Letters

To the Editor: Our recent study of patients' priorities for asthma care1,2 provides additional evidence supporting the concerns of Wilson and Robertson in their editorial questioning the possible overuse of inhaled corticosteroids.3 We have reported a qualitative study of 62 individuals who presented to an emergency department at either a central city, suburban or rural hospital, in which we explored individuals' perceptions about their asthma, its care and the impact of asthma on their lives.1,2 We also asked participants to complete a questionnaire on the use of medications and sought to amplify this information by further probing the use of medications in our qualitative data collection. Of the 82% of participants in our study currently using inhaled corticosteroid medication (51), 30% (16) were taking 1000 μg of fluticasone or equivalent daily and another 19% (10) were taking more than 1500 μg or equivalent. Current product information for fluticasone suggests a maximum dose of 1000 μg twice daily, whereas National Asthma Council (NACA) guidelines recommend that 500 μg fluticasone or equivalent daily may be the upper limit of useful effect.4,5 We also asked patients how long their medication lasted. Eleven (18%) stated that inhaled corticosteroid devices lasted three weeks or less. Use above recommended doses did not only occur for inhaled corticosteroids, but also for symptom controller medications. Twenty-four (35%) of the 31 (50%) patients receiving this medication reported that a device lasted three weeks or less, indicating use above usual recommended doses. Most patients in our study voiced concerns about the cost of asthma and drug side effects; some adjusted their medication use to manage these issues.1 In such individuals, high use or overuse of preventive and controller medication would increase both costs and side effects, partly explaining these patients' concerns. Doctors may be overprescribing inhaled corticosteroid medication because there is a discrepancy between dosages recorded in published drug information and newer recommendations for optimal inhaled corticosteroid dose.4,5 Our findings show that, in some patients, the risks associated with the use of inhaled corticosteroids are likely to be compounded by using them at higher doses than those recommended. Doctors need to be aware of this in managing patients with asthma who have severe symptoms, in whom overuse, rather than underuse, is likely to be a problem.

Dianne P Goeman · Susan M Sawyer · Michael J Abramson · Kay Stewart · Francis C K Thien · Rosalie A Aroni · Jo A Douglass

To the Editor: Halasz and Vance1 are correct to point out that there is a diversity of causes that can contribute to a child exhibiting symptoms of attention-deficit hyperactivity disorder (ADHD), as defined in the Diagnostic and statistical manual of mental disorders (DSM-IV).2 In their article, they describe a child who meets the DSM-IV criteria for diagnosis of ADHD and in addition has been affected by environmental factors including poor bonding (due to maternal depression), domestic violence and parental separation. The child also exhibits developmental disability, as exemplified by delayed language development. The message is that, by explaining his symptoms in terms of his early experiences and his developmental disability, a diagnosis of ADHD can be excluded. Children with ADHD frequently come from families with disharmonious parental relationships. This may be associated with ADHD in one of the parents, perhaps the violent father in the case described. As clinicians our aim is to ameliorate symptoms as promptly and effectively as possible, and I am frequently impressed by the dramatic improvement that stimulant medication can make to a child's functioning both at school and within the family, with follow-on improvements in mood and self-esteem. Behavioural interventions and family therapy are important adjuncts to medication, but families such as the one described can be difficult to work with and this can limit the effectiveness of such interventions. A carefully monitored one-month trial of stimulant medication, with behavioural rating scales completed by the class teacher, may be appropriate in cases such as the one described. On the other hand, to deny a child a trial of stimulant medication on the basis of adverse early experiences and developmental disability may be to keep from the child the treatment that would help most.

Alison Poulton

In reply: We believe the core symptoms of attention-deficit hyperactivity disorder (ADHD) in children reflect a behavioural "final common pathway" of developmental risk factors,1 which can include transgenerational associations of core symptoms, as Poulton notes. Current scientific evidence suggests both genetic and environmental contributions, such as verbal and visuospatial executive dysfunction2 and/or early patterns of attachment deficits.3 Increased levels of parental psychopathology, associated with (in the child) deficiencies in problem solving, affect regulation, emotional communication and secure attachment, may contribute to the child's symptoms. For this reason, we advocate that medical management be based on a thorough assessment, to ensure that appropriate psychological interventions (eg, parent and teacher management training) are offered alongside psychostimulant medication. In a recent speech at a scientific meeting of the Faculty of Child and Adolescent Psychiatry, Dr A Mawdsley, a distinguished child psychiatrist, expressed his belief that "prescribing medication in the absence of a careful emotional state assessment is inferior medical practice". He went even further to state that "prescribing medication in the absence of a behavioural modification program should be considered medical negligence".

