Methods for calculating Fever and Cough Incidence
We have answered some frequently asked questions regarding how the FluTracking team conduct their analysis.
Please contact us if your question is not answered below and we’ll get back to you as soon as we can.
What is Fever and Cough Incidence?
We define Fever and Cough Incidence as the number of new cases of fever and cough in participants over a specified period of time (usually the most recent reporting week).
How do you define Influenza-like illness (ILI)?
We define ILI as fever and cough. Thus, if a participant reports ‘fever’ and ‘cough’, they are said to experience “influenza-like illness”.
How do you calculate fever and cough incidence?
For fever and cough percentage calculations, the numerator includes all participants who completed a survey for the current week and reported new fever and cough symptoms, and the denominator includes all participants who completed a survey for that week. Prior to week ending 27 August 2023, ‘Don’t Know’ responses were excluded from the denominator in all analyses. From week ending 27 August 2023, ‘Don’t Know’ responses have been included in the denominator. Additionally, please note that from week ending 27 August 2023, weekly reports have been updated to include ‘Don’t Know’ responses for data preceding the week ending 27 August 2023.
Where there are consecutive weeks of reporting fever and cough symptoms, only the first week is used to determine attack rates. If a person reported fever and cough symptoms in one week, and then reported at least one week of no fever and cough symptoms, followed by another report of fever and cough symptoms, then this second symptom report is considered a new case of fever and cough.
What is age standardisation?
Age standardization is a method used to make it easier to compare data from different groups or populations where they have different age distributions e.g. one population may have an older or younger population than the other. By doing this, we can make more accurate comparisons because we’ve removed the influence of age variations.
Age standardisation methods used in FluTracking analyses
There are two ways to apply age standardisation: direct and indirect methods. FluTracking uses the direct method, which combines data from different age groups to get an overall picture.
Here’s how it works: We look at the percentage of participants who, for example, self-reported fever and cough symptoms in each age group. We then adjust these percentages based on how many people are in each age group in a standard population. In alignment with the guidance of the Australian Bureau of Statistics, the standard population in all FluTracking Australia age standardisation is the Estimated Resident Population on 30 June 2001.¹
Starting from the week ending on August 27, 2023, we group ages in ten-year intervals for this calculation. Before that, we used broader age groups: 0-4, 5-17, 18-64, and 65 and older. The final result is the combined percentage of self-reported fever and cough symptoms across all age groups.
How do you calculate time off work or normal duties associated with fever and cough?
Participant reports of time off work or normal duties at any time during an incident of fever and cough is captured, and recorded in the same survey week of the new fever and cough incident. This is captured even if the time off work was reported in the 2nd, 3rd or 4th week of illness.
How do you calculate seeking medical advice?
Participant reports of seeking medical advice at any time during an incident of fever and cough is captured, and recorded in the same survey week of the new fever and cough incident (even if the medical advice was sought in the 2nd, 3rd or 4th week of illness). For example, if a participant reports visiting a GP both in their 2nd and 4th week of having fever and cough symptoms, these visits are captured in the week of onset of illness, and only counted once in analysis.
How do you calculate influenza/COVID-19 tests?
Reports of being tested for influenza or COVID-19 are captured for all participants, regardless of symptoms. Influenza/COVID-19 tests (and results) reported at any time during a fever and cough incident are recorded in the same survey week of the new fever and cough incident (even if the test (and result of the test) are reported in the 2nd, 3rd or 4th week of illness).
How often do you update these figures?
A weekly interim report is uploaded to the FluTracking website every Tuesday and these figures are updated every Thursday. On Tuesday, we report on data received up until 9:00 AM (Local Sydney time) from the time the survey was sent on the Monday morning. On Thursday, we report on data received up until 9:00 AM (Local Sydney time) from the time the survey was sent on the Monday morning.
Why do you exclude the first survey week of data for each reporting year?
The first week of survey data collection for each year where there is a break in data collection over summer (e.g. surveys only collected from April – October each year) is likely to be an overestimate of fever and cough percentages. This is because the reported symptoms for the first week of data collection for the year (where there is a time series break) reflect prevalence rather than incidence. The case definition for fever and cough incidence requires the onset of new symptoms which cannot be assessed in the initial week. Where data is not continuously collected over the year (i.e. all survey years except 2009-10 and 2020-21), please consider excluding the first survey week of the year from analysis of fever and cough percentages.
References:
1) Australian Institute of Health and Welfare 2011. Principles on the use of direct age-standardisation in administrative data collections: for measuring the gap between Indigenous and non-Indigenous Australians. Cat. no. CSI 12. Canberra: AIHW.
