Administrative Data and HCUP Tools
Healthcare Cost and Utilization Project
Using Administrative Data to Answer Policy Questions
December 4-5, 2008
The Department of Health & Human Services and the AHRQ logos are located above the title of the presentation. Throughout the rest of the presentation, the AHRQ logo is located in the upper left corner and the HCUP logo in the upper right corner.
Slide 2
Introductions
Claudia Steiner, MD, MPH
Research Medical Officer
Agency for Healthcare Research and Quality
Slide 3
AHRQ - Agency within DHHS
The US Department of Health and Human Services is comprised of ATSDR, NIH, SAMHSA, HIS, HRSA, AHRQ, AOA, FDA, CMS, CDC, and ACF. HCUP can be found within AHRQ.
Slide 4
The HCUP Partnership: A Voluntary Federal-State-Private Sector Collaboration
This slide contains a map of the United States, indicating the level of HCUP participation of each state. A total of 40 states participate in HCUP and provide 90 percent of all discharges.
Non participant States: AK, NM, ID, MT, ND, LA, MS, AL, PA, DE
Partner States providing inpatient data only: WA, OR, NV, WY, TX, AR, IL, WV, VA, RI
Partner States providing inpatient and ambulatory surgery data: CO, OK, MI, KY, NC
Partner States providing inpatient and emergency department data: HI, AZ, MA
Partners States providing inpatient, ambulatory surgery, and emergency department data: CA, UT, SD, NE, KS, MN, IA, MO, WI, IN, OH, TN, SC, GA, FL, MD, NJ, NY, CT, VT, NH, ME
Slide 5
Current HCUP Partners
• Arizona Department of Health Services
• Arkansas Department of Health
• California Office of Statewide Health Planning & Development
• Colorado Hospital Association
• Connecticut Integrated Health Information (Chime, Inc.)
• Florida Agency for Health Care Administration
• Georgia Hospital Association
• Hawaii Health Information Corporation
• Illinois Department of Public Health
• Indiana Hospital &Health Association
• Iowa Hospital Association
• Kansas Hospital Association
Slide 6
Current HCUP Partners
• Kentucky Cabinet for Health and Family Services
• Maine Health Data Organization
• Maryland Health Services Cost Review Commission
• Massachusetts Division of Health Care Finance and Policy
• Michigan Health & Hospital Association
• Minnesota Hospital Association
• Missouri Hospital Industry Data Institute
• Nebraska Hospital Association
• Nevada Division of Health Care Financing and Policy, Department of Health and Human Services
• New Hampshire Department of Health & Human Services
• New Jersey Department of Health and Senior Services
Slide 7
Current HCUP Partners
• New York State Department of Health
• North Carolina Department of Health and Human Services
• Ohio Hospital Association
• Oklahoma Health Care Information Center for Health Statistics
• Oregon Association of Hospitals and Health Systems
• Rhode Island Department of Health
• South Carolina State Budget & Control Board
• South Dakota Association of Health Care Organizations
• Tennessee Hospital Association
Slide 8
Current HCUP Partners
• Texas Department of State Health Services
• Utah Department of Health
• Vermont Association of Hospitals and Health Systems
• Virginia Health Information
• Washington State Department of Health
• West Virginia Health Care Authority
• Wisconsin Department of Health and Family Services
• Wyoming Hospital Association
Slide 9
HCUP Is a Family of Databases, Tools, and Products
HCUP is comprised of HCUP databases, Software tools, Research publications, User support, and EQUIPS. This slide contains corresponding images to the components of HCUP.
Slide 10
Administrative Data
This slide contains an image of a technician writing on a clipboard.
Slide 11
The Foundation of Administrative Data is Billing Data
This slide contains Billing UB-04 Form. The top third contains demographic data and the bottom two-thirds contains diagnoses, procedures, and charges.
Slide 12
The Flow of Inpatient Admissions
There are two perspectives on this slide, the patient perspective and the data perspective. The patient perspective starts with a scheduled admission or a transfer from the Emergency department, which leads to a reception, admission, administration of care, and discharge. From the data perspective, the patient record is comprised of the reception, the admission, and administration of care, which leads to a discharge summary, a medical coder, and is transferred to the billing department. The billing department then generates the bill.
