I’ve spent a bit more time on Flu reports this year than in past years, so I thought I might compare the data that the public can see. There are no trade secrets in this post – I don’t talk about the specifics of my work in public. These are my personal observations based on reports that are in the public domain.

A couple of quick points: a laboratory-confirmed influenza infection is referred to as a ‘notification’ below. I’ve set today as the reference point against which I define how many weeks old a given report’s data are.

I had a quick look at each site for the raw de-identified data (perhaps to be found as a single large, curated line list akin to what the World Health Organisation now do for MERS-CoV?) to download and play with. It was not an exhaustive search. I did come across links where data can be requested but no actual data. I think it would be nice to have these data, in a non-re-identifiable basic format, linked to an appropriate license to recognise the source and hard work that went into curating them for download by data nerds. The Australian Government has a Data Access and Release Policy that outlines some of these uses, and the phrase “Data that can be made public should be made public” is a good one here.

So, clockwise from Queensland, we have:

Queensland reports (QLD)…[1]

Published weekly (with data up to the week before the current week) but only during the Flu season, these Queensland Government, Queensland Health reports list lab-confirmed Flu cases from public and private testing laboratories.

The current reports graph:

• public lab year-to-date (YTD) weekly notifications by type and percentage positive
• YTD weekly notifications by State region
• YTD weekly notifications by age group and gender
• YTD weekly notifications of a few other respiratory viruses (RSV, AdV, hMPV and PIV1-3) as a percentage of samples tested
• YTD weekly notifications of public hospital admissions, type of admissions, Flu type and number per 1,000 notifications
• YTD weekly notifications of public hospital admissions and Flu type, subtype and coinfection status
• YTD weekly bed occupancy for patients with confirmed Flu by admission type
• YTD notifications of public hospital admissions by age group and gender
• YTD plus past 5 years of weekly notifications as an appendix.

There is also an appendix map of Queensland Hospital and Health service districts. Ten figures and 3 tables.

The site does not keep these reports up – they are replaced by the next weeks’ – nor does it host the data in an open-access downloadable format. If you want to compare anything between weeks – make sure you save yourself a copy each week!

New South Wales (NSW)…[2]

We have 10 figures in this NSW Government weekly report (with data up to the week before the current week).

Graphs include:

• YTD and past 5-year weekly non-specific “influenza-like illnesses” (ILIs) in NSW hospitals
• YTD and past 5-year weekly emergency department visits for ILIs that were admitted to hospital
• YTD and past 5-year weekly emergency department visits for ILI and pneumonia that were admitted to a critical care ward
• YTD weekly FluCAN (The Influenza Complications Alert Network; three sentinel hospitals) notifications by Flu type, subtype and coinfection
• YTD weekly sentinel laboratory (a selection of different labs) notifications by Flu type, subtype and percentage of samples testing Flu positive
• YTD and past 5-year percentage of lab tests positive for FluA or FluB, Flu season weekly reported Flu outbreaks in different types of NSW institutions (hospital, aged care, other)
• YTD and last year’s weekly Australian Sentinel Practices Research Network (ASPREN) ILI cases comparing NSW to the national rate, Flu season weekly
• 5-YTD mean FluTracking self-reported ILI and self-reported Flu vaccination status.

There are also 4 tables with various details.

The site does keep these reports up; monthly prior to the Flu season and weekly during the season. The site does not appear to host these data in an open-access downloadable format.

Australian Capital Territory (ACT)…[3]

We have 2 figures and a table in this ACT Government fortnightly report (with data about a month old and reporting only starting from Week 28 of 2017).

The graphs include:

• YTD and past 4-year weekly notifications
• YTD and last year notifications by age group and total

The site does keep these reports up but does not appear to host these data in an open-access downloadable format.

Victoria…[4]

There are 3 graphs and 1 table in this Victorian Government weekly report (with data from two weeks before the current week) but only during the Flu season.

The graphs include:

• YTD and past 7-year weekly notifications by Flu type along with Victorian Sentinel Practice Influenza Network (VicSPIN) and National Home Doctor Service (NHDS) ILI tested Flu-positive proportions
• Flu season weekly VicSPIN ILI patients per 1,000 total consultations
• Flu season weekly notifications for a comprehensive (#kudos) list of respiratory viruses, including Flu types and proportion Flu positive,

The site does keep these reports online for the current year but does not appear to host these data in an open-access downloadable format.

Tasmania…[5]

There are 4 graphs and 2 tables in this Tasmanian Government monthly report, but only during the Flu season.

• YTD and past 4 year weekly notification
• YTD weekly Flu lab test result and proportion positive
• YTD weekly very good respiratory virus and bacteria total positives (proportion of those tested would be useful too) panel (Flu subtypes, rhinovirus, AdV, hMPV, RSV, HPIVs, Mycoplasma pneumoniae, Bordetella pertussis)
• YTD FluTracking self-reported ILI and Flu notifications. 

The site does keep these reports online for current and past years but does not appear to host these data in an open-access downloadable format.

South Australia (SA)…[6]

There is 1 graph in a PDF and separate PDFs with text tables or a text summary. Maybe I’ve missed something among these SA Government reports? Data seem to include up to the week before the current week, but I cannot tell if these are only published during the Flu season or every week of the year.

The graph is:

• YTD and past 2 year weekly notification with Flu types, also plotting emergency department and ASPREN diagnoses.

The site does appear to be keeping these reports online for current or past years and does not host these data in an open-access downloadable format.

Western Australia (WA)…[7]

There are graphs galore in this general “Virus WAtch” report with 10 pertaining to respiratory infection and 0 tables in this weekly WA Government report. Data include up to the two weeks previous to the current one.

