Director, neonatal screening service (northern sector). JANET STEVENSON. Senior registrar in public health medicine. Neonatal Screening Laboratory,.
to reduce crime and prevent violence. Focusing on violence and crime in public health reports can move these policies forward. J D MIDDLETON Director of public health J N RAO Consultant in public health J A REID Public health development manager Sandwell Family Health Services Authority, West Bromwich, West Midlands B70 9LD 1 Zwi A, Rifkin S. Violence involving children. BMJ 1995;311:1384. 2 Richardson J, Feder G. Domestic violence against women. BMJ 1995;311:964-5. 3 Shepherd JP, Farrington DP. Preventing crime and violence. BMY 1995;310:271-2. 4 Middleton J, ed. Safer Sandwell. West Bromwich: Sandwell Health Authority, 1995.
Minimum standards should be set for near patient testing EDITOR,-Richard Hobbs's editorial on near patient testing in primary care is welcome,' as a large expansion in the use of such testing can be predicted. However, the importance of collaboration and discussion with hospital laboratories, and of adequate quality control, needs to be
highlighted. The following recommendations should be useful. Firstly, there should be a formal training programme for staff performing the tests. This could include training in the collection of specimens, the principles of the analysis, use of the machines, how to document results correctly, calibration and quality assessment, expected values of the analyte in health and disease, and the safe disposal of samples. Secondly, the users of near patient testing apparatus should have to show their competence at regular intervals. Thirdly, patients should be tested only by certified users. Fourthly, a quality assurance programme, including both internal and external quality control, should be in operation and preferably should involve the local pathology laboratory. Fifthly, well defined user manuals, which should include standard operating procedures, should be instigated. Sixthly, apparatus and associated equipment should be adequately maintained and cleaned regularly. Seventhly, results should be documented adequately and an equipment logbook kept. Finally, the laboratory may be able to advise about what equipment to purchase. This list is not exhaustive but could form a minimum standard of expectation and is compiled from several sources.2 3 In this way near patient testing would provide the best possible results for the patient. MARTIN CROOK Senior lecturer Chemical Pathology,
Guy's Hospital, London SEl 9RT 1 Hobbs R. Near patient testing in primary care. BMJ 1996;312:263-4. (3 February.) 2 Anderson JR, Linsell WD, Mitchell FM. Guidelines on the performance of chemical pathology assays outside the laboratory. BMJ 1981;282:743. 3 Price CP, Burrin JM, Nattrass M. Extra-laboratory blood glucose measurement: a policy statement. Diabetic Medicine 1 988;5:705-9.
Psychosis in Afro-Caribbean people Further data should have been obtained EDITOR,-Kwame McKenzie and colleagues report on psychosis with a good prognosis in
BMJ VOLUME 312
4 MAY 1996
Afro-Caribbean people now living in the United Kingdom.1 We appreciate the difficulties encountered in conducting a prospective study with a considerable number of confounding variables but wish to make the following comments. There is a contradiction in the method used to limit the data studied. In the study recent onset was defined as within five years. Patients were excluded every third month in rotation to limit the number studied, yet the same result could have been achieved by defining recent onset as within two years. Excluding patients every third month rather than excluding every third patient is baffling as seasonality was not an issue in the study. We think that illicit drug use should have been considered during the recruitment of patients. For instance, cannabis has been shown to be a prognostic factor in psychotic illnesses.2 In an attempt to reduce the risk of misclassification associated with schizophrenia in ethnic groups the authors seem to have erred on the side of overinclusion. For example, the prevalence of affective psychosis was about 50% higher in the Afro-Caribbean group, which may have contributed to the good outcome. While using periods of unemployment and imprisonment as outcome variables it would have been more meaningful to compare these with the premorbid assessment in the groups themselves or in the general populations from which the groups were derived rather than just between the two groups, as was done in the study. Also, premorbid personality and rates of all admissions rather than rates of involuntary admission should have been considered when the outcome was measured. The authors hypothesise that the better prognosis in the Afro-Caribbean group may have been due to the higher prevalence of illness with social precipitants. No attempt was made in the study, however, to measure life events or social precipitants. We suggest that, in addition to life events, family involvement and support systems may have considerably affected the outcome.3 NITIN B PURANDARE Senior registrar in psychiatry
Queen's Park Hospital, Blackburn BB2 3HH PRIYADARSHAN NEELKANTH JOSHI Senior registrar in psychiatry
Withington Hospital, Manchester M20 2LR 1 McKenzie K, van Os J, Fahy T, Jones P, Harvey I, Toone B, et al. Psychosis with good prognosis in Afro-Caribbean people now living in the United Kingdom. BMJ 1995;311:1325-8.
