Using National VA Databases for Health Services Research:

1
Using National VA Databases for Health Services
Research

• Denise M. Hynes, PhD
• Patricia A. Murphy, MS
• Linda Kok, MA
• Philip M. Colin, AAS, CSP6
• VA Information Resource Center (VIReC)

2
Workshop Organization

• Part I
• Understanding the Medical SAS Datasets
• Part II
• Using the Medical SAS Datasets

3
Part I Objectives

• Medical SAS Datasets
• Source Data
• Organization Contents
• Strengths Limitations
• Data Quality Information

4
Medical SAS Datasets

• National VHA patient care data
• Administrative/workload purposes
• SAS datasets at Austin Automation Center (AAC)
• Utilize via time-sharing on VHA mainframe computer

5
Data FlowTransferred to Austin nightly loaded
to national files twice monthly

AAC
SAS Datasets
VHA Facilities
6
Focus on Two Groups of Medical SAS Datasets

• Inpatient Care SAS Datasets
• Discharges
• Outpatient Care SAS Datasets
• Ambulatory encounters ancillary services

7
Not Addressing Other Useful Medical SAS Datasets

• Extended Care
• Non-VA Care
• Observation Stays

8
Inpatient SAS Datasets

• Referred to as PTF (Patient Treatment Files)
• Patients Discharged
• Admission may be previous year

9
Inpatient SAS Datasets
10
Data Elements In AllInpatient Datasets

• Patient identifier (SCRSSN)

11
Special Topic SCRSSN

• Scrambled Social Security Number
• Same algorithm used for all Inpatient and
  Outpatient Care SAS Datasets
• Real SSN available with authorization

12
Data Elements In AllInpatient Datasets contd

• Patient identifier (SCRSSN)
• Facility VISN identifiers of where care
  provided
• Admission Discharge Date Time
• Discharge Type (e.g., Regular, Death-Autopsy,
  Non-bed Care)
• Primary Diagnosis for stay (DXLSF)

13
Special Topic Inpatient Diagnosis

• DXLSF Primary Diagnosis
• Diagnosis responsible for the major part of the
  full stay
• DXPRIME Principal Diagnosis
• Condition determined to be chiefly responsible
  for the admission (not in Inpatient Procedure
  file)

14
Inpatient Main Dataset
15
Inpatient Main DatasetOverview

• Record inpatient stay (episode of acute care)
• Most complete inpatient dataset for patient
  characteristics

16
Inpatient MainSelected Data Elements

• Patient Characteristics
• Date of Birth
• Home County, State, Zip Code, VISN
• Race
• Sex
• Marital Status
• Income

17
Inpatient MainSelected Data Elements contd

• Max. 10 diagnoses for the stay (DXLSF
  DXF2-DXF10)
• No order for 2-10
• Discharge destination (e.g., VA domiciliary,
  community hospital)
• Means Test Indicator

18
Special Topic Means Test Indicator

• Some veterans receive Mandatory Care (e.g.,
  POWs, service- connected disabled)
• For others, Means Test is used to determine
  eligibility for care

19
Special Topic Means Test Indicator contd

• Indicators are, e.g., Non-vet, Category A,
  Category B
• Income limits for eligibility categories set each
  January 1
• See VIReC Inpatient Data Resource Guide, FY00,
  Appendix C, for details

20
Inpatient Bedsection Dataset
21
Overview

• Record Bedsection stay
• Bedsection treating service of physician, not
  physical location
• Maximum of 25 per inpatient stay

22
Special Topic Examples of Bedsections
23
Inpatient Bedsection Dataset Selected Data
Elements

• Date time of transfer into out of Bedsection
• Physical Location Code (PLBED)
• Bedsection Diagnosis (max. 5) Bedsection DRG

24
Special Topic DRGs

• Diagnostic Related Group
• Basis Length of Stay, Diagnoses, Procedures
• Not comparable to private sector
• VHA bedsection transfer may private sector
  discharge
• DRG for full stay vs. DRG for bedsection

25
Inpatient Bedsection Dataset Selected Data
Elements contd

• Service-Connected Treatment
• Is the condition being addressed in the
  bedsection a service-connected one?
• Not the same as a veterans service-connected
  eligibility for mandatory care (e.g., Agent
  Orange exposure)

