Letters

To the Editor: Moynihan and colleagues1 make a good case for Australia responding to the complex challenge of reducing the overdiagnosis of clinical disease. However, this challenge should not lead to confusion with the early diagnosis of and the early intervention in preclinical disease, which are the mainstay of secondary prevention. While reducing diagnosis creep and expanding disease definitions are predominantly the realm of the specialist disciplines, primary care is uniquely placed to embrace preclinical disease diagnosis, increasing early detection and intervention, while specialists endorse reducing overdiagnosis, both collaborating in lowering long term costs to the health system. Preclinical disease and its impact are emerging as the logical next challenge. Prediabetes, for example, is almost always present before the onset of diabetes.2,3 Both the American Diabetes Association3 and Diabetes Australia4 have published recommendations on the diagnosis and screening for diabetes and prediabetes. The American Diabetes Association also leads the way in recommending that screening should be considered in children and adolescents who are overweight or obese and who have additional risk factors for diabetes.3 In 2015, the most comprehensive undertaking since Medicare's inception in the 1980s was established to consider how the more than 5700 items on the Medicare Benefits Schedule (MBS) could be aligned with contemporary clinical evidence and practice and improve health outcomes for patients with clinical disease (tertiary prevention).5 Although, the evidence base around early diagnosis and intervention is relatively recent, it is now opportune for the federal government to initiate a second tier to the MBS Review to consider new MBS items, where appropriate, for the early diagnosis of and early intervention in preclinical disease (secondary prevention). Further, this would align well with the new global awareness and endorsement of lifestyle medicine.6 Reversing the underdiagnosis of preclinical disease would be the perfect partner to reversing the overdiagnosis of clinical disease in delivering better health outcomes for patients, while reducing long term costs to the health system.

Eugen Molodysky

To the Editor: I read with interest the article by Moynihan and colleagues1 and commend the authors on their timely review of this important topic. While the article referred to research that had highlighted concerns about overdiagnosis in relation to several medical and surgical conditions, the only psychiatric condition mentioned was attention deficit/hyperactivity disorder. The authors referred to the problem of medicalisation as one of the possible drivers of overdiagnosis in one of the figures within the article, but not in the text of the article — medicalisation is the process by which non‐medical problems become defined and treated as disorders. In 2005, the medical expenditure on identified medicalised conditions in the United States was estimated to be about US$77 billion.2 Medicalisation in psychiatry has been a particular concern because of the problem created by ever‐expanding definitions of mental disorders and lowering of diagnostic thresholds.3 This has been well illustrated by the changes in the diagnostic criteria within the successive editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) published by the American Psychiatric Association, now in its fifth edition (DSM‐5). The effect of a change in the DSM diagnostic criteria was reported in 2001: changes from the third edition (DSM‐III) to the fourth edition (DSM‐IV), with an increase of the variety of stressors (events) considered capable of leading to post‐traumatic stress disorder (PTSD), led to a finding that 38% of DSM‐IV PTSD cases resulted from its wider definition.4 In the DSM‐5 the diagnostic criteria for PTSD have been further relaxed, so that “emotional reactions to the traumatic event” are no longer part of the diagnostic criteria. Another example of overdiagnosis in psychiatry is that of adjustment disorder. A 2008 article stated that there are “huge [numbers] of false‐positives for depressive and anxiety disorders because the context of symptoms is not taken into account,” and that the diagnosis of adjustment disorder “seems, by definition, ideally suited to apply to healthy people in dangerous and uncertain circumstances”.5 It is to be hoped that the 7th international Preventing Overdiagnosis conference in 2019 will include a session on overdiagnosis also in psychiatry.

George Mendelson

In reply

Raymond Moynihan · Alexandra L Barratt · Paul P Glasziou

To the Editor: I commend the Medical Journal of Australia for supporting high quality qualitative research with clear criteria for acceptance for publication1 on the background of increasing concerns these manuscripts are being rejected for reasons not based on the quality of the article submitted.2 However, I am concerned about the article by Chanchlani and colleagues,3 which involves randomised controlled evaluation of a peer mentoring program for new medical interns using qualitative interview‐based methodology.3 The Journal's Editor‐in‐Chief hoped this methodology would “encourage others to use comparable approaches when investigating similar topics”.4 A well conducted randomised controlled trial is considered among the highest level of evidence base for clinical practice; randomisation minimises bias from known and unknown confounders. However, other biases (selection, recall, measurement etc) also need to be controlled for randomised controlled trials to provide valid results and conclusions. Chanchlani and colleagues3 suggested their primary outcome was to assess psychosocial wellbeing and job satisfaction using inductive thematic analysis of data collected in semi‐structured interviews and focus groups at 12 months. This is different from the Australian New Zealand Clinical Trials Registry (ACTRN12618000455268, retrospectively registered) description which is “to determine the perception of the effectiveness of peer support on anxiety and depression;” the psychiatric training of the interviewers and the formal screening or post hoc assessment of the participants’ mental health are unknown. Qualitative research in randomised controlled trials is increasingly common, with new innovative purposes.5 Chanchlani et al3 reported a novel approach to qualitative research incorporating randomised controlled methodology; no quantitative data are apparent in outcome analysis, even though such measurement can be obtained from a post‐program feedback survey for both groups. Interview questions published in the online appendix cannot properly compare satisfaction rate nor assess the state of mental health. More is needed to justify comparative comments such as “participants with mentors reported high satisfaction with the program and a positive impact on stress levels, morale, sense of support, job satisfaction, and psychosocial wellbeing compared with participants without mentors”. It is desirable that the MJA supports innovative qualitative research. Important information may be omitted due to editorial requirements. Nevertheless, vigorous peer review and academic integrity are still needed. Care should be taken when comparative conclusions are made without adequate explanation.