Slide 13
From Patient to Data
This slide is a visualization of how hospital administrative data is entered into HCUP databases. There are corresponding images to the following steps:
- Patient enters hospital
- Billing record created
- Hospital sends billing data and any additional data elements to Data Organizations
- States store data in varying formats
- AHRQ standardizes data to create uniform HCUP databases
Slide 14
General File Structure for Hospital Administrative Data
Range of file sizes
• ~ 55,000 to 4.0 million records, depending on state
Core set of variables
• Generally available across most states
State-specific variables
• Vary by state
• Allow specific analyses to be done
• Patient race/ethnicity - to examine disparities
• Encrypted patient identifier - to examine readmissions
Slide 15
What Core Data Elements Are Included in Administrative Data?
This slide contains the image of UB-04 billing form. The following information can be found on this form:
• Patient demographics (age, sex)
• Diagnoses & procedures (ICD-9-CM, DRG)
• Expected payer
• Length of stay
• Patient disposition
• Admission source & type
• Dates of admission and discharge
• Hospital identifiers
Slide 16
What are Some State-Specific Data Elements?
• Race/Ethnicity
• Patient county
• Patient ZIP Code
• Severity of illness
• Birthweight
• Procedure date (days from admission)
• Primary payer details
• Secondary payer
• Detailed charges
• Patient identifiers, encrypted
• Physician identifiers, encrypted
• Physician specialty
Slide 17
Example: Payer Detail Varies by State
This slide contains a chart comparing HCUP standards and to data from States. A red circle highlights an example. In the HCUP standard, the description is Private insurance and the value is 3. Conversely, the submitted data from the state has three different descriptions (Blue Cross and Blue Shield, Other Insurance Company/Self Insured, and HMO-PPO) and four different values (B, I/S, H).
Slide 18
Hospital Billing Data Have Benefits and Limitations
Benefits
• Large sample size
• Uniformity of coding
• Routine, regular collection
• Ease of access
• All-payer
• Available at local, state, regional, national level
Limitations
• Differences in coding across hospitals
• No data on individuals outside of hospital system
• May not show complete episode of care
• May not include all hospitals
• Lack revenue information
• Sparse clinical details
Slide 19
Accessing Data
Your own state data organization can provide:
• Most recent data
• Data elements not released outside government
• e.g., patient identifiers, physician identifiers, linkage variables
To compare your data with other states:
• Obtain HCUP-formatted, uniform files through HCUP Central Distributor
• Obtain other states' data directly from other states
• Can access summary statistics through HCUPnet
National benchmarks
Slide 20
Obtain HCUP Data - Two Methods
On this slide, there are two screen shots of the HCUP data base. The top image is of the HCUP Central Distributor www.hcup-us.ahrq.gov/tech_assist/centdist.jsp. The bottom image is a list of HCUP Partner States http://www.hcup-us.ahrq.gov/partners.jsp.
Slide 21
States Releasing SID through HCUP Central Distributor
1990 - 2007*
• Arizona
• Arkansas
• California **
• Colorado
• Florida
• Hawaii
• Iowa
• Kentucky
• Maryland
• Massachusetts
• Michigan
• Nebraska
• Nevada
• New Jersey
• New York
• North Carolina
• Oregon
• Rhode Island
• South Carolina
• Utah
• Vermont
• Washington
• West Virginia
• Wisconsin
* Not all states participate in all years.
** Special application process.
Slide 22
Nationwide Inpatient Sample ( NIS )
Slide 23
What Is the Nationwide Inpatient Sample ( NIS )?
This slide contains an image of the United States. Super imposed is a box with the following text: State Inpatient Databases (SID), Comprehensive hospital discharge data from States. This leads to Nationwide Inpatient Sample (NIS).
Slide 24
Purpose of the NIS
• Allows national and regional studies of inpatient hospital utilization and charges - generates national estimates
• Not recommended for state-level analyses
Slide 25
Statewide Data Systems Participating in NIS
This slide contains a chart listing the data year certain states began participating in NIS.
Data year of:
1988: CA, CO, FL, IA, IL, MA, NJ, WA
1989-1992: AZ, PA, WI
1993-1994: CT, KS, MD, NY, OR, SC
1995-1996: MO, TN
1997-1998: HI, UT, GA
1999: ME, VA
2000: KY, NC, TX, WV
2001: MI, MN, NE, RI, VT
2002: NV, OH, SD (AZ not available)
2003: AZ, IN, NH (ME not available)
2004: AR (PA not available)
2005: OK (VA not available)
2006: VA
There are a total of 38 states participating in NIS.
Slide 26
The NIS Can Be Used for Many Purposes
• Use of and charges for hospital services
• Medical practice variation
• Medical treatment effectiveness
• Quality of care and patient safety
• Impact of health policy changes
• Diffusion of medical technology
Provides national and regional benchmarks to compare your own experience
Slide 27
Coming Soon... National Emergency Department Sample (NEDS)
This slide contains an image of an ambulance.
