NATIONAL CLAIMS HISTORY (NCH) FILES

The first file group in the Claims and Utilization Data chapter is the NCH 100% Nearline File.

The NCH 100% Nearline File contains all institutional and non-institutional claims from the CWF. It provides records of every claim submitted, including all adjustment claims (which amend claims originally submitted with errors). For analyses that do not require raw claims data, the SAFs offer final action claims data in the form of a netted (adjustment-resolved) file.

National Claims History (NCH) 100% Nearline File

The NCH 100% Nearline File contains all CWF-processed institutional provider and non-institutional claims record data.

File Creation

The NCH 100% Nearline File is created from claims records submitted to CMS (previously HCFA) by CWF host sites. A description of the steps leading up to the creation of the NCH File can be found in the Overview of the Health Care Financing Administration chapter in the Orientation Section. The NCH 100% Nearline File is the only file that provides access to 100 percent of physician/supplier claims data.

Daily transmissions of claims data from the nine CWF host sites are released in weekly batches for uploading into the NCH 100% Nearline File. Each month four to five weekly files are merged and 99 files based on state of beneficiary residence are created (larger states are divided into several segments based on Health Insurance Claim (HIC) sequence segmentation). After 18 months, all files are merged to create an annual file which is approximately 98% complete at that time. Claims continue to trickle into the Nearline File for the next 30 months until the file is "closed" for that particular year. For example, the file for calendar year 1993 was closed in December 1996.

Enterprise Databases Group (EDG) eliminates some fields from the claims data transmitted to CMS by the CWF hosts. Most of these fields contain CWF administrative information. A few data fields are added to NCH 100% Nearline records. Some of the added fields contain administrative data (e.g., CMS claim process date), and others are derived variables (e.g., NCH Beneficiary Discharge Date--on an inpatient and HHA claim, the date the beneficiary was discharged from the facility or died).

Claims data are subjected to the NCH/CWF Medicare Quality Assurance System (NCH/CWF MQAS). NCH/CWF/MQAS is a three tier sequential data editing process. The first two tiers are applied to daily transmittals of CWF data from the host sites. The third tier is applied to NCH 100% Nearline data.

• Tier 1:Daily files are examined to ensure that they have been properly and completely transmitted.
• Tier 2:Edits performed at the CWF host sites are replicated to ensure adherence to contractor's specifications. A five percent sample of data is subjected to a more intensive screening process.
• Tier 3:A sample of institutional and physician/supplier claims is extracted from the NCH 100% Nearline File and analyzed to identify Medicare contractors that exhibit unusual or inconsistent coding practices and to indicate data elements that are generally reliable or particularly unreliable. Information on unusual coding patterns is made available to CMS components.

The NCH 100% Nearline File includes data from 1991 to the present stored on tape at the HCFA Data Center (HDC). The annual files are organized by claim through date. Current year data are available to the user within one month of being received by HCFA from the CWF host sites.

File Maintenance

The NCH 100% Nearline File is updated monthly with records received from the CWF host sites as described in File Creation.

File Structure and Usage

The NCH 100% Nearline File is divided into six record types. Four of the record types are for institutional claims and two are for non-institutional claims. The institutional record types include inpatient/SNF, outpatient, HHA, and hospice. The non-intitutional claims record types include physician/supplier and Durable Medical Equipment Regional Carrier (DMERC) processed claims. The DMERC-processed claims record was phased in beginning with October 1993 with complete implementation by July 1994. Prior to the new DMERC processed claims record type, the DME data was reported as a physician/supplier claim. Migration of Data provides more information about the DMERC-processed claims record and other changes to the NCH record format.

The NCH 100% Nearline File contains detailed information on beneficiary demographics, claims processing, diagnoses, procedures, and payment. It is the only file that contains line item detail for 100 percent of physician/supplier claims data. Prior to 1991 when the CWF was implemented, the BMAD Beneficiary File was the only source of line item detail from physician/supplier claims. The BMAD Beneficiary File, however, contained data on only a five percent sample of all Medicare beneficiaries plus 100 percent of beneficiaries with End Stage Renal Disease (ESRD).

The unit of analysis in the NCH 100% Nearline File is the claim. More than one claim may be associated with a Medicare service such as a hospital stay or visit to the doctor. For example, a long hospital stay may generate one or more interim claims and a final claim. In addition, a claim containing incorrect information may be superseded by an adjustment claim, or a claim may be denied. The NCH 100% Nearline File contains all CWF-processed claims including adjustment, interim claims, and physician/supplier denials.

Data Structure and Usage

The NCH 100% Nearline File consists of six billion records: over 806 million records were created in 1998 alone. Annual NCH 100% Nearline files are organized by year of service based on claim through date. Annual files are segmented into 99 subgroups based on state of beneficiary residence and sequenced by the Beneficiary Claim Locator Number (Beneficiary Claim Account Number and NCH Category Equatable Beneficiary Identification Code (BIC)). The NCH 100% Nearline File data are stored on tape at the HDC in sequential flat files comprised of variable length records.

