VUHDDS Frequently Asked Questions

Return to Top

What is the Vermont Uniform Hospital Discharge Data Set?

Vermont has been collecting and reporting hospital discharge data since the early 1980’s. All fourteen of Vermont's general acute care hospitals currently contribute records for Vermont residents and non-residents to the Vermont Uniform Hospital Discharge Data Set (VUHDDS). Under the statutory authority to collect these data (18 V.S.A. §§§ 9410, 9456 and 9457), the Vermont Green Mountain Care Board (GMCB) administers the program and the Vermont Department of Health (the Health Department) manages the data set under an agreement with GMCB.

The Vermont Association of Hospitals and Health Systems-Network Services Organization (VAHHS-NSO) contracts with the state of Vermont to collect the Vermont hospitals data that are then incorporated into the VUHDDS following a series of quality assurance checks performed by the Health Department. The Veterans Administration Hospital in White River Junction has not submitted data since June 30, 2006. The Vermont psychiatric hospitals, including the Brattleboro Retreat and the Vermont State Hospital, do not submit data to VUHDDS.

In addition to the data collected from Vermont hospitals, GMCB obtains records for Vermont residents using hospital services in the bordering states of New Hampshire, New York and Massachusetts under data use agreements with the New Hampshire Department of Health and Human Services, the New York Department of Health, and the Massachusetts Division of Health Care Finance and Policy. While records for Vermont residents using hospitals in New Hampshire, New York, and Massachusetts include inpatient discharges, not all states collect comparable data sets for outpatient and emergency department discharges. GMCB is prohibited from releasing records obtained from these bordering states per provisions of the data use agreements with each state.

The Vermont hospital subset of the VUHDDS includes records for inpatient, outpatient, emergency department, observation bed, and series patient discharges for both Vermont residents and non-residents. The collection of emergency department records from Vermont hospitals began with reporting year 2001. From 1989-2000, outpatient data collection was limited to records that each hospital identified as surgical procedures performed in operating rooms. Starting in reporting year 2001, the outpatient definition was revised to include ICD-9-CM procedure codes 00.0-86.99 that occurred in all hospital-based outpatient settings. Starting in reporting year 2006, while all Vermont hospitals continued to report the core outpatient set specified by code range, most hospitals also started to submit all records considered outpatient procedures beyond the specified code range. This includes a wider range of diagnostic procedures and treatments. GMCB is working with the Health Department, VAHHS-NSO, and Vermont hospitals to determine how to define, edit, and use the expanded outpatient data set.

Return to Top

What are the conditions of use for the public use file?

Release of public use data is subject to the following conditions, which the requestor agrees to comply with upon accepting copies of the data:

Hospital Data Release Policy (Last updated November 2002) PDF

The data may not be used in any manner that attempts to or does identify,   directly or indirectly, any individual patient or health care practitioner (18   V.S.A. § 9457).

The requestor agrees to incorporate the following, or a substantially similar, disclaimer in all reports or publications that include public use data:

“Hospital discharge data for use in this study were supplied by the Vermont Association of Hospitals and Health Systems-Network Services Organization (VAHHS-NSO) and the Vermont Green Mountain Care Board (GMCB). All analyses, interpretations or conclusions based on these data are solely that of [the requestor]. VAHHS-NSO and GMCB disclaim responsibility for any such analyses, interpretations or conclusions. In addition, as the data have been edited and processed by VAHHS-NSO, GMCB assumes no responsibility for errors in the data due to coding or processing by hospitals, VAHHS-NSO or any other organization, including [the requestor].”

Return to Top

What data elements are included in the Public Use File?