George Halasz · Alasdair LA Vance

To the Editor: I wish to draw attention to a draconian anomaly in the National Health Act 1953 (Cwlth). All GPs encounter patients "shopping" for narcotics and/or tranquillisers. The Doctor Shopper phone line (which enabled GPs to rapidly obtain information from the Health Insurance Commission to identify non-genuine patients) was a boon in guiding GPs' management of such situations. Concern over the new private sector amendments to the Privacy Act 1988 (Cwlth) led to an examination of the legal standing of the Doctor Shopper phone line, and it has now been cancelled. Concerned GPs are now limited to requesting that a "patient" sign a voluntary release of their Pharmaceutical Benefits Scheme record. This tells the "patient" that they have been rumbled, and they move on to the next practice on their list. If they have signed the Privacy Release Form, then the GP will receive a printout of the drugs they have received under the Pharmaceutical Benefits Scheme in the previous six months. This is accompanied by a letter informing the doctor that he or she "cannot make a record of, divulge or communicate to any person, any information with respect to the affairs of the person whose information has been released. To do so attracts a penalty of $5000 and/or imprisonment for a period not exceeding two years". So, under the provisions of the National Health Act (subsection 135A), even putting this information in the medical records of a multidoctor practice would appear to be illegal. It is clearly illegal to warn other doctors outside the practice. There is no corresponding legislation which affects doctor shoppers. So the "right-doers" can finish up in jail, while the "wrong-doers" can, with impunity, continue to play their dissembling, time-consuming, and sometimes harassing, games.

Max Kamien

To the Editor: We are writing to add our wholehearted support to the plea made by Carapetis et al1 for a simplified ethical approval process for multicentre studies. We are conducting two national case-controlled studies of cancer in Australia, funded by grants from the National Institutes of Health and the Department of Defense in the United States, as well as the National Health and Medical Research Council (NHMRC). Our ability to achieve full population coverage was a major competitive advantage in terms of securing international funding. However, to realise this objective, we have spent more than a year obtaining ethics approval from the myriad institutions controlling access to patients and the public. We have been required to make more than 60 separate ethics applications, lodging about 550 copies of the proposal (a total of 40 000 sheets) at a cost of more than $7000 for paper and printing alone. When labour is included the cost of the initial submissions escalates to $16 000 (excluding substantial investigator time). Other costs include the extraordinary requirement of one ethics committee in Victoria that an investigator from Queensland personally attend a 10-minute interview at which no substantive issues were raised. The ethical benefit of this investment must be questioned when the majority of changes required by committees have dealt with minor issues such as grammatical style that have little to do with patient protection. Inevitably, such directives are inconsistent across institutions. It is thus impossible to comply with all requests while maintaining a standard set of study documents. These problems are accentuated for the increasing number of Australian researchers relying on overseas funding. For example, regulatory authorities in the United States insist that all ethics committees reviewing US-funded projects involving humans must have US federal approval to do so. In our experience, few Australian hospital ethics committees have this approval. Therefore, in addition to fulfilling standard institutional ethics requirements, we have also had to help several committees go through the lengthy process of securing US accreditation simply to approve our study! For all the above reasons, we strongly believe that Australian researchers and patients would be best served by a single national ethics committee for large multicentre studies. This would also reduce the enormous burden currently placed on the individual committees. In the meantime, we thank Breen and Hacker2 for their reminder to institutional ethics committees that the NHMRC national statement "empowers ethics committees to minimise unnecessary duplication".

David C Whiteman · Penelope M Webb · David M Purdie · Adèle C Green

To the Editor: Both adult and paediatric hospital emergency departments (EDs) are subject to inappropriate use.1,2,3 Some families use the ED as a primary care provider,4,5 often claiming that they have no regular general practitioner.6 Such families may experience poorer overall health.7,8 We hypothesised that providing such families with information about GPs in their area and emphasising the benefits of having a GP responsible for their long term healthcare might: facilitate the establishment of ongoing relationships between patients and GPs; and encourage families to use GPs more as their primary source of care. We conducted a controlled trial (week-on, week-off randomisation) of families identified as having no regular GP who presented to the Royal Children's Hospital ED over four months. Information about the GPs interested in seeing children was located on a computer database. Medical staff were able to search for a GP whose surgery was close to the patient's street address. Families were provided with detailed information about the GP's practice (eg, opening times, languages spoken, etc). Parents were given a list of GPs and a map showing the locations of their surgeries, together with a letter of introduction; the families decided which GP they would attend. Families in the control group were just treated as usual. Families were then contacted after two months to see if they had visited a GP and whether regular contact had been established. Over the four months, 216 families were enrolled; 96 were allocated to the intervention group. Despite our active encouragement, the ED medical staff provided the intervention material to families in the intervention group on just 49% of occasions. We found that, two months after the initial ED visit, intervention-group families were no more likely to have established an ongoing relationship with a GP than control families (46 [38.3%] and 41 [42.7%], respectively; P = 0.5), irrespective of whether or not they received the intervention material. In summary, this single intervention was not sufficient to alter healthcare-seeking behaviour of families with no regular GP. It seems the motivation to obtain a GP lies with the family. Thus, it would seem necessary to design and deliver an intervention that addresses the beliefs of families about the roles of various facets of the healthcare system. With time and work pressures, ED medical staff may not be in the best position to provide such intervention.

