Liz Soracco
Case StudyCare Coordination

Reducing friction in the therapy journey.

A mental health systems case study about what happened when fragmented scheduling, communication, and provider workflows quietly disrupted access to care — and what changed when we started designing them as one journey instead of three.

What looked like a scheduling problem was a coordination problem — seams, workflows, and continuity quietly deciding who reached care.

  • Role
    Systems design · care coordination
  • Domain
    Digital mental health
  • Scope
    Scheduling, communication, continuity
  • Surface
    Provider workflow + patient journey
Care journey with momentum driftA horizontal journey with six points — Onboarding, Assessment, Recommendations, Between appointments, Re-engage, Drift — and a thin overlay curve labelled "momentum" that descends gradually from left to right, indicating that most users did not quit but drifted away one small uncertainty at a time.MOMENTUMIntakeAssessmentProvider matchSchedulingFirst apptOngoing careBEFORE FIRST APPOINTMENTcoordination gap
Most people didn't disengage — they lost momentum, one small uncertainty at a time, in the gaps between systems.

Access to care breaks down quietly — through fragmented systems and lost momentum.

  • Disengagement happened gradually.

    People rarely abandoned care in a single moment. Momentum was lost in small intervals — a delayed reply, an unclear next step, a scheduling gap that quietly stretched.

  • Reducing patient friction shifted load.

    Every step removed from the patient side surfaced somewhere in the provider workflow. Operational coordination, not patient effort, became the constraint.

  • Timing mattered more than volume.

    A message in the right moment carried more weight than several in the wrong one. Communication cadence read as noise; communication timing read as care.

Static workflows compared to a responsive support systemTwo panes. The left pane shows three cohort blocks paired one-to-one with three sequence blocks — rigid pairing with no feedback. The right pane shows three signal blocks — behavior, context, progression — flowing into a responsive logic node, then into a comms and recommendations node, with a curved sage-colored feedback line returning to the signals.
Before · static workflows
cohort 1sequence 1cohort 2sequence 2cohort 3sequence 3

rigid pairing · no feedback

After · responsive support systems
behaviorcontextprogressionresponsive logiccomms + recsFEEDBACKin the momentself-initiated

behavior-aware comms · self-directed support moments

Four moves followed from this. None are surprising in isolation; the work was in coordinating them across systems that had been designed separately.

Coordination seams across the care journeySix rectangular nodes — Patient, Scheduling, Provider availability, Communication timing, Appointment confirmation, Ongoing care — connected by short vertical seam ticks marking handoffs. Two seams are dashed and carry a small failure-point caret, indicating fragile coordination edges. A segmented continuity rail runs beneath the sequence, broken at each seam.PatientSchedulingProvider availabilityCommunication timingAppointment confirmationOngoing care
Coordination seams — where continuity is held, and where it breaks.

  1. Tighten the path from intake to first appointment.

    Reduced the number of decision points and waiting states between assessment and a confirmed first session — the window where most early drop-off was concentrated.

  2. Redesign communication around timing, not cadence.

    Replaced the standing reminder schedule with messages triggered by behavior, progression state, and gaps in coordination. Volume dropped; relevance rose.

  3. Coordinate scheduling and communication as one system.

    Treated availability, outreach, and follow-up as a single operational surface rather than three adjacent ones. Conflicts resolved before they reached the patient.

  4. Reduce friction inside the provider workflow.

    Surfaced patient activity, missed steps, and pending coordination inside the tools providers already used — so absorbing the operational load didn't mean absorbing new context-switching.

The brief was friction reduction. The work, in practice, was deciding where the friction was allowed to live.

Patient flexibilityvsProvider operational complexity

Each option offered to the patient — reschedule, switch provider, message between sessions — created a coordination cost on the other side. Each one was worth it or it didn't ship.

Behavioral relevancevsCommunication fatigue

Behavior-aware messaging is more relevant, but also more frequent if left unbounded. We capped cadence at the system level, not the message level.

Every reduction in patient friction landed somewhere else in the system.

Patient, provider, operations dependency triangleThree nodes — patient at the apex, provider at lower left, operations at lower right — connected by six directed edges in three pairs. Each edge carries a short label naming the tension on that direction.appointment requestschedule offeredcoordination loadscheduling logiccommunication surfacecoordination frictionPatientProviderOperations
  • Improved progression through scheduling
    fewer stalls between intake and first appointment
  • Reduced pre-first-appointment drop-off
    the largest source of early disengagement
  • More reliable communication patterns
    fewer messages, better-timed
  • Improved operational coordination
    scheduling + comms as one surface
  • More sustainable care-journey systems
    load distributed, not displaced

The work fundamentally changed how I think about access to care. It isn't usually denied at the door — it erodes across the small operational gaps between intake and engagement.

Small moments of uncertainty and friction compound. A delayed confirmation, an unclear next step, a missed message: individually trivial, in aggregate the shape of who stays in care and who quietly falls out.

The deeper problem wasn't scheduling. It was continuity and coordination across systems that had each been designed reasonably on their own but never as one journey.

I keep returning to this pattern in adjacent systems work — wherever momentum, timing, and operational reliability decide whether a person experiences a service as present or absent.