26
Inpatient Procedure Dataset
27
Overview

• Record a days procedures (Max. 5 per record)
• gt5 procedures on a day ? additional record for
  the day
• 6 procedures on Day 1 and 2 procedures on Day 8
  ? 3 records

28
Overview contd

• Procedure Non-operative surgery-like procedure
  not performed in an operating room or under
  general anesthesia

29
Special TopicProcedures vs. Surgeries

• Surgery Procedure performed in main or
  specialized operating room
• Procedure in Facility A may Surgery in
  Facility B
• Depends on where performed
• Look at both datasets

30
Inpatient Procedure Dataset Selected Data Elements

• Procedure, coded in ICD-9-CM (vs. CPT for
  outpatient procedures)
• Dialysis type number of dialysis treatments
• Physicians specialty (bedsection)

31
Inpatient Surgery Dataset
32
Overview

• Record a days surgeries (Max. 5 per record)
• gt5 surgeries on a day ? additional record for
  the day
• 1 surgery on Day 1 and 6 surgeries on Day 8 ? 3
  records

33
Inpatient Surgery Dataset Selected Data Elements

• Category of chief surgeon first surgical
  assistant
• Transplant status (donor source live or cadaver)

34
Inpatient Care SAS DatasetsOutpatient Care SAS
Datasets
35
Outpatient Care SAS Datasets
36
Outpatient Datasets

• Referred to as OPC or NPCD (National Patient Care
  Database)
• 4 datasets Visit
• Procedure
• Diagnosis
• Event

37
Outpatient Datasets
38
Note Slides 38-47 were not included in the
workshop oral presentation due to time limits
39
Data Elements In BothVisit Event Outpatient
Datasets

• Patient identifier (SCRSSN)
• Patient demographics (Age, date of birth, race,
  marital status)
• Patient Zip Code, County, State of Residence
• Date of encounter

40
Data Elements In BothVisit Event Outpatient
Datasets contd

• Means Test Indicator
• Patient eligibility code (Separate vet and
  non-vet categories)
• Agent Orange exposure claimed
• Radiation exposure claimed

41
Outpatient Visit Dataset

• Record One days encounters (max. 15 Clinic
  Stops) for a patient

42
Special Topic Clinic Stops

• Clinic Stop Code DSS Identifier
• 1st 3 digits, Primary 2nd 3 digits,
  Secondary/Credit
• Secondary Stop Codes are facility- specific
  modifiers, range specified
• See VHA Directive 2000-009
• http//vaww.va.gov/publ/direc/health/ direct/vhadi
  r009.html

43
Outpatient Visit DatasetSelected Data Elements

• Clinic Stop Code(s) (Max. 15 per day)
• Additional patient characteristics
• Sex
• Home VISN Primary Service Area
• Income
• Number of Dependents
• Insurance Coverage

44
Outpatient Event Dataset

• Record Ambulatory encounter, coded as DSS
  Identifier (Stop Code)
• No limit on number of encounters per day

45
Outpatient Event DatasetSelected Data Elements

• Primary Clinic Stop Code
• Secondary/Credit Clinic Stop Code
• Appointment Type (e.g., Employee, Research,
  Regular)
• Station, Sub-station, VISN of stop
• Procedures (max. 15 per stop)
• Diagnosis (max. 10 per stop)

46
Outpatient Event DatasetSelected Data Elements
contd

• Provider Type (max. 10 per stop)
• Location of Visit (e.g., Clinic, Home Based
  Care, Ward)
• Service-Connected Condition Indicator
• Homeless code (Y/N)

47
Inpatient Care SAS DatasetsOutpatient Care SAS
Datasets
48
Strengths Limitations of Inpatient and
Outpatient SAS Datasets
49
Strengths

• Centralized data source
• Large groups of patients
• Given good coding, reflective of general clinical
  status
• Unique identifier (SCRSSN) allows linking records
  across files/years

50
Limitations

• Not all care dimensions
• Retrospective discharge abstracts
• Incentives to coding
• Limitations of ICD-9-CM coding

51
Data Quality Information

• VHA Data Quality Journey
• http//vaww.va.gov/med/dataquality
• Historical information
• Data Quality Summit papers
• Work Groups

52
Data Quality Information contd

• VHA Coding Council
• http//vaww.va.gov/health/him/coding_folder/Codin
  g_Council.htm
• Newsletter Close Encounters
• VHA Coding Handbook
• HSRData Listserv

53
Data Quality Information - contd

• Office of the Inspectors General. Report No.
  5R6-G07-109. Washington, DC September 29, 1995
• Kashner,T.M. et al. Agreement between
  administrative files and written medical records
  A case of the Department of Veterans Affairs.
  Medical Care 1998 36(9), 1324-1336.