Shyan Lii Goh

In reply

Sonia Chanchlani · Jeremy SL Ong

To the Editor: Pharyngitis, a common childhood illness, accounts for around 3% of presentations to general practice in Australia.1 Although usually benign and self‐limiting, group A streptococcus (GAS) pharyngitis, isolated in up to 20% of symptomatic children,2 can lead to infectious and autoimmune sequelae. Despite Australia being a high income country, acute rheumatic fever (ARF) and rheumatic heart disease (RHD) still cause significant morbidity and mortality in Aboriginal and Torres Strait Islander people.3 Prompt treatment of GAS pharyngitis has been shown to reduce the incidence of ARF by two‐thirds in high risk individuals.4 Low risk individuals require supportive management only.5 Clinical diagnosis of GAS pharyngitis is unreliable4 and culture results take time. As such, clinicians must balance the competing priorities of appropriate treatment of patients at high risk of ARF or RHD with prudent antimicrobial stewardship. Clinical practice guidelines play an important role in decision making at both a population and individual level. We undertook a search to identify Australian and New Zealand pharyngitis guidelines and compared these with previously published criteria.6 Nine guidelines were identified. Inconsistences in diagnosis, definition of high risk groups, analgesia, antibiotic rationale, agent, therapy duration, and tonsillectomy indications were observed (Box). Australia's multitude of heterogenic guidelines coupled with the transient workforce in remote Australia, where ARF burden is the highest,7 predispose to management confusion and potential poor patient outcomes, including higher rates of ARF and RHD, and also fail to address the growing worldwide problem of antimicrobial resistance. Australia needs a single national pharyngitis guideline to assist in providing rational, consistent and timely antibiotic treatment to patients at high risk of ARF, while minimising inappropriate antibiotic usage and resistance in individuals at low risk of sequelae. We call for an evidence‐based guideline that includes the following: a clear, succinct approach to diagnosis and management; a definition of individuals at high risk of ARF, and rationale for antibiotic treatment; clear guidance regarding throat culture and point‐of‐care testing for GAS; rationale for first‐ and second‐line empirical antibiotics, with alternatives for penicillin allergy; capacity to adapt management in different clinical settings; and supportive care recommendation including analgesia, tonsillectomy and school exclusion. Box – Summary of selected criteria:6 are criteria addressed by each sore throat guideline? Guidelines NZ HF BPAC NZ CH QLD NSW ICCPG CARPA RHD Aust PCH RCH eTG Number of criteria addressed 6/12 7/12 2/12 10/12 7/12 6/12 10/12 9/12 10/12 Diagnostic criteria × × × ✓ ✓ × ✓ × ✓ Routine throat culture/rapid antigen detection testing ✓ ✓ × ✓ × ✓ ✓ ✓ × Antibiotics to reduce symptoms × ✓ × ✓ × × ✓ × ✓ Antibiotics to prevent complications ✓ ✓ × ✓ ✓ ✓ ✓ ✓ ✓ NZHF = New Zealand Heart Foundation (http://www.heartfoundation.org.nz/shop/heart-healthcare/non-stock-resources/sore-throat-algorithm.pdf); BPAC = Best Practice Advocacy Centre (https://bpac.org.nz/antibiotics/guide.aspx#sore-throat); CH QLD = Children's Health Queensland Hospital and Health Service guidelines (http://www.childrens.health.qld.gov.au/chq/health-professionals/antimicrobial-stewardship/guidelines/ent-infections); NSW ICCPG = New South Wales infants and children clinical practice guidelines (http://www1.health.nsw.gov.au/pds/ActivePDSDocuments/GL2014_021.pdf); CARPA = Central Australian Rural Practitioners Association's standard treatment manual (https://docs.remotephcmanuals.com.au/review/g/manuals2017-manuals/d/20321.html?page=115); RHD Aust = rheumatic heart disease Australian guidelines (https://www.rhdaustralia.org.au/arf-rhd-guideline); PCH = Perth Children's Hospital emergency department guidelines (https://pch.health.wa.gov.au/For-health-professionals/Emergency-Department-Guidelines/Tonsillitis); RCH = Royal Children's Hospital Melbourne guidelines (with support of the Victorian Paediatric Clinical Network) (www.rch.org.au/clinicalguide/guideline_index/Sore_throat); eTG = electronic therapeutic guidelines (https://tgldcdp.tg.org.au/searchAction?appendedInputButtons=sore%20throat). ◆

