benefits.''3 When the data from all published trials of intraportal treatment are combined they show an overall reduction in the odds of death of 26%. '4 Whereas some of the trials of intraportal fluorouracil showed a reduction in the incidence of liver metastases" 1' others showed no such reduction,'2 15 and the effect of this treatment might be systemic rather than a local effect on liver metastases.'2 The perioperative approach to adjuvant chemotherapy does have the considerable advantage of a very short treatment period lasting only a few days -compared with one year of treatment in the intergroup studies.7 The United Kingdom Coordinating Committee for Cancer Research is currently conducting AXIS, a new trial to attempt to confirm or refute the value of this approach, and this study will be crucial in determining the future of intraportal administration of fluorouracil. Where do we go from here? The clear statement from the National Cancer Institute that adjuvant chemotherapy is now standard practice in stage C carcinoma of the colon and that ''no treatment" controls are no longer justifiable has resulted in most doctors in the United States using fluorouracil and levamisole. 16 17 The attitude of European oncologists, however, is much more variable.8 9 Though most are now more optimistic in their attitude to adjuvant therapy for cancer of the colon, many believe that studies have not yet established that the reduction in mortality with systemic chemotherapy outweighs the toxicity and inconvenience of the treatment and consider that trials should still include an untreated control arm. What is clear is that the pessimistic attitude to adjuvant therapy for cancer of the colon - common in the past - can no longer be justified. The recent intergroup trials from the United States were rigorous, and on the basis of these studies many clinicians in Britain are likely to conclude that those patients with stage C cancer of the colon who are treated outside a clinical trial should be offered adjuvant therapy as used in the intergroup studies. Much remains to be done-as the consensus report made clear. Newer treatments (such as the combination of fluorouracil with folinic acid, which has shown substantial activity in advanced cancer of the colon) should be tested as adjuvant therapy. The places of levamisole in combination with fluorouracil and of short perioperative infusions of fluorouracil into the portal vein need to be defined. Though it seems unlikely that the effect of adjuvant therapy in stage B cancer of the colon should be qualitatively different from that in stage C, the number of events and the follow up are not yet sufficient for any conclusions to be drawn from the published studies to date. Recent reports of adjuvant therapy in rectal

cancer now suggest a degree of benefit similar to that in the colon cancer studies.20 The priority must now be to enter many more suitable patients with cancer of the colon into randomised trials of adjuvant therapy. Patients who are eligible for clinical trials should be encouraged to take part in such studies. Only by this means will progress be made. Individual doctors will have to judge whether or not they think a "no treatment" control arm is still justifiable, and patients will need to be informed of the issues before they are treated or offered randomisation. MAURICE L SLEVIN Consultant Medical Oncologist, ICRF Department of Medical Oncology, St Bartholomew's and Homerton Hospitals, London EClA 7BE RICHARD GRAY

Research Officer, ICRF Cancer Studies Unit, Radcliffe Infirmary, Oxford OX2 6HE I Earlv Breast Cancer Trialists' Collaborative Group. Effects of adjuvant tamoxifen and of cytotoxic therapy on mortality' in early breast cancer. An overview of 61 randomised trials among 28 896 women. N Engl_] Med 1988;319:1681-92. 2 Early Breast Cancer Trialists' Collaborative Group. Treatment of earlv breast cancer. Vol 1.

t'orldnstde i-uidence 1985-1990. Oxford: Oxford University Press, 1990. 3 Buvse M\, Zeleniuch-Jacquotte A, Chalmers TC. Adjuvant therapy of colorectal cancer. Why we still don't know. JAMA 1988;259:3571-8. 4 Bancewicz J, Calman K, MacPherson SG, M\cArdle C, McVie J, Soukop M. Adjuvant chemotherapy and immunotherapy of colorectal cancer. ] R Soc Med 1980;73:197-9. 5 Windle R, Bell PRF, Shaw D. Five year results of a randomised trial of adjuvant 5-fluorouracil and levaamisole in colorectal cancer. Brj Surg 1987;74:569-72. 6 Laurie JA, Moertel CG, Fleming TR, et al. Surgical adjuvant therapy of large-bowel carcinoma: an evaluation of levamisole and the combination of levamisole and fluorouracil. Clin Oncology 1989;7:1447-56. 7 Moertel CG, Fleming TR, MacDonald JS, et al. Levamisole and fluorouracil for adjuvant therapy of resected colon carcinoma. V Engi M' ,Wed 1990;322:352-8. 8 NIH Consensus Conference. Adjuvant therapy for patients with colon and rectal cancer. JAMA

