Letters

To the Editor: A number of commentaries and articles have been published recently about the ability of doctors working in primary care skin cancer clinics to diagnose and manage skin cancer.1-3 However, limited information has been published comparing the patient populations that attend the different service providers (ie, “traditional” general practitioners versus doctors at skin cancer clinics). In 2005, we conducted a large population-based survey of Queensland residents aged 20–75 years to examine the prevalence of behavioural risk factors for cancer and current cancer screening practices.4 Using data from our study, we examined the prevalence of clinical skin examination and identified factors associated with choice of service provider. A total of 9419 respondents completed the interviews (response rate, 45.6%). Complete data for this analysis were available for 5499 of the respondents, of whom 48.2% were men. Thirty per cent of respondents reported they had had a general check of all or nearly all of their body in the previous 12 months. Factors associated with an increased likelihood of having a whole-body skin examination in the previous 12 months included being male (odds ratio [OR], 1.15 [95% CI, 1.00–1.31]), being 60–75 years of age (reference group, 20–39 years) (OR, 1.73 [95% CI, 1.45–2.07]) and having an annual gross income of ≥ $60 000 (reference group, < $20 000 annual gross income) (OR, 1.42 [95% CI, 1.18–1.71]). The strongest predictors were a self-reported history of melanoma (OR, 2.68 [95% CI, 2.01–3.57]) or non-melanoma skin cancer (OR, 2.01 [95% CI, 1.65–2.45]). No associations were seen between choice of service provider and any sociodemographic variables, including sex and age group. Additionally, skin cancer risk factors (such as having highly sensitive skin or a history of melanoma) did not make respondents any more or less likely to attend either a GP or a skin cancer clinic doctor. Various reasons were given by respondents for their choice of service provider (Box). Skin cancer clinics appeared to be chosen primarily because they offered bulk-billing or because respondents just wanted a general skin check. Traditional GPs were more likely to be chosen for convenience or because of concern about a specific spot or mole. Skin cancer is a major public health issue, and the provision of adequate and appropriate clinical services is a continuing and growing challenge. We found that a significant proportion of the Queensland population had undergone a whole-body skin examination by a doctor within the previous 12 months, and that those attending appeared to be the group most at risk of developing skin cancer. We did not find any significant differences in the profiles of those who chose a skin cancer clinic or a general practice for their skin examination. Reasons given for choice of service provider* by 2895 respondents who had had some type of skin check in the previous 12 months† * General practitioner or skin cancer clinic doctor. † Percentages do not total 100 due to multiple responses.

Philippa H Youl · Peter D Coxeter · David C Whiteman · Joanne F Aitken

To the Editor: Extended-spectrum-β-lactamases (ESBLs) are enzymes capable of hydrolysing penicillins, broad-spectrum cephalosporins and monobactams. Worldwide, ESBL-producing organisms are posing an increasing challenge for empirical antibiotic use and infection control. We recently carried out a review of microbiological isolates from clinical specimens taken from 2003 to 2007 at the Alfred Hospital, Melbourne. From 15 917 gram-negative bacilli, we identified 234 ESBL-producing organisms (1.5% of isolates) using double-disk synergy testing. Over the 5-year period, we noted three apparent changes in ESBL epidemiology relating to Escherichia coli isolates. First, E. coli became the most frequent organism in which ESBL production was observed, making up 55.6% of all ESBL-producing organisms in 2007 (up from 23.5% in 2003) (P = 0.03). Second, while the total number of E. coli isolates remained essentially constant over the study period, there was an increase in the proportion of E. coli isolates found to produce ESBLs: 1.8% of E. coli isolates in 2007 compared with 0.36% in 2003 (P < 0.001). The third and perhaps most striking change was in the epidemiology of ESBL-producing E. coli. In 2003, ESBL-producing E. coli infections were largely hospital-acquired, with 87.5% of isolates acquired after 48 hours in hospital or after a hospital admission in the previous 12 months. However, by 2007, ESBL-producing E. coli infections were found to be predominantly community-acquired, making up 62.2% of ESBL-producing E. coli isolates (P = 0.01). The increased proportion of community-acquired infections occurred despite a parallel increase in the frequency of hospital-acquired ESBL-producing E. coli infections (Box). Community-onset infections with ESBL-producing organisms have become increasingly recognised as important clinical entities.1 ESBL-producing E. coli bacteraemia is associated with higher mortality than bacteraemia caused by non-ESBL-producing organisms,2 a finding that has also been specifically demonstrated in the setting of community-acquired infections.3 Although local epidemiological data for infections with ESBL-producing organisms are not readily available, it appears that rates of community-associated infection vary greatly worldwide, with some regions of China reporting rates of ESBL-producing E. coli as high as 34% of all isolates.4 Although our study was limited by being a single-centre review, our findings are consistent with the emergence of multiresistant Enterobacteriaceae noted in Australian surveillance reports.5 It is not clear whether the change in our ESBL-producing isolates is reflective of local resistance patterns, or perhaps associated with travel to regions where ESBL-producing E. coli are known to be prevalent. Corroboration of these changes in other regions will be important for assessing the magnitude of this issue and responding appropriately, particularly in considering empirical antibiotic therapy for community-acquired gram-negative infections. Hospital-acquired (HA) versus community-acquired (CA) ESBL-producing E. coli isolates, Alfred Hospital, 2003–2007 E. coli = Escherichia coli. ESBL = extended-spectrum-β-lactamase.

Justin T Denholm · Michael Huysmans · Denis Spelman

To the Editor: Reported adverse reactions to food, which are common in many developed countries, can be produced by a wide variety of mechanisms. However, a low proportion of these are true food allergies.1 Recent Australian data show an increase in hospital presentations for food-induced anaphylaxis,2,3 but there are no Australian population data on the prevalence of either food allergies or adverse reactions to foods. Waiting lists for allergy services continue to remain long, and it is not known whether this is due to an increase in the prevalence of true food allergy or simply an increase in perceived food allergy. In October 2007, we undertook an Internet-based survey to assess the prevalence of self-reported perceived food allergies in Australian households. Participants were drawn from a consumer research panel of 8385 people (solicited through Internet-based marketing) who were proportionally representative of the Australian population with respect to age, sex and state. Cohort members were invited to participate in an Internet-based “health survey”, with no mention of food allergy during recruitment. Within 24 hours we had 1386 respondents and the required quota of participants was deemed to have been reached. Of the 1386 respondents, 406 (29.3%) reported at least one household member who believed he or she had a food allergy (Box). Of these, 250 (61.6%) reported at least one doctor-diagnosed allergy and 56 (13.8%) reported that the allergy was allergist-diagnosed. In addition, 42 respondents (3.0% of all respondents) reported that the person with the allergy had an EpiPen (Dey, LP, Napa, Calif, USA). Although there will be some selection bias in our sample because people without Internet access could not be sampled, we believe this bias is likely to be low, as at least 64% of the Australian population currently has home access to the Internet.4 Our questionnaire did not attempt to distinguish between true food allergy, sensitisation to foods, food intolerance or adverse reactions to food, although the majority of allergies had been diagnosed by a doctor or allergist, and foods such as peanut are more likely to be associated with allergies than intolerances. The high rate of perceived allergy to fruit and vegetables in an Australian context was surprising, although allergic reactions to fruit and vegetables are well documented.5 This may reflect either a rising prevalence of birch-pollen syndrome, as has been reported in Europe,6 or a community poorly informed about the true nature of food allergy reactions. Our data add to the evidence that there may be an increasing, largely unmet demand for health care information for patients with adverse reactions to food, including allergies. More formal evaluation should be undertaken to assess the type and prevalence of food allergy in the Australian context in order to facilitate future workforce planning and better community education. Proportion of Australian households in which at least one member believed they had a food allergy, and the individual foods nominated* Incidence of allergy (%) Food All households surveyed Households with perceived food allergy Cows milk 8.3 28.3 Peanut 6.9 23.4 Shellfish 5.9 20.2 Wheat 5.6 19.2 Fruit 5.3 20.9 Egg 3.4 11.6 Vegetables 2.7 6.7 Fish 2.5 8.4 Tree nuts 2.2 7.4 Soy 1.7 5.7 Other 6.3 21.4 * 40% had more than one food allergy.

