Archived snapshot
Directory accuracy report — 2026-09-14
These numbers are historical. This is the report exactly as it stood on 2026-09-14, kept so that anyone who cited it can check what it said. It is not updated and it does not describe the directory today — read the current report.
—Inquiries that got any response within 48 hours (not published yet — 0 of 3 observations needed)
—Inquiries the therapist replied to within 48 hours (not published yet — 0 of 3 observations needed)
—Inquiries declined within 48 hours with a pointer to other therapists (not published yet — 0 of 3 observations needed)
—Ghost rate (inquiries with no response of any kind after 7 days) (not published yet — 0 of 3 observations needed)
—Inquiries clients reported as unanswered (one-tap reports) (not published yet — 0 of 3 observations needed)
—Visible profiles confirmed taking clients in the last 14 days (not published yet — 0 of 3 observations needed)
—Active licenses re-verified in the last 90 days (not published yet — 0 of 3 observations needed)
—Clients confirming the listed fee was accurate (not published yet — 0 of 3 observations needed)
Computed from operational tables (timestamps only; message content is never read). Response outcomes are segmented: replied_48h_pct = the therapist wrote back; referred_out_48h_pct = an automated decline that pointed the client to other therapists; answered_48h_pct = both, over the same denominator (each rounded independently, so the segments can differ from the total by 0.1). ghost_rate_pct = inquiries sent 7+ days ago with no response of any kind within 7 days; a decline is a response and is not counted as a ghost. reported_unanswered_pct = inquiries a client actively reported as unanswered. The accuracy pulse (fee/insurance/availability) is asked only of clients whose therapist replied — a client who received an automated decline is never asked, because the questions presuppose an engagement that did not happen. That gate is applied when the pulse is sent and recorded on the pulse itself; only pulses carrying it are counted here. Each pulse question has its own denominator and its own evidence floor: a client can answer one question and ignore the others, so fee_answered (not responses) is both the divisor and the evidence behind fee_accurate_pct. No rate is published on fewer than 3 observations; below that the field is null, and so are the counts that would let the rate be recomputed from this payload. Benchmarks: a 2023 Senate Finance secret-shopper study found mental-health directory listings yielded an appointment 18% of the time.
Window: trailing 90 days. Generated 2026-09-14. No rate was published on fewer than 3 observations; a dash means the evidence floor was not met.
Kairn — Directory accuracy report, 2026-09-14. https://kairn.health/reports/accuracy/2026-09-14. Free to cite with attribution. This URL is stable and the figures behind it never change.