Synthetic demonstration · No patient, employer, or client data
Eight events. Five appointments.
A small worked analysis of two reports that appear to disagree. Every record below is invented to make the reasoning inspectable.
Download the workbook and templates →Try the problem · Invented records
What is each report counting?
Inspect the same records in two ways, then choose an explanation. You can read the complete analysis whenever you like.
8 status events
| Event ID | Appointment ID |
|---|---|
| E01 | A01 |
| E02 | A02 · Repeated ID |
| E03 | A02 · Repeated ID |
| E04 | A03 |
| E05 | A04 · Repeated ID |
| E06 | A04 · Repeated ID |
| E07 | A04 · Repeated ID |
| E08 | A05 |
| Appointment ID | Event IDs |
|---|---|
| A01 | E01 |
| A02 · Repeated ID | E02, E03 |
| A03 | E04 |
| A04 · Repeated ID | E05, E06, E07 |
| A05 | E08 |
Source B: scheduling export
A01, A02, A03, A04, A05
Each appointment ID appears once. Both exports use the same observation period and “scheduled” filter.
A scheduled status cannot establish attendance or completed care. This sample has no audit history explaining why extra events were recorded.
The decision
Should operations report eight appointments scheduled during the observation period?
No. In this sample, eight is the count of “scheduled” status events. Those events refer to five distinct appointments. The scheduling export also contains those same five appointment IDs. Agree on the unit of analysis before comparing totals.
Source A · Status-event export
The evidence
Observation period: September 1, 2026, 00:00 inclusive to September 2, 2026, 00:00 exclusive, UTC. Both exports use the same period and filter to “scheduled.”
Every event ID is unique. A02 appears twice; A04 appears three times. These are repeated statuses for an appointment, not duplicate event IDs.
| Event ID | Appointment ID |
|---|---|
| E01 | A01 |
| E02 | A02 |
| E03 | A02 |
| E04 | A03 |
| E05 | A04 |
| E06 | A04 |
| E07 | A04 |
| E08 | A05 |
Source B · Scheduling export
Compare the same unit.
The appointment export contains A01, A02, A03, A04, and A05, each once. All five appear in the event export, and no additional appointment ID appears there.
8 event rows − 3 additional occurrences = 5 distinct appointment IDs.
Reconciled difference: 5 − 5 = 0 appointments.
This supports agreement between the supplied samples after alignment. It does not establish that either export contains every real appointment.
What remains unresolved
- Why were the extra status events recorded? The sample has no audit history to answer that.
- Were any appointments missing from both exports? An independent source check is needed.
- Did the appointments result in attendance or completed care? A scheduling status cannot answer that.
Decision memo
Recommendation: Use five distinct appointments when describing these supplied records. Keep event counts as a separate workflow measure.
Confidence: High for the arithmetic and ID match in this synthetic sample; no inference about live-system completeness or patient outcomes.
Next action: Have the reporting owner document the unit and filters, and ask the system owner to explain repeat events before applying the same reconciliation to an operational report.