Katrina J Allen · Jennifer J Koplin · Carmen Gould · Nicholas J Osborne

To the Editor: So a large proportion of women experience pelvic pain, often over years. What’s new? Of course they do. Pitts and colleagues1 fail to mention that virtually every normal, physiological event that occurs within a woman’s pelvis is associated with pain. Clearly, such pains vary in duration and intensity and are associated with events such as ovulation, menstruation, pregnancy, labour and childbirth. We men have it easy by comparison. But to conclude by saying that “only about a third of women who experience chronic pelvic pain seek advice from a health professional” gives the impression the authors are trying to medicalise yet another essentially normal event. One can get into long, philosophical discussions as to why such normal events should be so painful, but it remains a fact. I have spent my career urging general practitioners and fellow specialists to avoid surgery and “silver bullets” in most cases of pelvic pain and follow a conservative approach.2 It would have been more helpful if the authors had gone on to discuss what type of pain is suffered by what type of woman and who is treated by what type of doctor. This truly would have assisted in determining who would benefit from the attention of a health professional and who would not.

Jules S Black

To the Editor: We read the recent article by Pitts and colleagues1 with interest, given the rising trend of diagnosed chronic pelvic pain (CPP) in Australian women. The article identified three types of CPP, but did not differentiate pain into the two major categories of nociceptive (visceral and somatic) and neuropathic. In pain management settings it is considered essential, where possible, to make this differentiation, as it significantly alters management strategies, particularly in relation to medication. While the true incidence of neuropathic pain is unknown, it is believed to be underdiagnosed and inadequately treated. A 2008 French study based on a nationwide postal survey revealed a 6.9% prevalence of neuropathic pain in the general population, with 5.1% of respondents reporting pain levels as moderate to severe.2 Neuropathic pain results from damage to the nervous system. Specifically, this can be from damage to, or pathological changes in, the axons of peripheral nerves or from damage to the central nervous system, probably as a result of deafferentation. This is the process whereby neurones in the central nervous system lose their accustomed afferent input, either from a peripheral nerve or from an ascending sensory tract. Furthermore, neuropathic pain can and does cross neuroanatomical boundaries, often presenting viscerally as referred pain and eliciting pain descriptors such as burning, shooting, stabbing, and searing. For this reason, CPP is often wrongly assumed to be visceral in origin.3 In such cases, awareness that CPP may in fact be neuropathic may avoid inappropriate surgical interventions. Moreover, an association between CPP and neuropathy has been demonstrated in studies of sacral nerve and percutaneous tibial nerve stimulation in women presenting with CPP.4,5 Differential diagnosis of pain of neuropathic origin has been shown to be pertinent for the accurate implementation of pain management strategies.6 Therefore, we suggest that future studies on the epidemiology and/or prevalence of pain include tools to determine the proportion of pain of neuropathic, nociceptive and mixed origin. There are a number of tools available, including questionnaires such as painDETECT, DN4 (Douleur Neuropathique en 4), LANSS (Leeds Assessment of Neuropathic Symptoms and Signs) and NPS (Neuropathic Pain Scale). Some of these, such as the self-assessed LANSS (S-LANSS), do not require clinical examination and thus can be worked into population-based questionnaires. The ability to identify neuropathic pain should lead to individualised treatment, resulting in improved pain control for patients with CPP.

David Vivian · Adele Barnard

In reply: We are pleased to see our article about chronic pelvic pain in Australian women has provoked interest.1 Black’s suggestion that virtually every normal physiological event that occurs within a woman’s pelvis is associated with pain is surprising, and not supported by our evidence. Of the women in our sample, 23% were totally pain free, and most of the chronic pelvic pain reported was mild. A parallel study showed that men also suffered chronic pelvic pain — a smaller proportion than women, but still significant.2 We are not medicalising normal events; rather, we are alerting general practitioners to the normal range of pelvic pain experience to help them assess its clinical significance. A GP who says to a female patient “it’s normal, love, just grin and bear it” denies the psychosocial complexity of her experience. Vivian and Barnard suggest we might have differentiated between two major types of pain, nociceptive and neuropathic. It would not be practical to collect this information in a broad survey on sexual and reproductive health. Certainly, a study of the prevalence of neuropathic pain in the Australian population that mirrors recent studies overseas would be informative. However, our study concerned pelvic pain only. The pelvis is not a common site for neuropathic pain.3

Marian K Pitts · Jason A Ferris · Anthony M Smith · Julia M Shelley · Juliet Richters

To the Editor: With ageing and population growth, the number of people with disabilities is rising. By 2010, it is projected that 1.5 million Australians will have severe or profound limitations to core activity in the areas of self-care, mobility and/or communication.1 In 1976, the United Nations proclaimed that 1981 would be the International Year of Disabled Persons, with the aim of promoting the rights of all people with disabilities to participate fully in society and to enjoy living conditions equal to those of other citizens.2 Yet today, more than 30 years later, even in a country as affluent as Australia, families struggle to care for children with disabilities, ageing parents look after adult children, and middle-aged people have their lives put on hold while caring for partners or parents with disabilities. Every aspect of life is affected, including physical and psychological health, employment prospects, and the ability to take holidays and maintain social relationships. While the physical and emotional demands can never be underestimated, an added pressure is the huge financial impact on every family affected by disability. Homes are mortgaged, endless time is spent securing funds for essential equipment, and some individuals reluctantly share their plight with the media. The time is right to reform the disability sector. We need to shift from the current crisis-driven welfare approach to a planned and fully funded national disability insurance scheme (NDIS) that will underwrite sustained, significant, long-term improvements in meeting the needs of people with disabilities. The models for an NDIS already exist, with fully funded no-fault insurance schemes for people injured in the workplace and in car accidents in several states and territories. At a forum in October 2007, National Disability Services (the national industry association for disability services) supported the extension of coverage provided by road accident insurance schemes to new groups, a proposal that was endorsed at the recent Australia 2020 Summit.3 The NDIS could be funded from a number of sources, including a supplement to the Medicare levy, third-party car insurance or workplace insurance. The costs would be modest and the benefits significant, because an NDIS would be more equitable than current arrangements (whereby a few achieve multimillion-dollar payouts and many others in similar situations receive nothing), and would enable people with disabilities and their families to be in control, make choices and plan their lives with confidence.