Adrian J Tarca · Robert M Hand · Rosemary Wyber

To the Editor: Patients may require long term anticoagulation for reasons that commonly include deep vein thrombosis, pulmonary embolism or atrial fibrillation.1 In these circumstances, heparin is commonly used for bridging and is discontinued after the effects of warfarin result in a therapeutic international normalisation ratio. It takes approximately 5 days for this to occur because warfarin inhibits the production of vitamin K‐dependent clotting factors II, VII, IX and X.2 The time to therapeutic anticoagulation is a reflection of the half‐lives of the circulating clotting factors and the time for them to diminish from the plasma. This has been our mindset for decades from the perspective of warfarin use. The introduction of direct‐acting oral anticoagulants (DOACs), such as apixaban, dabigatran and rivaroxaban, has resulted in important logistical and clinical benefits in patients admitted to hospital. For example, bridging with heparin is no longer needed3 because DOACs directly inhibit circulating clotting factors resulting in a relatively quick onset of anticoagulation. The maximum concentration in the plasma is achieved within a few hours, which also mirrors its anticoagulant effect.4 Unfortunately, based on internal audits at our institution, our medication safety committee has identified cases in which DOACs were combined with heparin or low molecular weight heparin. During a 38‐day audit period, there were 14 cases involving such duplication. Based on anecdotal discussions, this duplication is partly related to the prescribers’ lack of knowledge with regards to DOACs. Nurses or pharmacists usually intercepted these events such that the overlap occurred for a few doses and no patients were harmed. In the shift from warfarin to DOACs, inadvertent and unnecessary duplicate anticoagulation due to bridging increases the risk of bleeding. Although we seldom use absolutes, we can confidently endorse that there are no circumstances in which DOACs should be combined with another anticoagulant. Prescriber education and clear institutional guidelines may help deal with this issue. In addition, institutions with electronic medical records could optimise clinical decision support systems to prevent such duplication. The combination of DOACs with heparin is a bridge to nowhere.

Mark A Sheppard · Russell Levy · Asad E Patanwala

To the Editor: Recent Australian surveys of gluten content in gluten‐free labelled foods purchased from supermarkets or restaurants are reminders of the difficulties faced by patients with coeliac disease.1,2,3 Despite trying to adhere to a gluten‐free diet, significant inadvertent gluten exposure is common, leaving about 30% of patients with incomplete intestinal mucosal healing.4,5 In Australia, a “no detectable gluten” standard applies to food labelled gluten‐free. However, surveys published in the Journal reported detectable gluten in 14% of imported gluten‐free foods (0.5–1.1 parts per million [ppm]),2 in 9% of gluten‐free marketed restaurant foods in Melbourne (5.2 to > 80 ppm),3 and in 2.7% of “commonly purchased” gluten‐free foods (5–49 ppm), including foods manufactured in dedicated gluten‐free factories.1 The governance of the compliance with the gluten‐free food code is unsatisfactory; the testing of gluten‐free foods is done by the food industry. Despite a multilayered food code bureaucracy, there is no federal or state oversight of testing, and test results are not published. State authorities have not investigated the non‐compliance reported for imported gluten‐free foods in 2016.2 Local governments are responsible for implementing state food laws, yet, they cannot coordinate oversight of gluten testing nationally. The federal Department of Agriculture and Water Resources is responsible for imported foods, but there is no evidence they test gluten‐free foods imported from jurisdictions that permit up to 20 ppm gluten. The Australian Competition and Consumer Commission is responsible for the Australian Consumer Law, and Food Standards Australia and New Zealand establishes the food code standard; however, there has been no indication by either agency that they consider the problems with the gluten‐free standard or its governance a sufficient public health issue to warrant changes to current practices. Inadvertent gluten exposure may occur by cross‐contamination from known gluten‐containing foods, or from foods considered free of gluten by listed ingredients but not labelled gluten‐free. The very least that patients with coeliac disease should expect is negligible additional contamination from foods that are labelled gluten‐free. Transparent testing of gluten‐free labelled foods is therefore critical. It is unlikely that the government will implement regular testing programs in place of the current ad hoc and unreported industry‐based testing. However, mandating the regular publication of laboratory test results is a simple measure to reassure consumers with coeliac disease, and would likely be a positive initiative for local gluten‐free food exporters seeking to take international advantage of the tight Australian gluten‐free standard.

