by the bill would have been able to recoup money paid out to victims of accidents attributable to third parties-for example, the drug companies. It is impossible to put an exact figure on the cost of a no fault scheme because one cannot cost the present system, but savings in legal fees and legal aid payments would be substantial. I can scarcely believe the assertion by Mr Capstick and colleagues that a no fault scheme would not ensure proper accountability for the medical profession. I would not have let a bill that did not deal adequately with this issue go forward in my name. The bill allowed the Medical Injury Compensation Board to investigate claims fully, give explanations, and where appropriate seek apologies and refer matters to other authorities, including disciplinary bodies. Lessons would have been learnt and accountability ensured. Mr Capstick and colleagues are also somewhat critical of my decision not to prevent claimants from choosing the current legal process if they desire. Under my bill, those dissatisfied with the board's offer could have refused it and pursued a claim in the courts. The board's offer would then have fallen. My view is that most claimants would rather pursue a claim with the board than entrust themselves to the vagaries of the legal process. One of the major deterrents to pursuing a tort action in the courts is the time it takes to complete the process. I am sure that most people would rather receive "less" now than take the risk of "more" later. As the government has implicitly accepted the concept of no fault compensation in their award to those haemophiliacs who contracted HIV from infected factor VIII there is no good reason not to extend the principle to all those suffering as a result of medical accidents. If £4bn can be found to save the government's face over the fiasco of the poll tax then the estimated £1 OOm needed to establish a no fault compensation scheme can, I am sure, be found immediately-financed perhaps by an increase in VAT? ROSIE BARNES House of Commons, London SW IA OAA
1 Capstick B, Edwards P, Mason 1). F^or debate: compensatiorn for medical accidents. BMJ 1991;302:230-2. (26 January.)
Re: VAMP revamp SIR,-Ms Linda Beecham's article on the crisis in VAMP' reminds me that the history of computing is littered with company failures and subsequently unsupported systems. The problem is that the failed company usually retains the software source code and documentation, thus rendering the system almost unsupportable even by a computer expert. A resource is available, however, that would overcome many of the traumas of software companies going out of business: an escrow agreement. None of the general practitioner computing companies seem to offer their customers the facility of placing their software in escrow; neither have any of the guides to buying a general practitioner computer system that I have seen mentioned this facility. Briefly, in an escrow agreement the supplier deposits its program's source code with a third partv, such as a solicitor or the National Computing Centre. The supplier contracts to supply the third party with all software updates and documentation.
If a supplier stops trading the escrow agreement is enforceable and the users who have subscribed to it have access to the source code and documentation to support and maintain their systems. Suppliers retain the copyright of their software while they are trading. Both users and suppliers normally contribute towards the costs of escrow.
912
In the case of general practitioner computer systems, the user group would seem ideally placed to negotiate an escrow agreement with the supplier. Should the supplier cease to trade, the user group could hire computer experts to support and maintain general practitioners' systems. I have only once had to enforce an escrow agreement when a system supplier ceased trading; the transition was not painless, but it enabled us to maintain the software until a suitable successor system was available. JOHN L BERRY Cheadle, Cheshire SK8 5HJ 1 Beecham L. Re: VAMP revamp.
March.)
