@vjacob These types of errors may hint at a data problem (or a data flow problem) resulting in an unexpected drop-off of patient diagnosis information. Alternatively, it could just mean that the real ebb and flow in the dataset has a bunch of patients with something one month and very few the next.
The main point is: Look at your raw data and see if it makes sense that this concept would drop off from one month to the next. If it makes sense, then it is not an error, just a helpful guide from Achilles Heel.