Referral-Out Rate: Count People or Referral Records?

What does referral-out rate count?

Open Dental separates referrals into incoming, outgoing and custom types. Its procedure tracking report filters on the referral date. A separate option includes referred procedures that have a recorded completion date. Those controls matter when comparing extracts. They do not establish a clinical outcome. Open Dental referral types and procedure tracking explain the fields.

Start the review with a written question: what share of a fixed group of people had at least one outbound referral recorded during this month? A count of referral records cannot answer that question until each row is linked to a person. Also confirm that each person belongs to the chosen group. Referrals for people outside that group belong in a separate reconciliation.

Worked example: twelve records, eight people

This is an invented reporting exercise, with no recommended rate or clinical judgment. The eligible group contains 100 distinct people who had a completed visit during the month. All referral rows below were dated in that month and belong to that group. Include rows whether or not they later received a completion entry.

The person-based result is 8 divided by 100, or 8%. The file also contains 12 referral rows per 100 eligible people. Calling that second number “12% of patients referred” would count some people twice. Both totals are useful if their labels remain clear.

Keep the denominator fixed while testing two independent changes. First, add one row for person A. Rows rise to 13, but the count of people remains eight and the rate stays 8%. Next, reset to the original case and add one row for person I, who already belongs to the 100-person group. There are now 13 rows, nine referred people and a 9% rate. The same increase in rows gives a different change in the person-based measure.

Preserve the evidence behind the rate

Save the dates, eligibility rule and counting unit beside the result. In a review copy, use controlled person keys approved for the review. Replacing names alone does not establish de-identification. Keep the privacy review, key handling and access process separate from the worksheet. A reviewer should be able to reproduce eight distinct keys without receiving a list of patient names.

Ask the report owner to explain missing or repeated rows before comparing periods. A rising count could reflect a different extract rule or a different supplied population. The arithmetic alone does not tell you whether referral decisions were appropriate, whether care was completed or whether future demand will continue.

For a sale, keep initial outreach seller-blind: omit owner and practice identity, exact location, financials and staff details. Obtain an NDA before an authorized buyer receives identifying reports. Involve staff when their actual reporting or transition duties require it, with access limited to the task. Patient-data access requires a separate lawful basis.

Summary: retain both totals

The example produces eight people and twelve rows. An extra row can leave the person-based rate unchanged. Save the row count, distinct-person count and eligible group together so a reviewer can explain the result.

For related decisions, review active patient counts, incoming endodontic referrals, production reports and procedure mix. The buyer guide and DSO overview place those reports in a transaction. Discuss a confidential review.