Dinah S Reddihough · Bruce P Bonyhady

To the Editor: Impacts of alcohol consumption on health and wellbeing in remote Indigenous communities are well documented. In response, governments have applied supply and demand reduction programs, including the “Meeting Challenges, Making Choices” program, which has lowered the rate of serious injury. Although a pattern of heavy, episodic drinking has been documented, the nature of physical dependence in relation to acute alcohol withdrawal syndrome is uncertain. We report the results of sudden, temporary removal of alcohol in a small Indigenous community. In a remote Queensland Indigenous community (population, 1021) with one licensed premises, patterns of extreme drinking (30 standard drinks per session) are commonly seen around paydays. In 2008, the Queensland Government withdrew the sole liquor trading licence with 72 hours’ notice because of a breach of licensing laws. The licence was subsequently renewed after several months. During this time, there was no significant access to alternative (illegal) sources of alcohol within the community, as the prohibition against bringing alcohol into the community (initiated in 2003) was strictly enforced by police. Health services in this community comprised a primary health care centre (PHCC) with Queensland Health resident nursing staff, Royal Flying Doctor Service medical staff on weekdays (the principal doctor was S A M), and visiting specialists provided by both organisations, including a psychiatrist (E H). When the closure was being arranged, Queensland Government authorities requested that these health providers establish a process to treat any patients who developed acute alcohol withdrawal syndrome; this was monitored by S A M. Four weeks after the sudden cessation of alcohol availability, PHCC staff did not notice any outmigration of regular drinkers, and no patients presented with acute alcohol withdrawal syndrome. These findings are consistent with the anecdotal experience of E H, who has not encountered any cases of withdrawal delirium in this community over the past 16 years. By contrast, in recent years E H has observed several cases of withdrawal symptoms from cannabis use in this community, as seen previously in remote Northern Territory Indigenous communities.5 Our results suggest that people can develop physiological or psychological tolerance for heavy episodic drinking, which may be a function of adaptation to the intermittent nature of financial resources. This finding removes a potential health-related impediment preventing governments from considering sudden cessation of legal alcohol supply in these or similar environments.

Stephen A Margolis · Valmae A Ypinazar · Alan R Clough · Ernest Hunter

To the Editor: Urinary stones are very common, with a cumulative lifetime incidence of 5%–15% and a recurrence rate of about 50%.1 Many new treatment techniques have been developed, but availability, particularly in public hospitals, is variable. The Greater Metropolitan Clinical Taskforce2 assessed patterns of treatment in patients requiring urological consultation who presented to the emergency departments (EDs) of 12 New South Wales public teaching hospitals in major centres that had a specialty urology registrar. Between February and September 2007, the urology registrar or specialist completed a survey on consecutive patients presenting with urolithiasis who agreed to participate. The survey contained questions on patient demographics, the position and size of the stone, and the preferred treatment option. One of us (J W H M) conducted a telephone interview with each patient to obtain details of treatment, and follow-up interviews at 3-monthly intervals (until treatment was completed or the study ended) to determine the outcome. Ninety-two patients entered the study: 64 men (mean age, 50.4 years) and 26 women (mean age, 47.8 years) (sex was not reported for two patients). Thirty-seven patients were subsequently treated in the public system, and the remainder in the private system, either using private health insurance or at their own expense. The preferred treatment option of the treating medical officer, usually the urology registrar, was nominated: non-operative (spontaneous stone expulsion) with or without calcium-channel blockers, 13 patients (received by 6); rigid ureteroscopy with grasper or lithoclast, 21 patients (18); rigid ureteroscopy with laser, 4 patients (4); flexible ureteroscopy with laser, 17 patients (2); percutaneous nephrolithotomy, 3 patients (3); extracorporeal shock wave lithotripsy, 6 patients (2); or “other”, 28 patients — of whom stent was specified in 24 (23). The preferred treatment option was not used for 34% of patients because it was not available at the hospital. The mean duration of treatment (defined as the period between initial ED presentation and final treatment episode) for patients with pelvi-ureteric or upper ureteric stones requiring more than one treatment episode is shown in the Box. Thirty-nine patients had stents inserted in the ED, of whom four did not reach definitive management by the end of the study. Of the remaining 35, 20 were public patients and 15 were private patients. Fourteen had stents in situ for more than 3 months and required a change of stent before initiation of definitive treatment to avoid encrustation; 12 of these patients had treatment in the public system. Despite the relatively small number of participants in this study, its findings on access to timely treatment for public patients should not be ignored. Management of kidney stones was heavily influenced by insurance status. Ureteric stents are intended to be temporary, but patients treated in the public system who had a stent inserted at initial presentation had a 60% (12/20) chance of still having it 3 months later, thus requiring a change of stent before definitive intervention — an unnecessary procedure that increases hospital re-admissions. Patients would be treated more efficiently and effectively with more timely access to appropriate resources. This is an unacceptable burden of morbidity for patients. Urgent action is required to improve the current state of care for public patients with kidney stones in NSW. Duration of treatment of public and private patients with a pelvi-ureteric junction or upper ureteric stone who required more than one treatment episode* Public patients (n = 18) Private patients (n = 18) Mean duration of treatment in weeks (95% CI) 18.3 (12.9–23.7) 6.2 (3.0–9.4) Range (weeks) 3.0–49.5 0.6–25 Difference in weeks (95% CI) 12.1 (5.5–18.7) P < 0.001 * Up to four treatment episodes.

