R Duncan, G Barlow*, A C Smith*
West of Scotland Regional Epilepsy Service, Southern General Hospital, Glasgow * The Crescent Medical Practice, 12 Walmer Crescent, Glasgow
Correspondence to: Dr R Duncan, West of Scotland Regional Epilepsy Service,Southern General Hospital, Glasgow G51 4TF Email: r.duncan@clinmed.gla.ac.uk
SMJ 2005 50(3): 114-117
Abstract
Background: SIGN has recommended annual review of all patients with epilepsy. Annual review is rewarded in the new GMS contract. There is no information on how or by whom reviews should be carried out, nor on resource implications for secondary care. Aims: To determine whether a practice nurse can deliver annual review of patients with epilepsy, and to estimate the resource implications of such review. Methods: Evaluation of a practice nurse checklist against review by neurologist in 62 patients with epilepsy identified from a practice list of 6240 from Southwest Glasgow LHCC, and audit of case records in 1259 patients with epilepsy identified from the whole LHCC population of 96,565. Results: There were 8 discrepancies between nurse and doctor reviews in a first iteration, but none in the second. Changes suggested a training effect. The review process generated 19 epilepsy nurse appointments, 7 requests for cerebral imaging and 3 requests for video EEG. Twelve patients required continuing follow up. The LHCC audit identified a large number of patients who had inadequate documentation of information and advice (over 90% in some domains). 28.6% had not been seen by a specialist, 40.7% had not had cerebral imaging, and only 37.4% were seizure free. Conclusion: Annual reviews of patients with epilepsy can be carried out by practice nurses, but some training is required. The review process is likely to increase the burden on secondary care and have a significant adverse effect on neurology waiting times.
Key words: Epilepsy, annual review, epilepsy nurse
Introduction
Epilepsy has a prevalence of approximately 0.7% of the adult population,1 but some data suggest that the prevalence may be higher in some areas of Scotland.2 The new GMS contract for general practitioners rewards practices who deliver annual review of patients with epilepsy. The SIGN and NICE guidelines on the diagnosis and management of epilepsy3,4 also recommend annual review of patients with epilepsy, to ensure that problems such as drug side effects, sub-optimal control of seizures, the need for continuing treatment and other issues can be addressed. Recent studies have shown that up to 25% of patients thought to have epilepsy do not in fact have the disease,5,6 so reviews also need to address the issue of the security of the diagnosis.
While the necessity for regular reviews is widely accepted, there is little information in the literature that can tell us how and by whom yearly reviews should be carried out. In addition, yearly reviews may identify the need for specialist review of the diagnosis, for counselling, for investigation or other types of intervention that may have implications for secondary care. The resource implications of this are also unknown.
We have therefore carried out a study aimed at:
a. determining whether a checklist for annual review of patients with epilepsy can enable a practice nurse to accurately identify patients with problems, and
b. estimating the resource implication of annual review of patients with epilepsy.
The study falls into two parts. Firstly, we carried out a study based in a single practice, with the aim of evaluating a checklist approach to annual review carried out by a practice nurse, and of estimating the resource implication of annual review. Secondly, we carried out an LHCC-wide audit of epilepsy care in the local healthcare cooperative (LHCC) the practice belonged to, to assess the number and type of problem that might be identified by annual review in a larger population.
Patients and methods
Evaluation of practice nurse/checklist review This part of the study involved a single general practice in the Southwest Glasgow LHCC with a list of 6240 patients. Patients with epilepsy were identified by a combination of computer diagnostic search (epilepsy, seizures) and a search for anti epileptic drug (AED) prescriptions, followed by examination of case records to ensure that AED prescriptions were for epilepsy, and to exclude patients with seizures due to alcohol or drug withdrawal. This process identified 62 patients over the age of 16 with a diagnosis of epilepsy, giving a point prevalence of epilepsy for the practice of 1.0%.
