Issues in Coding Hospital Intensive Care for Data Analysis

44
Issues in Coding Hospital Intensive Care for Data Analysis Adeline J. Wilcox Department of Veterans Affairs Veterans Health Administration Office of Informatics and Analytics Office of Analytics and Business Intelligence Clinical Analytics and Reporting Performance Evaluation Performance Measurement Federal Committee on Statistical Methodology Research Conference Washington DC 04 November 2013 Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Transcript of Issues in Coding Hospital Intensive Care for Data Analysis

Page 1: Issues in Coding Hospital Intensive Care for Data Analysis

Issues in Coding Hospital Intensive Carefor Data Analysis

Adeline J. Wilcox

Department of Veterans AffairsVeterans Health Administration

Office of Informatics and AnalyticsOffice of Analytics and Business Intelligence

Clinical Analytics and ReportingPerformance Evaluation

Performance Measurement

Federal Committee on Statistical MethodologyResearch Conference

Washington DC04 November 2013

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 2: Issues in Coding Hospital Intensive Care for Data Analysis

Disclaimer of Endorsement

The views and opinions of the author expressedherein do not necessarily reflect those of theOffice of Informatics and Analytics or theDepartment of Veterans Affairs. Referenceherein to any specific commercial products,process, or service by trade name, trademark,manufacturer, or otherwise, does notnecessarily constitute or imply its endorsement,recommendation, or favoring by the Departmentof Veterans Affairs.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 3: Issues in Coding Hospital Intensive Care for Data Analysis

VTE-1H

Clinical Trial = YVTE-1

BVTE-1

X Missing

=2, 3, 4

ComfortMeasures Only = 1

VTE-1B

VTE-1X Missing

N

VTE-1I

Non-UTD Value

ICUAdmission

Date

Initial ICU Day = 0 days OR = 1 day

Initial ICU Day (in days) = ICU Admission Date – Admission Date

� 2 days

ICU Admission or Transfer

VTE-1X

= 1

ICU Discharge

Date

Non-UTD Value

Missing

VTE-1D=UTD

ICU LOS = ICU Discharge Date - ICU Admission Date

ICU LOS

= 0 days

>= 1 day VTE-1

B< 0 days VTE-1

X

Missing = UTD

= 2, 3 Missing

VTE-1X

VTE-1X

VTE-1D

< 0 daysVTE-1

B

Specifications Manual for National Hospital Inpatient Quality Measures Discharges 01-01-12 (1Q12) through 06-30-12 (2Q12) VTE-1-5

Part of The Joint Commission’s Specification for VTE-1:Venous Thromboembolism Prophylaxisfrom the NQF-Endorsed VoluntaryConsensus Standardsfor Hospital Care.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 4: Issues in Coding Hospital Intensive Care for Data Analysis

Wall Street’s Prescription forHospitalsby Devin Leonard

Bloomberg BusinessweekSeptember 2 – September 8,2013

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 5: Issues in Coding Hospital Intensive Care for Data Analysis

An Electronic Intensive Care Unit (eICU) is a form oftelemedicine that uses state of the art technology to provide anadditional layer of critical care service. An eICU may also bereferred to as a tele-ICU.http://searchhealthit.techtarget.com

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 6: Issues in Coding Hospital Intensive Care for Data Analysis

”tele-I.C.U.” described in Annals of Health Care

BIG MED

The New Yorker, August 13 & 20, 2012

Of ICUs, Atul Gawande gave his qualitative assessment:Nowhere in health care do we expend more resources.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 7: Issues in Coding Hospital Intensive Care for Data Analysis

Medicare UB-04 Revenue Codes Revised Aug 2010

Accessed 15 November 2012

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 8: Issues in Coding Hospital Intensive Care for Data Analysis

Hacking Healthcare

Trotter and Uhlman open their second chapter of their book withWhen you’ve seen one medical practice, you’ve seenone medical practice.

A variation goesIf you’ve seen one VA hospital, you’ve seen one VAhospital.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 9: Issues in Coding Hospital Intensive Care for Data Analysis

Hospital Level Classification

Intensive care units (ICUs) vary significantly fromhospital to hospital with respect to structure, servicesprovided, personnel and their level of expertise, andorganizational characteristics.

