VIReC Cyber Seminar Series 2006 VA Databases and Methods

1
VIReC Cyber Seminar Series 2006 VA
Databases and Methods

• Using VA Pharmacy Data
• Presented by
• Todd A. Lee, PharmD, PhD
• Senior Investigator, MCHSPR COE
• VIReC, Senior Scientific Expert Pharmacy Data
• Research Assistant Professor, Northwestern
  University Feinberg School of Medicine

2
Session Objectives

• Measurement Issues with Pharmacy Data
• Key Data Source Review PBM Data DSS Pharmacy
  Data
• Measurement Issues and Use of Pharmacy Data
• Where To Go For More Help

3
Pharmacy Data Measurement Issues

• Comparability of data sources
• Do PBM and DSS contain the same data?
• Medication utilization
• Recent year? Longer historical view? Does policy
  change impact medication use?
• Exposure to specific medications or medication
  classes
• Are specific drugs associated with better/worse
  outcomes?
• Medication adherence
• How much of a prescribed medication are patients
  using?
• Combining outpatient and pharmacy data to
  identify events
• Can we identify acute exacerbations of COPD with
  outpatient and prescription data?
• Assessing comorbidity or case-mix with medication
  data
• Diagnoses-based measures vs. pharmacy-based
  measures

4
Session Objectives

• Measurement Issues with Pharmacy Data
• Key Data Source Review PBM Data DSS Pharmacy
  Data
• Measurement Issues and Use of Pharmacy Data
• Where To Go For More Help

5
Pharmacy Data Sources

• Local Databases
• VistA
• VISN Warehouses
• National Data Sources
• PBM
• DSS NDE Pharmacy SAS? Datasets
• FCDM

6
VistA Pharmacy Data

• Veterans Health Information Systems and
  Technologies Architecture
• All Prescription Orders and Fills
• Inpatient and Outpatient
• CMOP (Consolidated Mail Outpatient Pharmacy)
• in VistA system for site where fill was
  requested
• Local Files
• At each VistA installation

7
VistA Pharmacy Data

• Prescription Orders Dispensed
• Prescription File (FILE 52) - Outpatient
• Pharmacy Patient File (FILE 55) - Inpatient
• IV Orders (FILE 55.01)
• Unit Dose Orders (FILE 55.06)
• Local Drug File (FILE 50)
• Years covered
• 1997 forward
• Varies by site

8
VistA Pharmacy Data

• Accessing
• FileMan hierarchical database management system
• MUMPS
• SQL
• VISN Warehouses
• Some contain prescription data
• Relational databases

9
PBM Database

• Pharmacy Benefits Management Database
• FY1999 forward (October 1, 1998)
• Maintained by PBM/SHG at Hines VA Hospital
• Researchers must request extract

10
National Extracts - PBM
11
PBM Database

• Information in the Database
• Outpatient Prescriptions Dispensed
• Inpatient Prescriptions Dispensed (IV Unit
  Dose)
• Selected Labs
• Controlled Substance Use
• Automatic Replenishment/Ward Stock
• Procurement and Accounting
• Provider Information
• Patient Information

12
PBM Database Variables

• Outpatient Prescription
• Dispensing Details
• Fill Date
• Drug Name
• Station Name
• Quantity
• NDC National Drug Code
• Dosing Instructions
• VA Drug Class
• Dispense Unit and Price per Dispense Unit

13
PBM Database Variables

• Outpatient Prescription
• Provider Information
• Provider ID
• Provider Service
• Cardiology, Dental, Nursing, Surgery, etc.
• Provider Specialty Subspecialty
• Provider Type
• Staff, Fee, or Non-VA (TPB)
• Patient Information
• Patient Prescription Status

14
DSS NDE Pharmacy SAS? Datasets

• Decision Support System National Data Extract
  Pharmacy SAS? Datasets.
• FY2002 forward
• Located on the host at the Austin Automation
  Center
• Directly accessible by Researchers