Slide 28
Software Tools for Use with Administrative Data
This slide contains an image of a medical professional holding a newborn.
Slide 29
HCUP Software Tools
This slide lists the 15 HCUP-related software tools:
- Clinical Classification System (CCS)
- ICD-9-CM CCS
- ICD-10 CCS
- CPT CCS
- Mental Health CCS
- Comorbidity Software
- Chronic Condition Indicators
- Procedure Classes
- Cost-to-Charge Ratios
- HCUPnet
- Hospital Market Structure (HMS) Flags
- Utilization Flags
- Quality Indicators
- Prevention QIs
- Inpatient QIs
- Patient Safety QIs
- Pediatric QIs
Slide 30
Most AHRQ HCUP Tools Can Be Applied to Any Administrative Database
On the left of the slide, there are 3-dimensional discs denoting SID, NIS, KID, SASD, SEDD, and Other Administrative Databases. A bracket encompassing these images point to the following list:
• CCS Tools
• Comorbidity
• Procedure classes
• Chronic Condition Indicator
• AHRQ QIs
Slide 31
Most AHRQ HCUP Tools Aid to Summarize Data
This slide contains the same images from the previous slide on the left. The bracket encompassing these images point to the following list:
• CCS Tools
• ICD-9-CM CCS
• ICD-10 CCS
• CPT CCS
• MH CCS
• Procedure Classes
• Chronic Condition Indicator
Slide 32
Most Tools Based On Medical Coding Classifications
• ICD-9-CM
• CPT
• HCPCS
• DRGs
• MDC
• CCS
On the right side of the slide are cover images of the 2006 ICD-9-CM, the 2006 CPT, and the 2006 HCPCS.
Slide 33
Multiple Coding Systems
Individual Codes
• ICD-9-CM
• CPT
• HCPCS
Groupers
• DRGs
• MDC
• CCS
Which coding system is appropriate for your policy analysis?
Slide 34
ICD-9-CM
• ICD-9-CM Diagnosis Codes
• ICD-9-CM Procedure Codes
• Included in both inpatient and outpatient databases
On the right is the cover image of the 2006 ICD-9-CM report.
Slide 35
Common Procedural Coding System - CPT & HCPCS
• CPT
• HCPCS
• Local Codes
An image of the 2006 CPT report is located on the left side of the slide and the 2006 HCPCS report is located on the right side of the slide.
Slide 36
Clinical Classifications Software (CCS)
This slide contains a visual of ICD-9-CM Diagnosis Codes (12,600 Diagnosis Codes and 3,500 Procedure Codes) being converted into Clinical Classifications.
The CCS collapses ICD-9-CM codes into a smaller number of clinically meaningful categories that can be more useful for presenting descriptive statistics than are individual ICD-9-CM codes.
Slide 37
CCS for ICD-9-CM
This slide contains the diagram showing how CCS groups ICD-9-CM codes into clinically meaningful categories.
A combination of CCS for ICD-9-CM, your data, and ICD-9-CM codes lead to CCS Codes such as CCS 2 (Septicemia) and CCS 6 (Hepatitis).
Slide 38
Most AHRQ HCUP Tools Aid to Summarize Data
This slide contains the same images from slide 30 on the left. A bracket encompassing these images point to the following list:
• CCS Tools
• ICD-9-CM CCS
• ICD-10 CCS
• CPT CCS
• MH CCS
• Procedure
• Classes
• Chronic Condition Indicator
Slide 39
Kansas - 2006
This slide contains a chart ranking ICD-9-CM based on total discharges by in Kansas in 2006.
Slide 40
Colorado - 2006
This slide contains a chart ranking CCS principal diagnosis based on total discharges in Colorado in 2006.
Slide 41
U.S. - 2006
This slide contains a chart ranking CCS principal diagnosis based on total discharges in the U.S. in 2006.