Michael K Marks · Daniel Steinfort · Peter LJ Barnett

In reply: We agree with Johnson and Lassere about the value of longitudinal studies, especially clinical trials, in assessing healthcare benefits and costs. They are critical to informing clinical practice and health policy. If it were possible to conduct these studies across a wide variety of settings, representing the range of "real life" practice, then there would be little need for epidemiological modelling. However, this is not possible. Clinical trials (with or without cost components) will only ever be conducted over relatively short durations, on circumscribed populations and under tightly controlled conditions. A key, but often overlooked, issue is whether the results of studies are externally valid (generalisable). Indeed, the evidence base that dictates clinical practice and health policy should comprise data that are both internally and externally valid. We do not suggest that epidemiological modelling replace longitudinal studies (in fact, modelling depends critically on robust prospective data); rather, it complements these studies by providing a means to assess their external validity. We are also mindful of the limitations of epidemiological modelling, as outlined in our article,1 and acknowledge the importance of ensuring rigour in the methods. Our article dealt with generating the data needed for sound economic evaluation, by taking into account the long-term benefits, risks and costs of treatment strategies, and "real-life" health service conditions. This is distinct from the issue of whether "conditional listing" on the Pharmaceutical Benefits Scheme should be implemented for drugs that are yet to be proven cost-effective.

Danny Liew · John J McNeil · Anna Peeters · Stephen S Lim · Theo Vos

To the Editor: In a recent letter in the Journal,1 Rodger writes that breast screening is unlikely to affect overall mortality and notes that this "gives the lie to the conclusions of Olsen and Gøtzsche's overview, which are based only on overall mortality". English is not my first language, but according to my English–English dictionary "give the lie to" means either "to disprove" or "to accuse of lying", and a related adjective is "mendacious". In actual fact, however, in our Cochrane Review,2 we carefully analysed both breast cancer mortality and all-cancer mortality. We found breast cancer mortality to be an unreliable outcome that is biased in favour of screening. For deaths ascribed to any cancer, including breast cancer, we found a relative risk of 1.02 (95% CI, 0.95–1.10) for the two trials with medium-quality data,3-5 and a relative risk of 1.00 (95% CI, 0.91–1.10) for the only trial with poor-quality data that reported all-cancer mortality.6 If it were true that screening reduced breast-cancer mortality by 30%, as some Swedish researchers have claimed,7 then the expected relative risk for all-cancer mortality should not be greater than 0.95. These findings should raise concern rather than complacency. Another, recent indication that things are not what they purport to be is provided by the results of the large Two-County study. A Swedish overview of the randomised trials reported a 10% reduction (95% CI, 0.73–1.11; absolute reduction, 5.0/1000 to 4.5/1000) in breast-cancer mortality for one of the two counties,8 whereas the authors of the Two County study reported a 24% reduction (95% CI, 0.62–0.93; absolute reduction, 5.7/1000 to 4.3/1000),9 with the same type of statistics, within the same age group of women (40–74 years), and after a similar follow-up (1.2 v 1.3 million women-years). The conclusion in our Cochrane Review is: "The currently available reliable evidence does not show a survival benefit of mass screening for breast cancer (and the evidence is inconclusive for breast cancer mortality)." I would not have expected Rodger, as an editor of the Cochrane Breast Cancer Group that approved and published our Cochrane Review, to talk about "giving the lie" to our results.

Peter C Gøtzsche

In reply: In his letter, Gøtzsche is clearly under the misapprehension that, in using the figure of speech "gives the lie to", I am accusing him of lying. Nothing could be further from the truth. As he quotes, I applied that phrase to his conclusions. In my Australian Oxford Dictionary1 to "give the lie to" can mean — and it is this meaning that I was applying — "serve to show the falsity of a supposition". I was replying to Gough's response2 to my editorial3 on breast screening. He clearly showed that breast screening was unlikely to reduce overall mortality. I agree. I believe, therefore, that Olsen and Gøtzsche are wrong in supposing that analysis on the basis of breast cancer mortality is inappropriate, and that only overall mortality should be considered. Gough argues for this better than I can. Being an editor of the Cochrane Breast Cancer Group does not require me to accept every supposition or conclusion in a Cochrane review. The whole point in publishing a scientific paper — as part of the Cochrane Library or in a peer-reviewed journal — is to open it, after appropriate review, to public scrutiny, scientific comment and even criticism. Their Cochrane review4 has succeeded in achieving all of this.5 Lastly, I reiterate my comments in the editorial3 that mammographic breast screening detects breast cancers that are "smaller, less likely to involve nodes and, if node positive, more likely to involve fewer nodes." In other words, if the TNM (tumour–node–metastasis) system means anything, there is a better prognosis with such breast cancers than with those detected clinically. Perhaps Gøtzsche needs to add a clinical oncology perspective to his undoubted expertise in the finer details of trial methodology analysis.

