Monitoring patients on high-dose antipsychotics

Aimsand method Followingthe publicationof the 1994 Consensus Statement on the use of high-dose antipsychotic medication, we identified our high-dose patients and undertook an audit of the recommended physical investigations. The patients were identified by scrutiny of prescription records. Data were collected retrospectively at six-monthly intervals for four audit cycles. Results were fed back to clinicians at the hospital journal club. Results The percentage of patients identified in the high-dose category fell from 35 to 23%over 18months. Electrocardiograph monitoring of the group Increased from 5 to 63%.Other testsshowed a similarimprovement. A very small number of abnormal resultswas shown. Therewere no untoward cardiac events. Clinical implications Introductionof the standardsof physical monitoring advised in the ConsensusStatement has implications of cost to the NHS to be balanced against risk avoidance for patients. The number of abnormalities detected in our population was low. Changing prescribing behaviour in response to reported abnormal findings proved unexpectedly slow and a new systemwas required.

In 1994, the Royal College of Psychiatrists produced a consensus statement on the use of high-dose antipsychotic medication (Thompson, 1994).This document described concerns that some patients were given medication at doses in excess of the British National Formulary upper limits and that some sudden deaths might be due to medication.When discussing unexpected sudden cardiac deaths, it was acknowledged that the link with neuroleptic medication was suspected but as yet unproven.Mechanisms by which neuroleptics might be implicated in sudden cardiac deaths included cardiac con duction abnormalities and a negative inotropic effect.Physical monitoring of patients on highdose antipsychotics was suggested as a means of safeguarding them from the risks of high-dose treatment.
Using the guidelines, attempts at audit in this area (Cornwall et ai 1996, Krasucki & McFarlane, 1996) have revealed a substandard level of mon itoring at initial survey and logistical difficulties in obtaining some tests on remote sites.Reassuringly, audit can bring about improvements in clinical practice, physical monitoring of patients and a reduction in the proportion of patients on high-dose medication.
The State Hospital, Carstairs, provides inpatient care in a medium and maximum secure setting for the forensic psychiatric population of Scotland and Northern Ireland.Thomson & Bogue (1994)

The study
Definitions of high-dose antipsychotics and a standard for physical monitoring against which to audit were based on the Consensus Statement (Thompson, 1994)  Conversion to equivalent doses of chlorproma/ine was performed using standard tables (Bazire, 1996).This method was used on each occasion that data were examined during the project.

Psychiatric Bulletin (1998). 22, 671-674
The initial survey took place in May 1995.Patients identified as being on high-dose antipsychotics had their case notes examined retrospectively for evidence of physical monitor ing in the preceding three months.The results were presented to the medical staff at a journal club meeting and it was subsequently decided to implement the guidelines in respect of patients who were prescribed high-dose antipsychotics.The survey was repeated using the same method in December 1995.
By the time of the third survey in June 1996.a formalised system of patient identification and monitoring had evolved.The pharmacist labelled the drug charts of patients on high-dose anti psychotics, alerting prescribers and nursing staff to the patient's status.The pharmacist also supplied a list of the names of these patients to the medical centre coordinator so that appro priate monitoring could be arranged.A letter of explanation to patients was introduced in an attempt to reduce the refusal rate, particularly for ECGs.The survey method was altered at this point so that the medical centre records of monitoring measures were examined rather than patients' case notes.The final survey, of the results from December 1996.took place in March 1997.

Findings
Men out-number women by approximately 10 to one (see Table 1).Three-quarters of the patients were prescribed regular antipsychotic medi-Table 1.General results of the four surveys cation.The proportion of patients on high-dose antipsychotics, expressed as a percentage of the total number of patients on medication, de creased during the time of the audit.On average, females were prescribed larger doses of medi cation (see Table 2).Average daily doses fell progressively from May 1995 to December 1996, with overall a 20% reduction.
The level of physical monitoring (see Table 3) improved from June 1996 onwards, coinciding with the introduction of a system of identifying patients.The improvement continued into the December 1996 survey with the exception of measurement of serum urea and electrolytes, which fell, but was still better than the level at initial survey.
Electrocardiograph abnormalities which were detected (see Table 4) were concerned with conduction and evidence of prolonged Q-T(.was found.As the level of ECG monitoring improved, the percentge of abnormalities found increased.

Comment Dosage
The results obtained for daily dosage potentially underestimate the dose a patient receives.Only routine medication was considered.Patients could become 'high-dose' through use of discre tionary medication and would not be counted by this method.This group of patients, on high-dose medication for brief periods during deterioration    in their mental state or behaviour, seems to be a group targeted by the Consensus Statement.
Modification of the flagging system may be needed so that prescribers are alerted to poten tial high-dose patients.
The proportion of patients on high-dose antipsychotics remained the same in the first two surveys but began to fall by June 1996.Over the period of the audit cycles, the percentage fell from just over one-third to under a quarter.Even the latter figure is a high proportion of the patients on antipsychotics, and merits further investigation.There is relatively little written about prescribing practices in forensic settings, but it is a widely held perception that large doses of neuroleptics are used for longer periods.
Fraser & Hepple (1992), who surveyed pre scribing habits in Broadmoor and compared their data with previous surveys in Oxford, Birmingham and Malvern, found that the pre scribing habits of doctors in Broadmoor did not differ markedly from those in other psychiatric hospitals serving similar patient groups.Thirtyeight per cent of their male patients were prescribed total doses in excess of 1000 mg per day, a finding comparable with our initial results.
Increasingly, evidence is emerging that treat ment-resistant patients do not benefit from ever larger doses of neuroleptics.This study attempted to examine the monitoring of the physical rather than the clinical state of patients on high-dose antipsychotics.It did not look at the response of patients to high-dose treatment.Our results would suggest that the clinical justification for prolonged high-dose prescribing should be further investigated, a question merit ing a study in its own right.
The prescribing survey showed a significant use of new antipsychotics, the safety of which has been more rigorously tested in clinical trials.The survey revealed a tendency to introduce new antipsychotics while continuing previous antipsychotic medication (with the exception of clozapine).This combination of antipsychotics may not only increase the risk of complications but also not maximise benefit from using drugs with a lower incidence of extrapyramidal sideeffects.These results perhaps indicate a need to survey the prescribers involved to clarify the reasoning behind such practice.