Guidelines of critical care services and personnel:Recommendations based on a system of categorization ofthree levels of care Crit Care Med 2003 Vol. 31, No. 11

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 10: Issues in Coding Hospital Intensive Care for Data Analysis

VA ICU Levels and VA Facility Grouping by ICU Level

VA Inpatient Evaluation Center (IPEC) Data Definitions

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 11: Issues in Coding Hospital Intensive Care for Data Analysis

The location of patients does not determine theirlevel of care .

Levels of Critical Care for Adult PatientsThe Intensive Care SocietyUK 2009Document describes Levels 0, 1, 2 and 3 by increasing intensity

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 12: Issues in Coding Hospital Intensive Care for Data Analysis

VHA Directive 1009 August 28, 2013

When a patient requires admission to a critical care unit and noIntensive Care Unit (ICU) bed is available, it is an absoluterequirement that the patient receive ICU-level care in analternative location including monitoring, staffing, andtreatment consistent with ICU standards.b. Temporary Bed Location. A temporary bed location is adesignated place where a patient awaiting inpatient care can becared for until a bed in the destination unit is available.Temporary bed locations may include but are not limited to thePost Anesthesia Care Unit for ICU overflow patients; theObservation Unit; and the ED or UCC for newly admittedpatients. It may also include short-term use of a higher level ofcare (for example an ICU bed for atelemetry inpatient admission) whileawaiting the appropriate location .

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 13: Issues in Coding Hospital Intensive Care for Data Analysis

Beneficiaries Use Location-Based Concept ofIntensive Care

Inspector General faults VA’s CareBaltimore patients spent long hours in emergency roomThe Baltimore Sun, Friday, September 27, 2013

”The concern is that they’re not able to get up tothe ICU...” said Jacob Gadd, deputy director forhealth care for the American Legion, whichmonitors veterans’ health care quality.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 14: Issues in Coding Hospital Intensive Care for Data Analysis

VistA, VA’s Legacy EHR

VA’s famous legacy EHR is called the Veterans HealthInformation Systems and Technology Architecture (VistA).”The Veterans Health Information Systems and TechnologyArchitecture (VistA) is an integrated system of both nationallymandated and locally adapted software applications thatdirectly support patient care at VHA facilities.””The electronic health records (EHRs) for VA patients are in theVistA files.”http://www.virec.research.va.gov/VistA/Overview.htm

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 15: Issues in Coding Hospital Intensive Care for Data Analysis

”Twenty years ago, the Veterans Administration (VA) led thehealthcare industry in the creation of VistA, its comprehensiveelectronic medical record system.”Oct 22, 2009

http://www.hsrd.research.va.gov/for_researchers/vinci/program.cfm

”VistA contains approximately 168 packages, 26,000+ routines,12 million lines of code.”

Stephen Hufnagel December 23, 2011

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 16: Issues in Coding Hospital Intensive Care for Data Analysis

VA Corporate Data Warehouse (CDW)

• Dimension and fact tables.

• Primary data source housed is VistA.

• Holds data from the entire nation.

• Transactional level data without aggregation.

”The DW/BI system is a highly political effort.”Mundy, Thornthwaite and Kimball.The Microsoft Data Warehouse Toolkit:With SQL Server 2008 R2 and theMicrosoft Business Intelligence Toolset.2nd ed. 2011

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 17: Issues in Coding Hospital Intensive Care for Data Analysis

Some Challenges Facing VistA Data Users

”... the fragmentation of data. VistA files are maintained in 128centers and data may not be compatible among sites becauseof local modifications.”

”Of the 57,000 data fields in VistA, 34% contain data stored asunstructured, free text.”

http://www.hsrd.research.va.gov/for_researchers/vinci/program.cfm

”Field descriptions that are available within the VistA systemare extracted to the CDW Metadata... Unfortunately, VistAfield descriptions tend to be cryptic and/or technical.”