The first draft of the checklist was prepared by agreement among the neurologist specialising in epilepsy, the general practitioner and the practice nurse, using the SIGN guidelines. The checklist was divided into four domains:
• Diagnosis and Classification
• Investigation
• Treatment
• Information and Advice
The checklist took the form of a simple list of factors that should signal a requirement for an action on the part of the nurse or the general practitioner. It was written with the aim that it should be usable by a practice nurse with a basic knowledge of epilepsy, and with access to standard information material (e.g. Epilepsy Scotland information sheets). (Table I)
The patients were randomised into two groups (1 and 2). Each group was in turn randomised to be seen first by the neurologist or first by the nurse. The neurologist provided a standard clinical assessment of the epilepsy, while the nurse followed the checklist. The doctor and the nurse both produced a problem list at the time of each consultation. After all the patients in the first group had been seen, the problem lists generated by the nurse and by the neurologist were compared for each patient. Discrepancies between the two lists were identified. With the list of discrepancies in mind, the checklist was modified for the second group. The patients in group 2 were then seen in the same way, and problem lists again compared. This then fed into producing the final version of the checklist.
At the time of each consultation, the neurologist recorded requests for tests, referrals, follow up appointments etc that were indicated by the problems that had been identified.
The LHCC-wide audit
At the time of the study, the Southwest Glasgow LHCC had a population of 96,565 patients, from which 1259 patients with epilepsy were identified using the same criteria and methods as above. This gave a point prevalence for the LHCC of 1.30%. The audit was of what was documented in the practice notes, and patients themselves were not interviewed. The categories audited were: who made the diagnosis, whether the patient had evidence of having had cerebral imaging, seizure control, number and type of AED and evidence of information and advice in a number of domains. (Table II) Differences between groups were analysed as categorical variables using 2X2 Chi Squared tests, with Fisher’s exact test being used when expected frequencies fell below 5.
Results
Evaluation of practice nurse/checklist review
Attendance at appointments
Four patients were known to be under review at an epilepsy clinic. The remaining 58 patients each had an appointment to see the neurologist and an appointment to see the nurse. Eight patients did not attend because they were in residential care, had moved house or had died. Nurse appointments were slightly better attended than neurologist appointments (38% vs 29.3%: p=ns). Thirteen patients (22.4%) did not attend either appointment, and only 21 patients (33.8%) attended both.
Comparison of epilepsy specialist and checklist based practice nurse assessments
The fact that only 21 patients attended both appointments meant that there was only limited scope for the comparison of the abilities of the nurse and the epilepsy specialist to identify potential problems. In these 21 patients, there were 18 discrepancies between nurse and doctor problem lists, out of a possible 84 (four domains (diagnosis and classification, investigation, treatment, information and advice, see methods), for each of the 21 patients). Eleven of these 18 discrepancies related to information and advice issues. The others related to deficiencies in diagnosis and classification (three patients - diagnosis or classification not well documented), investigation (one patient – imaging not carried out) and treatment (three patients – in two patient seizure free for many years, requiring consideration of withdrawal, in one case patient had drug side effects).
Only three of the discrepancies, all relating to information and advice, occurred in the second visit. This suggests a trend toward nurse improvement between visits, though difference between the visits did not reach statistical significance (p=0.057, Fisher’s exact test).
Changes made in the checklist after the first group
On discussing the reasons for the discrepancies between nurse and doctor review in the first visit, it became clear that most were the result of minor issues relating to lack of experience on the part of the nurse. Only detail changes in wording were made to the checklist.
Problems identified
Of the 45 patients that attended one or other appointment, no problems were identified in 11(24%) (rated as No Problems at both appointments, or did not attend one and rated as No Problems at the other). In the remaining 34 patients(89%), 47 problems were identified. These consisted of 19 deficiencies in information and advice (40%), seven diagnostic problems(14%), 14 treatment problems and seven deficiencies in investigations(15%). The diagnostic problems in most cases concerned lack of information on the original diagnosis. All were picked up by the first item in the checklist (diagnosis made by non epilepsy specialist). In one patient, the original diagnosis was thought to be erroneous (the patient probably had vasovagal syncope), but the patient was off treatment and attack free long term. In three other patients, the diagnosis of pseudoseizures was questioned. Treatment problems consisted of AED side effects in one case, and incomplete or poor seizure control in the rest. The investigational deficiencies all concerned lack of structural brain imaging.