Finlay Macneil · James W H Macneil · Kylie L Fraser · Andrew J Brooks

To the Editor: The Australian Transport Council should be commended for approving the National Transport Commission’s revised road rules for the safety of children in motor vehicles.1 A key requirement is that children aged 4–7 years are to be restrained in an approved forward-facing child restraint or booster seat. It is expected that such changes, once enacted by states and territory governments, will result in fewer children being injured and killed. In a letter in the 4 August issue of the Journal, Zurynski and colleagues argued that these changes will bring Australian rules closer to — but nevertheless fall short of — overseas jurisdictions, where children up to 12 years of age or 145 cm in height must be restrained in booster seats.2 Notably, a wider selection of booster seats is available in these jurisdictions, including seats suitable for children weighing up to 36 kg. In contrast, the Australian/New Zealand Child Restraint Standard (AS/NZ 1754) stipulates that an “approved booster seat” is one that has a maximum design weight threshold of 26 kg.3 Mandatory consumer information notes that booster seats are “to be used only with lap-sash seatbelt or with a seatbelt and child harness for a child weighing from 14–26 kg”, and that they are not to be used if the child’s eye level is above the top of the booster back, or above the top of the car seat back or headrest when restrained in the booster seat.3 The matter of children exceeding the maximum weight threshold of 26 kg while failing to meet the transition height to adult seatbelts is far from trivial. Previous research published in the Journal found that about 50% of 7-year-olds whose height fell between 100 cm (the upper recommended height for child car seats) and 145 cm (the recommended seatbelt transition height) exceeded 26 kg, with only 27% having an age- and sex-adjusted body mass index > 25 kg/m2.4 Consequently, due to the current range of approved booster seats available in Australia, children may be placed at some unquantified risk in the event of a crash, as optimal protection above this weight threshold cannot be guaranteed. Further compounding this scenario is that pursuant to r266 of the Australian Road Rules, if a child “cannot safely be restrained as required . . . because of his or her height or weight”, the use of a seatbelt may be deemed acceptable.1 To avoid this scenario, it is essential that booster seats with a higher maximum weight threshold be made available in Australia as soon as practicable.

Michael P Fitzharris · Diana M Bowman

To the Editor: Despite expert recommendation, Australian states have yet to enact legislation requiring use of child restraints beyond the age of 12 months. Our analysis of police crash records and linked hospital separation data for the period July 2000 to June 2001 in New South Wales found that a large proportion of children who were hospitalised following motor vehicle accidents were reported by police as having used adult (standard) seatbelts at the time of injury (Box). It is of particular concern that over 80% of 5–8-year-olds in this cohort were using standard seatbelts rather than child restraints (eg, booster seats). This pattern of premature “graduation” to seatbelts has also been reported in general populations of child motor vehicle passengers both overseas1 and in Australia,2 and also in presentations of child motor vehicle passengers after a crash to a NSW hospital’s emergency department.3 Child restraints are specifically designed to provide crash protection for children’s anthropometrical dimensions. Standard seatbelts are not designed to accommodate children, so they are unlikely to achieve the good fit to rigid body parts required for safety. Consequently, use of standard seatbelts by young children allows more head excursion during a crash, thereby negating their primary goal of protecting against central nervous system injury, and potentially causing Chance fractures and abdominal injuries.4 Our results provide further evidence that such seatbelt use may not protect, or may even cause injuries, during a crash. Child road trauma is largely preventable or controllable with the use of appropriate child restraints, including booster seats. A cost–benefit analysis showed that the use of booster seats produced a benefit–cost ratio for road trauma prevention of 9.4 (US$1854/US$197).5 To prevent child road trauma in Australia, all child motor vehicle passengers should use appropriate child restraints. Australian child road safety stakeholders recommend that child motor vehicle passengers use appropriate restraint systems according to their height, weight and age when travelling on road. Our findings provide further justification for proposed legislative changes that would require the compulsory use of appropriate child restraints for child motor vehicle passengers. Child restraint use among children (0–8 years) hospitalised for injury after a motor vehicle accident, compared with that of the general population, New South Wales Hospitalised children* General population† Restraint use 0–4 years 5–8 years Total 0–4 years 5–8 years Total Child restraint 9 (22%) 0 9 (10%) 295 (94%) 70 (31%) 365 (68%) Adult seatbelt 11 (27%) 42 (82%) 53 (58%) 12 (4%) 153 (67%) 165 (31%) Unknown/ no restraint 21 (51%) 9 (18%) 30 (33%) 6 (2%) 4 (2%) 10 (2%) Total 41 (100%) 51 (100%) 92 (100%) 313 (100%) 227 (100%) 540 (100%) * Linked hospital and police data from July 2000 – June 2001 were accessed from the NSW Injury Risk Management Research Centre. Case selection is based on corresponding codes in the International Classification of Diseases, 10th revision, Australian modification, 2nd edition. † Based on a telephone survey conducted in NSW during 2005–2006.

Wei Du · Caroline F Finch · Lynne E Bilston

To the Editor: Samuel and colleagues’ article on the ethics of umbilical cord blood (UCB) banking1 reflects much of the misinformation and bias that bedevil this debate. They argue against storage of autologous stem cells from UCB for a variety of contradictory, paternalistic or ideologically driven reasons. The indication for autologous UCB storage is in anticipation of developments in regenerative medicine, an exciting field that holds great promise; it is not in anticipation of transplantation for malignancy, which is unlikely to occur. Samuel et al acknowledge these points, yet base their criticism of autologous storage on the latter indication. Next, they state that autologous stem cells can be easily harvested at any time in life. Perhaps, for use in transplantation; but cells obtained later in life are likely to be of little use in regenerative medicine. Capacity for in-vitro manipulation declines with age. Third, they argue that the two alternatives of public and private UCB storage are mutually incompatible — “public good or private benefit” (my emphasis). Australia’s three public banks are close to achieving the desired number of cords (about 20 000) needed for transplantation medicine in the non-Indigenous population.2 There is no shortage, and certainly no shortage of potential donors. Why argue against autologous storage as if there is? Fourth, the authors disapprove of the for-profit motive in private-sector medicine. That may be their ideological position, but it is paternalistic to impose that view on the rest of us. Can parents not make up their own minds on the value of autologous storage? At $2000 upfront and $150 a year, storage is not so expensive that “only a small proportion of the population are able to afford [it]”. I remind readers of media reports that the Australian Government’s Baby Bonus (now $5000) was often used to purchase luxury items such as flat-screen televisions rather than being spent on the baby’s needs.3 The authors are correct in one respect: some (but not all) private UCB banks have been deliberately deceptive and misleading in their marketing, and, in so doing, have been predatory and exploitative. However, it does not then follow that parents cannot access sound and sober health advice in the marketplace. The public versus private UCB storage debate does have an ethical dimension, but not this one. This debate is really just a turf war.