(18 November.) 2 Martinez-Avevalo MJ, Calcedo Ordoficz A, Varo-Prieto JR. Cannabis consumption as a prognostic factor in schizophre-
nia. Brj7Psychiatry 1994;164:679-81. 3 El-Islam MF. A better outlook for schizophrenics living in
extended families. BrJ Psychiatry 1979;135:343-7.
"Afro-Caribbeans" could have been of Chinese, Indian, European, or African extraction EDITOR,-I wish to point out serious flaws in Kwame McKenzie and colleagues' study of the prognosis of psychosis in Afro-Caribbean people.' It is courageous of the authors to tackle this comparative study, but their failure to clarify their sampling leaves their work with glaring methodological flaws. Two groups of people were compared. The first group, the "white" group, was selected by skin colour, place of birth (United Kingdom), and place of parents' birth (United Kingdom). The second group, the "Afro-Caribbean" group, was selected simply on the basis of the place of parents' birth (the Caribbean islands): there is no reference to skin colour or to the patients' place of birth. Besides, McKenzie and colleagues assert that white skinned people born in the
United Kingdom of parents also born in the United Kingdom form a "culturally homogeneous" group. This is misinformed: it takes a brave person to say that people such as the white Irish, Scottish, Welsh, and English are of the same culture. What is more, being white skinned is a property of both the Caucasoid and the Mongoloid divisions of humankind. The Caribbean islands are home to people of all the main races except Australian Aborigines. So what is the "Afro-" in the authors' Afro-Caribbean group? The Afro-Caribbean group could have been made up of people of Chinese, Indian, European, or African extraction. And if the subjects were born in the Caribbean but now live in the United Kingdom then environmental factors that acted on them in infancy or childhood could account for the difference in the outcome of their illness compared with that of people born in the United Kingdom. This is either poor science or poor reporting of research. It is interesting to note that, in the November issue of the Psychiatric Bulletin, McKenzie and an associate argued that the term "Afro-Caribbean" should no longer be used to describe any group of people as it was too imprecise2-yet here we find McKenzie and other associates applying in the BMJ the very term he repudiates elsewhere. This is the reason why the study reported in the BMJ is so unsatisfactory: McKenzie is struggling, like all of us, with the elusive concepts of race, culture, and ethnicity and how these relate to the origins and outcome of human disease. IKECHUKWU 0 AZUONYE Visiting consultant psychiatrist
Grovelands Priory Hospital, London N14 8RA 1 McKenzie K, van Os J, Fahy T, Jones P, Harvey I, Toone B, et
al. Psychosis with good prognosis in Afro-Caribbean people now living in the United Kingdom. BMY 1995;311:1325-8. (18 November.) 2 Hutchinson G, McKenzie K. What is an Afro-Caribbean? Implications for psychiatric research. Psychiatric Bulletin 1995;19:700-2.