Physical monitoring
The physical monitoring of patients on high-dose antipsychotics was minimal in the surveys of May and December 1995.Patients are routinely offered an annual physical examination, which some refuse.The bulk of the monitoring recorded in the first and second surveys was attributable to this examination rather than to a perceived need for monitoring because of a patient's highdose status.
By June 1996.the more formalised system showed an increase in the number of patients receiving some form of physical monitoring.The biggest improvements appeared in the area of ECG recordings.By this stage it was also possible to examine the reasons why monitoring had not taken place.Problems such as the unavailability of an ECG technician, patient refusal and recent increases in medication featured as the most common reasons for failure to obtain an ECG within the target time period.

Abnormalities detected
Blood tests detected minor electrolyte abnormal ities, many of which had been detected in previous tests.ECG tracings revealed Q-Tr prolongation (440 msec or more) in some cases.All of these patients were on combinations of two or more antipsychotics.Follow-up of the four patients with prolonged Q-Tr in the survey of June 1996 showed that two had their dosage reduced through monotherapy with clozapine.One had been discharged on high-dose anti psychotics and one had a further increase in medication and refused monitoring in the three months prior to December 1996.Of the nine patients with Q-TCprolongation in the survey of December 1996, three had their medication reduced, while six were unchanged -had the Consensus Statement been fully adhered to, the dosage would have been reduced.This may imply a need to notify prescribers more effectively about abnormalities.It remains to be seen how commonplace Q-TC, prolongation is in a population on high-dose medication and also if this abnormality progresses with time.Con tinuation of the monitoring process will provide information on both questions.
In this study to date, no adverse cardiac events have been observed in patients on high-dose antipsychotics.Just over 100 patients have been identified as being on high-dose antipsychotics at some point during the two years of the study.Eleven of these appear to have been on high-dose antipsychotics throughout the two-year period.It may be that patients stabilised on high-dose antipsychotics do not need monitoring as fre quently as patients on high doses do acutely, but until there is sufficient evidence to this effect, a case cannot be made for change.

Other implications
The standard for physical monitoring included measurement of urea and electrolytes, as the Consensus Statement argued that electrolyte disturbance made a patient more likely to develop a cardiac arrhythmia and increased the risk of sudden death.As antipsychotics can affect the liver and have been associated with blood dyscrasias, future monitoring could be extended to include three-monthly full blood count and liver function tests.At present all patients have an annual physical examination including these tests.It is questionable whether the added costs involved would be justified.
As ECG and blood monitoring have become more commonplace for high-dose patients, pro blems of increased workload and financial cost have arisen.The State Hospital is located in a rural area, routine blood samples are sent for analysis to a hospital 18 miles away, while the ECG service is provided by another hospital nine miles away, sending a technician on an as required basis.With the introduction of regular ECG monitoring, the number of requests in creased 20-fold and the availability of the technician declined owing to departmental staff ing problems in the general hospital.Delays occurred in the receipt of reports.Problems also exist with the organisation of blood sampling and the acceptability of all tests to patients.This had led to a search for a more efficient service.Options considered include the training of existing staff in phlebotomy and ECG tracing and the use of a fax or modem link to facilitate more rapid ECG reporting.

Conclusion
Introducing audit of the physical monitoring of patients on high-dose antipsychotics adopting the standards suggested in the Consensus Statement and the development of a system of patient identification, has increased the amount of monitoring taking place.
It may eventually provide information on the physical effects of long-term use of high-dose antipsychotic.It has also raised questions of how to continue the monitoring process in the most cost-effective manner possible, how best to convey information to responsible medical officer so that results can be acted upon, and how to reduce the level of refusal of tests.Having found that ECGs with Q-TC prolongation were not always acted upon timeously, we have now devised a method of flagging up these results to responsible medical officers.This will be audited in the next cycle.
Changing prescribing habits is well known to be difficult (Clark & Holden, 1987).Clinical audit is widely supposed to be effective, although scientific evidence, including controlled trials, for the effectiveness of what is now national policy remains relatively sparse.We have found that introducing changes in prescribing for individual patients in response to feedback has perhaps been slower than anticipated and suggests the need to heighten the impact of the findings.However, the general trend towards lower doses of neuroleptic medication over the period of the survey has been maintained.There may be several reasons for this, including the possibility that increasing awareness of a patient's highdose status leads to more cautious prescribing.A causal link between our audit results and any change in practice cannot be proven from the methodology adopted and other sources of in formation may have had an impact on prescribing over the period of the audit.We did not examine the reasons for the trend and we hope to amend the data collection to include this in future.

Table 3 .
Physical monitoring of high-dose patients (number of recordings)