David E Roberts

In reply: Roberts asserts that we are biased, paternalistic and driven by an ideological objection to private umbilical cord blood (UCB) storage. While this is impressive rhetoric, it bears little resemblance to the points made in our article.1 Contrary to Roberts’ assertion, we do not object to private UCB storage. Our primary concern is that marketing campaigns make misleading or grandiose claims about the possible application of privately stored UCB in cancer care and regenerative medicine. For, although stem cell research does have great promise, it remains clinically unproven in the management of degenerative conditions. Offering hope of cure or amelioration of illness based on scientific speculation is enormously problematic, especially when directed at vulnerable parents concerned about their unborn child. While some private UCB banks take great care to avoid deliberate deception, many do not.2,3 We agree that parents should be able to decide for themselves how and if to store their child’s UCB, but they need accurate information to do so. Finally, Roberts asserts that there is no shortage of public UCB units or donors. In reality, there is a vast shortage of UCB units available to ethnic minority and Indigenous patients.4 North Caucasian donation also needs to be maintained, particularly as the impact of double-cord transplantation becomes apparent.4

Gabrielle N Samuel · Ian H Kerridge · Tracey A O’Brien

To the Editor: I have read every issue of the Journal since graduating in 1953. In my opinion, one of the best articles I’ve seen is Dawn DeWitt’s story, “I went to work with a ‘cold’ ...”.1 Dr DeWitt’s dilemma mirrors that of the average doctor precisely. Her “solution” gives us a good dose of commonsense that we seldom hear. In my time as a general practitioner, I have given thousands of doses of influenza vaccine; but I have never had the flu or pneumonia shot myself (nor has my wife — a registered nurse). I took 2 weeks off duty in 1971 with a “cold”, when I simply flaked out on the floor. A corneal transplant and transurethral resection of the prostate were done while “on vacation”. Luckily, I no longer have the dilemma of whether or not to go to work when I am sicker than my patients, but I will have the flu shot this week anyway. Thanks Dawn.

John F Knight AM

To the Editor: There is an established link between salt intake and blood pressure. The public health impact of a 1–3 mmHg reduction in blood pressure by lowering salt intake could be substantial. An American study found that a projected reduction in diastolic blood pressure of 2 mmHg would result in a 17% decrease in the prevalence of hypertension, a 6% reduction in the risk of coronary artery disease events, and a 15% reduction in risk of stroke and transient ischaemic attacks.1 In Finland, a one-third decrease in average salt intake achieved over 30 years was accompanied by a fall of more than 10 mmHg in the population averages of systolic and diastolic blood pressure.2 However, in the absence of active measures to reduce salt in the food supply, public health messages to reduce salt intake have largely been unsuccessful. The National Health and Medical Research Council (NHMRC) has recently revised its recommendations and now states that an adequate sodium intake for adults is 460–920 mg/day (20–40 mmol/day), with a suggested dietary target for chronic disease prevention of 1600 mg/day (70 mmol/day).3 The most recent Australian sodium intake data are from the 1995 Hobart Salt Study, in which the then national target of 6 g/day was achieved by only 6% of men and 36% of women, with an average salt intake of 7–10 g/day.4 Volunteers in weight-loss studies at the Commonwealth Scientific and Industrial Research Organisation (CSIRO) Human Nutrition unit over the past 5 years continue to have a high salt intake of 8–11 g/day (urinary sodium/24 h: men [n = 85], 181 ± 95 mmol; women [n = 189], 136 ± 61 mmol; reference range, 40–100 mmol).5 Recent data from another Australian study report similar urinary sodium concentrations.6 The average salt intake of Australian adults appears to be 7–12 g/day, which is little changed from 10 years ago. Achieving a low salt intake in the present food supply is difficult, as more than 80% of intake is from salt added to food during processing. As well as the more obvious foods in which salt is a preservative, staple foods such as bread also contain salt. The variety of foods has increased considerably, and many of the numerous processed foods now available contain very high salt levels. Processed and convenience foods frequently have salt added that the consumer cannot avoid. It is clear that large changes to the food supply are needed to reduce salt intake. We believe that salt reduction in the food supply should be the first line of attack to reduce the risk of cardiovascular disease in the population.

Jennifer B Keogh · Peter M Clifton

To the Editor: It’s all very difficult isn’t it? Teasing out the issues around impartiality, weighing evidence and competing interests? Wain wrote a recent editorial for the Journal, and included a list of his “competing interests”: Chair of the CSL Gardasil Advisory Board; speaker fees, travel assistance and consultancy fees from CSL Biotherapies and from Merck and its affiliates in relation to Gardasil; and shares in CSL Limited.1 He helpfully advises that Gardasil (Merck) “is available at no cost to Australian girls and women between the ages of 12 and 26 as part of the National Immunisation Program. The bivalent vaccine, Cervarix (GlaxoSmithKline), has to date not been included in the program, having initially been rejected by the Pharmaceutical Benefits Advisory Committee (PBAC) on the basis of uncertain cost-effectiveness, but subsequently recommended for inclusion.”1 This is disingenuous at best. Initially, the PBAC also “rejected the application for [Gardasil] . . . based on unacceptable and uncertain cost-effectiveness at the price requested.”2 At the request of the then Health Minister, the PBAC reviewed its decision after the company made some small changes to its submission, including a change in pricing. At about the time Wain’s editorial was published, a hard copy of the previous Medical Journal of Australia article on human papillomavirus (HPV) arrived on my desk,3 courtesy of GlaxoSmithKline. That article on HPV vaccination listed the “competing interests” at the end. Clearly, all the authors have received some sort of funding through GlaxoSmithKline, CSL and/or Merck. I am not impressed by authors who receive funding from pharmaceutical companies that market the drugs they are discussing. It seems to me that the problem lies with interests not competing, or at least not competing with the author’s intent. Perhaps a little healthy competition would bring out some more thoughtful, articulate articles, unaffected by any commercial pressures. “A plague o’ both your houses.” Let’s consider where we could best spend our money without the help of the competing interests of various pharmaceutical marketing mechanisms. “HPV vaccination will not prevent all cases of cervical cancer, therefore vaccinated women should continue to have two yearly Pap smears.”4 Given that HPV vaccination (in this country) will not change the rate of cervical screening required in the near future, perhaps the money would be better spent on ensuring that all Australian women are screened in a timely manner. That is, ensuring that poor women, Indigenous women, rural women, and immigrant women are part of “Cervical screening in Australia . . . one of the great public health success stories, as witnessed by a continuing dramatic fall in the incidence of carcinoma of the cervix and mortality from this disease since the introduction of the National Cervical Screening Program (NCSP).”5

Lilon G Bandler

To the Editor: We would like to express our disappointment with the Journal’s decision to publish an editorial on human papillomavirus (HPV) vaccines that demonstrated significant bias.1 It seems that simply documenting an author’s conflicts of interest exonerates the author and relieves the Journal of the responsibility of considering whether or not the article is biased. In his article, Wain states that the bivalent vaccine, Cervarix (GlaxoSmithKline), has not been included in the National Immunisation Program, having initially been rejected by the Pharmaceutical Benefits Advisory Committee (PBAC) on the grounds of uncertain cost-effectiveness. He omits to mention that Gardasil (Merck), was initially also rejected, and only funded after political intervention, an emergency meeting of the PBAC and further price negotiations with CSL.2 Wain claims that Cervarix is being promoted to older women despite the absence of efficacy data and the uncertain population benefits in this age group. Surely a balanced argument would include the fact that Gardasil is licensed for boys aged from 9–15 years based on immunogenicity data only, and that efficacy has not been established in this population. Why does the author consider this to be acceptable, but that licensing the vaccine for older women where the indication is to prevent cervical cancer is unacceptable? Many vaccines are licensed on the basis of immunogenicity data provided these have been shown to predict efficacy. While the efficacy of HPV vaccines in older women is being established, there are good data to show that an immune response to HPV vaccine is predictive of efficacy.3 Women aged over 26 years produce a robust immune response to HPV vaccines, similar to levels achieved in women aged 15–25 years, for whom efficacy has already been demonstrated. The Therapeutic Goods Administration has licensed Cervarix for women aged 26–45 years, before efficacy data became available, based on the assumption that efficacy will be demonstrated in seronegative older women. Women of all ages have shown interest in benefiting from a vaccine to protect against cervical cancer.4 Women of any age have the right to be informed, and to have the opportunity to discuss with their treating physician the relative benefits and risks of receiving the HPV vaccine for prevention of cervical cancer. Many vaccines are initially available only if purchased by individuals and, although this may result in inequity, ultimately, this is a decision of priority (individual financial and public health funding priorities) and not a reason to withhold licensing a vaccine with proven benefit.