54
Part I Understanding the Medical SAS Datasets
55
On to Part II!
56
Part II

• Using the Medical SAS Datasets

57
Part II Objectives

• Getting Ready
• Connecting to Austin
• Allocating a JCL library
• Research Application
• Running a Job at Austin

58
Getting Ready

59
Who Can Use These Data?

• VA employees (incl. WOC) whose work requires
  their use.
• Other non-VA researchers who have obtained
  authorization.

60
Request Access to Austin

• Contact local Information Resources Management
  (IRM) department or Information Security Officer
  (ISO) or system administrator
• Identify needed functional task codes
• Submit ACRS Form 9957 to get a
• Time Sharing Option (TSO) logon ID

61
Prepare Your PC

• 1. Install 3270 terminal emulator software on
  your PC

62

• 2. Direct Network Connection
• Web based, e.g., Host On Demand
• Telnet based, e.g., Hummingbird
• 3. Remote Network Connection
• - Outside a VA facility
• Remote Access Server (RAS) phone .
• RAS permission on your Austin Account
• ltcall mike re how to get permissiongt

63
Connecting to Austin
64
Welcome to VACCESS
65
TSO Logon ID Screen
66
TSO Password Screen
67
TSO Ready Prompt
68
Interactive System Productivity Facility (ISPF)

• To Manage
• Libraries Data Sets
• Program development
• Program submission
• Output

69
SPF Main Menu Page 1
70
Allocating A JCL Library
71

SPF Screen Option 3.2

• 3. Utilities
• 2. Data Set
• A. Allocate new data set
• Project your logon id
• Group JCL
• Type LIB

72
Allocate A Data Set
73
Allocate Partitioned Data Set
74
Research Application
75

• What is our research question?

76
Sample Research Question

• Can we identify and describe veterans diagnosed
  with Amyotrophic Lateral Sclerosis (ALS) during
  the fiscal year 2000 by period of service?

77
Selection Criteria

• 1. Specific diagnosis ALS patients
• 2. Veterans only
• 3. By period of service
• 4. Time frame FY2000

78
Which Variables?

79
Variables for Identifying ALS Patients

• DXLSF DXF2 through DXF10
• Contain ICD-9-CM codes.
• Values for our search
• 3352
• 33520
• 33529

80
Variable for Identifying Veterans

• MEANS
• Values for our search
• AN Category A Not Service Connected
• AS Category A Special Category
• C Category C
• CO Category C

81
Variables for Identifying Period of Service

• PSRCD - Inpatient File
• PS Outpatient File
• This will be our SAS BY variable. Our patient
  count will be shown for each value of period of
  service.

82
Which Files?

83
Files with diagnosis variables

• Inpatient MAIN
• Outpatient EVENT

84
Files with Period of Service

• Inpatient MAIN
• Outpatient VISIT

85
Files with Means variable

• Inpatient MAIN
• Outpatient VISIT or EVENT

86
Time Frame

• Use Fiscal Year 2000 files
• Begins October 1, 1999 and ends September 30,
  2000.

87
How to Proceed?
88
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
89
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
90
1. Inpatient Main File

• Select ALS patients

91
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
92
2. Outpatient Event file

• Select ALS patients.

93
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
94
3. Outpatient Visit file

• Merge with ALS patients
• from the Event file by
• SCRSSN to get period of
• service and means test
• variables.

95
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
96
4. Combine inpatients and outpatients.
97
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
98

• 5. Keep one record per
• ALS patient.

