Letter: comment on swine flu algorithm
2 October 2009. Related: Correspondence.
To the editor:
Following the publication of the editorial in the July/August 2009 edition of HIV Treatment Bulletin (Vol 10, No. 7/8), we would like the opportunity to respond to the content and also some of the comments made.
Firstly, the flow diagram (referred to as Example 2: Flow Diagram B) included in the editorial, is not the version that is currently being used by our Directorate. The flow diagram was designed to assist in the triage of patients who contacted our unit by telephone, within the context of increasing demand during an emerging influenza pandemic. It is not being used for patients who have actually attended our unit, all of whom are seen by a clinician and undergo assessment and diagnostic testing where appropriate.
Furthermore, the flow diagram represents only part of our pandemic (H1N1) 2009 clinical guidance package for our HIV clinicians. Thus, without including our clinical guidelines which clearly remind clinicians to consider pandemic (H1N1) 2009 infection whilst not attributing all presentations of flu-like or non-specific febrile symptoms to infection with influenza virus the triage tool can be taken out of context and perhaps appear to oversimplify the assessment of such patients.
In comparison with the Birmingham Heartlands Hospital flow diagram (Example 1), we elected not to include a history of contact with a case of swine flu in our flow diagram. In line with national and international guidelines during a period of sustained transmission of pandemic (H1N1) 2009 in the community, the absence of contact with a known or suspected case is no longer considered to be helpful in assessing the likelihood of a patient being infected with pandemic (H1N1) 2009. At out unit, only one patient in our existing cohort of patients with confirmed pandemic (H1N1) 2009 had an appropriate contact/exposure history.
We feel that the comment “a caution with this option is that a patient with a life-threatening condition who thought they had flu could end up going down the left-hand no further action route”, is inaccurate and may alarm our patients. Every endpoint in our telephone triage flow diagram results in the patient obtaining medical advice, rather than “no further action”. Potentially life-threatening conditions that are not flu-related can occasionally present with early symptoms suggestive of uncomplicated influenza, and this is a risk associated with existing local, national and international pandemic influenza management algorithms, including both of the flow diagrams published in the editorial. In the context of an established pandemic, for our patients who make contact by telephone and dont have risks for, or symptoms of, severe pandemic (H1N1) 2009 disease, but in whom uncomplicated influenza is thought to be a possibility, we advise them to contact their GP or the National Pandemic Flu Service for further assessment, with the additional advice that they should attend hospital if they deteriorate or fail to improve.
The risk factors for developing severe pandemic (H1N1) 2009 disease are not yet completely defined; therefore we chose to adopt the list of medical conditions used by the Department of Health and several other health advisory bodies around the world. Such lists are subject to change as more information on cases from the first wave becomes available. For example, although pregnancy and asthma appear to be significant risk factors, reanalysis of initial data from the United States, which suggested that morbid obesity is a significant risk for severe pandemic (H1N1) 2009 infection, demonstrates that morbid obesity may not in fact be a significant risk in their population overall. As accurate risk identification is often a retrospective exercise, we chose to follow the best availableevidence and consensus expert opinion on possible risk factors for severe disease during the first wave of the pandemic.
Algorithms such as these are difficult to develop and are subject to review in order to adapt to an evolving pandemic with predicted separate waves of varying disease activity. They are designed to facilitate the rapid assessment of a potentially large number of patients presenting with symptoms that may be consistent with pandemic influenza, whilst recognising that the same symptoms can also occur with many conditions seen in HIV-infected individuals. As we come to the end of the first wave of the pandemic, we have managed to assess the majority of our patients presenting with flu-like symptoms in person. However, the situation may clearly change during the predicted second wave, which is expected to affect a greater proportion of the UK population and place significantly increased demands on healthcare services. As such, we would fully support the refinement of existing guidelines and algorithms, along with a collaborative effort to pursue a uniform and safe approach to the assessment of possible pandemic (H1N1) 2009 infection in HIV-infected patients across the UK.
Yours Sincerely,
Dr Jake Dunning and Professor Brian Gazzard
Department of HIV and Genitourinary Medicine
Chelsea and Westminster Hospital Foundation Trust.
editors reply
The inclusion of the C+W diagram was only included as an exxample of a telephone triage tool for other clinics to start from, and was seen in context by Professor Gazzard in draft form prior to publication.
The provision of two examples was not meant to reflect any difference in clinical responses resulting from the two different solutions, and we are sorry if this impression was given. They were both included to highlight how two clinics, both with expertise in HIV and the swine flu epidemic, interpreted differently the limited available data.
It would clearly be preferable for clinics to agree on one protocol, with appropriate caveat where difference remain, and we hoped this might have been possible prior to this issue of HTB going to press. This would be the most useful outcome for HIV-positive patients on a national level.
The latest version of the C+W diagram is posted to the i-Base website.