BMJ7
1991;302:489-90. (2
SIR,-That practices "receive the basic computer system and £500 a month on average in return for supplying anonymised morbidity and prescription information"' is a common misconception, probably because the scheme was originally called a "no cost option." In reality, practices either buy the computer system or lease it from a leasing company independent of VAMP, paying the standard purchase and maintenance costs. In a separate agreement they then sell anonymised data for VAMP in return for a regular monthly payment. This payment is not for the computer, it is in respect of the increased work incurred in collecting data of a quality that is useful for morbidity analysis and research. It is important to get this matter straight for three reasons. Firstly, it is not the hardware and software development side of VAMP that was in financial trouble but the data payments side of the company-largely because the level of data payments had been extremely generous in the early stages. Secondly, it should be clear that, because the payment is for the work involved in producing data and not for the computers, practices on the research panel should not be excluded from direct reimbursement for computer purchase and maintenance costs. Thirdly, the scheme shows the value that the company has put on high quality data. For these routinely collected data to be useful they have to have a high level of completeness and accuracy, and this requires record keeping by the participating practices far above that normally required for the provision of general medical services. The profession is fortunate that independent companies have taken the initiative to build up such a research database, which is currently being validated; it might otherwise never have been realised. Although general practitioners will always prefer large fees with no risk, it may well be that the new profit sharing scheme will provide even greater incentives to good record keeping than the original one. M S LAWRENCE
West Steet Surgery. Chipping Norton, Oxfordshire OX7 5AA 1 Beecham L. Re: VAMP revamp. B.M7 1991;302:489-90. (2 March. )
Large computer databases SIR,-The recent editorial entitled "Large computer databases in general practice" is highly misleading as it relates to the VAMP database.' The VAMP validation study we reported in the BMJ was begun in mid- 1989, at a time when only a small fraction of practices enrolled in the VAMP research program had completed the 12 months of training for proper recording of clinical data.2 As noted in our paper, most practices entered the plan
between July 1989 and June 1990 and had not completed their training when the study was initiated. Thus the reported results related only to the practices enrolled earliest in the program. Currently over 500 VAMP practices encompassing some three riillion patients are considered to be up to standard, and our most recent studies indicate that the quality of the information available from practices which have come up to standard since the time of our validation study have continued to show a high standard of data quality. It is most important to understand that practices which come up to standard differ in principle from practices which do not only in regard to how thoroughly they record relevant data, not in terms of the effects of the drugs they prescribe. In studies of drug safety based on computer data it is critical that relevant data items be recorded routinely. The a priori standardised exclusion of data of inferior quality in no way influences the results of studies that are based on data of high quality. The suggestion that drug safety studies should be based on inadequately collected data as well as properly collected data simply to achieve the purpose of including a representative sample of doctors is a nonsense. It is simply not correct to state that drug safety studies should include a representative sample of doctors because such studies evaluate how drugs behave, not how doctors behave.3 We have no experience with the AAH Meditel data resource, but there is no a priori reason to assume that it is of the same quality as the VAMP data resource simply because they are both derived from general practice. The editorial concludes, based in part on our report, that "Early hopes for large databases have not been fulfilled" and that the VAMP data are of "poor quality";. but we conclude, based on our considerable firsthand experience with the VAMP data resource, that reasonable hopes for its utility have been surpassed by the high quality and size of the data available for research. HERSHEL JICK SUSAN S JICK LAURA E DERBY Boston Collaborative Drug Surveillance Program, Boston University Medical Center, Lexington, Massachusetts 02173-5207, United States 1 Pringle M, Hobbs R. Large computer databases in general practice. BMJ7 1991;302:741-2. (30 March.) 2 Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in United Kingdom. BMJ7 1991;302:766-8. (30 March.) 3 Jick H, Vessey MP. Case-control studies in the evaluation of drug-induced illness. Am7 Epidemiol 1978;107:1-7.