Geoffrey M Forbes

To the Editor: Bateson and Black1 do a great service in encouraging clinicians to discuss pre‐conception care with women of reproductive age.1 However, in relation to infection prevention, one area not discussed was cytomegalovirus (CMV) infection, which is the most common infectious cause and the second most common aetiology of all causes of severe congenital malformations.2 Mother to child transmission of CMV can result in prematurity, stillbirth, cerebral palsy and neurodevelopmental delay and is the most common infectious cause of hearing loss.2 Discussions about CMV prevention should ideally commence before pregnancy, as maternal CMV infection in the first trimester poses the greatest risk of harm to the fetus if mother to child transmission occurs. Such discussions should continue throughout pregnancy, as secondary maternal infection with a different strain of CMV can also result in mother to child transmission of virus,2 although the risk per infectious event is lower. Women can adopt simple hygiene strategies to reduce risk of CMV infection and thus reduce mother to child transmission of virus during pregnancy. These recommendations have been published3 and referenced in consensus recommendations2 and other sources.4 Strategies preventing women acquiring CMV (usually from children)3 are acceptable and inexpensive — handwashing, not sharing food or objects covered with children's saliva, not kissing children on the lips and wearing disposable gloves during nappy changes. These strategies reduce the risk of infection before pregnancy and of mother to child transmission during pregnancy;2 they do not affect reactivation of latent virus, although this is associated with lower mother to child transmission. Universal serological screening with CMV IgG to determine previous immunity is not recommended, as congenital CMV can still occur as a result of non‐primary maternal infection and reactivation during pregnancy. Women should be advised to use hygiene strategies regardless of their serological status.2 In Australia, only one in six women who are pregnant know about CMV,5 and only one in ten maternity clinicians routinely discuss CMV prevention with pregnant women.6 It is likely fewer discuss CMV prevention before conception. We encourage clinicians, women considering pregnancy and parents to increase their knowledge about CMV and its prevention.4

Antonia Shand · Pamela Palasanthiran · William D Rawlinson

In reply

Deborah J Bateson · Kirsten I Black

In reply

Carmelle Peisah · Tiffany Jessop · Henry Brodaty

To the Editor: We note with interest the research letter by Taft and colleagues regarding unplanned pregnancy in Australia.1 An audit of terminations of pregnancy in our Australian metropolitan hospital service (Austin Health, Melbourne) supports the findings of the national survey while highlighting additional areas of educational need. Of 309 women (aged 15–46 years) who underwent terminations for unwanted pregnancy from January 2016 to October 2018, 145 chose surgical termination and 164 women chose medical termination. Compared with the national survey, our cohort had similar rates of not using contraception (48.2% v 56.5%) and contraception failure (51.8% v 41.4%).1 Failed barrier and oral contraception respectively accounted for 52.1% and 29.3% of contraception failures. This differs from the survey findings, where the most common contraception failure was oral contraception, followed by condoms.1 There is a paucity of Australian literature reporting long‐acting reversible contraception (LARC) usage following termination; however, a Swedish study found a rate of 34%2 and a New Zealand public hospital service reported a rate of 44.5%.3 Our post‐termination LARC rate was 58.2% (72.5% following surgical termination and 43.2% following medical termination). We attribute the higher rate to our conscientious use of tiered counselling4 (ie, presenting the most efficacious method first) to enable women to make informed decisions regarding contraception. The most common choice of contraception following surgical termination was a levonorgestrel‐releasing intrauterine device (Mirena [Bayer]) (56.0%); however, oral contraception was more popular (38.8%) following medical termination, followed by Mirena (25.9%). A British study also found that women more often chose an intrauterine device or implant after surgical termination,5 suggesting that ease of insertion under anaesthetic influences choice. Of the 134 women documented as not using contraception in our audit, three had received inaccurate medical advice regarding contraception, or had been inappropriately labelled as infertile due to age, endometriosis or polycystic ovary syndrome. A further three pregnancies occurred as a result of removing LARC without arranging alternative contraception. We support the researchers’ conclusion regarding the need to explore reasons for poor contraception uptake and determine where education would be most helpful.1 In addition to increasing access to and promoting informed choice of LARC in primary health care settings,6 greater efforts to improve practitioner knowledge of contraception management, as well as common gynaecological conditions to aid appropriate diagnosis and evaluation of fertility impact, are vitally important.

