Impacts often don’t make sense and the percentages are misleading. We should be able to mark these as omit or incorrect. Percentage should be of max score, not change score (.8 to 1.6 should not be 50%, if max symptom score is 5 it should be less than 20%, at least, changes of.1 should be ignored). Weighting should be clarified, if there is 3 data points or 500, the latter is more likely to be real. And include an info button to clear explanation how three calculations are done.