VIReC Resource Guide June 2012

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 18: Issues in Coding Hospital Intensive Care for Data Analysis

CDW Tables Where I Started Looking for ICUs

Dim.RoomBed

RoomBedSIDRoomBedDescription

Inpat.Inpatient

AdmitRoomBedSIDAdmitWardLocationSID

Inpat.PatientTransfer

RoomBedSIDWardLocationSID

Inpat.Lodger

RoomBedSIDWardLocationSID

Dim.WardLocation

WardLocationSIDBedSection

WardLocationName

1

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 19: Issues in Coding Hospital Intensive Care for Data Analysis

WardLocationName Joined to RoomBedDescriptionthrough Fact Table

WardLocationName RoomBedDescription

1N HOSPICE ICUICUM OVERFLOWZZ2BSICU PRIVATE ROOMSICUWCICU PRIVATE3 ICU (OBS) NO DESCRIPTIONIMICU STEPDOWN4MICU PSEUDO BEDOBS-MICU PRIVATE1N MED ICU2D-SICU NONE

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 20: Issues in Coding Hospital Intensive Care for Data Analysis

Columns That Joined to Inpatient Records From FY13

Division 648Table Column Distinct Values

Dim.WardLocation WardLocationName 12Dim.WardLocation BedSection 0Dim.RoomBed RoomBed 28Dim.RoomBed RoomBedDescription 1Dim.Location WardLocation 12Dim.Location LocationName 118Dim.Location LocationAbbreviation 5Dim.Location PhysicalLocation 2Dim.Location TreatingSpecialtyName 2Dim.Specialty Specialty 2

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 21: Issues in Coding Hospital Intensive Care for Data Analysis

select distinct TreatingSpecialtyNamefrom [CDWWork].[Dim].[TreatingSpecialty]where TreatingSpecialtyName like’%[iI][cC][uU]%’or TreatingSpecialtyName like’%[iI][nN][tT][eE][nN][sS][iI][vV][eE][cC][aA][rR][eE]%’

1956 unique strings found when all ten columns unduplicated.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 22: Issues in Coding Hospital Intensive Care for Data Analysis

Portland, OR VAMC

Greatest Number of Different WardLocationNameValues Containing String ICU.

DivSta6a=648 WardLocationName

3DCARD-ICU 3DEN-ICU3DGS1-ICU 3DGU-ICU3DINTS-ICU 3DKID-ICU3DNS-ICU 3DOR-ICU3DPLS-ICU 3DTRA-ICU3DTS-ICU 3DVAS-ICU

In this division WardLocationNamewas redundant to WardLocation.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 23: Issues in Coding Hospital Intensive Care for Data Analysis

Portland, OR VAMC

DivSta6a=648 RoomBed

3DICU-1 3DICU-18 3DICU-263DICU-10 3DICU-19 3DICU-273DICU-11 3DICU-2 3DICU-283DICU-12 3DICU-20 3DICU-33DICU-13 3DICU-21 3DICU-43DICU-14 3DICU-22 3DICU-53DICU-15 3DICU-23 3DICU-63DICU-16 3DICU-24 3DICU-73DICU-17 3DICU-25 3DICU-83DICU-9

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 24: Issues in Coding Hospital Intensive Care for Data Analysis

Portland, OR VAMC

DivSta6a=648 TreatingSpecialtyName

MEDICAL ICUSURGICAL ICU

DivSta6a=648 Specialty

MEDICAL ICUSURGICAL ICU

DivSta6a=648 PhysicalLocation

MICUSICU

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 25: Issues in Coding Hospital Intensive Care for Data Analysis

Portland, OR VAMC

DivSta6a=648 RoomBedDescription

ICU

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 26: Issues in Coding Hospital Intensive Care for Data Analysis

Do these values from the column namedTreatingSpecialty describe Intensive Care?

ICU UROLOGYNEUROLOGY ICUOPHTHALMOLOGY SICUPLASTIC SURGERY ICUPOST ICUPSYCHIATRY(PICU)CARDIOLOGY TELEMETRY ICUHEAD AND NECK ICUMEDICAL ICU/OBSERVATIONORTHO (INCL HAND) SICUL-MED MICU HOSPICEMICU/PALLIATIVE CAREJC-ICU/CCUColumn named TreatingSpecialty couldnot be joined to patient records.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 27: Issues in Coding Hospital Intensive Care for Data Analysis

