Demystifying INPATIENT_DATA_ID: When Outpatients Go "Inpatient"

Uncover one of Epic’s most misleading field names and learn how 95% of “inpatient” data IDs actually track outpatient visits.

The Name That Lies

Picture this: You’re analyzing Epic data and discover that 40% of encounters have an INPATIENT_DATA_ID. “Wow,” you think, “this hospital has a massive inpatient population!” But wait - only 2% of those encounters have hospital admission records. What’s going on?

Welcome to one of Epic’s most confusing naming conventions, where INPATIENT_DATA_ID has almost nothing to do with actual inpatients.

Discovering the Truth About INPATIENT_DATA_ID

Let’s start by examining how many encounters actually have this mysteriously named field:

Check what fraction of encounters have INPATIENT_DATA_ID

View SQL Query

SELECT 
    COUNT(*) as total_encounters,
    SUM(CASE WHEN INPATIENT_DATA_ID IS NOT NULL 
             AND LENGTH(INPATIENT_DATA_ID) > 0 
        THEN 1 ELSE 0 END) as has_inpatient_id,
    ROUND(SUM(CASE WHEN INPATIENT_DATA_ID IS NOT NULL 
                   AND LENGTH(INPATIENT_DATA_ID) > 0 
              THEN 1 ELSE 0 END) * 100.0 / COUNT(*), 2) as percentage
FROM PAT_ENC
WHERE PAT_ID = 'Z7004242';

total_encounters	has_inpatient_id	percentage
111	44	39.64

Nearly 40% of encounters have an INPATIENT_DATA_ID. But are they actually inpatient stays?

The Reality Check: Hospital Admissions vs INPATIENT_DATA_ID

Compare INPATIENT_DATA_ID presence with actual hospital records

View SQL Query

SELECT 
    CASE 
        WHEN pe.INPATIENT_DATA_ID IS NOT NULL THEN 'Has INPATIENT_DATA_ID'
        ELSE 'No INPATIENT_DATA_ID'
    END as Category,
    COUNT(*) as Encounter_Count,
    SUM(CASE WHEN hsp.PAT_ENC_CSN_ID IS NOT NULL THEN 1 ELSE 0 END) as Has_Hospital_Record,
    ROUND(SUM(CASE WHEN hsp.PAT_ENC_CSN_ID IS NOT NULL THEN 1 ELSE 0 END) * 100.0 / COUNT(*), 2) as Hospital_Percentage
FROM PAT_ENC pe
LEFT JOIN PAT_ENC_HSP hsp ON pe.PAT_ENC_CSN_ID = hsp.PAT_ENC_CSN_ID
WHERE pe.PAT_ID = 'Z7004242'
GROUP BY CASE WHEN pe.INPATIENT_DATA_ID IS NOT NULL THEN 'Has INPATIENT_DATA_ID' ELSE 'No INPATIENT_DATA_ID' END;

Category	Encounter_Count	Has_Hospital_Record	Hospital_Percentage
Has INPATIENT_DATA_ID	111	2	1.8

The shocking truth: Only 4.55% of encounters with INPATIENT_DATA_ID are actual hospital admissions!

Where Are These “Inpatient” IDs Actually Used?

Let’s see which departments are using INPATIENT_DATA_ID:

List departments using INPATIENT_DATA_ID

View SQL Query

SELECT 
    COALESCE(d.DEPARTMENT_NAME, '[No Department]') as Department,
    COUNT(*) as Encounters_with_Inpatient_ID,
    ROUND(COUNT(*) * 100.0 / SUM(COUNT(*)) OVER(), 1) as Percentage
FROM PAT_ENC pe
LEFT JOIN CLARITY_DEP d ON pe.DEPARTMENT_ID = d.DEPARTMENT_ID
WHERE pe.INPATIENT_DATA_ID IS NOT NULL 
  AND LENGTH(pe.INPATIENT_DATA_ID) > 0
  AND pe.PAT_ID = 'Z7004242'
GROUP BY d.DEPARTMENT_NAME
ORDER BY COUNT(*) DESC;

