Binomial probability tool

The tool can be accessed at the following link: binomial probability tool

The tool can only support a certain number of users at once. If you cannot access the tool, please try again later or use the previous version of the tool. The previous, Excel spreadsheet version of this tool can be accessed here: Excel binomial probability tool. If you still cannot access the tool, try clearing your browser’s cookies and cache.

Purpose

The Food Exposure Analysis Tool facilitates the comparison of case exposure data against population reference values from the Foodbook Report.

Compare your case exposures to typical population exposures from Foodbook to prioritise hypotheses during outbreak investigations.

Data Sources

Foodbook is a population-based survey conducted in all Canadian provinces and territories. It provides essential data on food, animal and water exposure used to understand, respond to, control and prevent enteric illness in Canada.

Foodbook 2.0 (2023-2024): Online and telephone survey with ~21,000 respondents across Canada
Foodbook 1.0 (2014-2015): Telephone survey with ~10,000 respondents (exposures marked with * are from this survey only)

Reference percentages are calculated using survey weights to ensure population representativeness.

How references are computed

References use Foodbook microdata with survey weights.
If multiple PTs are selected, a single combined reference is computed across them.
You can optionally limit the reference by Age Group and Month.

Statistical Methodology

The tool uses a one-sided binomial test to compare observed case exposure rates against population reference values:

Null hypothesis: Case exposure rate ≤ Population reference rate
Alternative hypothesis: Case exposure rate > Population reference rate

Interpretation Guide

Results are classified based on statistical comparison (Binomial Exact Test):

Alert : p-value ≤ 0.05. Observed proportion is significantly higher than reference.
Borderline : p-value ≤ 0.10. Observed proportion is marginally higher than reference.
Not Significant : No significant difference from reference (p > 0.10)
Insufficient Data : Too few cases to calculate statistics (< 5 total responses)
No Reference Value : Exposure not found in Foodbook database

How to use

Starting out

The drop-down menu in the top right of the tool can change the language of the tool between English and French.
- Start by selecting your reference population. It is auto-set to the entire Foodbook population (all P/Ts, all age groups, and all month of response).To filter, in the ‘reference settings’ pane on the left of the screen, click on the box you would like to filter and select the group(s) you would like included in the reference dataset. Multiple groups from each category can be selected. Please be careful not to overanalyse the data. Limiting the data to a small subset of respondents (for example, respondents ages 0-9 from PEI in March) can result in small sample sizes and make the data less reliable. This is especially important for exposures that are rare within the population.
- If you are only interested in a certain category or categories of food exposures, that can be filtered in the ‘Filter category’ dropdown menu

There are two different methods to select which exposures you would like included in the tool;

Manually add in exposure names and data

To add exposures to the table, start typing in the ‘Select exposures’ box. You should see a menu appear below with a list of matching exposures. You can also click on the box, and the exposure list will appear.
Select your exposure of interest. If it is not available in the list, type in your exposure of interest in full. This will add it to the table as a new variable (i.e., you can add exposure data, but there will not be a reference value).Once the exposure shows up below, add in your case exposure data. This is to be entered as the number of cases responding ‘Yes’, ‘Probably’, ‘No’ or ‘DK’ (don’t know) to each of the exposures.
When you enter in the case data, the table on the right should automatically update

2. Upload an exposure data file

Download the template by clicking ‘Upload Exposure counts’ and then ‘Download template’
Using the template provided (or another compatible template), fill in the case exposure data.
Once done, upload the file by clicking ‘Upload Exposure counts’ and then browsing for and selecting the file

Final steps

Once the exposure data is entered, you will see the results in the table on the right on the screen.
This will show which exposures fall into the different result categories (i.e., Alert, Borderline, Not Significant, Insufficient Data, or No Reference Value).
You may sort the results table as you wish, using the up and down arrows next to each column header.
You can export the results to a csv file by clicking ‘Export’
You can also print the results, or copy the results, using the corresponding buttons
By clicking on ‘Visualization’ (next to results) you will see a visual representation comparing the case exposure data to the Foodbook reference data. This can also be exported as a PNG or SVG file.

Additional notes

The ‘reference’ data tab includes reference values for each exposure by province/territory of residence
The ‘Data info’ tab includes information on the reference population. For example, if you filter the data to only include those from your P/T, the ‘Data info’ tab will also update to show a snapshot of that population.

Good Practices

Select reference population filters that match your case demographics (PT, age, season).
Focus on Alert and Borderline classifications for hypothesis generation.
Consider multiple testing correction when examining many exposures.
Custom exposures require you to provide the expected reference percentage.
Please be careful not to overanalyse the data. Limiting the data to a small subset of respondents (for example, respondents ages 0-9 from PEI in March) can result in small sample sizes and make the data less reliable. This is especially important for exposures that are rare within the population.

Limitations

Survey data may not reflect current food consumption patterns (data collected in 2014-2015 and 2023-2024)
Self-reported exposure data is subject to recall bias
Some exposures may have seasonal variations not captured when using annual data
Small sample sizes in specific PT/age/month combinations may yield unstable estimates
Exposures from Foodbook 1.0 (*) use different survey weights than Foodbook 2.0

Frequently Asked Questions

Why is my exposure showing ‘No Reference Value’?
- This means the exposure was not asked in either Foodbook survey, or the variable name doesn’t match. Try searching for a similar exposure name.
What does the * mean next to some exposures?
- Exposures marked with * are only available from Foodbook 1.0 (2014-2015). They are included for completeness but may not reflect current consumption patterns.
Why do reference values change when I select different PTs?
- Food consumption varies by region. The reference is recalculated using only respondents from the selected province(s)/territory(ies).
How should I interpret ‘Borderline’ results?
- Borderline results (p-value between 0.05 and 0.10) suggest a possible association that warrants further investigation but doesn’t meet conventional significance thresholds.