Clinical Integration Data Needs for Assessing a Project
JMH Background Two acute care hospitals and one behavioral health hospital One acute hospital on Meditech Second acute hospital and behavioral health on McKesson STAR Foundation-model Physician Network uses McKesson Plus for billing and Cerner for EHR Numerous other departmental and specific-purpose systems throughout the Health System with varying degrees of integration (Home Health, Lab, Blood Bank, Surgery, ED, etc.) 2
Overall Project Goal Evaluate need for a dedicated Observation unit Assess financial impact of establishing such a unit Establish baseline and criteria for monitoring results 3
Request: How many Observation patients are there now, where are they cared for, and what are their characteristics (physicians, diagnoses, length of stay, etc.) Data Issues: Observation patients located throughout the house Two main types: Recovering from a procedure or have a medical condition that needs to be watched for some period Patients frequently move between Inpatient and Observation, sometimes retroactively, occasional conflict between physician orders and billing regulations Outcome: HPM Data Warehouse contains the data needed to answer these questions Must have a zero charge code for Recovering patients 4
Request: What is the number of Urgent Care patients that are sent to the ED and/or become direct admits? How many of the OBS patients return or bounce back at 7, 15 and 30 day increments? Data Issues: Urgent Care is part of the Physician Network not the hospitals Physician Network billing data is passed to HPM, Cerner is not Need a unique person identifier that allows tracking a patient from one platform to another and across multiple encounters within each system Outcome: McKesson Passport system provides an Enterprise Identifier (HNE) that is used in many JMH systems but not all HNE is pretty good but not perfect; duplication exists 5
Request: List of Patients admitted through ED to either Inpatient or Observation status and who stay 2 or fewer days. Categorize patients into Diagnosis groupings. Data Issues: DRGs typically used for Diagnosis groupings, but may not be available for Observation and other Outpatients Inpatient stays are measured in full days, Observation stays are measured in hours Outcome: Develop a user-defined grouping of Principal Diagnosis codes Calculate LOS based on difference between Admit Date/Time and Discharge Date/Time 6
Request: How many Urgent Care to ED transfers were after-hours (between 8 pm and 8 am)? How many of those were admitted to Inpatient or Observation status? Data Issues: ED check-in times maintained in ED system, not hospital billing system When a patient is admitted, Admission Date/Time is when they were admitted to the hospital, not when they were registered in the ED; could be a difference of several hours Outcome: Get a separate file of check-in and discharge date/time from the ED system and match up with both Hospital and Urgent Care data 7
Conclusions Standard HPM data warehouse provides most of the data needed for analysis Some data must be extracted from other systems to support specific questions Fluidity between Inpatient and Observation is problematic Difficult to assess cost differences due to different ways of measuring LOS and utilization Project is ongoing 8