Improving outpatient services SIR,-As reported, the National Audit Office recently published a review of NHS outpatient services, which was then the subject of discussion by the parliamentary Public Accounts Committee on 27 February.' In our review we examined the management of outpatient services in a sample of 10 hospitals. These hospitals also undertook surveys of a number of clinics for us, recording among other criteria the staffing levels, patients' waiting time in clinic, and the incidence of missing patient records. We found that many patients had to wait a long time after their arrival at the clinic before they were seen by the doctor-in 53% of clinics patients' average waiting time was less than 30 minutes, but in 40% it was between 30 minutes and an hour, and in 7% it exceeded an hour. We were therefore interested to read Dr M Jennings's account of his introduction of a new appointments system and its beneficial impact on patients' waiting times in clinics.2 Realistic appointment arrangements are clearly a great step towards achieving improvements. Studies like ours cannot prescribe ideal
BMJ VOLUME 302
13 APRIL 1991
blueprints for individual clinics because clinic arrangements must reflect the nature of the work carried out at that clinic. But we found signs in some clinics that practices had not been realistically planned: clinics did not start on time; patients were seen in order of arrival rather than in appointment order, thus encouraging patients to arrive early and increasing their waiting time; appointments were not evenly spread throughout the clinic period; the number of patients attending a clinic varied significantly from one week to the next owing to patients not attending for their appointment and the clinic being overbooked to compensate; overbooking of clinics meant that additional patients would not have had individual appointment times. It was clear from our survey that some factors which contributed to long waiting times (such as doctors being called away from the clinic) may have 'reflected emergencies elsewhere in the hospital or staffing problems and may have been unavoidable. The points listed above were avoidable, however, and action to deal with them would help to reduce patients' waiting time in clinic. We were encouraged to find that some of the hospitals visited were beginning to do this. JILL GOLDSMITH National Audit Office, London SWI'W;9SP I Smith J. Improv-ing outpatient services. BAR7 1991;302:435. (23 i'chruarv.' 2 Jettnittgs A. Audit of a new appointments system in a hospital ouitpatient clinic. BMU 1991;302:148-9. 19 January.
bill for erythropoietin has been halved. Assay of erythropoietin before treatment can indicate patients' requirements and can largely predict dosage and subsequent response. Optimum therapeutic schedules and avoidance of overdosage are readily ascertained, and patient non-compliance can be highlighted. Modern immunoassays for serum erythropoietin cost a fraction of the weekly cost of the drug to a single patient, and the establishment of the assay service will repay the outlay within a month. This assay is, however, not generally available in NHS laboratories, in spite of the clear need for this service. We hope that erythropoietin assays will soon be available on a regional or supraregional basis, to the benefit of all concerned -not least the Exchequer. ANDREW HILLIS JOANNE MARSDEN Dulwich South Hospital, London
TIMOTHY j PETERS ROY SHERWOOD King's College Hospital, London SE5 9RS I Correspondence. Picking up the tab for ervthropoietin. BMJ 1991;302:407-8,592. (16 February, 9 March.) 2 Gabriel R. Picking up the tab for ervthropoietin. BMJ 1991;302:
248-9. (2 February.) 3 Taylor JE, Henderson IS, Mactier RA, Stewart WK. Effects of withdrawing erythropoietin. BMJ 1991;302:272-3. 2 February.)
secondary prevention of myocardial infarction has been shown only for U adrenergic blockers. Our results suggest that the results of recent clinical trials may have had some influence in increasing the prescription rate of [i adrenergic blockers in these patients. They also suggest that this influence may have been limited, at least in part, by the strong promotion of calcium channel agonists for the secondary prevention of myocardial infarction, an indication for which their efficacy has not been proved.' J M ARNAU A AGUSTi X VIDAL J R LAPORTE
Servei de Farmacologia Clinica, C S Vall D'Hiebron, Unitat de Farmacologia, Universitat Autonoma de Barcelona, 08035-Barcelona, Spain 1 Eccles M, Bradshaw C. Use of secondary prophylaxis against myocardial infarction in the north of England. BMJ 1991;302: 91-2. (12 January.) 2 Agusti A, Arnau JM, Laporte JR, Soler J, Vidal X. Secondary prevention of myocardial infarction: the influence of clinical trials on cliriical practice. EurJ7 Clin Pharmnacol 1989;36(suppl): 253. 3 Held PH, Yusuf S, Furberg CD. Calcium channel blockers in acute myocardial infarction and unstable angina: an overview. BMJ 1989;299:1187-92.