Sophie L Yeates · Charlotte V Elder · Sonia R Grover

TO THE EDITOR: After a high‐pressure injection injury with an oil‐adjuvant vaccine many patients are triaged exclusively as a needlestick injury. This incomplete classification reduces the likelihood for early identification of local or systemic infections or injury, zoonoses or allergic or anaphylactic reactions.1 A review of European agricultural workers showed that of 59 patients who experienced needlestick injuries, 20 cases (34%) involved oil‐adjuvant vaccines.2 Surgical intervention was reported in only 25 patients (42%),2 contradicting product label directions, which instruct that the wound should be incised and irrigated to remove the vaccine. A similar need for an improved treatment plan was recently articulated after a high‐pressure injection injury from a ruptured hydraulic hose in an Australian farmer.3,4 Oil in water emulsions are commonly used by Australian farmers to vaccinate sheep and goats against Mycobacterium avium subsp. paratuberculosis, which causes the chronic wasting condition ovine Johne disease in ruminants and camelids. The Australian Pesticides and Veterinary Medicines Authority (APVMA) is the independent federal authority that regulates the safe supply and use of veterinary medicines and agricultural chemicals for sale in Australia. The APVMA maintains the Australian Adverse Experience Reporting Program, which is a post‐market program that monitors roughly 5500 reports received annually from product registrants, medical and veterinary professionals, and members of the public. An epidemiological review shows that the peak time of year for high‐pressure oil‐adjuvant injection injuries in Australian adverse experience reports is in spring and summer (October–February), when livestock vaccination programs for young animals are at their peak. Potentially serious long term adverse outcomes, including amputation, are possible without appropriate early intervention as described on the product label.2,3,4 Of the 210 adverse experience reports in humans related to mineral oil injections reported over the life of all registered products, the most common immediate reactions mimicked those of a sharps injury: needlestick injury, a reaction at the injection site, pain and swelling (Australian Adverse Experience Reporting Program data). In order to be prepared for the high volume season for livestock vaccination programs, the APVMA recommends that medical professionals, particularly those serving populations heavily involved with primary production, revisit best practices for the management of this type of injury (Box). Health services are encouraged to include “high‐pressure injection injury” as a triage entry option to reduce the potential misinterpretation of a “needlestick injury” entry.4 Appropriate early intervention reduces the risk of subsequent complications and adverse outcomes. Underpinning the entire process is a need for health care professionals to receive appropriate training to recognise injuries of this type and initiate early intervention as per the product label and the manufacturer's instructions. Box – An example of appropriate triage and treatment for a patient presenting with a high‐pressure injection injury* *The green box indicates the occurrence of an injection incident, based on data from the Australian Adverse Experience Reporting Program. Checkpoint steps in the process that can have a significant impact on patient outcomes if omitted are indicated by orange boxes. † This is the most common formulation for injections containing Mycobacteria

Elvira Currie · Rhian Cope · Margaret C Hardy

TO THE EDITOR: Skin disease contributes to the health disadvantage of Indigenous Australians, and the recent publication of the healthy skin guidelines is welcome.1 Scabies is a significant health problem for some remote Indigenous communities, and the healthy skin guidelines describe a series of community‐based scabies control programs. These programs resulted in initial reductions in prevalence of scabies, but they were not sustained, as human scabies was eventually reintroduced.1 The guidelines used a systematic literature review to ensure that all relevant research was included. However, since the review of scabies was limited to studies of human scabies, implications of canine scabies may have been overlooked.2,3 Therefore, a statement such as “dog control programs are of no benefit to the community control of human scabies infestations” requires scrutiny. Canine scabies mites are distinguishable from human scabies mites by genotyping.4 Clinically, canine scabies can cause a transient human infestation, with no ongoing transmission cycle. The lesions of canine scabies occur primarily on body areas that have been in contact with dogs, and are intensely itchy after a shorter period compared with lesions of human scabies. The infestation is self‐limiting unless the person is reinfested.5 As with human scabies, the intense itch from infestation by canine scabies can lead to scratching and skin trauma, providing an entry point for bacterial infection. Complications such as post‐streptococcal glomerulonephritis and chronic renal disease can also arise from canine scabies. Moreover, outbreaks of scabies in humans can be caused by repeated transmission of canine scabies.5 Management of people affected by canine scabies includes treatment of affected dogs and their contacts.3 Comprehensive dog health programs provide broad‐based community benefits, including reduced injuries from dog attacks, improved community and workplace safety, reduced sleep disturbance from barking and fighting dogs, and enhanced dogs’ appearance, behaviour and wellbeing. Dogs are considered family members in some Aboriginal and Torres Strait Islander communities, sharing housing, bedding and food; hence, human and dog health and wellbeing are intimately linked. No trials have yet examined the impact of dog health programs on scabies transmission in humans or other health outcomes in the remote Indigenous communities where scabies is a public health problem.2 Without evidence from trials, the impact, or lack thereof, of dog health programs on human health is speculative.

Rosalie Schultz

To the Editor: Norman and Teede outline the new international guidelines on polycystic ovary syndrome (PCOS), led by Australia and involving 37 societies and patient support groups and 71 countries.1 These guidelines highlight gaps in evidence and emphasise the critical need for more research into PCOS.2 In the United States, a recent analysis of National Institutes of Health (NIH) research funding from 2006 to 2015 for PCOS concluded that PCOS research may be underfunded by the NIH.3 In Australia, the National Health and Medical Research Council (NHMRC) is the premier funder of medical research and its main funding mechanism is by way of project grants, with over 500 granted annually. Using NHMRC online data (https://www.nhmrc.gov.au/grants-funding/outcomes-funding-rounds/previous-outcomes-project-grants-funding-rounds) and searching for “polycystic” or “PCOS” in the titles of funded project grants, we found only nine grants associated with PCOS from 2003 to 2018. Additionally, while there are many not‐for‐profit organisations raising funds for medical conditions, there are none for PCOS. We have considered the issues that may affect funding for PCOS. It could be that the name PCOS does not accurately describe the condition because having polycystic ovaries is neither needed nor sufficient in order to diagnose PCOS, and the name does not indicate any of the condition's important metabolic symptoms.4 This could potentially lead to grants being assigned to panels without the full expertise required to handle such grants. Barriers to funding in Australia could also potentially include a lack of internationally competitive researchers in PCOS in Australia, but this is not the case. Australia has established an international network in PCOS and led the world by producing the first evidence‐based guidelines.2 Three individual Australian researchers are listed in the top ten in the world in PCOS research (http://expertscape.com/ex/polycystic+ovary+syndrome). This attests to the calibre of Australian researchers in PCOS. We acknowledge that the international guidelines were funded in part by the NHMRC via a Centre for Research Excellence in PCOS. The Centre's efforts have positioned Australia at the forefront of international PCOS activities and have highlighted the vital need for specific dedicated research funding. However, given Australia's leading global role in the development of PCOS guidelines and identification of knowledge gaps, we, along with patient support groups, believe that greater efforts are required to recognise the prevalence, diverse clinical impact and health and economic burdens of PCOS, and to prioritise funding for research into PCOS.

