Inpatient and Outpatient Costs from DSS Jean Yoon Paul Barnett March 25, 2009

Health Economics Resource Center 2 Outline Differences between HERC and DSS cost data Differences between HERC and DSS cost data How DSS gets costs How DSS gets costs DSS data files and file structure DSS data files and file structure How to obtain costs of care for cohort of patients How to obtain costs of care for cohort of patients Specific issues in using DSS for research studies Specific issues in using DSS for research studies

Health Economics Resource Center 3 Overview of VA Cost Data Sets Two possible sources: Two possible sources: –DSS cost  Activity-based cost allocation system  Implemented on a local level –HERC average cost  Assigns costs for each VA encounter based on diagnoses and length of stay  Directly comparable to Medicare and other payers

Health Economics Resource Center 4 White board question What are you interested in using DSS cost data for? What are you interested in using DSS cost data for?

Health Economics Resource Center 5 Costing Methods Direct measurement Pseudo-bill Reduced list costing Clinical cost function Average cost per visit More preciseLess precise Inpatient HERC Med/Surg Outpatient HERC AC Costs Inpt. Rehab, HERC MH, LTC DSS

HERC Average Costs Datasets

Health Economics Resource Center 7 HERC Average Cost Datasets HERC method of distributing costs to hospital stays and outpatient visits HERC method of distributing costs to hospital stays and outpatient visits Created to merge easily with clinical files Created to merge easily with clinical files Acute medical surgical stays Acute medical surgical stays –Estimate of what stay would have cost in a Medicare hospital, based on a regression model Other inpatient care Other inpatient care –Length of stay Outpatient care Outpatient care –Hypothetical Medicare payment based on procedure codes assigned to visit

Health Economics Resource Center 8 Directly comparable to non-VA providers (Medicare) Directly comparable to non-VA providers (Medicare) Costs identical for all stays with same characteristics Costs identical for all stays with same characteristics HERC has file with average cost for each person in each fiscal year HERC has file with average cost for each person in each fiscal year HERC Average Cost Datasets Con’t

DSS National Data Extracts

Health Economics Resource Center 10 Where Do DSS Data Come From? VISTA workload, clinical, & financial data National Data Extracts of DSS Time allocation Relative values DSS VISN Level Production Databases

Health Economics Resource Center 11 DSS Determines Costs of Products Cost assigned to cost center Cost assigned to cost center –Staff activities reports and financial data Cost of overhead departments distributed Cost of overhead departments distributed Products of each department tabulated Products of each department tabulated Relative values assigned to products Relative values assigned to products Unit cost of each product determined Unit cost of each product determined

Health Economics Resource Center 12 DSS Assigns Cost to Encounters = Total cost of encounter X Unit cost of each intermediate product ∑Workload Count of each intermediate product

Health Economics Resource Center 13 DSS National Data Extracts Inpatient files Inpatient files –Treating specialty file –Discharge file Outpatient Encounter Files Outpatient Encounter Files Pharmacy files Pharmacy files Intermediate Product Department files Intermediate Product Department files Clinical Files Clinical Files ALBCC files ALBCC files

Health Economics Resource Center 14 DSS Cost File: Inpatient Discharge File Care of patients discharged in fiscal year Care of patients discharged in fiscal year One record per discharge One record per discharge Includes cost incurred in prior fiscal years Includes cost incurred in prior fiscal years May exclude stays that began before DSS implementation May exclude stays that began before DSS implementation

Health Economics Resource Center 15 DSS Data Only in Discharge File Discharge day Discharge day Total days of stay Total days of stay Discharge treating specialty (bed section) Discharge treating specialty (bed section)

Health Economics Resource Center 16 DSS Cost File: Inpatient Treating Specialty File Treating specialty is synonymous with bed section Treating specialty is synonymous with bed section One record per bed section (treating specialty) per month One record per bed section (treating specialty) per month –All care provided during fiscal year –Include stays not yet over

Health Economics Resource Center 17 DSS Data Only in Treating Specialty File Treating specialty Treating specialty Census indicator Census indicator Date of entry and exit from treating specialty Date of entry and exit from treating specialty –No discharge date Treating specialty length of stay Treating specialty length of stay –No total length of stay

Health Economics Resource Center 18 DSS Data in Both Inpatient Files Admit day Admit day Admitting DRG Admitting DRG Principal diagnosis Principal diagnosis Admitting diagnosis Admitting diagnosis

Health Economics Resource Center 19 Comparison of Record Structure ADMITDAY 3/15/04 TRTIN 3/22/04 TRTOUT 4/12/04 FP 6 ADMITDAY 3/15/04 TRTIN 3/22/04 TRTOUT 4/12/04 FP 7 March Rehabilitation April Rehabilitation ADMITDAY 3/15/04 DISDAY 4/12/04 March General Medicine ADMITDAY 3/15/04 TRTIN 3/15/04 TRTOUT 3/22/04 FP 6 Treating Specialty- 3 records Discharge– 1 record in FY 04

Health Economics Resource Center 20 Comparison of Record Structure (Overlapping fiscal year) September General Medicine October General Medicine ADMITDAY 9/22/03DISDAY 10/8/03 ADMITDAY 9/22/03 TRTIN 9/22/03 TRTOUT 9/30/03 FP 12 CENSUS=Y In FY03 File ADMITDAY 9/22/03 TRTIN 9/22/03 TRTOUT 10/8/03 FP 1 CENSUS=N In FY04 File Treating Specialty- 2 records Discharge– 1 record in FY 04 file

Health Economics Resource Center 21 DSS Cost Files: Outpatient Files One record per patient per day per clinic stop One record per patient per day per clinic stop –NPCD events file allows more than 1 record per clinic stop per day –DSS includes care not in NPCD events file, e.g., prosthetics Primary DX and primary CPT Primary DX and primary CPT