Alan Rodger

Comment: The expression "give the lie to" has shifted its emphasis over the centuries, from the very direct "accuse (someone) of lying" to the much more abstract "show or imply (something) to be false". Some modern dictionaries, such as the Macquarie Dictionary (1997) and Merriam-Webster (2000), still give both meanings; others, such as the New Oxford Dictionary (1998), only the second. Large British and American databases, such as the British National Corpus, show that the phrase is usually used abstractly: one "gives the lie to" propaganda/a claim/an argument/a theory — whether in the context of academic discussion or political debate. The validity of an intellectual position is questioned, not the integrity of the person(s) associated with it. Yet, the simplicity of the phrase "give the lie to" probably gives the lie to the complexity of the challenge it expresses.

Pam Peters

To the Editor: The articles by Reid1 and Paterson,2 former bureaucratic leaders of the New South Wales and Victorian health systems, respectively, on the process for developing the 2003–2008 Australian Health Care Agreements (ACHAs) are disappointing. They offer few original conceptual insights or clear proposals. Reid's dream is that the 2003–2008 ACHAs will see "a new expression of national health policy on which funding decisions can be based". However, he presents only old ideas, such as "ACHAs will need to extend beyond public hospital issues to incorporate primary care", and, on the perennial cost-shifting between the two levels of government, "clearer lines of financial management of care and appropriate incentives are needed". Reid laments that the focus of all previous agreements has been "narrowly limited to one aspect of healthcare . . . the maintenance of universally accessible public hospital care free of charge". Paterson does propose something radical, and the core of his proposals is that "the payer must stand behind the patient and not between the patient and the provider". The way to Paterson's "outcome-enabled health system" is to "relieve the constraints that bind inputs and distort the 'production' system". Does he mean we need more doctors and nurses, or does he mean substitutes should perform some of their current activities? Patterson proposes more investment in "information and communications technology" to facilitate a gradual move to "patient-based funding". Does this mean capitation, medical savings accounts, or is he proposing non-insurable copayments? Whatever it means, there will be "no outcome-driven healthcare until the system recognises the whole patient", and this will only be achieved with "electronic patient record systems in routine and ubiquitous daily use by providers". Given their experience as senior health system administrators, it is a pity neither Reid nor Paterson provides any explicit suggestions that recognise the key factor that will determine the outcome of the ACHAs. This is the policy gridlock that any federal system almost inevitably imposes. A recent issue of the Journal of Health Politics, Policy and Law was devoted to health politics and policy in a federal system. The editor, Petersen, concludes with a view relevant to Australia: "You can love it, you can hate it, but . . . federalism thwarts uniformity and universalism, frustrates responsiveness and policy analysis, limits large scale innovation while churning more localized mills of idea generation and promotion, and offers a permanent employment plan for health policy researchers".3 Parts of Australian health arrangements certainly need an overhaul. An example is general practice. This sector, differently organised and financed, could deliver much more to the community, the rest of the healthcare system, the Federal Government and to general practitioners themselves. Change in this sector would not depend on improbable cooperation between levels of government, and would be more manageable than the multifarious whole-of-system reforms about which Reid and Paterson speculate.

William Coote

To the Editor: In their article describing an outbreak of Bordetella pertussis infection, Spearing and colleagues report an adult contact who was infected with B. pertussis and was treated with roxithromycin.1 In our experience, this is common practice in Australia, where roxithromycin is a frequently used macrolide antibiotic. We are currently preparing a systematic review (registered with the Cochrane acute infections group) of the effectiveness of antibiotic therapy for treating pertussis. We have found no studies of the effectiveness of roxithromycin for either treatment or contact prophylaxis for pertussis infection. B. pertussis is sensitive in vitro to roxithromycin but 2–4-fold less so than to erythromycin. While relying on the class effect of macrolides in eradicating B. pertussis and using roxithromycin in preference to erythromycin because of its lower side-effect profile may seem logical, there is no evidence to support this practice. In contrast, there is at least one study showing the efficacy of clarithromycin as an alternative to erythromycin for the treatment of pertussis.2

R John Massie · Sultan Altunaji · Renata Kukurozovic · Nigel Curtis

In reply: As we mentioned in our article, erythromycin is the drug of choice for treatment and prophylaxis of pertussis in people of all ages.1 Several statements can be made about the use of macrolides in this condition.2 Firstly, erythromycin has been shown to decrease the duration of illness when given early in pertussis infection and to eliminate Bordetella pertussis from the nasopharynx. Secondly, erythromycin therapy for index cases has been shown to reduce the rate of secondary cases of pertussis in households in uncontrolled studies.3 However, in a recent randomised placebo-controlled study, while erythromycin reduced the incidence of culture-positive pertussis in household contacts there was no reduction in respiratory symptoms.4 Thirdly, clarithromycin and azithromycin have been shown to be at least as effective as erythromycin in treating pertussis in two small comparative studies.5 As Massie et al point out, there are no clinical studies of the use of roxithromycin in this condition, and laboratory in-vitro sensitivity studies suggest roxithromycin may be inferior to erythromycin. Roxithromycin is the most widely used macrolide in Australia, but it requires a leap of faith and extrapolation to prescribe this drug for prophylaxis or treatment of pertussis. In the one case in our series where roxithromycin was prescribed, this was not our decision. We would have recommended erythromycin on the basis of the available clinical evidence. Vaccination is preferable to antibiotic prophylaxis for long-term control of pertussis.