Raymond J Rodgers · Jodie Avery · Veryan McAllister

In reply

Helena J Teede · Robert J Norman · Rhonda M Garad

TO THE EDITOR: In 2016, there were an estimated 490 000 cases globally of multidrug‐resistant (MDR) tuberculosis exhibiting resistance to isoniazid and rifampicin.1 The first case of MDR tuberculosis diagnosed in Tasmania occurred in 2016 in a Vietnamese‐born person. Vietnam was the second highest reported country of birth for overseas‐born patients with tuberculosis notified in Australia in 2014.2 The patient had previously tested positive for tuberculosis infection in an interferon‐γ release assay test performed in Tasmania in early 2016, but at the time, the patient was asymptomatic and had a normal chest x‐ray and a negative sputum culture. After an episode of colitis, a colon tissue biopsy specimen isolated Mycobacterium tuberculosis. Whole genome sequence of the isolate (TASMDR1), identified high confidence mutations for isoniazid, rifampicin, ethambutol and pyrazinamide, in accordance with the culture‐based drug susceptibility testing, and, in addition, it identified a mutation associated with streptomycin resistance.3 We became aware that a household contact of the Tasmania‐located patient with MDR tuberculosis had been diagnosed with pulmonary tuberculosis in Vietnam in 2012 and requested the drug susceptibility testing data for this isolate (VTB1) from the treating hospital in Ho Chi Minh City. VTB1 was resistant to isoniazid, rifampicin, ethambutol, pyrazinamide and streptomycin in culture‐based drug susceptibility testing. We therefore obtained a genomic DNA preparation of VTB1 to enable direct comparison with the TASMDR1 isolate collected in Tasmania. Next generation sequencing of VTB1 was performed on an Illumina platform and paired‐end reads were mapped to the M. tuberculosis H37Rv reference genome (NC_000962.3). The Box shows variants associated with drug resistance. In addition to drug resistance mutations, VTB1 shared all previously described variants in TASMDR1 with respect to H37Rv3 and, therefore, the two isolates were genetically indistinguishable. This is strongly indicative of transmission involving the two patients based on established single nucleotide polymorphism thresholds.4 It is most probable that the patient diagnosed in Tasmania contracted the MDR strain of M. tuberculosis during the episode of pulmonary disease diagnosed in the household contact in 2012 and that the infection remained latent until reactivating as extrapulmonary MDR tuberculosis in 2016. In conclusion, the global burden of latent tuberculosis infection has been estimated to be 23% of the world's population, which corresponds to about 1.7 billion people.5 Despite the immense prevalence of latent tuberculosis infection, there are few reports in the literature that confirm using genome variant analyses for the translocation of the MDR form of tuberculosis from one jurisdiction to another as a latent infection and its subsequent emergence as active MDR tuberculosis in a new host country. This type of transit of tuberculosis is difficult to detect with pre‐immigration screening practices that are reliant upon a diagnosis of pulmonary tuberculosis based on a chest x‐ray. The international movement of MDR tuberculosis in latent form, as has been determined in this case, is an area of concern and could be a significant challenge for future tuberculosis eradication. The growing application of genome sequencing in tuberculosis diagnostics and surveillance will help establish the level of MDR tuberculosis cases due to reactivation of latent tuberculosis infection. Box – Drug resistance determining mutations in a contact who presented with tuberculosis in Vietnam in 2012 (VTB1) and in the first confirmed patient with multidrug‐resistant (MDR) tuberculosis in Tasmania in 2016 (TASMDR1). The isolate from VTB1 and TASMDR1 share identical drug resistance mutations Isolates Drug Gene Function Mutation Substitution VTB1 and TASMDR1 Rifampicin rpoB (Rv0667) RNA polymerase β‐subunit gAc/gGc, tCg/tTg D435G, S450L Isoniazid katG (Rv1908c) Catalase‐peroxidase aGc/aCc S315T Pyrazinamide pncA (Rv2043c) Pyrazinamidase/nicotinamidase cCg/cTg P62L Ethambutol embB (Rv3795) Arabinosyltransferase B Atg/Gtg M306V Streptomycin rrs (MTB000019) 16S ribosomal RNA a/c a514c* * Substitution located in a non‐protein coding gene. ◆