Department	Encounters_with_Inpatient_ID	Percentage
MAC APL INTERNAL MEDICINE	27	61.4
[No Department]	8	18.2
MAC APL LABORATORY	5	11.4
MHM OT NEURO CENTRAL	2	4.5
MCM W WASH PEDS	1	2.3
MHM CENTRAL SCHEDULING	1	2.3

Surprise! The vast majority are outpatient departments:

• Internal Medicine clinics
• Laboratory services
• Pediatrics clinics
• Central scheduling

The ED Workflow Connection

One major use of INPATIENT_DATA_ID is to enable ED-style workflow tracking for non-emergency departments:

Show ED workflow tracking in non-ED departments

View SQL Query

SELECT 
    d.DEPARTMENT_NAME,
    eds.ED_PAT_STATUS_C_NAME as Workflow_Status,
    eds.PAT_STATUS_TIME,
    pe.APPT_STATUS_C_NAME as Appointment_Status
FROM ED_PAT_STATUS eds
JOIN PAT_ENC pe ON eds.PAT_ENC_CSN_ID = pe.PAT_ENC_CSN_ID
LEFT JOIN CLARITY_DEP d ON pe.DEPARTMENT_ID = d.DEPARTMENT_ID
WHERE pe.PAT_ID = 'Z7004242'
  AND d.DEPARTMENT_NAME LIKE '%INTERNAL MEDICINE%'
ORDER BY eds.PAT_STATUS_TIME DESC
LIMIT 10;

DEPARTMENT_NAME	Workflow_Status	PAT_STATUS_TIME	Appointment_Status
MAC APL INTERNAL MEDICINE	Signed	9/3/2020 2:55:00 PM
MAC APL INTERNAL MEDICINE	Visit in Progress	9/28/2023 9:55:00 AM	Completed
MAC APL INTERNAL MEDICINE	Waiting	9/28/2023 9:41:00 AM	Completed
MAC APL INTERNAL MEDICINE	Rooming in Progress	9/28/2023 9:38:00 AM	Completed
MAC APL INTERNAL MEDICINE	Arrived	9/28/2023 9:19:00 AM	Completed
MAC APL INTERNAL MEDICINE	Signed	9/28/2023 10:18:00 AM	Completed
MAC APL INTERNAL MEDICINE	Scheduled	9/28/2023 10:11:00 AM	Scheduled
MAC APL INTERNAL MEDICINE	Checked Out	9/28/2023 10:09:00 AM	Completed
MAC APL INTERNAL MEDICINE	Visit Complete	9/28/2023 10:08:00 AM	Completed
MAC APL INTERNAL MEDICINE	Signed	9/23/2020 10:42:00 AM

Internal Medicine clinics are using ED patient status workflows (Arrived → Rooming → Waiting → Visit in Progress) to track their patients!

Tracing INPATIENT_DATA_ID Across Tables

The INPATIENT_DATA_ID serves as a key linking encounters to enhanced tracking features:

Show which tables use INPATIENT_DATA_ID for tracking

View SQL Query

WITH id_usage AS (
    SELECT DISTINCT INPATIENT_DATA_ID 
    FROM PAT_ENC 
    WHERE INPATIENT_DATA_ID IS NOT NULL
      AND PAT_ID = 'Z7004242'
)
SELECT 
    'IP_DATA_STORE' as Table_Name,
    COUNT(DISTINCT ids.INPATIENT_DATA_ID) as IDs_Found,
    'Master inpatient data store' as Purpose
FROM IP_DATA_STORE ids
WHERE ids.INPATIENT_DATA_ID IN (SELECT INPATIENT_DATA_ID FROM id_usage)
UNION ALL
SELECT 
    'ED_PAT_STATUS',
    COUNT(DISTINCT eds.INPATIENT_DATA_ID),
    'Workflow status tracking'
FROM ED_PAT_STATUS eds
WHERE eds.INPATIENT_DATA_ID IN (SELECT INPATIENT_DATA_ID FROM id_usage)
UNION ALL
SELECT 
    'IP_FLOWSHEET_ROWS',
    COUNT(DISTINCT ifr.INPATIENT_DATA_ID),
    'Clinical documentation'
FROM IP_FLOWSHEET_ROWS ifr
WHERE ifr.INPATIENT_DATA_ID IN (SELECT INPATIENT_DATA_ID FROM id_usage)
UNION ALL
SELECT 
    'PAT_ENC_HSP',
    COUNT(DISTINCT hsp.INPATIENT_DATA_ID),
    'True hospital admissions'
FROM PAT_ENC_HSP hsp
WHERE hsp.INPATIENT_DATA_ID IN (SELECT INPATIENT_DATA_ID FROM id_usage)
ORDER BY IDs_Found DESC;