Milk, butter, and heart disease SIR,-The National Audit Office review showed that patients still too often experience long waits in outpatient clinics.' In Princess Royal Hospital, I have established 'a new patient clinic at which patients often do not wait and occasionally see the doctor ahead of their booked time. This has been achieved by assigning appointments of varying lengths when the general practitioner's referral letter is initially reviewed. The computerised patient administration system used by Shropshire district, however, is unable to cope with a booking system in which patients variously need 5, 10, 15, or 20 minute appointments. In the initial stages clinics had to be booked by hand until we found ways of fiddling the computer. Staff in records administration need to bear this in miind when rewriting outpatient administration systems, which I gather they are now doing. Patients attending my follow up fracture clinics do, however, wait, and applying a timed system to these is much harder because patients are booked from several different places: three or four wards, new fracture clinics, previous old fracture clinics, etc. 'A computer program that could cope with appointments of different lengths would enable a timed system to be introduced tomorrow. P C MAY
Princess Royal Hospital, 'I'elford, Shropshiie TF6 6TF I Smith J. Improving outpatient services.
B&M7 1991;302:435. (23
Februarv.)
Picking up the tab for erythropoietin SIR,-There have been a number of papers, leading articles, and correspondence recently regarding the use, availability, and cost of erythropoietin in the treatment of anaemia in chronic renal failure." In the debate about the cost versus the benefits-of the treatment little attention has been paid to the assay of serum ervthropoietin in diagnosis and monitoring treatment. Since introducing this assay at King's, the drug
BMJ
VOLUME
302
13
APRIL
1991
Secondary prophylaxis against myocardial infarction SIR,-Drs Martin Eccles and Colin Bradshaw' do not mention other drugs for secondary prophylaxis against myocardial infarction that might account for the underuse of D adrenergic blockers. We analysed the use of acetylsalicylic acid and adrenergic blockers in 736 patients with myocardial infarction discharged from our centre between 1982 and 1988.2 During this period we saw an increase of the prescription rate of both drugs. In 1988 the prescription rate was 34% (42/122) for D adrenergic blockers, 33% (40/122) for acetylsalicylic acid, and 7% (9/122) for both. For [i adrenergic blockers our results are similar to those of Drs Eccles and Bradshaw. The participation of 13 hospitals in their study in the third international study of infarct survival (ISIS-3) may have induced a higher prescription ofacetylsalicylic acid at hospital discharge. Thirty four per cent (253/736) of our study population had relative contraindications to the use of li adrenergic blockers. We also analysed the prescription of other drugs. In 1988 the prescription rate of calcium channel antagonists amounted to 44% (54/122). They were prescribed more often for patients with previous cardiovascular disease, such as hypertension or angina (table). In addition, we observed that the use of f adrenergic blockers was lower in patients treated with calcium channel agonists. f3 Adrenergic blockers and calcium channel agonists have some common therapeutic uses, such as hypertension and angina, but efficacy for the
SIR,-Professor A G Shaper and colleagues' have referred to data on milk, spreading fats, and heart disease in our Caerphilly prospective heart disease study2 and have presented data on this same topic from their regional heart study. We can certainly confirm what Professor Shaper and colleagues show in their own data. Milk drinking is confounded with many other factors of relevance to heart disease risk, and men who use butter differ in many ways from those who use margarine. The effects of some of these differences on relations with heart disease are substaniial, and difficulties arise because their interrelations are complex. As we said in the progress report, the data are being explored further and we will publish a detailed report in due course. In retrospect, it was unwise for a brief description and preliminary analysis of these data to be included in a privately published report that attempted to summarise a wide range of work on cardiovascular disease. It was particularly unfortunate that the one small section on milk and fats was widely reported and interpreted uncritically in the popular press. I can only apologise for not having foreseen that this was likely to happen, and I now greatly regret not having withheld the data until a more adequate analysis could be submitted to the scientific press. PETER C ELWOOD MRC Epidemiological Unit (South Wales),
Llandough Hospital, I'enarth, Sotith Glamorgan CF6 IXX I Shaper AG, Wannamethee G, Walker M. Milk, butter, and heart disease. BUM7 1991;302:785-6. (30 March.) 2 MRC Epidemiology Unit. Eptdemiological studies of cardiovascular disease. Penarth, South Glamorgan: Llandough Hospital, 1991. (Progress report VII.)
