Focus on Trends
-
When interpreting wastewater data, we advise you to focus on trends, not individual data points in isolation.
-
We adjust the axes to the maximum data point on your graph.
-
So, focus less on whether the clinical data and wastewater data lines intersect, for example, and more on how their trends are (or are not) tracking with each other.
- Additionally, what qualifies as a "high" or "low" concentration can depend on your locality, catchment population, and previous data.
- Contact support@biobot.io if you have specific questions about your particular data.
-
Analyzing Trends
- Doubling or halving between only two samples is not abnormal
- 50k ➡️ 100k does not necessarily mean disease activity has doubled.
- When the change is 2x or less, wait for the next sample to look for trends
- 50k ➡️ 100k does not necessarily mean disease activity has doubled.
- But, more major % changes are more significant…
- 100k ➡️ 400k (4x increase) is more meaningful than 1 million ➡️ 2 million (2x increase)
- …especially at higher concentrations (and especially for larger sewersheds)
- 500k ➡️ 2 million (4x increase) is more meaningful than 10k ➡️ 40k (4x increase), as it suggests that more people in total are infected.
- When concentrations are low, just a small cluster of disease activity may lead to a significant percentage-change
- 500k ➡️ 2 million (4x increase) is more meaningful than 10k ➡️ 40k (4x increase), as it suggests that more people in total are infected.
Why can wastewater data fluctuate?
- One data point can reliably indicate disease presence, but avoid over-interpreting individual data points. Focus more on trends.
- Any individual sample is subject to multiple sources of variation:
- Actual variation in disease activity
- Lab process variability (minimized by our lab’s rigorous QC protocols)
- Fluctuating environmental, industrial, or chemical influences on your wastewater matrix
- Catchment size: small catchments tend to yield more variable results
- So: more frequent sampling enables a quicker identification of trends
- Trend: 4 data points, preferably; a 2-3x change two times in a row is also meaningful