Helen S Marshall · David Isaacs

To the Editor: Wain’s criticism of the Australian Therapeutic Goods Administration (TGA) approval of the bivalent vaccine, Cervarix (GlaxoSmithKline), suggesting it did not adhere to World Health Organization guidelines,1 should not detract from the potential benefits of human papillomavirus (HPV) vaccination in women over 26 years of age. An immunogenicity study showed all women up to the age of 55 years seroconverted to both HPV types and, while mean antibody concentrations at Month 7 were lower than in the younger age group, they were still three to four times higher than those observed in 15–25-year-old women in the long-term follow-up study (up to 4.5 years after vaccination), where continued efficacy was demonstrated.2 HPV infection is most prevalent in younger age groups, with one study showing a prevalence of 44.8% in women aged 20–24 years.3 As indicated by Wain, the vaccine has diminished efficacy in populations with high rates of previous exposure. Thus, 20–24-year-olds would benefit least, and if we extrapolate his argument, should not be included in any catch-up vaccination program. The United Kingdom has, in fact, recommended a catch-up campaign for girls aged up to 18 years only. The prevalence of HPV infection decreases after 26 years of age, and these “older” women should benefit from vaccination as supported by preliminary efficacy data of the quadrivalent vaccine Gardasil (Merck) in an older population.4 This vaccine may also be protective for women who have been previously exposed to the same subtypes of HPV as the vaccines, as shown by 100% efficacy against cervical intraepithelial neoplasia (CIN) grade 2/3 or adenocarcinoma in situ (AIS) among people who are seropositive but HPV-DNA-negative to the relevant HPV type.5 The Pharmaceutical Benefits Advisory Committee initially did not recommend funding for Gardasil on the basis of cost-effectiveness until after an extraordinary meeting to consider a revised submission, following an “unusual” request from the Health Minister.6,7 CSL agreed to reduce the price, undertook to make a substantial contribution to any booster program if it became necessary in the next 20 years, and also to the costs of setting up a national register to link vaccination data to later cervical screening records. It is obvious that the efficacy of HPV vaccines will inevitably be lower after commencement of sexual activity, but we believe that it is the medical practitioner’s responsibility to offer women aged over 26 years the current, albeit incomplete, information on vaccine efficacy, and allow women to make the choice.

Jeffrey H J Tan · Michael A Quinn

To the Editor: Recent articles in the Journal describing endeavours to measure and classify the tasks of doctors are indicative of the re-emergence of work analysis and time and motion studies.1,2 While commending the authors on their endeavours, it is concerning that the articles state almost contradictory findings. Westbrook and colleagues reported that professional communication, social activities and meal breaks represented the greatest proportion of observed time.1 Zhu and colleagues reported that direct patient-related tasks accounted for 86% of intern time.2 Acknowledging the different contexts of the individual studies, the collective picture is one of confusion and may lead to misrepresentation of the work of doctors. We have also analysed the work of doctors using observational techniques.3 We built on work done in the United States4 that is underpinned by functional job analysis (FJA)5 to produce a list of tasks (task taxonomy) that describe the work in the acute-care setting, and coordination of roles between hospitals and the community.3,6 The task is the fundamental unit of work, and FJA describes each task in terms of behaviours and interdependencies between people, data and things for the achievement of the task. The method seeks to achieve quality information through adopting precise language descriptions and benchmarks for levels of tasks required for jobs.5 Data are recorded by means of a simple electronic tool.6 Our method captures contextual information about the service (eg, location of work) and rigid details about the observed tasks. We have presented our findings at the 5th Health Services and Policy Research Conference of the Health Services Research Association of Australia and New Zealand,6 and the Change Champions Skill Mix and Workforce Development conference, both in 2007. We found that doctors in the units studied spent about 11% of time on education and training, between 50% and 60% on direct clinical activities (depending on context and role), and less than 10% of time on non-clinical administration. We suggest that if observations are recorded according to the purpose of the output, what may appear to be “socialising” may, in fact, be waiting for something or someone. It is far more important to measure what the impediment to getting on with the job is, rather than inferring that socialising is the main activity. There is a need for a consistent task classification system that can be used across units and across professions to describe the work that is being performed. A common unit of measure would provide a strong foundation for collaboration and learning in work redesign projects across the nation. Without such a system, planning for the future and evaluation of new roles will continue to be hindered. To this end, we are happy to share the task taxonomy that we have developed, and welcome contact via email.

Mark Mackay · Pamela J Castle

To the Editor: The study of intern activities in Melbourne emergency departments (EDs) highlighted gaps in the ED training of interns.1 In particular, it was shown that interns undertake a low number of procedural tasks. Most did not perform urinary catheterisation, nasogastric tube insertion or reductions of fractures and dislocations, just some of the skills highlighted as important in the Australian Curriculum Framework for Junior Doctors (ACFJD).2 Time pressure on supervisors was cited as a reason for this. New South Wales will experience a doubling in the number of medical graduates in the next few years. Increased capacity in emergency terms will be needed. The reported capacity (available supervision, clinical workload and funding) in NSW EDs indicates that a shortage of terms could develop from 2011. The NSW Institute of Medical Education and Training (IMET) recently explored the educational validity of retaining the emergency term as a mandatory requirement for general registration3 in view of the reported limited capacity. IMET-accredited emergency term descriptions were reviewed and mapped against the “Common presenting problems and conditions” listed in the ACFJD. The study found that an ED term would likely provide exposure to most of the common conditions, many of which were unlikely to be encountered in other medical or surgical hospital rotations. The acute phase of key conditions, that all medical practitioners are expected to be able to recognise and treat appropriately, were often unique to the ED term. Further, most of the “Skills and procedures” for junior doctors were likely to be learned in the ED, and around 50% were not commonly experienced elsewhere. Many of the broader competencies of the curriculum such as “Doctor and society” are also covered in an ED term. The ED provides a unique context for learning, bridging community and hospital situations. The approach to acutely ill patients, with as yet undifferentiated problems, for whom JMOs learn to initiate treatment and appropriate investigations concurrently, cannot be reliably replicated elsewhere. Emergency departments bear the brunt of the growing demand for acute care from an ageing population with multiple comorbidities, and also of the growing demand for excellent and accountable postgraduate medical training. In NSW, about 30% of directors of prevocational training are ED physicians, reflecting their commitment to education. They typically supervise numerous trainees at a time. The educator role of ED physicians and other senior doctors in EDs is essential. Resources should be directed to supporting this role and enhancing the capacity of EDs to train medical graduates.