Slide 42
North Carolina - 2000 - 2006
This slide contains a line graph entitled 'Total number of discharges CCS principal diagnosis category 203, Osteoarthritis.' Below are the corresponding dates and number of discharges on the graph:
2000: 11,630
2001: 13,337
2002: 14,658
2003: 16,088
2004: 18,967
2005: 21,083
2006: 22,007
Slide 43
Procedure Classes
ICD-9-CM Procedure codes
• Minor Diagnostic
• Minor Therapeutic
• Major Diagnostic
• Major Therapeutic
Groups ICD-9-CM Codes into One of Four Categories to Distinguish Between Diagnostic/Therapeutic Procedures
Slide 44
Procedure Classes
Cardiac Procedures
• Minor Diagnostic
• Electrocardiogram (ICD-9-CM: 8952)
• Minor Therapeutic
• Pacemaker (ICD-9-CM: 3778)
• Major Diagnostic
• Pericardial Biopsy (ICD-9-CM: 3724)
• Major Therapeutic
• CABG (ICD-9-CM: 3610)
Slide 45
Chronic/Non-Chronic Indicator
ICD-9-CM Diagnosis codes
• Chronic
• Non-Chronic
Groups ICD-9-CM Diagnosis Codes into Chronic or Non-Chronic Categories
Slide 46
Chronic/Non-Chronic Indicator
• Chronic
• Diabetes (ICD-9-CM 25000)
• Non-Chronic
• Food Poisoning (ICD-9-CM 0059)
Slide 47
Probability of Readmission 6 HCUP SID, 2002
This slide contains a chart of data listing various illness categories with unlabeled data.
Slide 48
Comorbidity Software*
This slide contains a visual representation of the elements of Comorbidity software. A combination of ICD-9-CM Codes and DRGs on Administrative Data are the basis of Comorbidity software, which offers 30 comorbidity classifications.
The Comorbidity Software is based on the ICD-9-CM coding scheme. This software creates about 30 variables that identify major comorbidities.
* Elixhauser A, et al. Medical Care, Jan. 1998.
Slide 49
Appends Indicator Flags for Each Comorbidity to Records
• Congestive heart failure
• Valvular disease
• Pulmonary circulation disorders
• Peripheral vascular disorders
• Hypertension (uncomplicated and complicated)
• Paralysis
• Other neurological disorders
• Chronic pulmonary disease
• Diabetes without chronic complications
• Diabetes with chronic complications
• Hypothyroidism
• Renal failure
• Liver disease
• Chronic peptic ulcer disease
30 flags are created and appended to each record.
Examples:
The flag for Congestive Heart Failure is CM_CHF= 0 or 1
The flag for Valvular disease is CM_VALVE = 0 or 1
Slide 50
Appends Indicator Flags for Each Comorbidity
• HIV and AIDS
• Lymphoma
• Metastatic cancer
• Solid tumor without metastasis
• Rheumatoid arthritis/collagen vascular diseases
• Coagulation deficiency
• Obesity
• Weight loss
• Fluid and electrolyte disorders
• Blood loss anemia
• Deficiency anemia
• Alcohol abuse
• Drug Abuse
• Psychoses
• Depression
Slide 51
Utilization Flags
Reveals additional information about use of health care services.
This slide also contains a visual showing that a combination of Utilization Flag software, your data, ICD-9-CM codes, and UB-92 codes are used in departments such as the emergency room, chest x-ray/CT scan, and the intensive care unit.
Slide 52
30 Utilization Flags
On the left of this slide, there is an image of a chart listing Utilization Flags in specific categories. The box next to the image contains the following text: ICD-9-CM codes are not available for all services - there is some concern that some diagnostic procedures may be under-reported.
Slide 53
HCUPnet - On-line Query System.
This slide contains an image of surgeons working on a patient.
Slide 54
HCUPnet: Quick, Free Access to HCUP Data
• Free, interactive online query system
• Users generate tables of outcomes by diagnoses and procedures
• Data can be cross-classified by patient and hospital characteristics
• Includes national, regional, and state data (for states that choose to participate)
http://hcup.ahrq.gov/hcupnet
Slide 55
HCUPnet Can Answer a Variety of Questions
• What percentage of hospitalizations for children are uninsured, by state?
• What are the most expensive conditions treated in U.S. hospitals?
• What is the trend in admissions for depression?
• Will there be sufficient cases to do my analysis?
• How do my estimates compare with HCUPnet (validation)?
Slide 56
HCUPnet Demonstration
This slide contains a screenshot of the HCUPnet homepage.
Slide 57
HCUP User Support Website
• Find detailed information on HCUP databases, tools, and products
• Access HCUPnet
• Find comprehensive listing of HCUP-related publications, database reports, and fact books
• Access technical assistance
http://www.hcup-us.ahrq.gov
A screen shot of the home page of the HCUP User Support Website is located on the right of the slide.
Slide 58
Using HCUP Technical Assistance
Active Technical Assistance
• Responds to inquiries about HCUP data, products, and tools
• Collects user feedback and suggestions for improvement
E-mail: hcup@ahrq.gov
Phone: (866) 290-HCUP
|