Treatment with ji adrenergic blocking agents (r3AB) or with calcium channel antagonists (CaCA) at hospital discharge after myocardial infarction in 736 patients (49 patients treated with both drugs were excluded) No (%) of patients treated with CaCA No (¼/.) of patients treated with liAB Total With previous cardiovascular disease Without previous cardiovascular disease
154 (64-7) 46 (41 8)
84 (35-3) 64 (58-2)
238 110
Total
200 (57 5)
148 (42-5)
348
Z.= 16- 12 (p-- 0-0006). 913
1 Capstick B, Edwards P, Mason 1). F^or debate: compensatiorn for medical accidents. BMJ 1991;302:230-2. (26 January.)
Re: VAMP revamp SIR,-Ms Linda Beecham's article on the crisis in VAMP' reminds me that the history of computing is littered with company failures and subsequently unsupported systems. The problem is that the failed company usually retains the software source code and documentation, thus rendering the system almost unsupportable even by a computer expert. A resource is available, however, that would overcome many of the traumas of software companies going out of business: an escrow agreement. None of the general practitioner computing companies seem to offer their customers the facility of placing their software in escrow; neither have any of the guides to buying a general practitioner computer system that I have seen mentioned this facility. Briefly, in an escrow agreement the supplier deposits its program's source code with a third partv, such as a solicitor or the National Computing Centre. The supplier contracts to supply the third party with all software updates and documentation.
If a supplier stops trading the escrow agreement is enforceable and the users who have subscribed to it have access to the source code and documentation to support and maintain their systems. Suppliers retain the copyright of their software while they are trading. Both users and suppliers normally contribute towards the costs of escrow.
912
In the case of general practitioner computer systems, the user group would seem ideally placed to negotiate an escrow agreement with the supplier. Should the supplier cease to trade, the user group could hire computer experts to support and maintain general practitioners' systems. I have only once had to enforce an escrow agreement when a system supplier ceased trading; the transition was not painless, but it enabled us to maintain the software until a suitable successor system was available. JOHN L BERRY Cheadle, Cheshire SK8 5HJ 1 Beecham L. Re: VAMP revamp.
March.)
BMJ7
1991;302:489-90. (2
SIR,-That practices "receive the basic computer system and £500 a month on average in return for supplying anonymised morbidity and prescription information"' is a common misconception, probably because the scheme was originally called a "no cost option." In reality, practices either buy the computer system or lease it from a leasing company independent of VAMP, paying the standard purchase and maintenance costs. In a separate agreement they then sell anonymised data for VAMP in return for a regular monthly payment. This payment is not for the computer, it is in respect of the increased work incurred in collecting data of a quality that is useful for morbidity analysis and research. It is important to get this matter straight for three reasons. Firstly, it is not the hardware and software development side of VAMP that was in financial trouble but the data payments side of the company-largely because the level of data payments had been extremely generous in the early stages. Secondly, it should be clear that, because the payment is for the work involved in producing data and not for the computers, practices on the research panel should not be excluded from direct reimbursement for computer purchase and maintenance costs. Thirdly, the scheme shows the value that the company has put on high quality data. For these routinely collected data to be useful they have to have a high level of completeness and accuracy, and this requires record keeping by the participating practices far above that normally required for the provision of general medical services. The profession is fortunate that independent companies have taken the initiative to build up such a research database, which is currently being validated; it might otherwise never have been realised. Although general practitioners will always prefer large fees with no risk, it may well be that the new profit sharing scheme will provide even greater incentives to good record keeping than the original one. M S LAWRENCE
West Steet Surgery. Chipping Norton, Oxfordshire OX7 5AA 1 Beecham L. Re: VAMP revamp. B.M7 1991;302:489-90. (2 March. )
Large computer databases SIR,-The recent editorial entitled "Large computer databases in general practice" is highly misleading as it relates to the VAMP database.' The VAMP validation study we reported in the BMJ was begun in mid- 1989, at a time when only a small fraction of practices enrolled in the VAMP research program had completed the 12 months of training for proper recording of clinical data.2 As noted in our paper, most practices entered the plan