Belinda Doherty · Mark A Brown

To the Editor: Learning in the clinical setting remains central to the development of well trained health care professionals. The issue is whether that learning should occur through formal or informal learning opportunities. Westbrook and colleagues define “supervision or education” in a way that focuses mainly on formal experiences,1 possibly because trained observers could accurately classify such experiences. As noted by Brown and Arnold, much learning in the hospital setting is largely informal in nature.2 Although learning is likely to be occurring during the many discussions that junior doctors have with consultants or during the procedures they perform in an emergency department,3 it can be difficult to describe, and may not be recognised as learning by the individuals involved.4 There is a divide between the perceptions of teachers and trainees about how much learning is occurring, whether teaching has occurred and feedback has been given. Consultants believe they are providing a great deal, but junior doctors do not recognise it. Although junior doctors perceive they have adequate informal contact with registrars, and some (but not enough) with consultants, what they want is more teaching in “formal” sessions.5 Additionally, supervisors think they give detailed feedback, but junior doctors view it as less than adequate.6 The answer might lie in upskilling both junior doctors and teachers to make teaching and learning more effective, in part by making it more explicit to all involved. At Sir Charles Gairdner Hospital in Perth, an innovation has been to create new positions known as “medical education registrars” who, as supernumerary staff at a senior registrar level, have time to advise on patient management, supervise and teach skills.7 Much of this is provided at the patient’s bedside. The very nature of their job title makes it explicit that they are there to help learning during daily activities. The staff development program, “Teaching on the Run”, developed by the Education Centre at the Faculty of Medicine and Dentistry, University of Western Australia,8 aims to make teachers more effective. More recently, we have piloted another program, “Learning on the Run”, for junior doctors and students, to provide them with the skills to recognise opportunities and drive their own learning agendas. We agree with Brown and Arnold that learning and service are not mutually exclusive.2 By providing both junior doctors and senior medical staff with the necessary skills, many tasks within a day in the life of a new doctor could become a valuable learning experience. Whether this will ultimately translate into improved teaching and learning outcomes is a question we are continuing to explore.

Lisa Caputo · Fiona R Lake · Margaret Potter · Ian Rogers

To the Editor: In his exploration of the health advocacy potential of modern clinicians, Gruen1 observes that the public first needs to be convinced that “the profession has its own house in order”. Unfortunately, one room in that house accommodates one of the serious health threats identified by the author: terrorism. Currently, over 95% of the world’s radiopharmaceuticals are generated from highly enriched (bomb-grade) uranium (HEU), an unnecessary nuclear weapons proliferation hazard.2 Prompt conversion of the global medical isotope supply chain to low enriched uranium (LEU, containing less than 20% uranium 235, so not viable for weapons production) is technically feasible.3 Clinicians are thus uniquely placed to advocate conversion to the use of LEU, while pressuring their imaging and isotope providers to end reliance on HEU, thereby blocking one of the most vulnerable pathways to producing a “terrorist bomb”. But, as Gruen suggests, we can do even more through “collective advocacy” to address the much larger nuclear threat: that is, the 26 000-plus nuclear weapons remaining in the arsenals of Russia, the United States, the United Kingdom, France, India, Pakistan, Israel, China and North Korea. While a sophisticated terrorist group armed with home-manufactured nuclear weapons could devastate a few cities, the existing nuclear-armed states have the capacity to destroy between tens and thousands of urban centres and their populations within a few short hours. Worse still, recent research indicates that 100 Hiroshima-sized (ie, “small”) nuclear weapons exploded on major cities would be capable of precipitating a “nuclear winter” that could persist for 10 years.4 The dispersal of carbonaceous material into the stratosphere from major urban firestorms could dramatically reduce terrestrial sunlight, lower surface temperatures by several degrees, shorten the growing season, reduce rainfall and trigger global famine. One billion deaths from starvation is a realistic assessment of the consequences.5 Such a catastrophic scenario is within the firepower capacity of all currently nuclear-armed nations except North Korea. A new generation of medical students and young physicians has launched several initiatives over the past few years to challenge this threat, including the Nuclear Weapons Inheritance Project and Target X (http://www.ippnw-students.org). Most recently, International Physicians for the Prevention of Nuclear War launched the International Campaign to Abolish Nuclear Weapons (http://www.icanw.org), whose goal is to establish a nuclear weapons convention to eliminate all nuclear weapons once and for all. By ending our reliance on bomb-grade HEU in medical imaging, we can certainly begin to put our own house in order. But let’s also follow Virchow’s lead: let’s “engage with the broader social concerns that cause illness and harm”,1 get active for our patients’ — and our own — wellbeing, and help prevent a global nuclear pandemic.

Bill Williams

To the Editor: We note with interest the recent commentary by Cretikos et al on the predictive value of a high respiratory rate for adverse outcomes.1 We wish to provide empirical evidence from Australian patients with pneumonia in support of their view that simple clinical parameters are good predictors of adverse outcomes. We examined data from a prospective cohort of consecutive patients presenting to the Royal Melbourne Hospital Emergency Department with radiologically confirmed, community-acquired pneumonia between 2003 and 2006.2 In an earlier study of a subset of these patients,3 we found that hypotension and tachypnoea were strongly associated with death and/or the need for respiratory/inotropic support (odds ratios, 8.0 and 3.5, respectively). In the full cohort (n = 740), we examined factors associated with either admission to the intensive care unit (ICU) or mortality (106 patients were in one of these two categories). Respiratory rate was documented in 712 patients (96%). A combination of tachypnoea (≥ 24 breaths/min) and/or hypotension (systolic blood pressure ≤ 90 mmHg) had similar predictive value for the risk of ICU admission and/or death to the recommended system of risk stratification, the Pneumonia Severity Index (PSI)4 (Box). The combination of respiratory rate and systolic blood pressure performed better than either sign alone in ruling out at-risk patients if both were normal (ie, a high negative predictive value), although almost a quarter of patients with either clinical sign had adverse outcomes. The PSI is based on 20 individual clinical and laboratory parameters, and evidence suggests that it is poorly documented in patients’ records.5 Our data relate to patients with community-acquired pneumonia from a single centre and thus have limited statistical power for making comparisons. However, they suggest that these two routinely measured clinical parameters can be used to stratify patients at risk of adverse outcomes at the time of presentation. We support efforts to incorporate simple clinical indicators into systems that can identify seriously unwell patients early in the course of illness. Summary statistics for tachypnoea, hypotension and PSI class as predictors of ICU admission and/or death from community-acquired pneumonia* Number of patients Sensitivity Specificity PPV NPV RR ≥ 24 breaths/min 712 82% (73%–89%) 48% (44%–52%) 22% (18%–26%) 94% (91%–96%) RR ≥ 27 breaths/min 712 70% (60%–78%) 67% (63%–71%) 27% (22%–32%) 93% (90%–95%) RR ≥ 24 breaths/min and/or systolic BP ≤ 90 mmHg 713 93% (86%–97%) 45% (41%–49%) 23% (19%–27%) 97% (94%–99%) PSI class IV or V† 740 90% (82%–95%) 49% (45%–53%) 22% (19%–27%) 97% (94%–98%) BP = blood pressure. ICU = intensive care unit. NPV = negative predictive value. PPV = positive predictive value. PSI = Pneumonia Severity Index. RR = respiratory rate. * Figures in parentheses represent 95% CIs calculated using the exact binomial distribution. † Severe pneumonia.