Appointments and investigations generated by reviews
The review process generated 19 nurse appointments for information and advice. Twelve patients required further follow up appointments by the neurologist at the epilepsy clinic, for further diagnostic work or for changes in drug therapy. Seven patients required brain imaging, one patient required measurement of plasma phenytoin levels and three patients required video EEG.
The non-attenders
Further information was available for eight of the 13 nonattenders. Five were known to be seizure free long term, one of whom was not on AED. One had a known alcohol problem and had a history of defaulting general practitioner appointments. One was thought to have psychogenic nonepileptic seizures (pseudoseizures).
The LHCC-wide audit
Of the 1259 patients with a diagnosis of epilepsy, 71.4% had been seen by an epilepsy specialist at some time. 59.3% had evidence of cerebral imaging. Only 37.4% were documented as being seizure free. Only a minority of patients (9-23%) had documented information on driving and the other domains that were audited.
Discussion
Practices have a number of options for dealing with the issue of annual review of patients with epilepsy. They can consider the option of simply referring all their patients to secondary care. Given the number of patients concerned and the length of existing neurology waiting times, it seems unlikely that this option will be tenable. Carrying out reviews within primary care has the potential to reduce the number of patients referred on, so that problems identified can be dealt with in a realistic time frame.
The present study shows that a checklist approach in the hands of a practice nurse can identify deficiencies in the diagnosis and treatment of patients with epilepsy. Since the checklist changed only minimally between the first and second groups, it seems likely that the improved agreement between doctor and practice nurse was due to an educational effect (experience plus the effect of discussion after the first group). This suggests that practice nurses should have some training before applying the checklist.
Depending on what services are available within the LHCC, the checklist could be extended to give some guide as to what action to take once a problem has been identified (i.e. developed into a protocol). Not all problems necessarily require a referral. For example, if a patient is on Phenobarbitone and is free of seizures and side effects then this could normally be dealt with by a brief discussion with an epilepsy nurse, as could most problems relating to information and advice. If there was a willingness to allow some practise nurses to acquire a degree of specialisation (the possibility of full time epilepsy nurses based in LHCCs might also be considered), then potentially a substantial proportion of the problems identified by annual review could be dealt with at primary care level. In some areas of England, general practitioners have trained to be able to have a special interest in epilepsy, and have contributed usefully to this process.
The lower non-attender rate seen for the nurse suggests that a practice-based review clinic might expect more attenders than a hospital clinic. Assuming a non-attender rate of 30% and extrapolating from the practice population of 6240, our data suggest that for every 100,000 practice patients, the review process might identify approximately 1000 patients with epilepsy, of whom up to 700 might attend for review. The process would generate 304 nurse appointments for information and advice, 192 requests for review in secondary care, 48 video EEG slots (average duration five days), 112 requests for cerebral imaging and 16 requests for estimation of plasma anticonvulsant levels. These figures can only be compared to those of the LHCC audit to a limited degree, the main difference being that the audit was of patient documentation rather than of the patients themselves. Additionally, the practice in which the evaluation of practice nurse/checklist review took place had a known epilepsy interest hence, for example, the high (89.9%) proportion of patients who had been seen by an epilepsy specialist.*
However, it is worth considering the number of patients in the LHCC who had not been given their diagnosis by an epilepsy specialist, at 28.6%. Assuming a 30% non-attender rate, this item alone would generate 252 requests for specialist review from a population of 100,000. A checklist allowing non referral of patients in whom there was a clear documentation of seizure descriptions might mitigate this, but it seems unlikely that it would make a major difference. Additionally, Glasgow is somewhat atypical in being relatively well provided with epilepsy specialists: in other areas the review process might well throw up a much higher proportion of patients in whom the diagnosis could not be regarded as secure.