Sanjay S Gautam · Greg Haug · Louise A Cooley · Micheál Mac Aogáin · Ronan F O'Toole

TO THE EDITOR: The article by Stevens and Schweitzer1 contains many useful teaching pearls for frontline physicians and generalists. I concur with the final diagnosis reached and the excellent teaching points emphasised in the article. I would like to make two important points regarding localisation. First, binocular diplopia can be caused by diseases in a variety of locations in the neurological axis, not just the extraocular muscle. Other possible locations include the brainstem, the cranial nerves (III, IV and/or VI), the neuromuscular junction and the orbit itself. Accompanying features will help distinguish the location; for example, the presence of hemiparesis, cerebellar signs, or lower cranial nerve abnormalities would localise to the brainstem. Second, localisation is critical before organising neuroimaging, as knowing where the lesion is will prove useful for the radiologist. The likelihood of subtle pathologies being detected by the radiologist increases significantly when the site of potential localisation is included in the clinical request information. This helps to reduce the possibility of false negatives or misinterpretation if clinical correlation is not applied to test requests.2 Finally, there are two differential diagnoses to consider in this case. Myasthenia gravis, an uncommon disorder of the neuromuscular junction, is known to be a great mimicker and can cause various patterns of ophthalmoparesis and diplopia3 and it should be considered in all cases of binocular diplopia. In the article by Stevens and Schweitzer,1 this diagnosis was less likely in the patient given the painful nature of the ophthalmoparesis. The other rare differential diagnosis with serious implications to consider is a dural (or indirect) carotid‐cavernous fistula. Patients with this condition can present with painful diplopia without significant visual loss; they are frequently misdiagnosed by specialists and are eminently treatable endovascularly.4 The absence of orbital signs (eg, proptosis) and ocular signs (eg, arterialised “corkscrew” conjunctival vessels) assisted in excluding this condition as a differential in this case.

Benjamin Nham

TO THE EDITOR: We commend Heldens and colleagues1 for publishing their data on the incidence and in‐hospital mortality of sepsis and septic shock among patients admitted to Australian intensive care units (ICUs). The incidence of sepsis and septic shock in ICUs is estimated to be 101.8 and 19.3 per 100 000 patient‐years, respectively, at an attributable cost of $32 421.2 We concur that sepsis cases captured using the Australian and New Zealand Intensive Care Society Centre for Outcome and Resource Evaluation database criteria, compared with prospective clinical diagnoses,3 has poor sensitivity for sepsis case ascertainment. Notwithstanding, we propose that the application of a third surveillance metric using coded discharge data could be a viable alternative for sepsis case ascertainment and monitoring in ICUs. International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD‐10‐AM) diagnostic coding data are feasible to collect with a reduced risk of sampling bias and minimal loss to follow‐up. Using tandem dataset comparison following the implementation of a hospital‐wide sepsis pathway,4 we explored the utility of coding data for sepsis surveillance. We noted that 78% and 74% of ICU cases were designated an ICD‐10‐AM code denoting sepsis at admission and patient level, respectively (Box). Alarmingly, the concordance rate between coded administrative data and clinically verified sepsis diagnoses was even lower in non‐ICU settings. These data are in keeping with international reports.2 Robust and reproducible data are required to evaluate quality improvement regarding sepsis management. Given the poor sensitivity of research criteria and coding data, used in isolation for sepsis identification, a multifaceted approach is required. We hypothesise that the combination of administrative coding data and electronic medical record data, augmented with sepsis screening algorithms, may improve the sensitivity for sepsis case ascertainment in both cancer and non‐cancer settings.5 We encourage Heldens and colleagues to consider these suggestions as an alternative reproducible method needed to elucidate the incidence of sepsis and septic shock in Australian ICUs. Box – Relationship between sepsis cases satisfying clinical criteria and designated coded discharge data in intensive care unit (ICU) and non‐ICU settings, 2012–2014 Year Admission level ICU Non‐ICU All new admissions* ICD‐10‐AM captured cases Concordance All new admissions* ICD‐10‐AM captured cases Concordance 2012 38 27 71% 70 62 89% 2013 39 34 87% 175 103 59% 2014 81 61 75% 331 149 45% Mean (± SD) – – 78% (± 8.3%) – – 64% (± 22%) ICD‐10‐AM = International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification; SD = standard deviation. * Clinically diagnosed sepsis cases according to consensus diagnostic criteria.◆

