Keep the measurement process stable
If a program changes its definition of completion halfway through the chart, an apparent improvement may be a measurement change. The same issue arises when data coverage or the denominator changes.
Record those events beside the intervention dates.
A useful annotation explains context
| Date marker | Possible relevance |
|---|---|
| New enrollment process | Intended process change |
| Holiday closure | Temporary capacity change |
| Reporting-system update | Measurement discontinuity |
| Staff vacancy | Delivery constraint |
These notes help generate explanations that can be investigated.
Patterns need enough information
Formal run-chart rules depend on an appropriate baseline and sufficient observations. Do not infer a stable process shift from a few points or repeatedly move the reference line to favor the latest result.
Rare events and changing denominators may require a different analytic approach.
Separate learning from attribution
A chart can support local learning and reveal when something deserves attention. Stronger causal claims may require a comparison group, a suitable time-series design, or other methods that address alternative explanations.
Use the chart to ask a better question: what changed, when, and under what conditions? It is a record of a process over time, not an automatic certificate that the most recent initiative worked.