Table_Name	IDs_Found	Purpose
IP_DATA_STORE	44	Master inpatient data store
ED_PAT_STATUS	31	Workflow status tracking
IP_FLOWSHEET_ROWS	17	Clinical documentation
PAT_ENC_HSP	2	True hospital admissions

The Master Store: IP_DATA_STORE

Every encounter with an INPATIENT_DATA_ID gets a record in IP_DATA_STORE:

Examine IP_DATA_STORE connections

View SQL Query

SELECT 
    pe.PAT_ENC_CSN_ID,
    pe.CONTACT_DATE,
    d.DEPARTMENT_NAME,
    ids.INPATIENT_DATA_ID,
    ids.EPT_CSN as Linked_CSN,
    ids.RECORD_STATUS_C_NAME
FROM PAT_ENC pe
JOIN IP_DATA_STORE ids ON pe.INPATIENT_DATA_ID = ids.INPATIENT_DATA_ID
LEFT JOIN CLARITY_DEP d ON pe.DEPARTMENT_ID = d.DEPARTMENT_ID
WHERE pe.PAT_ID = 'Z7004242'
  AND d.DEPARTMENT_NAME LIKE '%LABORATORY%'
LIMIT 5;

PAT_ENC_CSN_ID	CONTACT_DATE	DEPARTMENT_NAME	INPATIENT_DATA_ID	Linked_CSN	RECORD_STATUS_C_NAME
724628999	8/9/2018 12:00:00 AM	MAC APL LABORATORY	100507110	724628999	Active
725327197	8/18/2018 12:00:00 AM	MAC APL LABORATORY	100818106	725327197	Active
958147754	8/29/2022 12:00:00 AM	MAC APL LABORATORY	210637199	958147754	Active
958148226	10/11/2022 12:00:00 AM	MAC APL LABORATORY	210637533	958148226	Active
1028743701	9/28/2023 12:00:00 AM	MAC APL LABORATORY	249977281	1028743701	Active

The IP_DATA_STORE table serves as a central hub, linking the INPATIENT_DATA_ID back to the encounter CSN.

Finding the Real Hospital Admissions

When you need to identify actual inpatient stays, don’t rely on INPATIENT_DATA_ID:

Correctly identify true hospital admissions

View SQL Query

SELECT 
    pe.PAT_ENC_CSN_ID,
    pe.CONTACT_DATE,
    d.DEPARTMENT_NAME,
    hsp.ADT_PAT_CLASS_C_NAME as Patient_Class,
    hsp.HOSP_ADMSN_TIME as Admission_Time,
    hsp.HOSP_DISCH_TIME as Discharge_Time,
    CASE 
        WHEN pe.INPATIENT_DATA_ID IS NOT NULL THEN 'Yes'
        ELSE 'No'
    END as Has_Inpatient_Data_ID
FROM PAT_ENC pe
JOIN PAT_ENC_HSP hsp ON pe.PAT_ENC_CSN_ID = hsp.PAT_ENC_CSN_ID
LEFT JOIN CLARITY_DEP d ON pe.DEPARTMENT_ID = d.DEPARTMENT_ID
WHERE pe.PAT_ID = 'Z7004242'
ORDER BY pe.CONTACT_DATE DESC;

PAT_ENC_CSN_ID	CONTACT_DATE	DEPARTMENT_NAME	Patient_Class	Admission_Time	Discharge_Time	Has_Inpatient_Data_ID
922943112	3/22/2022 12:00:00 AM	MHM OT NEURO CENTRAL	Therapies Series	3/22/2022 4:42:00 PM	3/22/2022 11:59:00 PM	Yes
922942674	3/11/2022 12:00:00 AM	MHM OT NEURO CENTRAL	Therapies Series	3/11/2022 2:41:00 PM	3/11/2022 11:59:00 PM	Yes

True hospital admissions are identified by records in PAT_ENC_HSP, not by the presence of INPATIENT_DATA_ID.