99
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
report
100
6. Produce frequency counts

• FY 2000
• ALS patients
• Veterans
• By period of service

101
Running a Job At Austin
102
Step by Step

• Editing a job
• Submitting a job
• Viewing output
• Extracting output
• Transferring output
• Logging off AAC

103
Editing a job
104
SPF 2 Edit
105
Edit Screen
106
Your SAS Job

• Job Control Language (JCL)
• Identifies you.
• Requests time and type of resource
• Requests access to specific files
• Identifies where output should be sent
• SAS Statements - your program

107
JCL

• //YourIDx JOB XXXUNKA8,YourID,MSGCLASSI,NOTIFYS
  YSUID
• //MAIN LINES(60,W)
• //STEP01 EXEC SAS,WORK'200,200'
• //PM00 DD DSNMDPPRD.MDP.SAS.PM00,DISPSHR
• //SF00 DD DSNMDPPRD.MDP.SAS.SF00,DISPSHR
• //SE00 DD DSNMDPPRD.MDP.SAS.SE00,DISPSHR
• //LIBRARY DD DSNMDPPRD.MDP.FMTLIB6,DISPSHR
• //SYSIN DD

108
Job Card

• //YourIDx JOB,XXXUNKA8,YourID,MSGCLASSI,NOTIFYS
  YSUID

Job ID
Job service level (6-9)
109
Job service levels
110
Job service levels

• Tape Files
• Service level 8 or 9
• Numeric volume identifiers
• Disk Files
• Service level 6 or 7 -- faster turnaround
• Alpha-numeric volume identifiers

111
SPF 3.4Data Set List

• 
  

112
Tape or Disk?
Numeric Tape
Alpha-numeric Disk
?????? File in Use

• 
  

113
Exec Card

• //STEP01 EXEC SAS,WORK'200,200'

114

Data definition (DD) cards

• //PM00 DD DSNMDPPRD.MDP.SAS.PM00,DISPSHR
• //SF00 DD DSNMDPPRD.MDP.SAS.SF00,DISPSHR
• //SG00 DD DSNMDPPRD.MDP.SAS.SG00,DISPSHR

Library references (libref)
SAS dataset libraries
115

DD Card For Format Library

• //LIBRARY DD DSNMDPPRD.MDP.FMTLIB6,DISPSHR

SAS permanent format library
Or use the NOFMTERR on a SAS Option Statement
116
The JCL and SAS code for this research question
are included in your supplemental hand-out.
117
Submitting A Job
118
Submit from Edit Screen
119
Viewing SAS Output
120
JES Spool Display Facility

• Type J on Edit Command gt line
• or
• Type J on SPF main menu

121
JES Spool Display Facility
122
JES Status Screen
123
JES Status Screen

• A list of jobs you submitted
• Completed jobs W- OUTPUT
• Type S on job line under CMD to select

124
Output From One Job
125
Output - SAS List

•  
• ------------------------- Patient Type
  Non-Veteran --------------------------
•  
• The FREQ
  Procedure
•  
• Number of Patients
  with ALS
• by
•   Period of Service (Recoded)
• 
  Cumulative Cumulative
• Period of Service Frequency Percent
  Frequency Percent
• --------------------------------------------
  -------------------------
• WWII 1 1.96
  1 1.96
• KOREA 1 1.96
  2 3.92
• POST-KOREA 2 3.92
  4 7.84
• VIETNAM 9 17.65
  13 25.49
• POST VIETNAM 4 7.84
  17 33.33
• OTHER 33 64.71
  50 98.04
• DES. STORM (VET) 1 1.96
  51 100.00

126
Output - SAS List

• ------------------------------- Patient Type
  Veteran --------------------------
•  
• The FREQ
  Procedure
•  
• Number of Patients
  with ALS
•   by
• Period of Service (Recoded)
• 
  Cumulative Cumulative
• Period of Service Frequency Percent
  Frequency Percent
• --------------------------------------------
  -------------------------
• WWII 393 26.81
  393 26.81
• PRE-KOREA 13 0.89
  406 27.69
• KOREA 290 19.78
  696 47.48
• POST-KOREA 148 10.10
  844 57.57
• VIETNAM 452 30.83
  1296 88.40
• POST VIETNAM 93 6.34
  1389 94.75
• OTHER 8 0.55
  1397 95.29
• DES. STORM (ACT) 1 0.07
  1398 95.36
• DES. STORM (VET) 68 4.64
  1466 100.00

127
Extracting Output

128
JES Status Screen Type E in the Cmd column
next to the job .
129
Extracting Output On the next screen, type a
file name to identify your output.