Problem Values Selected By Query for Intensive CareUnits Run On Ten Columns

BU-8A/MICU,CCU,CVU NUR 2B2 PAICUNUR NICU/6B BO NUR 4 EAST MICU/CCUELP-NSG-TESTICULAR SELF EXAM NUR PICUNUR RICU NUR PSY-WEST (PICU)NOL ORTHOTIC LAB HUNNICUTT NUR WCICUPHL INSOMNIA CVT CONNECTICUT SMICU virtual bedHRB-SMOKE CESS PRACTICUUM NUR MICU STEPL/NURSING PEDICURE CCU/MICU/TELE/ISOO/PICU RECREATION THERAPY GP pickup ICUZZMU-AURICULAR/ACCUPUNCTURE TLOVICUTO-REC TX-HORTICULTUREBU-MEDICAL/SURGICAL ICUICUBOARDER

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 28: Issues in Coding Hospital Intensive Care for Data Analysis

Not Found in the VA Acronym Lookup

TICU 11Oct2013NICU 15Oct2013STICU 17Oct2013DICU 24Oct2013HICU 24Oct2013RICU 24Oct2013CICU 24Oct2013PAICU 25Oct2013QICU 30Oct2013VICU 30Oct2013ORICU 30Oct2013PHICU 30Oct2013

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 29: Issues in Coding Hospital Intensive Care for Data Analysis

PACU: An Acronym With Duplicate Meanings

PACU defined as Post-anesthesia care unitin wikipedia Entry for Intensive Care Unit

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 30: Issues in Coding Hospital Intensive Care for Data Analysis

PICU: An Acronym With Many Meanings

PICU illustrates the lack of healthcare data standards.

On VA’s Acronym Lookup site,PICU is given as Patient Intensive Care Unit.

A VHA document dated 1994 spells out PICU asPsychiatric Intensive Care Unit.

PICU means Pediatric Intensive Care Unitaccording to SNOMED.

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 31: Issues in Coding Hospital Intensive Care for Data Analysis

VHA Medical Centers(Local VistA system)

NPCD data at AITCmonitored by NDS

Inpatient Encounters (IE) Dataset: The Inpatient Encounters Datasetprovides information about professional services received by patients during an inpatient stay (see 2.4.3).

Visit (SF) Dataset: The Visit Dataset reports services provided to a patient in a 24-hour period (see 2.4.1).

Event (SE) Dataset: The Event Dataset provides information about individual outpatient encounters (see 2.4.2).

Medical SAS® Outpatient Datasets/ Inpatient Encounters Dataset

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 32: Issues in Coding Hospital Intensive Care for Data Analysis

Stay Level Data on ICU Utilization from the Weekly FluReport

)LJXUH����3HUFHQWDJH�RI�9$�LQIOXHQ]D�FRGHG�KRVSLWDOL]DWLRQV�ZKLFK�LQFOXGHG�WLPH�LQ�DQ�LQWHQVLYH�FDUH�XQLW��,&8���E\�)OX�6HDVRQ���,&8V�UHSUHVHQWHG�LQFOXGH��0HGLFDO�,&8�DQG�6WHS�'RZQ��6XUJLFDO�,&8�DQG�6WHS�'RZQ��&DUGLDF�,&8�DQG�6WHS�'RZQ��Note: VA influenza-coded hospitalizations are not necessarily laboratory-confirmed. CDC data for comparison is available at:

Medical ICU and Step DownSurgical ICU and Step DownCardiac ICU and Step Down

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 33: Issues in Coding Hospital Intensive Care for Data Analysis

Step Down

High dependency unit (HDU): Many hospitals have atransitional high dependency unit (HDU) for patientswho require close observation, treatment and nursingcare that cannot be provided on a general ward, butwhose care is not at a critical enough level to warrantan ICU bed. These units are also called step-down,progressive and intensive recovery units and areutilised until a patient’s conditions stabilises enough toqualify them for discharge to a general ward.