1990;264: 1444-50. 9 Taylor 1, Miachin D, Mullee M, Trotter G, Cooke T, West C. A randomised control trial ofadjuvant portal vein cytotoxic perfusion in colorectal cancer. Br7 Surg 1985;72:359-63. 10 Gray BN, deZwart J, Fisher R, et al. The Australia and New Zealand trial of adjuvant therapy in colon cancer. In: Salmon SE, ed. Adjutant therapy of cancer. Philadelphia: Grune and Stratton,

1987:357-456. 11 Metzger U, Mermillod B, Aeberhard 1P, et al. Intraportal chemotherapy in colorectal carcinoma as an adjuvant modality. Worldj Surg 1987;ii:452-8. 12 Wolmark N, Rockette H, Wickerham DL, e. al. Adjuvant therapy of Dukes' A, B and C

13 14

15 16 17 18 19 20

adenocarcinoma of the colon with portal vein fluorouracil hepatic infusion: preliminary results of national surgical adjuvant breast and bowel project protocol C-02. . Clin Oncol 1990;8:1466-75. Wereldsma JCJ, Bruggink EDIM, Meijer WS, et al. Adjuvant portal liver infusion in colorectal cancer with 5-fluorouracil/heparin scrsus urokinase versus control. Cancer 1990;64:425-32. Gray R, James R, Mossman J. AXIS-a suitable case for treatment. UKCCCR Colorectal Cancer Sub-committee. Br] Cancer (in press). Beart RW, Mioertel CG, Wieand HS, et al. Adjuvant therapy for resectable colorectal carcinoma with fluorouracil administered by portal vein infusion. Arch Surg 1990;125:897-901. Vanchieri C. Colon cancer update triggers changes in practice. ] Natl Cancer Inst 1990;82:898. Mayer RJ. Does adjuvant therapy work in colon cancer? N Engli Med 1990;322:399-401 Anonymous. Mixed European reactions to American colorectal data. Annals of Oncologv 1990;1 :239-40. Wills J, Wagener DJTH. Adjuvant treatment of colon cancer: where do we go from here? Annals of Oncolotgy 1990;1:329-31. Krook JE, Moertel CG, Gunderson LL, et al. Effecti\e surgical ad'juant therapy for high risk rectal carcinoma. N Engl] Med 199 1;324:709-15.

How sick the baby? A new method of assessment suggested One of the most worrying problems that general practitioners face is an acutely ill baby. Should they admit the baby to hospital? If the baby is not admitted will the parents call the doctor if the baby's condition deteriorates? A team from the department of paediatrics in Cambridge has now devised a method for assessing sick babies. The Baby Check is intended to assess the severity of acute systemic illness in babies less than 6 months old. It consists of a checklist of 19 symptoms and signs, and the total score derived from these determines the recommended course of action. It is being promoted as a means whereby parents can BMJ

VOLUME

302

11

MAY

1991

decide when to call a doctor and for general practitioners to use to determine when specialist attention is needed. The Child Growth Foundation is currently distributing copies of Baby Check to some Family Health Service Authorities. Will the Baby Check live up to its claims? How it was developed and tested is described in a series of papers in the Archives of Disease in Childhood.'4 The symptoms and signs used to predict a poor outcome of illness were derived from one paediatrician's experience, but they accorded with those of other paediatricians. Because of the low incidence of severe illness in a community the instrument's sensitivity 1101

and specificity had to be calculated on a theoretical population. This was developed by extrapolation from the prevalence of symptoms and signs in a random sample of well babies examined at home at varying times after birth and from children who had been referred for admission to a paediatric unit. Two assumptions have been made in creating this theoretical population. The first is that those children who are referred for admission are the sickest. We know that general practitioners vary in their propensity to refer patients to hospital, and parents vary in their ability to cope with sickness at home. For this reason we cannot be certain, without assessing sick children who are being cared for at home, that there are not many children who never reach hospital but who have similar symptoms and signs to those of children who are referred. The second assumption is that the sickest children will form the 2% of the population who are seen in hospital. Again, without a detailed community survey of all children within a defined population it is difficult to know whether this assumption is correct. Common sense leads us to expect that admissions vary with the season and by the social class distribution of the population. The difficulties that the researchers faced in developing the Baby Check were considerable, and the sensitivity and specificity that they quote for their test should be regarded with caution. Is the Baby Check easy to use? Although it is dauntingly long, the explanations on how to interpret the symptoms and elicit the physical signs are excellent. Field trials report that mothers and their general practitioners were satisfied with the