130
Exit SPF

• On any command line type X

131
TSO Ready Prompt
132
Transferring Output to Your PC
133
TSO Ready Prompt Click on Transfer in the top
toolbar.
134
Download from TSO
135
Download from TSO -CONTINUED

• Click on Receive file from Host.

• In Receive File From... box, type file name
  you assigned to your output .

136
Download from TSO CONTINUED

• In the Scheme box, click on TSO_Text.
• In Local File Name box, type the disk drive,
  path and filename which will receive your output.

137
Transfer Complete
138
Type Logoff
139
Terminate Session
140
(No Transcript)
141
Open the extracted output on your local system
and edit as needed.
142
AAC Usage General Rules

• Use the appropriate job service level (6
  9) to reflect
• time (CPU seconds.)
• type of data storage (tape vs. disk).
• Always check your SAS output Log carefully.
• Adhere to common data security practices.

143
Some useful resources

• Resource material available on the VIReC website
  http//www.virec.research.med.va.gov/
• VIReC Insights publication Guide for First Time
  Users of VA Austin Automation Center (AAC)
• SAS Programming Efficiency Tips presentation.
• FY 1999 and FY 2000 Medical SAS Datasets Resource
  Guides.

144
Other Useful Resources

• Subscribe to HSRData, an e-mail listserv that
  provides a network of top level consultants for
  VA researchers. Members include researchers,
  policy makers and information managers.
• Work with your local COE, IRM, HSRD Resource
  Centers (VIReC, HERC MDRC) and the AAC
  HelpDesk.

145
VA Information Resource Center (VIReC)

146
Note Slides 92-xx were not included in the
workshop oral presentation due to time limits
147
Using National VA Databases for Health Services
Research

• Supplemental Materials

• Sample SPF Screens (Slides 94-97)
• SAS Sample Program with JCL (Slides 98-115)
• VIReC Toolkit for New Users of VA Data
  (http//virec.research.hines.med.va.gov)

148
SPF Menu Page 1
149
SPF Menu Page 2
150
SPF Menu Page 3
151
SPF Menu Page 4
152
SAS Program Statements
153

• Data INPAT (rename(PSRCDPS))
• Set PM00.PM00 (keepSCRSSN DXLSF DXF2-DXF10 PS
  MEANS)
• array DX DXLSF DXF2-DXF10
• do over DX
• if DX 3352' or dx'33520' or dx'33529 then
  FLAGY
• end
• if FLAGY

154
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
155

• Data OUT_Dx
• set SE00.SE00 (keepSCRSSN DXLSF DXF2-DXF10)
• array DX DXLSF DXF2-DXF10
• do over DX
• if DX 3352' or dx'33520' or dx'33529 then
  FLAGY
• end
• if FLAGY

156
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
157

• Data OUTPAT
• merge OUT_Dx (inA)
• SF00.SF00 (keepSCRSSN PS MEANS)
• by SCRSSN
• if A

158
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
159
SAS Program Statements
160

• Data INPAT (rename(PSRCDPS))
• Set PM00.PM00 (keepSCRSSN DXLSF DXF2-DXF10 PS
  MEANS)
• array DX DXLSF DXF2-DXF10
• do over DX
• if DX 3352' or dx'33520' or dx'33529 then
  FLAGY
• end
• if FLAGY

161
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
162

• Data OUT_Dx
• set SE00.SE00 (keepSCRSSN DXLSF DXF2-DXF10)
• array DX DXLSF DXF2-DXF10
• do over DX
• if DX 3352' or dx'33520' or dx'33529 then
  FLAGY
• end
• if FLAGY

163
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
164

• Data OUTPAT
• merge OUT_Dx (inA)
• SF00.SF00 (keepSCRSSN PS MEANS)
• by SCRSSN
• if A

165
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
166

• Data BOTH
• set INPAT OUTPAT
• length PATTYPE 11
• if MEANS in (AN, AS, C, CO) then
• PATTYPEVeteran
• else PATTYPEnon-Veteran
• Proc sort dataBOTH nodupkey
• by SCRSSN

167
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports
168

• Proc freq dataBOTH
• by PATTYPE
• table PSPATTYPE / list

169
Sample program Data flow

• Inpatient data Outpatient data

Event dataset
Visit dataset
Main dataset
analysis data
reports