations is not stated either, although this clearly should be a sensible length of time – in practical terms there should be several weeks between them. To fit these in before the end of the year means that an efficient system should be in place to send out the invitations throughout the 12 months of the QOF year. However, a system that is too efficient may attract the attention of the PCT.

Where exception codes are entered they may be for the individual indicator or for the whole of the clinical area. For instance, a patient may decline to have their blood pressure measured or to take any part in the diabetes indicators. They may have the maximum-tolerated cholesterol therapy or be unsuitable for all of the coronary heart disease indicators.

There is a limitation that codes for specific indicators are not always available. In the current version of the rules (v20) there is no code to say that a cardiology referral is not indicated in a new case of angina; NICE no longer recommends referral in clear-cut cases. In this case the exception will have to be entered for the whole of the CHD area, although achievement of blood pressure or cholesterol targets will still be counted positively. In this particular case, this code would also have to be re-entered in each subsequent year.

We can see from the above that exception reporting is a mixture of automatic processes, information routinely entered for patient care and codes that are specifically entered as exception reports. Understanding this can make understanding statistics and responding to challenges possible.

For many practices, entering specific exception codes is an exercise undertaken in the last couple of months of the QOF year. By this stage it is clear who is going to attend for review and which patients are likely to reach targets. Details of exception reporting are available on QMAS (Quality Management and Analysis System) for practices in England and Scotland, but this is only available from October or so.

For these reasons practices only tend to enter codes where there is a need. Exception reporting is only used where the practice is below the top threshold. So exception-reporting statistics may not be a true reflection of the numbers of patients who could potentially be excepted. There is some evidence that as thresholds rise so exception rates rise to meet them.

Some practices have very efficient and robust systems of exception reporting with reminders sent to patients throughout the year. Ironically this can sometimes lead to higher than expected rates of reporting and challenge by PCTs. It is sometimes unwise to be too keen.

Post-payment verification audits may require verification of exception reporting levels. This is really just a test that practices can demonstrate robust mechanisms for the identification of patients requiring exception reporting.

Exception reporting has been, and remains, a powerful and necessary part of the QOF. Like any powerful tool it must be used carefully but in doing so you can enhance the effectiveness of the QOF in incentivising quality care.

Your comments (terms and conditions apply):

"Excellent article, thanks very much. The surgery I work at is solely for the homeless and as such we have a disproportionate number of exception reports as patients simply disappear. This has set my mind at ease though, thanks again" – Name and address withheld

Twitter
LinkedIn