With the Version H upgrade, the NCH 100% Nearline File includes approximately 375 variables. Each record type contains a fixed portion consisting of variables that can occur only once per record and trailer groups consisting of variables that may occur multiple times. Some variables are common to all six record types while others are not. For example, the fixed portion of all record types contain demographic information such as beneficiary date of birth and sex, and transaction information such as the date the CWF host site adjudicated the claim. The fixed portion of all institutional records also contains the principal diagnosis and additional transaction information such as payment amounts.

The variables contained in the trailer groups portion of the record vary by record type. For example, all record types contain an edit group, composed of variables that indicate data quality problems. Only physician/supplier and DMERC-processed claims records contain line item variables associated with specific services; many services can be listed on the same claim. All institutional claims contain another (secondary) diagnosis group and a revenue center group.

Institutional data (inpatient/SNF, outpatient, HHA, and hospice) and non-institutional (Physician/Supplier and DME) are available beginning in 1991. DMERC-processed claims record in a special format are available beginning in 1994. (The special format and processing of these claims was phased in through the end of June 1994.)

The Division of Data Quality within the Office of Information Services (OIS) serves as the central repository of issues related to CMS's data. In addition, where there are inaccuracies or shortfalls they must be documented and shared with all users of our data.

It is important that all users report data quality issues to:

Centers for Medicare & Medicaid Services
Division of Quality Coordination and Data Distribution (DQCDD)
Enterprise Databases Group
7500 Security Boulevard
Mailstop: N1-15-03
Baltimore, MD 21244-1850

Methods of Access

Access to the NCH 100% Nearline File is available through the Decision Support Access Facility (DSAF) or the Menu-driven Access to the 100% Nearline Claims File (MANRLINE) System. MANRLINE is an online, menu-driven system for submitting retrieval requests for NCH 100% Nearline File data. MANRLINE allows users selections of groups of NCH claims records based on the values of one or more variables such as HIC number, beneficiary state, beneficiary county, provider type, provider number, Unique Physician Identification Number (UPIN), diagnosis, Diagnosis Related Group (DRG), procedure code (ICD-9-CM and Healthcare Common Procedure Coding System (HCPCS)), pricing locality, revenue center, and claim years.

DSAF is a menu-driven system for retrieving data from a variety of claims, utilization, and enrollment files including the NCH 100% Nearline File. It can be used for access to NCH 100% state-specific raw claims data, to obtain final action (adjustments resolved) claims data, and for assistance with MANRLINE requests. Users can also obtain state-specific physician/ supplier claims data in final action form. Final action claims data are produced by matching adjustment claims with initial claims and resolving all adjustments. More information on final action data can be found in the introduction to the SAFs.

Migration of Data

Prior to the CWF and the creation of the NCH 100% Nearline File, institutional claims data were stored in five annual bill files: inpatient, HHA, outpatient, hospice, and SNF. Physician/supplier data were stored in files containing payment records. Payment records did not contain line item detail for physician/supplier services. NCH physician/supplier claims records contain detailed information, including submitted and allowed charges and diagnoses, for every line item on the claim.

Since the NCH was implemented in 1991, several versions of NCH record formats have been produced. In 1998, the Nearline was upgraded to a new format (Version 'H') that includes most of the fields coming in from CWF, as well as a number of new derived data fields to enhance the efficiency of the file. Many of the new elements are redundant summary fields that were added to reduce access processing requirements and to simplify logic; e.g., non-institutional line items Allowed Charges, Submitted Charges, and Medicare Payment Amount are accumulated and the totals stored at the claim level. History (back to 1991 service dates) was populated with data in the new fields where derivable from existing information.

The expanded format contains a number of significant data fields, including placeholders for the National Provider Identifier (NPI) and PAYERID fields. Version 'H' also provides for up to two occurrences of Managed Care Organization (MCO) period information, including the MCO Plan Contract Number, which will be a key field for implementing the collection/storage of encounter data pursuant to the Balanced Budget Act. The revised format allows up to five occurrences of Demonstration Identifier information (alias Demo ID; recently renamed Special Processing Number or SPN). The Demo ID/SPN will facilitate the ready identification of demonstration claims in the Nearline File, which is not possible in the current format.

Additionally, fields have been renamed that have caused confusion among users; fields have been resequenced into more logical groupings; field sizes have been increased and filler positions added throughout to accommodate future expansion; and fields that no longer contain any data have been deleted from the record. Two major format changes have also been implemented: (1) all dollar amounts are increased to S9(9)V99 to provide a consistent standard format for all dollar amounts in the Nearline data; and (2) all dates are reformatted to 9(8) to conform to the International Standards Organization (ISO) standard format for millennium compliance and ease of use.

List of Data Elements

Lists of the data elements contained in version H of the NCH 100% Nearline File, along with brief definitions and coding schemes, will be included in the future on the CMS web site at www.cms.hhs.gov. The data elements are listed by each of the six version H record types: inpatient/SNF, outpatient, HHA, hospice, physician/supplier, and DME.