Public Use File Data Elements:

NOTE:  Significant changes in 2014 for reporting outpatient procedures

As of July 1, 2014, changes were made in the requirements for Vermont hospitals’ reporting of all outpatient procedures and services. Prior to that date, outpatient procedures were reported using the ICD-9-CM coding system; these codes were included as appropriate on each individual discharge record with a primary procedure code determined by the reporting hospital.  As of July 1, 2014, outpatient procedures are reported using the HCPCS/CPT coding system; these codes are included on each of one-to-many revenue records, and there is no determination by the reporting hospital of which CPT code is considered to be primary.  Therefore, a primary CPT code has been calculated and flagged for each outpatient discharge using an algorithm that includes relative value units, charges, and whether the CPT code is for an ambulatory surgical procedure.  Any CPT codes reported prior to July 1, 2014 are considered incomplete, as are any ICD-9-CM procedure codes reported on or after that date.

To ensure that analyses of outpatient procedures include complete data that are as consistent across time as possible, the 2014 files include two new fields (CCSPROC and CCSPROCGRP) that have been created using Clinical Classifications Software (CCS).  This software groups both ICD-9-CM procedure codes and HCPCS/CPT codes into the same single and high level categories based on body systems.  For the first half of 2014, groupings are based on the first procedure in ICD-9-CM range 00.00-86.99, and for the second half of 2014 groupings are based on the primary CPT.  In analyses that include prior years of data, these new fields are comparable to the earlier CCSPPX and CCSPPXGRP fields. Some discontinuities may exist due to differences between the two coding systems. 

No changes were made to the reporting of inpatient procedures, and ICD-9-CM codes were used throughout 2014.  CCSPPX and CCSPPXGRP have been renamed to CCSPROC and CCSPROCGRP in the inpatient data file for consistency across hospital settings in 2014.

Return to Top

Which Public Use Files are currently available for Downloading?

Brief Descriptions of Data Files

The Inpatient Data and Outpatient Data files combine to make the complete data set for each calendar year (based on discharge date). These files are separated for easier downloading and because they are frequently used individually. The third data file for each year includes only records relating to the Emergency Department (ED), and is extracted from both the Inpatient and Outpatient data files.

The fourth data file, starting in 2006, is the Expanded Outpatient file. The submission of "expanded" outpatient records by Vermont hospitals for calendar years 2006, 2007, and 2008 with principal procedures outside the ICD-9-CM range of 00.00-86.99 was incomplete and inconsistent across the hospitals. GMCB is working with the Vermont hospitals to ensure a complete and consistent filing of the expanded outpatient data set in more recent reporting years.

Data files from 2006 and later do not include records from the Veterans Affairs Medical Center, although these records are included in the 2005 and earlier files. Text files are .zip formatted for greater downloading ease. Zipped sizes shown.

For each calendar year, there are two Excel files in addition to data files. These files, Field Properties and Code Tables, include the file layouts, field properties and code tables for use with the Inpatient, Outpatient, Emergency Department, Expanded Outpatient and Revenue Code data. These files can be used to set up your tables prior to downloading or to reduce the size of your tables after downloading them.

Please note that the Inpatient data files include a number of fields that do not appear in the Outpatient, Emergency Department and Expanded Outpatient data files (SCU, SAME DAY FLAG, DRG, MDC, and DRG_GROUPER_VERSION). These fields are not valid across all hospital settings and should only be used with the Inpatient data. Please use the Outpatient Layout for the Outpatient, Expanded Outpatient and Emergency Department files for this reason.

The DRG field in the Inpatient data file changed as of October 1st, 2008 as a result of Federal regulation to use of MS-DRG codes. The field name is still referred to as DRG, but MS-DRG codes populate the data field.

Note that for each year from 2006 forward, the UNIQ field can be used to link each discharge record with its associated revenue records.  For the years prior to 2006, an ID field is not included:  you may want to add a unique ID to your data set if you are combining multiple text files into a master file.

Data Files Available Online

Annual data files available online are listed below. Please contact the Health Department for information about earlier VUHDDS data files, as listed in the Contacts section.

Return to Top

How do I request a research file with non-public data elements not included in the Public Use Files?