Health Economics Resource Center 22 DSS Data Only in Outpatient Files Date of encounter Date of encounter DSS identifier (clinic stop) DSS identifier (clinic stop) –DSS uses “pseudo stop” code for laboratory, pharmacy, etc. Flag variables identifying data source Flag variables identifying data source –NPCD, pharmacy, prosthetics, Vast CBOC, etc

Health Economics Resource Center 23 DSS Cost Variables in All Files Fixed direct Fixed direct Fixed indirect Fixed indirect Grand total Grand total Variable labor category 4 & 5 Variable labor category 4 & 5

Health Economics Resource Center 24 Additional Cost Variables in Inpatient Files Separate costs for lab, nursing, pharmacy, radiology, surgery, all other Separate costs for lab, nursing, pharmacy, radiology, surgery, all other –Variable, fixed direct, fixed indirect, supply (where applicable) IPD files released in 2005 with costs at product department level IPD files released in 2005 with costs at product department level

Health Economics Resource Center 25 Example: Finding Costs for a Cohort Study to compare costs of care for two types of CABG surgery Study to compare costs of care for two types of CABG surgery One is more labor intensive, more minutes in operating room One is more labor intensive, more minutes in operating room –DSS accounts for time allocation, so difference in surgical costs Step 1: Identify patients with surgery 1 and surgery 2 with CPT codes in PTF file Step 1: Identify patients with surgery 1 and surgery 2 with CPT codes in PTF file Step 2: Merge records from PTF with DISCH on SCRSSN, ADMITDAY, STA3N, DISDAY Step 2: Merge records from PTF with DISCH on SCRSSN, ADMITDAY, STA3N, DISDAY

Health Economics Resource Center 26 Finding Costs for a Cohort Con’t Step 3: compare mean surgery total for surgery 1 and surgery 2 for surgical cost differences Step 3: compare mean surgery total for surgery 1 and surgery 2 for surgical cost differences –Expect differences in surgery variable labor cat 4 Step 4: compare mean grand total for surgery 1 and surgery 2 for overall hospitalization differences Step 4: compare mean grand total for surgery 1 and surgery 2 for overall hospitalization differences –LOS differences contribute more to costs Step 5: combine pharmacy, outpatient, inpatient costs over 2 year period to compare long-term costs Step 5: combine pharmacy, outpatient, inpatient costs over 2 year period to compare long-term costs

Health Economics Resource Center 27 DSS Vs Medicare Costs Cost Type DSSMedicare Physician Services Included in hospital costs Excluded from hospital costs Indirect costs Includes VA central office and national operation costs plus hospital admin costs Only hospital admin costs Capital costs Financing costs excluded Financing costs included

Health Economics Resource Center 28 Advantages of Using DSS DSS costs estimate reflect facility differences in productivity, efficiencies, economies of scale, etc DSS costs estimate reflect facility differences in productivity, efficiencies, economies of scale, etc DSS has pharmacy data DSS has pharmacy data VA purchases about 4% of care from non-VA providers VA purchases about 4% of care from non-VA providers –Community nursing home costs in DSS outpatient file New DSS inpatient discharge costs by category of care New DSS inpatient discharge costs by category of care –HERC is creating new files with cost subtotals for each patient DSS is activity-based method and potentially more accurate than other methods DSS is activity-based method and potentially more accurate than other methods

Health Economics Resource Center 29 Treating specialty PTF Bed section files more difficult DischargePTF Main files easy OutpatientNPCD Outpatient Files moderate VA Utilization DataDSS Cost Data Ease of Merging DSS Cost Files with Utilization Files

Health Economics Resource Center 30 Cost Outliers in DSS Users should look for cost estimates that are unexpectedly high given characteristics of care Users should look for cost estimates that are unexpectedly high given characteristics of care Mismatch of cost and utilization can result in unit costs that are very high cost, or negative Mismatch of cost and utilization can result in unit costs that are very high cost, or negative DSS quality assurance efforts DSS quality assurance efforts –Audit that costs in DSS agree with general ledger –Extreme high outliers are filtered out when DSS national data extracts (NDE) are built

Health Economics Resource Center 31 Summary: DSS Vs HERC Average Costs HERC cost estimates HERC cost estimates –based on strong assumptions –reflect relative resource use in non-VA settings DSS cost estimates DSS cost estimates –reflect actual VA experience –have more variance –may be more prone to inappropriate outliers Both data sets rely on DSS distribution of costs to departments Both data sets rely on DSS distribution of costs to departments

Health Economics Resource Center 32 DSS Data Access See HERC guide on DSS See HERC guide on DSS VISN Support Services Center Web Site (KLF Menu) VISN Support Services Center Web Site (KLF Menu) –Summaries of DSS data –Documentation of DSS and new DSS datasets DSS Program Office Web Site DSS Program Office Web Site See HERC intranet web site for links See HERC intranet web site for links

Health Economics Resource Center 33 HERC Guidebooks Research Guide to Decision Support System National Cost Extracts (updated 12/08) Research Guide to Decision Support System National Cost Extracts (updated 12/08) HERC’s Average Cost Datasets for VA Inpatient Care HERC’s Average Cost Datasets for VA Inpatient Care HERC's Annual Person Level Cost Dataset User Guide: Fiscal Years HERC's Annual Person Level Cost Dataset User Guide: Fiscal Years HERC's outpatient average cost dataset for VA care: fiscal year 2007 update HERC's outpatient average cost dataset for VA care: fiscal year 2007 update

Health Economics Resource Center 34 Next Classes April 1, 2009 Pharmacy Costs Mark Smith, Ph.D. April 22, 2009 Estimating Non-VA Costs Mark Smith, Ph.D. Mark Smith, Ph.D.