between July 1989 and June 1990 and had not completed their training when the study was initiated. Thus the reported results related only to the practices enrolled earliest in the program. Currently over 500 VAMP practices encompassing some three riillion patients are considered to be up to standard, and our most recent studies indicate that the quality of the information available from practices which have come up to standard since the time of our validation study have continued to show a high standard of data quality. It is most important to understand that practices which come up to standard differ in principle from practices which do not only in regard to how thoroughly they record relevant data, not in terms of the effects of the drugs they prescribe. In studies of drug safety based on computer data it is critical that relevant data items be recorded routinely. The a priori standardised exclusion of data of inferior quality in no way influences the results of studies that are based on data of high quality. The suggestion that drug safety studies should be based on inadequately collected data as well as properly collected data simply to achieve the purpose of including a representative sample of doctors is a nonsense. It is simply not correct to state that drug safety studies should include a representative sample of doctors because such studies evaluate how drugs behave, not how doctors behave.3 We have no experience with the AAH Meditel data resource, but there is no a priori reason to assume that it is of the same quality as the VAMP data resource simply because they are both derived from general practice. The editorial concludes, based in part on our report, that "Early hopes for large databases have not been fulfilled" and that the VAMP data are of "poor quality";. but we conclude, based on our considerable firsthand experience with the VAMP data resource, that reasonable hopes for its utility have been surpassed by the high quality and size of the data available for research. HERSHEL JICK SUSAN S JICK LAURA E DERBY Boston Collaborative Drug Surveillance Program, Boston University Medical Center, Lexington, Massachusetts 02173-5207, United States 1 Pringle M, Hobbs R. Large computer databases in general practice. BMJ7 1991;302:741-2. (30 March.) 2 Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in United Kingdom. BMJ7 1991;302:766-8. (30 March.) 3 Jick H, Vessey MP. Case-control studies in the evaluation of drug-induced illness. Am7 Epidemiol 1978;107:1-7.
Improving outpatient services SIR,-As reported, the National Audit Office recently published a review of NHS outpatient services, which was then the subject of discussion by the parliamentary Public Accounts Committee on 27 February.' In our review we examined the management of outpatient services in a sample of 10 hospitals. These hospitals also undertook surveys of a number of clinics for us, recording among other criteria the staffing levels, patients' waiting time in clinic, and the incidence of missing patient records. We found that many patients had to wait a long time after their arrival at the clinic before they were seen by the doctor-in 53% of clinics patients' average waiting time was less than 30 minutes, but in 40% it was between 30 minutes and an hour, and in 7% it exceeded an hour. We were therefore interested to read Dr M Jennings's account of his introduction of a new appointments system and its beneficial impact on patients' waiting times in clinics.2 Realistic appointment arrangements are clearly a great step towards achieving improvements. Studies like ours cannot prescribe ideal
BMJ VOLUME 302
13 APRIL 1991
blueprints for individual clinics because clinic arrangements must reflect the nature of the work carried out at that clinic. But we found signs in some clinics that practices had not been realistically planned: clinics did not start on time; patients were seen in order of arrival rather than in appointment order, thus encouraging patients to arrive early and increasing their waiting time; appointments were not evenly spread throughout the clinic period; the number of patients attending a clinic varied significantly from one week to the next owing to patients not attending for their appointment and the clinic being overbooked to compensate; overbooking of clinics meant that additional patients would not have had individual appointment times. It was clear from our survey that some factors which contributed to long waiting times (such as doctors being called away from the clinic) may have 'reflected emergencies elsewhere in the hospital or staffing problems and may have been unavoidable. The points listed above were avoidable, however, and action to deal with them would help to reduce patients' waiting time in clinic. We were encouraged to find that some of the hospitals visited were beginning to do this. JILL GOLDSMITH National Audit Office, London SWI'W;9SP I Smith J. Improv-ing outpatient services. BAR7 1991;302:435. (23 i'chruarv.' 2 Jettnittgs A. Audit of a new appointments system in a hospital ouitpatient clinic. BMU 1991;302:148-9. 19 January.