Joseph G McCormack · Natalie M Spearing · Robert L Horvath

To the Editor: Van Der Weyden has asked a provocative question about the relevance of what he calls "medicine's homage to health".1 In so doing, he pays his own homage to a world where boundaries are sharp and healing becomes reduced to a matter of applying "bioscience to matters of mind and body". While I daresay many editors of biomedical journals would share his view, he is only highlighting an age-old tension. Indeed, Crookshank wrote in 1926 about the Ancient Greek schools of Cos and Cnidus, and of their debate about doctrines of the natural/descriptive and the conventional/academic approaches to medical knowledge.2 The Coans promoted the importance of the influence of the individual and society on the presentation of illness, and managed illness holistically with regimens oriented towards the needs of the individual, while the Cnidans oriented their practice around the distinctions between diseases, treating specific diseases with specific remedies.3 These differing approaches to illness and disease highlight the conflict between cybernetic and linear thought which has underpinned medical history ever since. It is rarely understood that the linear doctrine of biomedicine has only become dominant in the last century or so. Medicine's isolation from contemporary scientific thought explains our long delay in confronting the challenges of systems thinking — challenges that the basic and applied sciences took on in the early years of the 20th century.4 Our contemporary error is in assuming that the accumulation of data that passes for modern medical knowledge is sufficient to deal with the tasks of medicine. The disquiet expressed by Van Der Weyden is only a symptom of the continuing inability of a mechanistic view of medicine to deal with caring for patients in the real world.

Grant M Russell

In reply: Russell implies that my recent column on the boundaries of medicine1 conveys the premise that medicine is a defined and discrete discipline underpinned by biomedical science. So it did, but the contention is not mine, it is that of Seldin.2 However, no one would deny that biomedical knowledge is the best developed and most powerful component of modern medicine; nor would they deny that there are other intrinsic components, such as the art of caring.3 My argument is that the boundaries between medicine and health have become blurred with modernity's overwhelming homage to health.4 Furthermore, the very nature of health is difficult to define, shaped, as it is, by variable psychological, socioeconomic and cultural factors. If, in fact, health is the World Health Organization's utopian "the complete [my emphasis] physical, psychological and social well being", then most of us are unhealthy.5 My purpose was to question the boundaries of medicine and health. In particular, to question where the responsibility for attaining the modern Shangri-La of health rests — with the individual, society or medicine?

Martin B Van Der Weyden

To the Editor: Interpretation of the report describing the first death attributed to the Irukandji syndrome should be tempered by the fact that significant unstated assumptions have been made in attributing the cause of death to a jellyfish.1 While envenomation by a jellyfish remains the likely diagnostic possibility, no evidence is presented that unequivocally confirms a jellyfish as the lethal agent. Several methods could have been used to support or confirm the diagnosis of jellyfish envenomation, including sampling of nematocysts from the victim's skin (before or after death), jellyfish capture, or reports of other similar, but less severe, stings from the same beach around the time the victim was stung. In severe jellyfish envenomation, attempts are often made to harvest nematocysts from patients' skin, most commonly by skin scraping or by sticky tape sampling.2 Recovered nematocysts may help to identify the species, and confirm the diagnosis.3 Although successful nematocyst recovery is uncommon in Irukandji syndrome, it is disappointing that "no attempt was made to sample nematocysts"1 given the relative simplicity of the procedure and the importance of this case. The authors state that "no sting site was clearly delineated",1 but then go on to say that there were, in fact, areas of "skin flushing and intermittent diaphoresis"1 over a significant period of time. Sticky tape sampling of these areas may have yielded nematocysts, allowing positive species identification. Postmortem skin sections have also been employed in Chironex fleckeri fatalities, and have shown nematocyst barbs on the victim's skin.4 Postmortem examination may also have revealed other contributing factors. I am particularly interested in the assertion that almost every Irukandji syndrome patient in the Whitsundays develops a "rise in cardiac troponin levels".1 In fact, the cited article makes no mention of troponin, simply stating that CK-MB (creatine kinase isoenzyme) levels "can be abnormal",5 and that "some severe cases [of Irukandji syndrome] may have a CK-MB [level] well above the normal range".5 Many aspects of the diagnosis and treatment of jellyfish envenoming remain controversial. Accurate reporting of unusual cases is thus of the utmost importance.

Paul M Bailey

To the Editor: In their Notable Case on jellyfish envenomation causing Irukandji syndrome,1 Fenner and Hadok suggest that similar deaths may have occurred in the past, with the relationship to Irukandji syndrome not being recognised. However, their call for urgent research into developing an antivenom needs to be based on a clear understanding about the risk of death. Unfortunately, they have not presented data that clearly establish causation or support their conclusions about treatment and the need for further research. While they reported a history that supports envenomation, there was no confirmation by detection of nematocysts or autopsy to examine for other causes of death. An alternative explanation could be that the patient was overcoagulated and died from complications of an intracerebral haemorrhage. While it is clear that blood pressure must be monitored, to suggest that it must be treated with phentolamine is not supported by this case report. Clearly, supportive management and, in particular, the optimal treatment of cardiovascular complications needs to be defined and may obviate the need for antivenom.

