Increasing intake completion rates

The situation

Author Health's growth plan for the year rested on a target number of completed intakes and a targeted percent annual engagement rate for list referrals. Underneath those targets was a familiar behavioral-health funnel — reach the patient (majority cold outreach), get them scheduled, get them to complete the visit — and every stage of it was leaking.

In 2024, roughly a quarter of scheduled intake slots were lost to late cancellations and no-shows before a provider ever saw the patient. Underneath that number were basic data gaps: only 54% of patients scheduled for intake had a verified email on file and just 45% had documented consent to text or email - critical paths in sending and receiving appointment reminders. Appointment reminders were only supported in English, while roughly 12% of patients spoke Spanish. Caregivers — who internal data already showed improved both show rates and reschedule rates after a missed visit — were formally on file for only 2% of prospective patients, even though partner data suggested closer to 40% had one available. None of these were single-fix problems. They were a chain of small, connected gaps between "referred" and "showed up," and closing the chain meant shipping in a deliberate sequence rather than waiting on one large release.

Fixing the funnel one link at a time

I led product for a series of releases in Maya, our care platform, each one closing a specific gap before the next capability could be built on top of it:

• 01

  Contact data as a foundation

  Made Maya — not the EHR — the system of record for patient phone, email, and address, with verification status, source attribution, and channel preferences. Reminders would only ever be as reliable as the data behind them, so this shipped first.

• 02

  Outreach mechanics

  Tuned call-saturation logic to stop over-dialing unreachable patients, added callback date-and-time picking with agent self-assignment, and ingested a new partner data feed carrying preferred language and caregiver contacts we hadn't had before.

• 03

  Reminders that actually send

  Made verified contact info and consent the explicit gate for appointment reminders, and surfaced SMS/email deliverability errors directly in Maya so staff could troubleshoot instead of guessing. At time of scheduling, notified staff if any contact information or verification statuses were missing so they could resolve live with patient.

• 04

  Caregivers as a real channel

  Replaced an ad hoc workaround — staff quietly entering a caregiver's number as the patient's own just to get them reminders — with a proper caregiver entity: verbal-then-written consent capture, automated appointment reminders on the same cadence as the patient's, and manual sends for consent forms and visit links.

• 05

  Support for Spanish reminders

  Launched support for Spanish translated communications.

What changed

Intake completion rate — the share of scheduled intake appointments that were actually completed — moved from 34% in July 2024 to 42% by July 2026, an 8-point increase (about a 24% relative gain) as the above releases landed in sequence.

Contact data quality moved in step: among scheduled-for-intake patients, verified phone coverage rose from 72% to 98%, and verified email coverage (of those with an email on file) rose from 54% to 83%.

The caregiver channel is now driving a measurable completion-rate lift: as of July 2026, intake appointments where a caregiver reminder was sent completed at 57.1%, versus 44.4% for appointments without one — a 12.7-point gap that's held consistently since caregiver reminders launched, with completed-appointment volume carrying a caregiver reminder growing from 6/month in Nov 2025 to 20–21/month by mid-2026.

Figures reflect internal program metrics generalized for public sharing; specific system, vendor, and payer names have been omitted or simplified.