Jake C Valentine · Gabrielle Haeusler · Leon Worth · Karin Thursky

TO THE EDITOR: We congratulate Heldens and colleagues1 for their work investigating the prevalence and mortality of sepsis within a tertiary hospital intensive care unit (ICU). As the authors rightly point out in their article and podcast, this requires careful screening, consistent diagnostic criteria, and considerable time and effort. The Australian and New Zealand Intensive Care Society (ANZICS) registry provides a highly specific, widely applicable, cost‐effective, timely and practical epidemiological measure of the prevalence and outcomes of sepsis and septic shock as a primary cause of ICU admissions throughout Australia and New Zealand.2 Findings from the registry appear consistent with those of Heldens and colleagues.1 However, their article confirms that there are more patients with sepsis within our ICUs than those who enter with this as an admission diagnosis. Between January 2016 and June 2018, 11.6% (43 529/374 442) of the ICU admissions reported to the ANZICS Adult Patient Database were due to sepsis as defined by the international Sepsis‐3 taskforce.3 In tertiary hospital ICUs, this was slightly higher at 12.1% (19 204/159 067), which is between 14.0% (121/864) by clinical criteria and 11.3% (98/864) by database criteria found within the first 24 hours of ICU admission by Heldens et al. With the recent addition of information about vasopressors and lactate levels, ANZICS can also now confidently identify patients with septic shock. Of 59 069 ICU admissions with available information, 3.3% (1978) had septic shock, again similar to the findings of Heldens et al. ICU admissions in the ANZICS registry due to sepsis or septic shock were associated with mortality of 15.1% and 26.7%, respectively. Heldens et al reported a similar mortality of 24% (8/33) with septic shock identified in the first 24 hours of admission (Appendix, table 4). The study by Heldens et al and the findings from the ANZICS registry are complementary. Together they highlight how the measurement of sepsis depends on the exact definition applied, and the population under consideration, which may vary between clinical practice, prospective observational and interventional studies, and widespread epidemiological comparisons.

David Pilcher · Sue Huckson · Peter Hicks

To the Editor: Schilling and colleagues1 state that the Australasian Rehabilitation Outcomes Centre (AROC) — the national rehabilitation clinical quality registry for Australia and New Zealand — does not routinely collect data on post‐surgery outcomes for private total knee replacement (TKR) recipients. This statement is factually incorrect. All private inpatient rehabilitation services in Australia are members of AROC and routinely submit data (including functional outcomes as assessed by a functional independence measure) describing all episodes of rehabilitation they provide. More specifically, over the period described by Schilling and colleagues,1 AROC received data on outcomes for 93 278 TKRs receiving private rehabilitation. If we restrict the AROC data to match the study data (patients aged 40–89, single TKR, first admission), AROC received data describing 76 847 privately rehabilitated TKRs. In rehabilitation, the Australian National Subacute and Non‐Acute Patient Classification2 is routinely used to classify episodes into resource‐homogeneous groups. In interrogating the AROC TKR data, we concur with Schilling et al1 that the average length of stay in rehabilitation has been declining, with this decline accelerating over the past 5 years. Concurrent with the decline in length of stay, the functional change achieved (both absolute and relative) during rehabilitation has been increasing, and has in fact accelerated over the past 5 years. Achieving more functional change in a shorter length of stay shows that services are becoming more efficient while also continuing to produce positive outcomes for their patients. Moreover, it is also factually incorrect that AROC does not collect data outside of the inpatient setting. In fact, AROC also runs an ambulatory benchmarking initiative, and while coverage is not 100%, it is growing. There are currently 35 private ambulatory rehabilitation services that participate and routinely provide data describing their ambulatory rehabilitation outcomes. In conclusion, we suggest that while the authors provide an interesting analysis, it is incomplete, given that they did not include function — the key driver of cost and outcomes in rehabilitation — as one of the variables they used.

Frances Simmonds · John H Olver

To the Editor: In reply to Shilling and colleagues,1 the Rehabilitation Medicine Society of Australia and New Zealand refers the authors and readers to our position statement regarding referral for rehabilitation in the home after total knee replacement (TKR).2 Shilling and colleagues1 state that the most important determinant for referral to inpatient rehabilitation was the hospital where the TKR took place. Independent researchers might be more circumspect, considering there is no acknowledgement that Medibank Private did not fund rehabilitation in the home nationally during the study period nor whether their data included outpatient rehabilitation carried out as “same day rehabilitation”, usually coded as inpatient. Also, disturbingly, some of the literature is misrepresented. The unblinded Canadian randomised controlled trial3 comparing a publicly funded combination of rehabilitation in the home and hospital‐based outpatient therapy with inpatient rehabilitation is not generalisable to privately insured Australian patients. Moreover, the Australian randomised controlled trial4 showing equivalent outcomes for the same two groups excluded patients who were appropriately referred for inpatient rehabilitation on the basis of numerous patient factors. The present study included few patient factors and not clinically relevant factors, such as obesity, ability to walk after TKR, or complications.1 Finally, while no patient safety or outcome data were included, the choice to include the dollar value of the previous year's private hospital claims seems gratuitous — are those patients with higher cost to insurers more likely to use inpatient rehabilitation, or perhaps they were just sicker? It is interesting that no reference is made to the 2017 study that found that referrals to inpatient rehabilitation were directly influenced by preferences of the patient, the surgeon, therapists, discharge planners, insurers and others.5

Steven G Faux · Lee Laycock