Andrew H Dawson

In reply: While overcoagulation causing intracerebral haemorrhage could have caused the death of the patient we described,1 he was normotensive until developing signs and symptoms of Irukandji syndrome some 20 minutes after being stung. The Irukandji syndrome is, and always has been, a clinical diagnosis only. Biochemical and pathological test results become abnormal later, but are not diagnostic — actual cause and effect have been described only once, with the experiment unlikely to be repeated!2 Nematocyst studies, while established for Chironex fleckeri,3,4 have never identified species associated with Irukandji syndrome, except Carukia barnesi, which appears to occur in the Cairns area only. Other species probably cause the more severe syndrome seen in the Whitsundays and on the Great Barrier Reef, where these deaths occurred.5 One of us (P J F) is possibly the only person to have captured specimens likely responsible for causing Irukandji syndrome from the Whitsundays, and the species remain unidentified, as they are a new species and not described to date. Also, when the moribund patient was admitted, no obvious sting site was visible, and a negative skin scraping would not rule out a jellyfish sting. Phentolamine has previously proved effective for relieving distressing autonomic symptoms,6 and not just for cardiovascular complications, although it appeared ineffective at the lower doses used in our patient. However, nothing appears to prevent toxic cardiac dilatation occasionally occurring later in the syndrome.7 Further research is currently under way. Antivenom development may prevent some (possibly all) major symptoms of Irukandji syndrome. However, production is impossible until sufficient specimens of all species (some six to 10) causing the syndrome are caught and their venom assessed. Such advances are many years away and may never be achieved with current poor levels of funding. Cardiac markers for jellyfish envenomation have previously been identified.5,8 Since 1999 troponin level has replaced creatine kinase isoenzyme (CK-MB) level, and both are invariably raised in patients stung by the Whitsunday jellyfish. Thus, the words "cardiac markers" should have been used in the article and for not doing so I apologise. Despite C. barnesi stings being common at north Cairns beaches, it has taken six years of dragging the beaches, with nets to catch jellyfish of this species. The thought of trying to catch a 12 mm jellyfish that makes erratic and irregular appearances in several hundred square kilometres of ocean around the Whitsunday Islands is totally daunting, but the possibility is being assessed. Such a venture will depend on funding becoming available. Other stings were reported in the area at the time of our patient's death and are well known at the resorts where people who have been stung in surrounding areas are taken for treatment. However, stings remain erratic; they have no predictable patterns of appearance, and unfortunately prophecy is currently impossible.

Peter J Fenner · John C Hadok

To the Editor: The article by Tan and Fitzgerald1 raises numerous concerns. The authors report that their recommended personal protective equipment (PPE) conforms to standards "in a hospital environment where the chemical vapour concentration will not be high". At the same time, the authors acknowledge data indicating most patients from a disaster will present to the local hospital by private transport (ie, without triage, decontamination, or prehospital care). These two considerations are incompatible and further ignore the possibility of the hospital as a direct terrorist target. The authors' assertion that their three decontamination lines "allow mass casualties, as well as trolleys and equipment, to be decontaminated quickly, efficiently, and in an orderly fashion" is simply not evidence based. Of greater concern, the authors report "major considerations were policies and plans [referring to the hospital External Disaster Committee] and the emergency department response". Although this bottom-up approach to disaster planning is typical, it pays inadequate attention to interdisciplinary issues of proper hazard identification and management, environmental health, syndromic surveillance, and field outbreak investigation. Readers seeking robust emergency department templates are better referred to other sources for guidance.2,3 Of greatest concern, the authors report "our recommendations are similar to systems in the US and Israel, but much less intensive, as the threat of a terrorist attack here is perceived to be much lower". The three references cited for that statement date back to 1994, with none more recent than 1999. Moreover, the logic of the unreferenced threat assertion confuses hazard and risk. Although the absolute probability of a given hazard may be low, the risk attending that hazard encompasses vulnerability of the exposed population. With weapons of mass destruction, the conditional probability of catastrophic public health consequences is high — one event is the only number you will ever need. The current public health context of chemical–biological–radiological (CBR) incident management in Victoria is one of limited experience, performance improvement indicators, and budgetary support from public health authorities. Public health is at risk when authorities report that "faced with dozens of requests each day to attend sites to assess white powder, the stretch capacity did not exist and nor should it".4 As a result, the leading trauma centre in Australia extracts $20 000 from its existing operations budget to discharge its CBR responsibilities. This is not good enough. Nevertheless, the authors deserve credit for their initiative. Until cross-trained and disaster-experienced healthcare authorities reprioritise, this article shows the reader an excellent way to play a very weak hand.