The graphs include:

• Flu season weekly routine and ASPREN Flu notifications and ILI per 1,000 consultations
• YTD and past 4-year weekly ASPREN ILIs, Flu season weekly ILI emergency department presentations and admissions and those per 1,000 rates
• YTD and past 4-year weekly emergency department presentations for respiratory viruses
• YTD and past 4-year weekly Flu notifications with a seasonal threshold (defining epidemic levels of seasonal transmission)
• Flu season weekly Flu notification that have been hospitalised by Flu type and subtype with 5-year mean number of hospitalised cases
• Flu season weekly FLu cases per 1,000 hospital admissions with a weekly 5-year mean number of cases per 1,000 admissions
• YTD and past 4-year weekly PathWest lab percentage Flu-positive tests
• YTD and past year weekly Flu notification with subtype and coinfection
• YTD and past year weekly non-Flu PathWest lab respiratory virus positives.  

The site does keep these reports online for current and past years, but does not host these data in an open-access downloadable format.

Northern Territory (NT)…

I couldn’t find a regular influenza surveillance report for the NT and the federal webpage had no link but does include some data in its graphs (see below).

Federal Government (national) Department of Health…[8]

Australia’s Flu report is a biggie. Unfortunately, it’s also a latey – as I write this, the current report lists data from up to 3-weeks ago. In its defence though, this team have to get everyone else’s data and collate them and curate their own lists and as we saw above, some state’s produce fortnightly reports. This job would give me an ulcer. Nevertheless, this doesn’t help a journalist trying to satisfy their need for facts to add into weekly Flu stories-especially when we have a huge season like we’ve had this year. The site links to its downloadable seasonal Flu report (all the reports above are at least in PDF format) and to previous reports, but once again there is no obvious sign of downloadable open-access raw data.

The report hosts 17 graphs (winner!) and 1 table. The graphs include:

• a map that is not specifically referenced anywhere in the text (Grr) showing a change in Flu activity compared to 2-weeks ago
• YTD weekly proportion of sentinel lab tests positive for Flu from around the nation – excluding QLD (?)
• YTD weekly proportion of same sentinel lab tests that were Flu positive with Flu subtype
• YTD and previous 4-year national (now including QLD) weekly Flu notifications
• YTD weekly Flu notifications graphed separately for each state or territory (includes Northern Territory)
• YTD proportion of Flu notifications by type and subtype, coinfection and all states or territories
• YTD weekly national proportion of Flu notifications by subtype
• YTD weekly and cumulative Flu notifications by age groups broken into Flu subtypes
• YTD total Flu notifications by age groups with total and Flu type and subtype
• YTD weekly and cumulative notifications across multiple graphs by age groups with Flu type and subtype (nice one)
• YTD and previous 4-year weekly FluTracking self-reported fever and cough
• YTD and previous 4-year weekly call centre self-reported ILI
• YTD and previous 4-year weekly ASPREN and VicSPIN GP ILI surveillance reports
• YTD and previous 4-year weekly proportion of ASPREN and VicSPIN GP ILI surveillance data that test positive for Flu virus with type and subtype and consultation rate
• YTD and previous 4-year weekly notifications of sentinel hospitalisations
• YTD weekly notifications of sentinel hospitalisations by type and subtype, including ICU admissions
• YTD and past 5-year death rates from NSW registered death certificates listed as due to influenza and pneumonia.

This report also has a table of haemagglutination inhibition typed Flu virus isolates from around the nation (n=736) and some text noting the existence of a “small number” of low reacting (to the vaccine) strains. This section could be made more clear, especially about the significance of a larger number of weakly haemagglutinating H3 isolates.

Loootta work.

What’s brilliant…?

Queensland’s (I live and work here but do not contribute to the reports) reports are based on notifications (not ILIs or FluTracking self-reported illness). This is concrete. I like that. Queensland also has a great balance of explanatory text and graphics. Some states include other respiratory viruses – this is brilliant. They should all include rhinovirus – some don’t. Boo. 

The national report plays with the data – 17 graphs speak to that – and it’s great to see, I just wish it could come out more rapidly.

What’s missing…?

For me, there are a couple of standout missing things that would be nice to know during the season. I’m saying this both as a nosy member of the public and as a virologist who talks to the media to help with interpreting fast-moving outbreaks. Data are key to understanding what’s happening – being able to quickly access and interpret data to help the public understand what’s going around them is therefore essential. These are public data gathered using public funds and referring to public health. I’d like to see the following:

1. make the data available for download
2. keep past reports up. Most states do this. Queensland and South Australia do not (as far as I can tell). Northern Territory don’t report
3. better data and interpretation about Flu virus mutations that occur during the season and how this matches/mismatches with the current vaccine
4. more of No.3 – typing of only a small percentage of the Flu viruses detected is a common practice worldwide – but since I’m making a list, I’d like to see more done and more interpretation of the genetic sequencing data that accrues during the season. This all needs funds, of course.

These reports are a lot of work, and everyone involved should be congratulated.  

References…

1. https://www.health.qld.gov.au/clinical-practice/guidelines-procedures/diseases-infection/surveillance/reports/flu
2. http://www.health.nsw.gov.au/Infectious/Influenza/Pages/reports.aspx
3. http://health.act.gov.au/public-information/public-health/public-health-alerts/influenza-act
4. http://www.vidrl.org.au/surveillance/influenza-surveillance/
5. http://www.dhhs.tas.gov.au/publichealth/communicable_diseases_prevention_unit
6. http://www.sahealth.sa.gov.au/wps/wcm/connect/public+content/sa+health+internet/about+us/health+statistics/surveillance+of+notifiable+conditions
7. http://ww2.health.wa.gov.au/Articles/F_I/Infectious-disease-data/Virus-WAtch
8. http://www.health.gov.au/flureport

Views: 4023