booklet. It should be borne in mind, however, that the sample of mothers in whom these trials were conducted contained a high proportion of well educated people and even in this sample the uptake was not good. The results might not be nearly so favourable in a less advantaged population, and further field trials are needed in this group. The Baby Check may be of great value to a general practitioner faced with a sick baby whom he or she wants to look after at home. In such a case the parents would have a systematic way of assessing their child that would relieve anxiety, both their own and that of the general practitioner. The Baby Check does not address the problem of the sudden infant death syndrome, in which major and minor signs of illness are neither a sensitive nor specific indicator of sudden unexpected deaths of infants and have no predictive value.5 CHRIS WATKINS Consultant Senior Lecturer in General Practice, General Practice Unit, Department of Epidemiology and Public Health, University of Bristol, Bristol BS8 2PR I Morley CJ, Thornton AJ, Cole 1TJ, Hewson PH, Fowler MA. Babv Check: a scoring svstem to grade the sccrity of acutc systemic illness in babies under 6 months old. Arch Dis Child 1991;66:100-5. 2 Thornitoni AJ, Morley CJ, Green SJ, W'alker KA, Reninie JAI. Field trials of the Baby Check score card: mothers scoring their babies at home. Arch Dis Chhild 1991 ;66: 106-10. 3 Morle CJl, Thornton AJ, Green Sj, Cole TIJ. Field trials of the Baby Check score card in gcienral practicc. Arch Dis Child 1991;66:111-4. 4 Thornton AJ, Morley CJ, Cole TJ, Green SJ, Walker KA, Rennie JM. Field trials of the Baby Chcck score card in hospital. Arch Dis Child 1991;66:115-20. 5 Gilbert RE, Flemittg Pj1, Azaz Y, Rudd P. Signs of illness preceding suddeni unexpectcd dcaths in infants. BIM7 1990;300:1237-9.

The politics of change The take off was never going to be smooth The occupational disease of politicians, it is often argued, is myopia-that is, an inability to see beyond the next general election. It is not a criticism that can be levelled against the authors of the NHS reforms. On the contrary, the charge against them must be that their eyes were so firmly fixed on the distant horizon that they ignored the political pot holes on the road towards their objectives. Not surprisingly, ministers have stumbled. Not surprisingly, either, the opposition has gleefully seized on every slip: the announcement of redundancies by Guy's and Lewisham Trust' came providentially just as the poll tax issue was slipping away. It is a pattern that is likely to be repeated with increasing frequency-and stridency -as the general election approaches. Given the scale and complexity of the NHS, probably not a week has gone by since 1948 when there has not been some problem or scandal somewhere waiting for exposure. The difference now is that the NHS reforms provide a focus for discontent; everything can be blamed on them. Indeed, there may even be incentives to exaggerate them. This is not to argue that the implementation of the NHS reforms is unproblematic. It is to suggest, however, that if we are to identify and isolate their impact we must first try to understand the political dynamics of change: the reason why we might expect political turbulence even if the introduction of the reforms had been handled differently or with a more generous injection of funds to ease the pains of transition. The main reason for turbulence is simple. It is that the NHS reforms, quite deliberately, represent a challenge to the status quo. They are meant to force a re-examination of 1102

existing clinical practices and patterns of organisation. The point emerges strongly from Enthoven's influential paper,2 which helped to crystallise many of the ideas that shaped the reforms. In this he argued, following other commentators,3 that the NHS's problems stemmed from "rigidity." Organisational innovation had lagged behind clinical innovation; even in the case of clinical innovation, incentives to generalise best practices were weak. Given this diagnosis, implicitly endorsed by the government,4 it is not surprising that the cure is seen as threatening. The balance of power-between consultants and managers and between hospitals and general practice-is shifting. New ways of doing things have to be learnt; inherited investments in organisational knowledge have to be written off. The process is understandably painful for many of those concerned, just as it has been painful for teachers in schools and universities, where a similar process of change has been going on. Much of the reaction to the NHS reforms therefore tells us nothing (for good or bad) except that resistance to change is as inevitable as it is predictable. Indeed, it was predicted by Enthoven2: Politicians face powerful disincentives to attempting significant change. The benefit from any serious attempt to improve things would be gradual, not clearly visible before the next election.... It is hard for non-specialist voters to distinguish image from reality in the short run. But any serious efficiency-improving change risks being attacked as 'downgrading the quality of care' by threatened provider groups, or being blasted by the left as 'tampering with the NHS.' There are other grounds for predicting resistance to change if this entails any shift in policies or resources. Existing

BMJ

VOLUME 302

11 MAY 1991