David A Bradt

To the Editor: Tan and Fitzgerald's template for emergency department response to chemical–biological–radiological hazards appears to be based on a dubious assumption of a low level of risk.1 A recent report details exposure of emergency department staff to potentially fatal secondary contamination during a hazardous materials incident, highlighting the need for staff to have the appropriate training and equipment to deal with these events.2 I believe the level of Personal Protective Equipment (PPE) proposed by Tan and Fitzgerald is inadequate. The "facemask with filter" they describe is classified as Level C respiratory protection, and this level only conforms to the Australian Standard (for PPE) when the identity of the chemical and its vapour concentration are known, and when these do not exceed the filtering capacity of the particular filter mask being used.3,4 In the initial confusion of a hazardous materials incident, the identity of the chemical agent and its vapour will not be known. There may even be misinformation: during the 1995 Tokyo sarin attack, for example, initial advice to hospitals by the Tokyo fire service was that the incident was "a gas explosion in the Tokyo subway".5 Emergency department staff must be able to respond before the nature and severity of the chemical hazard can be determined. The only respiratory protection which conforms with the Australian Standard for PPE when the nature and severity of the chemical hazard has not been determined is the supplied gas respirator with full face shield of Level A (an encapsulating suit and self-contained breathing apparatus) or Level B (a non-encapsulating suit with self-contained breathing apparatus or a full face respirator on a gas line).3,4 Confronted with a hazardous materials emergency, potentially involving very toxic chemicals, emergency department staff need to have complete confidence in their own protection. This is only possible with the use of supplied gas respirators (Level A or B PPE), which provide complete respiratory protection. An additional problem with Level C air-purifying respirators is that their performance may be adversely affected by water ingress into the filter, which could occur during the decontamination procedures described in the template. Tan and Fitzgerald also propose having a clerk don PPE and enter the contaminated zone. Any stationery taken into a contaminated area would have to be decontaminated before being taken out to a "clean" area, and it is not clear what a clerk would add to the initial response within a contaminated zone.

Antony Nocera

In reply: We thank Bradt for his interest in our article and acknowledge his expertise in this field, which he has gathered in the United States and other countries. We also thank Nocera for his interest in our article. Our aim was to stimulate interest among the medical community in chemical–biological–radiological (CBR) response. The interdisciplinary issues mentioned by Bradt were mentioned in our article, but not in detail because of space limitations. Our personal protective equipment (PPE) conforms to Australian standards1-3 and the three decontamination lines are in keeping with other institutions. We are not aware of any simple decontamination system which, evidence-based, is superior. The choice of PPE in the ideal situation would be one that would provide adequate protection in all situations with a minimal amount of training, maintenance and expense. Nocera is correct in stating that the respiratory protection in an unidentified chemical hazard is Level A or B. These PPEs are expensive, bulky (which results in poor manual dexterity), and their use requires specialised training. The amount of chemical present on a victim surviving long enough to self-present to an emergency department is significantly less than that involved at the site of the incident. Therefore, the level of protection required for hospital staff would be less than that required by emergency rescue workers. Our PPEs were supplied by the Victorian health authorities. It is more important for staff to be familiar with their PPEs and for hospitals to have a CBR response that is regularly practised than having excessive protection that is limited to personnel who have undergone specialised training. The role of the clerk is to take patient details. These are radioed to staff in the hospital to help identify and correctly label patients, which is very important in mass casualty situations. Recent experience has demonstrated that terrorist acts are a worldwide phenomenon, and Australians are potential targets. This underlines the need for comprehensive training and maintenance of hospitals' CBR response. Since publication of the article, Victorian health authorities have reprioritised, and we therefore feel we have achieved the aims of our article.

Gim A Tan · Mark C B Fitzgerald

To the Editor: White emphasised problems that can arise if medical decisions are overly reliant on the results of laboratory tests.1 He relates the case of a patient who, because of a peculiarity of her immunology, consistently produced a false-positive test result. In probabilistic language, the issue is are there risks of both random errors and patient-specific errors? It is important to distinguish between them. Suppose a test has a false-positive rate of 10%. If this is truly random error, the probability of two false-positive results in the same person is 1%, and the probability of three false-positive results is tiny. But if it is due to there being 10% of healthy people for whom the test is invalid and who consistently give a positive result, the probability of two false-positive results in the same person is 10%, and the probability of three false-positive results is 10%! Major textbooks of medicine have excellent chapters on decision-making. These warn about limitations of sensitivity and specificity (eg, that data from the general population may not apply to people who have tested positive in screening). But, other than this, little is said about reasons for errors in testing, and the consequences for how sensitivity and specificity should be used. In most cases, the impression given is that errors occur completely randomly. However, it appears that White's example, in which repeated testing led to repeated errors, is not unique. Lee2 writes as follows: "Suppose a low-risk patient has an abnormal lung ventilation–perfusion scan. Obtaining that same test result over and over will not truly raise that patient's probability of coronary disease further and further." Perhaps Goldman3 had something similar in mind when writing, "It may be quite difficult to distinguish random laboratory errors from test results that might be falsely positive or negative because of coexistence of a process that can affect the test". Lists of possible reasons4,5 for errors include both short-acting (eg, distracting external noise, and biochemical effects of foods recently eaten) and long-term (eg, physical handicaps, and demographic factors) influences. I wonder if information about tests should routinely include separate random and patient-specific components of sensitivity and specificity. For example, it might be stated that a false-positive rate of 15% arises from 10% random errors and 5% patient-specific factors, or that a false-negative rate of 10% arises from 3% random errors and 7% patient-specific factors. This is the conclusion I have been led to by White's article.