15
National Extracts - DSS
16
DSS NDE Pharmacy SAS?Datasets

• Information in the Datasets
• Outpatient Prescriptions Dispensed
• Inpatient Prescriptions Dispensed
• IV
• Unit Dose

17
DSS NDE Pharmacy SAS? Datasets

• Files
• RMTPRD.MED.DSS.SAS.FYYY.VISNX.PHA
• YY year
• VISN V1TO5, V6TO10, V11TO16, V17TO22
• X I for inpatient, O for outpatient
• Based on patient status for encounter not type of
  prescription
• Safest to always use both files
• Inpatient
• RMTPRD.MED.DSS.SAS.FY03.V1TO5I.PHA
• Outpatient
• RMTPRD.MED.DSS.SAS.FY03.V1TO5O.PHA

18
PBM vs. DSS
19
Session Objectives

• Measurement Issues with Pharmacy Data
• Key Data Source Review PBM Data DSS Pharmacy
  Data
• Measurement Issues and Use of Pharmacy Data
• Where To Go For More Help

20
Measurement Issues Pharmacy Data Comparison

• CSP 456 Hernia Study
• Population
• 1,591 Patients in the CSP 456 Study
• Prescriptions
• Outpatient
• FY2002
• Fills and refills
• 42,469 prescriptions

Report Available at http//www.virec.research.va.
gov/References/TechnicalReports/VIReCTechnicalRepo
rt1.pdf
21
Measurement Issues Pharmacy Data Comparison

• Preliminary Results

22
Measurement Issues Pharmacy Data Comparison

• Limitations
• Outpatient only
• Cohort not representative of whole population
• Conclusions
• DSS and PBM Pharmacy extracts capture same
  prescriptions
• DSS or PBM?
• Future Comparisons
• Inpatient data?
• Representative Cohort
• Anecdotal evidence of other examples where match
  is not as good

23
Measurement Issues Medication Utilization

• Did change in prescription drug copayment impact
  medication utilization? (HSRD ECI 02-220, PI
  Kevin T. Stroupe, PhD)
• Examined 30-day equivalents of use of chronic
  medications in 3 groups of patients before and
  after copayment change
• Identified utilization in several categories
  essential vs. non-essential OTC vs.
  prescription high cost vs. low cost brand vs.
  generic
• Number of medications obtained from the VA
  decreased among those subject to copayments and
  biggest effects were in low cost and OTC
  medications

24
Measurement Issues Exposure to specific
medications

• Determine if the use of ICS is associated with an
  increased risk of non-vertebral fractures in
  patients with COPD in the VA
• Conducted a nested case-control study in a cohort
  of VA patients with COPD
• Found increased risk of fractures in COPD
  patients using high dose ICS
• Needed to quantify amount of use of inhaled
  medications
• Pharmacy data not always easy to work with
  particularly true with regard to inhaled products
• More straightforward to calculate cumulative
  exposure when dealing with tablets/capsules than
  with inhalers

25
Measurement Issues Exposure to specific
medications
ID
1
2
3
26
Measurement Issues Exposure to specific
medications

• VA_PRODUCT
• Used to determine specific product
• Used to determine dose strength
• Used to determine number of actuations
• SIG
• Used to determine dosing frequency
• Used to determine number of doses per day

27
Measurement Issues Exposure to specific
medications

• Calculation of cumulative ICS exposure
• Determine strength for each prescription
• Fluticasone 220?g
• Convert strength to beclomethasone equivalents
• BDP_Equiv gt 2200.5 110?g per dose
• Determine number of doses per prescription
• quantity dispensed doses per product
• 1 canister 120 actuations/canister 120 doses
• Calculate beclomethasone equivalents for each
  prescription and sum for cumulative exposure

28
Measurement Issues Medication Adherence

• Examine factors associated with non-adherence in
  patients with COPD
• Measured adherence to respiratory medications
  using Medication Possession Ratio (MPR)
• MPRi
• Found use of CMOP and hospitalizations in prior
  period associated with higher adherence
• Cautions day supply variable accuracy (oral meds
  vs. inhaled meds) accounting for medications and
  days supply at beginning and end of period of
  interest

29
Measurement Issues Combining Outpatient and
Pharmacy Data

• Identify acute exacerbations of COPD in the
  outpatient setting
• Use a combination of outpatient ICD-9 codes and
  Rx data
• Found many outpatient ICD-9 codes non-specific
  for identifying COPD exacerbation
• Most Rx for oral steroids or antibiotics
  dispensed within 5 days of ICD-9 code
• Used algorithm to disqualify ICD-9 codes and
  medication prescriptions
• SIGS with cellulitis, pharyngitis, sinusitis, etc.

