“How many new patients come back?” sounds like a simple question. Before putting a percentage on a dashboard, decide who counts as new, what counts as a return and how long you will wait for it.
Define the starting point
One workable definition is a patient’s first attended appointment recorded in the source data. That is different from their first booking: someone who cancels before attending has not yet had a first visit.
Check how much history you have. If an export begins in January, a long-standing patient’s first January appointment may look like their first-ever visit. Flag incomplete history instead of quietly classifying those patients as new.
Give each patient the same opportunity to return
Choose a return window that makes sense for the question and the practice. For illustration, you might examine a second attended visit within 30 days. That is a reporting choice, not a recommended clinical schedule.
If you run the report on 31 March, a patient who first attended on 28 March has not had 30 days to return. Either exclude that patient from the completed-window calculation or show their group separately as still in progress.
Separate bookings from attendance
A second appointment booked and a second appointment attended answer different questions. Keep both if they help, but label them clearly. Decide how cancellations, reschedules and duplicate records are handled before comparing months.
Use the number to ask a better question
A change in return rate is a prompt to investigate. Look at the size of the group, appointment availability and the types of visits involved. Returning more often is not automatically a better outcome, and this measure does not establish the quality or appropriateness of care.
Write the definition next to the report: first-visit rule, return event, observation window and data cutoff. That gives the next person a way to reproduce the answer.
See how Wedge builds business reporting around definitions you can explain.