T Paul Hutchinson

In reply: Many factors potentially contribute to error in generating a diagnostic test result, and include random pre-analytical errors arising from patient preparation and specimen collection, random errors associated with the act of measurement, and systematic errors caused by, for example, drug interference. Tested individuals may also harbour an interfering substance, such as a drug or immunoglobulin. The theoretical and practical description of these components of test error is generally well understood and documented by laboratories, and the basics of test error and diagnostic sensitivity and specificity are taught in medical schools. However, I think trying to apply probability data to a test result for a specific patient is of limited value to the treating doctor. The commoditisation and automation of much of pathology testing contributes to a perception that tests are 100% reliable, and there is also a perhaps related decline in communication between requester and provider. Most tests have limitations, many inconsequential, some important and patient-specific. Although Hutchinson draws a valid conclusion, I hope readers also concluded that communication with diagnostic laboratories remains important for safe patient care, and that test results still need to be interpreted in the context of other clinical information about a patient, and not accepted without question.

Graham H White

To the Editor: In reviewing the CD-ROM containing an integrated set of Therapeutic guidelines (eTG complete),1 Mann noted that, although utility was improved, the cost was high ($220 for a first user, and $110 for each subsequent user, compared with $264 for a set of the printed volumes which could be shared within a practice).2 Unfortunately, the cost of distilling evidence-based knowledge is also high, especially for publications that require regular review and update. In addition, there are extra costs involved in electronic conversion (text to HTML), reformatting material to fit computer screens, creating expandable and collapsible tables of contents (JavaScript) and searchable indexes. There is also a need to ensure that the electronic version works with a range of common operating systems (Windows, Macintosh and Linux) and with varied and ever-changing web browsers (eg, Netscape and Internet Explorer). In short, producing electronic versions of print publications requires additional resources, stringent quality control and considerable pre-release testing. In the near future, guideline producers face an additional technological challenge: the need to move from "static" HTML-tagged documents to "active" XML-tagged documents, in which the data elements and underlying logic are machine-readable and thus capable of interacting with the health worker and the emerging electronic medical record (computerised decision support). These developments (and Mann's concerns) raise the question as to who should pay for best-practice guidelines, including the evolving electronic conversion. Currently, the Australian government appears to lack a coherent approach to the provision of health information, at least with respect to best-practice clinical guidelines. For example, the Federal Government currently funds revision, production and distribution of The Australian immunisation handbook3 and the Manual of use and interpretation of pathology tests.4 More recently, the government has commendably funded a national subscription to the Cochrane Library. However, the government does not fund other equally valuable resources, such as the Australian medicines handbook5 or Therapeutic guidelines.1 The latter survive solely on a user-pays market model. It can be argued that it is time that the government adopted a more even-handed approach to stimulate the uptake of national best-practice guidelines and related services, perhaps by reimbursing health workers who subscribed to such services with practice incentive payments. This would lower the cost of practitioners acquiring national information resources, preserve the market model and assist guideline producers to meet the ongoing challenges of electronic conversion by improving their income stream.

Ken J Harvey

To the Editor: I was pleased to read the article on Ross River virus (RRV) disease by Mylonas and colleagues,1 because it included information on the cost of the disease. This makes it much easier to do something practical from a government and health economics perspective about the problem of RRV. The cost of $1018 per patient, including costs of negative tests looking for cases, sums to a total estimated cost to the nation of $5 million per annum (based on the reported average of 5000 cases per year in the study by Harley and colleagues2). Of note, $567 was spent per patient on diagnostic tests (56% of the total cost per patient), while the authors noted that in many cases the condition was self-limiting. I begin to wonder what is the use of spending $567 per patient diagnosed to prove a largely self-limiting condition that is treated symptomatically? Banning RRV testing could save $2.8 million per annum, which would be immediately available for mosquito control measures, and perhaps vaccine research, to reduce the burden of RRV disease. We could model the impact of a vaccination program — vaccine development cost, vaccine unit production cost, vaccine delivery, population target, and savings in disease prevented — to determine whether funding of vaccine research is worthwhile. The point is that without economic data we cannot make sensible "evidence-based" clinical management decisions. We are trapped in a scientific paradigm, and the health system implements unpopular cost–control interventions because it needs to control spiralling costs. We clinicians need to do better. If other researchers would follow the lead of Mylonas and colleagues and explore the economics of their subject, we would be able to make more rational choices about healthcare. It is up to clinicians to understand the economic agenda and suggest interventions that make both economic and clinical sense. We can then begin to make more efficient and rational use of our health dollars, relieving stress on a stressed system.

Ian R Cheong