At Risk Appointment Detection

Problem

Today approximately 5% of scheduled dialysis appointments are missed.  Missed dialysis appointments lead to serious health complications, including hospital admissions.

Solution

Detect when patients are in danger of missing scheduled dialysis appointments and communicate relevant information to provider teams, reducing the number of missed appointments, follow-on complications and costs.

Use Case

We will detect at risk appointments by creating an algorithm that uses:

  • dialysis clinic topology (lat/lng)

  • the upcoming dialysis appointment schedule (datetime/clinic)

in conjunction with the Patient’s current location 60 minutes and 30 minutes prior to each dialysis appointment. 

The algorithm will determine if a Patient is at risk of missing their scheduled dialysis appointment at a given clinic and, if so, notify the relevant provider team.

Design Notes

  • Patients within Position Health are de-identified.  We track patients anonymously, using a generated unique identifier.  All patient identifying information will reside in customer systems.  When notifications are sent from Position Health to our customers, they will contain our generated patient identifiers.  Customers will resolve the identity of the patient prior to executing downstream business workflows.

  • Times in the diagram below are drive times.  This can be configured per-patient, so we will compute walking; mass transit; drive; etc travel times based on how each patient typically travels to their appointments.

Setup

Dialysis Clinics

Import dialysis clinic locations (lat/lng) into Position Health.

Dialysis Appointment Schedules

For each patient, import dialysis appointment schedule into Position Health.

Preferred Transportation Modes

For each patient, import preferred appointment transportation mode (how they typically travel to their dialysis appointments: walk, mass transit, drive) into Position Health.

Workflow

Patient has an upcoming dialysis appointment at dialysis clinic A.

Examples