Authors' reply
EDrrOR,-Nitin B Purandare and Priyadarshan Neelkanth Joshi question the sampling in our study. We gathered data from several people and places for each patient. To ensure that the data were collected close to the time of admission and were of high quality, a break from assessments was necessary every third month. We doubt whether this introduced a systematic bias in terms of ethnicity and outcome. Use of cannabis did not differ between the two groups.' Premorbid personality was investigated but did not explain our findings. We reported that whether a patient of Caribbean origin was born in the United Kingdom or outside did not affect the result, that the number of admissions over the follow up period did not differ between the two groups, and that the better prognosis with respect to the course of the illness was not due to differences in diagnoses between the groups. Comparison with a premorbid assessment of unemployment and imprisonment would not have been more meaningful. A conclusion of the paper is that discrimination may have affected the prognosis. This could occur at any stage of the illness. It does not make sense to control for a variable under scrutiny in the study. Differences in family involvement as measured by whether the patients lived with their family did not explain our results. Moreover, research that has shown a better prognosis in South Asian patients and has hypothesised that this is because of family involvement has not shown the same effect in people of Caribbean origin.2 Ikechukwu 0 Azuonye confuses the issue of culture with the issue of discrimination. Our
1157
paper focused on discrimination in its widest sense and the hypothesis that this might explain differences in outcome. Thus we looked at two groups that had different risks of discrimination: people of Caribbean origin and white British people. Evidence suggests that Irish people are discriminated against in the United Kingdom,3 so we excluded them. We were interested in culture only as it might affect discrimination, which is consistent with our statement that we aimed to demarcate a white group that was "as culturally homogeneous...as possible to decrease difficulties in interpreting results." Azuonye highlights the problem of how to use data on the minority of patients of Caribbean origin who are not of African ancestry. Because we did not know what level of discrimination they would suffer and no systematic evidence exists, we set our criteria to include such patients. We were unable to assess this question as our sample contained only people ofAfrican ancestry. The paper in the Psychiatric Bulletin to which Azuonye refers relates to culture, not discrimination.4 It was aimed at helping researchers to choose the most appropriate variables when comparing ethnic groups. We believe that by using populations at differing levels of risk we chose the most appropriate variables. KWAME MCKENZIE
Honorary research fellow BRLAN TOONE Consultant psychiatrist TOM FAHY
Consultant psychiatrist ROBIN MURRAY Professor of psychiatry
Department of Psychological Medicine, Institute of Psychiatry and Kings College Hospital, London SE5 8AF
JIM VAN OS Senior lecturer Department of Psychiatry, Section of Social Psychiatry and Psychiatric Epidemiology, University of Umburg, Maastrict 6200 MD, Netherlands PETER JONES Senior lecturer
We audited the neonatal screening programme in the Northern region in 1993.2 We found initially that only six of the 16 districts had a timely, failsafe mechanism in place for ensuring that all babies were screened. Five other districts had mechanisms to ensure that babies were screened, but this was not always timely. After the audit all districts except two had a timely, failsafe mechanism for identifying babies who had not been screened. Those districts where the system worked well were those in which a consultant (usually a community paediatrician) took a lead in coordinating and monitoring it and in which a working group of all the professionals involved met regularly to discuss problems. In a circular on screening to detect phenylketonuria issued in 1969 responsibility for the screening programme in each district was clearly stated to lie with the medical officer of health3; the current equivalent is the director of public health. Directors of public health have an important role in ensuring the provision of a high quality screening programme that includes monitoring arrangements, either through direct involvement or by clearly delegating responsibility to another consultant. We agree that the neonatal screening service must continue to be provided regionally and not be fragmented. In addition, the overview of the service must be maintained centrally by the Department of Health so that any further screening programmes-for example, for cystic fibrosis-can be coordinated throughout Britain. ANGELA GALLOWAY
Director, neonatal screening service (northern sector) JANET STEVENSON Senior registrar in public health medicine Neonatal Screening Laboratory, Department of Pathology, Dryburn Hospital, Durham DHI 5TW 1 Cappuccio PP, Hickman M, Barker M. Neonatal screening. BMJ 1996;312:182. (20 January.) 2 Galloway A, Stevenson J. Audit improves neonatal (Guthrie) screening programme. BMJ 1994;309:878. 3 Department of Health and Society Security, Welsh Office.
Screening for early detection of phenylketonuria. Cardiff: DHSS, 1969. (HM(69)72.)
Department of Psychiatry,
University of Nottingham, Nottingham NG3 6AA IAN HARVEY
Consultant psychiatrist Towers Hospital, Leicester LE5 OTD 1 McKenzie KJ, Murray R, Jones P. The aetiology of schizophre-
nia in Afro-Caribbeans. Schizophrenia Research 1995;15: 194-5. 2 Birtchwood M, Cochrane R, MacMillan F, Copestake S,
Kucharska J, Carris M. The influence of ethnicity and family structure on relapse in first episode schizophrenia. BrJ Psychiatry 1992;161:783-90. 3 Pearson M, Madam M, Greenslade L. Generations of an invisible minority: the health and well-being of the Irish in Britain. A preliminary survey. Liverpool: Institute of Irish Studies, University of Liverpool, 1981. (Occasional papers in Irish studies No 2.) 4 Hutchinson G, McKenzie K. What is an Afro-Caribbean? Implications for psychiatric research. Psychiatric Bulletin 1995;19:700-2.