The relatively high proportion of patients on monotherapy may also be due to this factor. It was disappointing that such a small proportion of patients was seizure free. In the LHCC-wide audit, we did not record how many patients were under active review by secondary care but only 4/62 patients(6.5%) from the evaluation of practice nurse/ checklist review study were being seen in secondary care. Numerically speaking, the biggest problem identified by the LHCC audit was the lack of documented patient information, even on such basic and medico-legally sensitive items as DVLA regulations. Some patients might well nonetheless have information on these subjects, which might account for the lower prevalence of lack of information in the checklist development study.
The number of patients identified as having epilepsy both in the practice and in the LHCC was high with respect to known epidemiological data, despite our having excluded the paediatric population and seizures related to alcohol or drug withdrawal. Alcohol was identified as a current problem in only one out of the 45 patients who attended the checklist development study (though it may have been overrepresented in the non-attenders or not admitted or known in some cases).
It is unlikely that our prevalence was artificially overestimated by an excess of patients with other disorders such as psychogenic non epileptic seizures (PNES, pseudoseizures) that had misdiagnosed as epilepsy, given the unusually high proportion who had been diagnosed by epilepsy specialists – the studies of Smith et al and Scheepers et al5,6 provided evidence that specialist diagnosis readily identifies this type of misdiagnosis.
A previous audit of epilepsy in Glasgow2 using a different sample of 145,609 patients found a slightly high prevalence at 0.72%, while an audit of epilepsy in Ayrshire and Arran (a mixed urban and rural population) found 0.69% (unpublished data). A higher prevalence of epilepsy is associated with markers of deprivation7 and recent evidence suggests that the incidence of epilepsy may be up to twice as high in socio-economically deprived populations.8 It has been suggested that this association is because patients with epilepsy are more likely to have jobs not requiring qualifications or to be unemployed,9,10 though that mechanism would not explain higher incidence. We did not in any event collect socio-economic data on our patients.
Our study is too small to conclude that published epidemiological data underestimates epilepsy prevalence widely in the UK, but it suggests that generic data may not always provide a good basis on which to plan service developments, and that more work is needed in this area.
Overall, the results of this study suggest that the initial round of annual reviews of patients with epilepsy will identify a worryingly large number of patients with problems. The review process, and some of the work generated by it could be dealt with by appropriately trained nurses in primary care. However, the burden for secondary care, already underprovided with specialist epilepsy services, will be considerable and waiting times will lengthen significantly if no resources are provided.
REFERENCES
1 Hauser WA, Annegers JT, Kurland LT. Prevalence of epilepsy in Rochester Minnesota 1935-1984 Epilepsia 1991: 34;453-468
2 Muir TM, Bradley SF, Wood SF, Murray GD, Brodie MJ. An audit of treated epilepsy in Glasgow Seizure, 1996;5:41-46
3 Diagnosis and management of epilepsy in adults. SIGN Guideline No 70, revised version, SIGN, Royal College of Physicians of Edinburgh, April 2003
4 The epilepsies: diagnosis and management of the epilepsies in adults in primary and secondary care. National Institute for Clinical Excellence, Guideline 20, October 2004
5 Smith D, Defalla BA, Chadwick DW. The misdiagnosis of epilepsy and the management of refractory epilepsy in a specialist clinic Q J Med 1999;92:15- 23
6 Scheepers B, Clough P, Pickles C. The misdiagnosis of epilepsy: findings of a population study Seizure 1998;7:403-406
7 Morgan CLI, Ahmed Z, Kerr MP. Social deprivation and prevalence of epilepsy associated health usage J Neurol Neurosurg Psychiat 2000;69:13-17
8 Heaney DC, MacDonald BK, Everitt A, Stevenson S, Leonardi GS, Wilkinson P, Sander JW. Socioeconomic variation in incidence of epilepsy: prospective community based study in south east England Br Med J 2002;325:1013-1016
9 Callaghan N, Crowley M, Giggin T. Epilepsy and employment, mental, educational and social status Ir Med J 1992;85:17-19
10 Elwes RD, Marshall J, Beattie A et al. Epilepsy and employment. A community based survey in an area of high unemployment J Neurol Neurosurg Psychiat 1991;54:200-203