bill for erythropoietin has been halved. Assay of erythropoietin before treatment can indicate patients' requirements and can largely predict dosage and subsequent response. Optimum therapeutic schedules and avoidance of overdosage are readily ascertained, and patient non-compliance can be highlighted. Modern immunoassays for serum erythropoietin cost a fraction of the weekly cost of the drug to a single patient, and the establishment of the assay service will repay the outlay within a month. This assay is, however, not generally available in NHS laboratories, in spite of the clear need for this service. We hope that erythropoietin assays will soon be available on a regional or supraregional basis, to the benefit of all concerned -not least the Exchequer. ANDREW HILLIS JOANNE MARSDEN Dulwich South Hospital, London
TIMOTHY j PETERS ROY SHERWOOD King's College Hospital, London SE5 9RS I Correspondence. Picking up the tab for ervthropoietin. BMJ 1991;302:407-8,592. (16 February, 9 March.) 2 Gabriel R. Picking up the tab for ervthropoietin. BMJ 1991;302:
248-9. (2 February.) 3 Taylor JE, Henderson IS, Mactier RA, Stewart WK. Effects of withdrawing erythropoietin. BMJ 1991;302:272-3. 2 February.)
secondary prevention of myocardial infarction has been shown only for U adrenergic blockers. Our results suggest that the results of recent clinical trials may have had some influence in increasing the prescription rate of [i adrenergic blockers in these patients. They also suggest that this influence may have been limited, at least in part, by the strong promotion of calcium channel agonists for the secondary prevention of myocardial infarction, an indication for which their efficacy has not been proved.' J M ARNAU A AGUSTi X VIDAL J R LAPORTE
Servei de Farmacologia Clinica, C S Vall D'Hiebron, Unitat de Farmacologia, Universitat Autonoma de Barcelona, 08035-Barcelona, Spain 1 Eccles M, Bradshaw C. Use of secondary prophylaxis against myocardial infarction in the north of England. BMJ 1991;302: 91-2. (12 January.) 2 Agusti A, Arnau JM, Laporte JR, Soler J, Vidal X. Secondary prevention of myocardial infarction: the influence of clinical trials on cliriical practice. EurJ7 Clin Pharmnacol 1989;36(suppl): 253. 3 Held PH, Yusuf S, Furberg CD. Calcium channel blockers in acute myocardial infarction and unstable angina: an overview. BMJ 1989;299:1187-92.
Milk, butter, and heart disease SIR,-The National Audit Office review showed that patients still too often experience long waits in outpatient clinics.' In Princess Royal Hospital, I have established 'a new patient clinic at which patients often do not wait and occasionally see the doctor ahead of their booked time. This has been achieved by assigning appointments of varying lengths when the general practitioner's referral letter is initially reviewed. The computerised patient administration system used by Shropshire district, however, is unable to cope with a booking system in which patients variously need 5, 10, 15, or 20 minute appointments. In the initial stages clinics had to be booked by hand until we found ways of fiddling the computer. Staff in records administration need to bear this in miind when rewriting outpatient administration systems, which I gather they are now doing. Patients attending my follow up fracture clinics do, however, wait, and applying a timed system to these is much harder because patients are booked from several different places: three or four wards, new fracture clinics, previous old fracture clinics, etc. 'A computer program that could cope with appointments of different lengths would enable a timed system to be introduced tomorrow. P C MAY
Princess Royal Hospital, 'I'elford, Shropshiie TF6 6TF I Smith J. Improving outpatient services.