30
Measurement Issues Identifying Comorbidities
with Pharmacy Data

• Development of a VA-based version of RxRisk
  (Chronic Disease Score)
• Sloan KL, et al. Construction and characteristics
  of RxRisk-V a VA-adapted pharmacy-based case-mix
  instrument. Med Care 2003 41(6) 761-74
• Potential value in using pharmacy-based measures
  versus ICD-based measures
• RxRisk-V performed similarly to HCC and ADG
  case-mix adjusters when predicting costs
  prospectively
• Sales AE, et al. Predicting costs of care using a
  pharmacy-based measure risk adjustment in a
  veteran population. Med Care 2003 41(6) 753-60

31
Session Objectives

• Measurement Issues with Pharmacy Data
• Key Data Source Review PBM Data DSS Pharmacy
  Data
• Measurement Issues and Use of Pharmacy Data
• Where To Go For More Help

32
VIReC Help

• VIReC Webpage
• http//www.virec.research.va.gov
• Information on VA data sources and how to access
  data
• Documentation on some VA datasets, i.e., Medical
  SAS datasets
• http//www.virec.research.va.gov/DataSourcesName/M
  edical-SAS-Datasets/SASdocumentation.htm
• Includes lists of variables and their dataset
  locations
• Descriptions of each of the variables
• Values for selected variables

33
VIReC Help (contd)

• HSRData Listserv
• Join at VIReC Web site
• Discussion among gt 200 data stewards, managers,
  and users
• Past messages in archive (on intranet)
• VIReC Help Desk
• VIReC staff will answer your question and/or
  direct you to available resources on topics
• VIReC_at_va.gov
• [phone]

34
References

• Arnold N, Hynes DM, Stroupe KT. VIReC Technical
  Report 1 Comparison of VA Outpatient
  Prescriptions in the DSS Datasets and the PBM
  Database. Edward Hines, Jr. VA Hospital, Hines,
  IL VA Information Resource Center, January 15,
  2006.
• Lee TA, Weiss KB. Risk of non-vertebral fractures
  associated with inhaled corticosteroid use in
  obstructive lung disease. Am J Respir Crit Care
  Med. 2004 169(7) 855-859.
• Charbonneau A, Rosen AK, Ash AS, Owen RR, Kader
  B, Spiro A, III, et al. Measuring the quality of
  depression care in a large integrated health
  system. Med Care 2003 41(5)669-680.
• Sloan KL, et al. Construction and characteristics
  of RxRisk-V a VA-adapted pharmacy-based case-mix
  instrument. Med Care 2003 41(6) 761-74
• Sales AE, et al. Predicting costs of care using a
  pharmacy-based measure risk adjustment in a
  veteran population. Med Care 2003 41(6) 753-60

35
VIReC CyberSeminar Series 2006VA Databases
and MethodsSponsored by VA Information Resource
Center (VIReC)Every first Tuesday of the month
from 1 2 pm ET

• Next Cyber Seminar
• November 7, 2006
• VA-Medicare Data
• Presented by Kathy Mallin, PhD and Kristin
  Koelling, MPH (VIReC)
• This session focuses on assessing non-VA health
  care use using VA-Medicare data. The following is
  the session agenda
• Why use VA-Medicare Data?
• Learn about the VA-Medicare Data Merge
  Initiative and available data
• Understand how to request VA-Medicare data
• Learn where to go for help
• Schedule available at http//www.hsrd.research.va
  .gov/for_researchers/cyber_seminars/