Wikipedia entry for Intensive care unit

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 34: Issues in Coding Hospital Intensive Care for Data Analysis

Rooms Not Nested Within Wards

Dim.RoomBed

RoomBedSIDRoomBedDescription

Inpat.Inpatient

AdmitRoomBedSIDAdmitWardLocationSID

Inpat.PatientTransfer

RoomBedSIDWardLocationSID

Inpat.Lodger

RoomBedSIDWardLocationSID

Dim.WardLocation

WardLocationSIDBedSection

WardLocationName

1

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 35: Issues in Coding Hospital Intensive Care for Data Analysis

We Need an Information ModelOur database should be ”a faithful model of reality”

C. J. Date tells usA data model is a model of the persistent data ofsome particular enterprise.

He says a data model is a database design and...we might speak of the data model for some bank, orsome hospital , or some government department.

Database In Depth. Relational Theory for PractitionersO’Reilly Media, Inc.Sebastopol, CA, 2005

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 36: Issues in Coding Hospital Intensive Care for Data Analysis

How Can We Model Hospital Beds and HospitalWards?

The National Health Service (NHS) Data Model andDictionary for England explicitly nests Hospital Beds withinWards, describing a Ward as

”A group of Hospital Beds with associated treatmentfacilities managed as a single unit for the purposes ofstaffing and treatment responsibilities.””A critical care unit will comprise one WARD if the HospitalBeds and associated treatment facilities are managed as asingle unit.”

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 37: Issues in Coding Hospital Intensive Care for Data Analysis

NHS Data Definitions and Data Standards DefiningIntensive Care

A new CRITICAL CARE PERIOD starts when the PATIENTis admitted to a critical care location regardless ofCRITICAL CARE LEVEL.

A CRITICAL CARE PERIOD ends when the PATIENT isdischarged from the critical care location, or dies, or thecare that is being delivered in a non-standard location isCRITICAL CARE LEVEL National Code 00 Level 0 orNational Code 01 Level 1.

A CRITICAL CARE PERIOD does not include caredelivered in a cardiacor coronary care unit.

http://www.datadictionary.nhs.uk/Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 38: Issues in Coding Hospital Intensive Care for Data Analysis

2013-10-29-E, Stephen Hufnagel PhD

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 39: Issues in Coding Hospital Intensive Care for Data Analysis

Proprietary 3M HDD PromisesSemantic Interoperability

The 3M Healthcare Data Dictionary (HDD) is a controlledmedical vocabulary server that has been continuouslyexpanded and maintained for over 15 years. The 3M HDDmakes it possible to map and manage medical terminologies,integrate content and standardize healthcare data.

The 3M HDD provides a road map to the content and structureof patient data by defining and translating every data element ina computerized patient record across all domains includingICD-9, ICD-10, General Equivalence Mappings (GEMs),SNOMED CT®, LOINC®, RxNORMand more.Web Accessed 29 October 2013

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 40: Issues in Coding Hospital Intensive Care for Data Analysis

Unified Medical Language System ®UMLS Terminology Services

Result of search on ICU in SNOMED CT Browser

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 41: Issues in Coding Hospital Intensive Care for Data Analysis

Unified Medical Language System ®UMLS Terminology Services

Partial result of search on Intensive CareUnit in SNOMED CT Browser

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 42: Issues in Coding Hospital Intensive Care for Data Analysis

Unified Medical Language System ®UMLS Terminology Services

UTS Metathesaurus Browser

Concept Semantic Types

Level of Care – Critical Care Idea or Conceptcare by level of intensity Qualitative Concepttransfer to intensive care unit Health Care ActivityIntensive care Health Care Activityintensive care unit Health Care Related Organization

Manufactured Object

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 43: Issues in Coding Hospital Intensive Care for Data Analysis

ConfusionsWhat data from Electronic Health Records (EHRs) do wewant to analyze?Are these data collected for us?Are VistA modifications feasible?How do we know when an acronym has multiplemeanings?Is Intensive Care a Location or a Specialty or a Level ofCare?How will mapping free text fields from VistA into a datadictionary work?What would it take to persuade data architects anddata users that thehospitalized patient is aState-Dependent Object?

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox

Page 44: Issues in Coding Hospital Intensive Care for Data Analysis

Your comments and questions are valued and encouraged.

ÆAdeline J. WilcoxDepartment of Veterans AffairsVeterans Health Administration Central Office810 Vermont Avenue, NWWashington, DC

202-461-5673

adeline dot wilcox at va do gov

Novel Approaches to Coding, Editing, and Imputing Data Coding Hospital Intensive Care by Adeline J. Wilcox