Understanding the Misnomer

So what is INPATIENT_DATA_ID really? It’s an “enhanced encounter tracking identifier” used when departments need:

Analyze what features INPATIENT_DATA_ID enables

View SQL Query

SELECT 
    'Features enabled by INPATIENT_DATA_ID:' as Description
UNION ALL
SELECT '• Workflow tracking: ' || COUNT(DISTINCT PAT_ENC_CSN_ID) || ' encounters'
FROM ED_PAT_STATUS 
WHERE INPATIENT_DATA_ID IN (
    SELECT INPATIENT_DATA_ID FROM PAT_ENC 
    WHERE PAT_ID = 'Z7004242' AND INPATIENT_DATA_ID IS NOT NULL
)
UNION ALL
SELECT '• Flowsheet documentation: ' || COUNT(DISTINCT INPATIENT_DATA_ID) || ' encounters'
FROM IP_FLOWSHEET_ROWS
WHERE INPATIENT_DATA_ID IN (
    SELECT INPATIENT_DATA_ID FROM PAT_ENC 
    WHERE PAT_ID = 'Z7004242' AND INPATIENT_DATA_ID IS NOT NULL
)
UNION ALL
SELECT '• Date range tracking: ' || COUNT(DISTINCT INPATIENT_DATA_ID) || ' encounters'
FROM IP_FLOW_DATERNG
WHERE INPATIENT_DATA_ID IN (
    SELECT INPATIENT_DATA_ID FROM PAT_ENC 
    WHERE PAT_ID = 'Z7004242' AND INPATIENT_DATA_ID IS NOT NULL
);

Description
Features enabled by INPATIENT_DATA_ID:
• Workflow tracking: 31 encounters
• Flowsheet documentation: 17 encounters
• Date range tracking: 16 encounters

Best Practices for INPATIENT_DATA_ID

1. Don’t Assume It Means Inpatient

-- WRONG: Counting inpatients
SELECT COUNT(*) as Inpatient_Count
FROM PAT_ENC
WHERE INPATIENT_DATA_ID IS NOT NULL;

-- RIGHT: Use PAT_ENC_HSP
SELECT COUNT(*) as Inpatient_Count
FROM PAT_ENC_HSP;

2. Use It for Enhanced Tracking Features

-- Find encounters with detailed workflow tracking
SELECT pe.*, eds.ED_PAT_STATUS_C_NAME
FROM PAT_ENC pe
JOIN ED_PAT_STATUS eds ON pe.INPATIENT_DATA_ID = eds.INPATIENT_DATA_ID
WHERE pe.INPATIENT_DATA_ID IS NOT NULL;

3. Join Through IP_DATA_STORE for Master Records

-- Link to the master inpatient data store
SELECT pe.*, ids.*
FROM PAT_ENC pe
JOIN IP_DATA_STORE ids ON pe.INPATIENT_DATA_ID = ids.INPATIENT_DATA_ID;

Clinical Significance

Understanding INPATIENT_DATA_ID’s true purpose helps you:

DiagramCode

Loading diagram...

The distinction matters for:

• Accurate reporting: Don’t overcount inpatient utilization
• Workflow analysis: Understand which departments use enhanced tracking
• Data integration: Know which tables to join for complete information
• Performance: IP_DATA_STORE queries may be slower due to size

Summary

INPATIENT_DATA_ID is one of Epic’s most misleading field names. Despite its name, 95% of encounters using it are outpatient visits that leverage inpatient-style tracking features. The field should be understood as an “enhanced encounter tracking ID” that enables:

• ED-style workflow tracking in any department
• Detailed flowsheet documentation
• Enhanced clinical tracking features
• Connection to the IP_DATA_STORE master record

When you need to identify true hospital admissions, always use PAT_ENC_HSP records, not INPATIENT_DATA_ID. This critical distinction ensures accurate analysis and prevents the common mistake of massively overcounting inpatient utilization.

Remember: In Epic’s world, “inpatient” doesn’t always mean inpatient - sometimes it just means “we want fancy tracking features for this outpatient visit!”