B&M7 1991;302:435. (23
Februarv.)
Picking up the tab for erythropoietin SIR,-There have been a number of papers, leading articles, and correspondence recently regarding the use, availability, and cost of erythropoietin in the treatment of anaemia in chronic renal failure." In the debate about the cost versus the benefits-of the treatment little attention has been paid to the assay of serum ervthropoietin in diagnosis and monitoring treatment. Since introducing this assay at King's, the drug
BMJ
VOLUME
302
13
APRIL
1991
Secondary prophylaxis against myocardial infarction SIR,-Drs Martin Eccles and Colin Bradshaw' do not mention other drugs for secondary prophylaxis against myocardial infarction that might account for the underuse of D adrenergic blockers. We analysed the use of acetylsalicylic acid and adrenergic blockers in 736 patients with myocardial infarction discharged from our centre between 1982 and 1988.2 During this period we saw an increase of the prescription rate of both drugs. In 1988 the prescription rate was 34% (42/122) for D adrenergic blockers, 33% (40/122) for acetylsalicylic acid, and 7% (9/122) for both. For [i adrenergic blockers our results are similar to those of Drs Eccles and Bradshaw. The participation of 13 hospitals in their study in the third international study of infarct survival (ISIS-3) may have induced a higher prescription ofacetylsalicylic acid at hospital discharge. Thirty four per cent (253/736) of our study population had relative contraindications to the use of li adrenergic blockers. We also analysed the prescription of other drugs. In 1988 the prescription rate of calcium channel antagonists amounted to 44% (54/122). They were prescribed more often for patients with previous cardiovascular disease, such as hypertension or angina (table). In addition, we observed that the use of f adrenergic blockers was lower in patients treated with calcium channel agonists. f3 Adrenergic blockers and calcium channel agonists have some common therapeutic uses, such as hypertension and angina, but efficacy for the
SIR,-Professor A G Shaper and colleagues' have referred to data on milk, spreading fats, and heart disease in our Caerphilly prospective heart disease study2 and have presented data on this same topic from their regional heart study. We can certainly confirm what Professor Shaper and colleagues show in their own data. Milk drinking is confounded with many other factors of relevance to heart disease risk, and men who use butter differ in many ways from those who use margarine. The effects of some of these differences on relations with heart disease are substaniial, and difficulties arise because their interrelations are complex. As we said in the progress report, the data are being explored further and we will publish a detailed report in due course. In retrospect, it was unwise for a brief description and preliminary analysis of these data to be included in a privately published report that attempted to summarise a wide range of work on cardiovascular disease. It was particularly unfortunate that the one small section on milk and fats was widely reported and interpreted uncritically in the popular press. I can only apologise for not having foreseen that this was likely to happen, and I now greatly regret not having withheld the data until a more adequate analysis could be submitted to the scientific press. PETER C ELWOOD MRC Epidemiological Unit (South Wales),
Llandough Hospital, I'enarth, Sotith Glamorgan CF6 IXX I Shaper AG, Wannamethee G, Walker M. Milk, butter, and heart disease. BUM7 1991;302:785-6. (30 March.) 2 MRC Epidemiology Unit. Eptdemiological studies of cardiovascular disease. Penarth, South Glamorgan: Llandough Hospital, 1991. (Progress report VII.)
Treatment with ji adrenergic blocking agents (r3AB) or with calcium channel antagonists (CaCA) at hospital discharge after myocardial infarction in 736 patients (49 patients treated with both drugs were excluded) No (%) of patients treated with CaCA No (¼/.) of patients treated with liAB Total With previous cardiovascular disease Without previous cardiovascular disease
154 (64-7) 46 (41 8)
84 (35-3) 64 (58-2)
238 110
Total
200 (57 5)
148 (42-5)
348
Z.= 16- 12 (p-- 0-0006). 913