Performance of neonatal screening programme must be monitored EDITOR,-We agree with Francesco P Cappuccio and colleagues that monitoring the performance of the neonatal screening programme for phenylketonuria and congenital hypothyroidism is difficult at district level,' but our experience in the Northern region shows that it is not impossible.While neonatal screening is complex and involves many health professionals-midwives, health visitors, laboratory staff, consultants with responsibility for neonatal screening, child health officers, and the directors of the neonatal screening service-close communication between these groups can overcome difficulties. 1158
Postoperative deep vein thrombosis and surgery for varicose veins EDrroR,-In his critical editorial comment Bruce Campbell discusses the possible links between thrombosis, phlebitis, and varicose veins.1 While pointing out the evidence for an association between varicose veins and deep vein thrombosis in patients undergoing abdominal or pelvic surgery, he questions the need for a thromboembolic prophylaxis in patients undergoing surgery on their varicose veins. Admittedly, reliable data on the incidence of deep vein thrombosis after varicose vein surgery are not available from controlled, randomised studies. However, we reported a few years ago on the incidence of pulmonary embolism in a large retrospective series of 19 161 patients who were operated on over a 10 year period (1980-9) in our clinic of digestive surgery.2 Among the 1063 patients who underwent stripping of varicose veins in that population, four presented with clinically manifest and objectively confirmed pulmonary embolism during hospital stay, with two additional events occurring during the four weeks after discharge from hospital, giving an overall incidence of postoperative pulmonary embolism after this type of surgery of 0.56% (95% confidence interval 0.21 to 1.23), which was similar to the incidences of 0.40% (0.22 to 0.66)that was observed after biliary surgery and 0.60% (0.22 to 1.30) after laparotomy.
On the basis of these data and the established increased risk of deep vein thrombosis at the time of abdominal or pelvic surgery in patients with varicose veins, we believe that patients undergoing surgery for varicose veins are likely to be at special risk of developing postoperative venous thromboembolism and should thus receive prophylaxis until especially designed trials will confirm or negate this need. HENRI BOUNAMEALUX Chief, division of angiology and haemostasis
OLJVIER HUBER Consultant surgeon, clinic of digestive surgery University Hospital of Geneva, CH-1211 Geneva 14, Switzerland
1 Campbell B. Thrombosis, phlebitis, and varicose veins. BMJ 1996;312:198-9. (27 January.) 2 Huber 0, Bounameaux H, Borst F, Rohner A. Postoperative pulmonary embolism after hospital discharge. An underestimated risk. Arch Surg 1992;127:310-3.
Postoperative pulmonary complications Pain relief improves respiratory fimction
EDrrOR,-John C Hall and colleagues addressed the important issue of postoperative complications after abdominal surgery.' However, their evidence does not support their conclusion; in the light of relevant recent literature, we would like to offer some explanation. Major changes in respiratory function occur in all patients after anaesthesia and surgical incisions, especially on the thorax and upper abdomen, because of a decrease in the functional residual capacity with minimal change in the closing volume leading to airway closure during tidal breathing.2 These changes are most exaggerated in elderly patients, obese patients, smokers, and those with pre-existing cardiopulmonary disease. These changes in pulmonary mechanics are not as great after surgery of the lower abdominal cavity.2 Several randomised studies have shown that postoperative pulmonary complications correlate positively with the decrease in ventilatory efficiency.2 Hall and colleagues, by combining upper and lower abdominal surgical patients, have diluted the impact of physiotherapy and the use of incentive spirometer on the prevention of postoperative complications. The authors also suggest that the postoperative decline in respiratory functions is obligatory and is not reversible by effective analgesia. The most important cause of regional impairment of ventilation, ineffective cough, and impaired ability to sigh and to breathe deeply is incisional pain.2" Considerable evidence has accumulated since the early 1980s confirming that these changes can be minimised and even prevented by effective analgesic techniques, especially if analgesia is started preemptively and continued postoperatively until wound healing has taken place.3 Overreliance on systemic opiates rather than on regional analgesic techniques (only 31% in this study) contributed to the study's less than optimal result. We are intrigued by the term "floppy" diaphragm. The pathophysiology of postoperative diaphragmatic dysfunction after abdominal surgery is complex and multifactorial. A reduction of intrinsic diaphragmatic contractile properties is not the predominant factor.' Reflex inhibition of phrenic nerve activity, particularly after upper abdominal surgery, is the most attractive explanation.5 These changes might be expected to be considerable after upper abdominal surgery, but they are probably of little consequence after appendicectomy. Combining all surgical groups is therefore unhelpful. We contend that high quality pain relief, centred around optimal afferent blockade, BMJ voLuME 312
4 MAY 1996