Allen C Cheng · James F Black · Kirsty L Buising

To the Editor: Cretikos et al make a strong case for routinely recording patients’ respiratory rate (RR) in acute wards.1 In a prospective study designed to evaluate the prognostic value of RR in acutely hospitalised patients aged over 75 years, we analysed data from all admissions to a single 14-bed acute-care geriatric unit between 15 May and 15 November 2007. Clinical data were recorded on admission and blood tests were performed the next morning. The Charlson score was used to assess comorbidity.2 Features of the 195 admissions during the study period (13 patients were admitted twice) are reported in the Box. The main reasons for admission were falls (15%), left ventricular failure (11%), pneumonia (11%), cancer complications (9%), pyelonephritis (7%) and stroke (5%). Twenty-nine patients died in hospital, including six from cancer complications, five from pneumonia, four from pyelonephritis and four from left ventricular failure. Based on univariate logistic regression analysis at the 0.05 significance level, the following variables were predictive of death during hospitalisation: being male (odds ratio [OR], 2.42, Wald test P = 0.03); increased Charlson score (OR, 1.50 for each additional point between 3 and 13; P < 0.001); decreased systolic blood pressure (OR, 6.71 for systolic blood pressure < 100 mmHg; P =0.004); abnormal heart rate (< 60 beats/min or > 100 beats/min) (OR, 3.63; P = 0.003); increased RR (OR, 1.81 for each additional 5 breaths/min between 14 and 44; P < 0.001); abnormal blood sodium level (< 137 mmol/L or > 143 mmol/L) (OR, 2.82; P = 0.01) and raised C-reactive protein level (OR, 2.67 for C-reactive protein level > 45 mg/L; P = 0.02). After multivariate logistic regression analysis with stepwise backward elimination, the only remaining factors that were significant predictors of death during hospitalisation were Charlson score (OR, 1.53 for each additional point between 3 and 13; P < 0.001) and RR (OR, 1.83 for each additional 5 breaths/min between 14 and 44 breaths/min; P < 0.001). RR on admission was associated with an area under the ROC (receiver operating characteristic) curve of 0.73 (95% CI, 0.64–0.82) for the prediction of death during hospitalisation. An RR of ≥ 20 breaths/min had a sensitivity of 97% (95% CI, 80%–100%), a specificity of 28% (95% CI, 21%–35%) and a negative likelihood ratio of 0.12 (95% CI, 0.02–0.82) for prediction of death during hospitalisation. Only one patient (admitted for stroke) with an RR of < 20 breaths/min at admission died in hospital. Higher RR cut-off values increased specificity but reduced sensitivity. For instance, an RR of ≥ 30 breaths/min had a specificity of 90% (95% CI, 84%–94%), but a sensitivity of 38% (95% CI, 21%–58%), leading to a positive likelihood ratio of 3.68 (95% CI, 1.93–7.04) for prediction of death during hospitalisation. In conclusion, besides comorbidity (Charlson score), RR was the most useful predictor of death in acutely hospitalised patients aged over 75 years. Our results extend the evidence base promoting regular documentation of RR in acute-care departments.1 Features of the 195 admissions to the acute-care geriatric unit between May and November 2007 Age (years) 85 (81–91)* Women 123 (63%)† Charlson score 6 (5–8)* Dementia 88 (45%)† Body temperature (°C) 37.0 (36.8–37.5)* Systolic blood pressure (mmHg) 136 (120–152)* Heart rate (beats/min) 80 (70–92)* Respiratory rate (breaths/min) 20 (20–25)* Serum sodium level (mmol/L) 140 (137–143)* Serum creatinine level (μmol/L) 91 (76–120)* Serum protein level (g/L) 65 (59–70)* Serum C-reactive protein level (mg/L) 36 (12–108)* Blood haemoglobin level (g/L) 115 (100–125)* White blood cell count (× 109/L) 8.3 (6.3–11.5)* * Median (interquartile range). † Number of admissions (%).

Olivier Steichen · Gilles Grateau · Eric Bouvard

To the Editor: The concept that respiratory rate (RR) is a key vital sign is hardly new,1 but it is being re-emphasised.2,3 The recent article by Cretikos et al3 highlights the diagnostic relevance of a raised RR for serious adverse events. The authors make sensible recommendations regarding the need to educate hospital staff about the importance of measuring patients’ RR. However, they do not specify how this rate should be measured. Indeed, most textbooks of general medicine, and even respiratory medicine, fail to provide guidance on this or to define an abnormal rate. A widely used book on clinical examination4 suggests measuring RR while feeling the pulse, and quotes a normal (adult) resting range of 16–25 breaths/min, but no source for this information is provided. There are few reports of true normal resting RR measurements obtained by covert observation. Respiratory physiologists have long known that RR commonly increases and becomes more regular as soon as a subject becomes aware of the measurement. This is especially so if a mouthpiece is in place. Rates as low as 8 breaths/min may be seen at rest, and the normal adult range quoted by physiologists is 11–14 breaths/min.5 Bradypnoea is usually defined as a rate less than 8 breaths/min and tachypnoea as a rate greater than 18–20 breaths/min. There is no gold standard method for accurate measurement of RR in clinical practice. Of course, when patients are being monitored, particularly with a nasal cannula, it should be easy to obtain the rate. Oximetry is not a surrogate measure of RR, although it is often easier to record. Inductance bands around the chest provide a simple non-invasive way to measure RR. In addition to the absolute rate, an irregular and erratic rate is of concern. In settings in which formal monitoring is not being conducted, RR is the one “vital” sign that must be assessed when the patient is resting quietly, unaware of its measurement, and not conversing with staff. Duplicate measurements should be made over an interval of at least 1 minute. An RR of over 20 breaths/min, particularly if irregular, is noteworthy. Tachypnoea is intimately linked with the sensation of breathlessness, and hence the patient’s respiratory sensations should be assessed. In some patients with a normal RR at rest, marked tachypnoea may be precipitated by mild exertion, such as walking a few paces.

Simon C Gandevia · David K McKenzie