Please note that there is a moratorium on the release of research files with non-public data elements outside of Vermont state government. This webpage will be updated when the moratorium is lifted.

Return to Top

What do I need to know in order to download the VUHDDS?

This section has been prepared to assist you with loading and using the Vermont Uniform Hospital Discharge Data Set (VUHDDS) public use files. This section covers Installation Minimum Requirements, Concepts & Purpose of this Relational Data Set, Data Issues, and Getting Started Instructions. If you have questions about or run into issues with the loading of these data sets, please send an email requesting assistance to the Health Department, as listed in the Contacts section.

The Expanded Outpatient and Revenue Code public use files are very large and contain millions of records. These larger data files are too large to load into software packages such as Microsoft Excel or Microsoft Access.

To load the large files and take advantage of their features requires a relational database application, such as MYSql, Oracle, DB2, SAS, SPSS or SQLServer. It is also recommended that sufficient hard drive storage space be available to efficiently manage the processing of the data tables. Your Database Administrator (DBA) should be able to determine minimum storage requirements based on the size of the files that are extracted to your computer.

The primary purpose of this relational data set is to give research entities access to de-identified hospital discharge data that can be used to produce public research analyses.

The concept used in the creation of this data set was to create primary data tables that hold specific information and then provide code tables that define the values listed within primary data tables. A conceptual layout of the field properties and code tables is included on this web site.

Each line of information is a unique discharge where all information relating to a given discharge record has already been combined. Since there are a large number of variables in these files, the unique discharge is separated from the revenue code file, but can be linked together using the unique id for each record. There may be many revenue records that link to a single discharge record.

This method provides a useable dataset where all adjustments to the data have been applied to the extent possible within the given time period. The annual data sets provided on the website are based on the discharge date by calendar year. A given year’s data set will be posted as soon as the data are cleaned and become available for public use.

Discharge records may contain values that are blank or negative dollar charge values. Although the data are edited and cleaned, there may still be missing information or unresolved code issues in some cases. Timely availability of data requires the acceptance of a reasonable degree of errors or missing data.

The Vermont Public Use files on the web site are in a comma-delimited text file format and can be read by any software package/database loading process that can use this format. In each of the data files the first row contains the appropriate Data Dictionary field/column name. Each field is delimited with a ‘,’ or comma.

Return to Top

How are hospital service areas (HSA) defined in the VUHDDS?

The hospital service area (HSA) definitions used in the VUHDDS have been modified three times to reflect changing discharge patterns. Starting with the 2002 Monograph, HSAs are based on inpatient discharges where the diagnosis indicated the need for immediate care, for the cumulative reporting years spanning 1997-2001.

Prior definitions of HSAs had been based on all inpatient stays (except for newborns and transfers) and had used a strict decision rule for inclusion of ZipTowns in a hospital’s service area (towns or groups of towns: see definition below for ZipTowns). Use of a strict inclusion rule meant that there were always towns that were not assigned to any HSA and were labeled “contested.”

In preparing for the fourth version of HSAs, rules used to define the 3rd version of HSAs were applied to 1997-2001 data. The number of contested towns rose from 54 to 80, nearly a third of all Vermont towns. This resulted in the loss of one hospital service area and a serious erosion of others. As a result of this increase in contested towns, two major changes were made to the process of defining HSAs.

The first major change was that selection of records was limited to those discharges with diagnoses considered to require immediate hospitalization as described by Newton and Goldacre in their 1994 article, “How many patients are admitted in districts other than their own, and why?” (Journal of Public Health Medicine, 1994, Vol. 16, No. 2, pp. 159-164). Newton and Goldacre used records of patients with acute conditions that require immediate hospitalization, because patients with these conditions tend to go to the hospital nearest to them when they become ill. The authors excluded newborns and patients transferred from another hospital. The authors included records with the following acute conditions as their primary diagnoses: appendicitis, abscess of anal and rectal regions, peritonitis, acute myocardial infarction, pneumococcal pneumonia, and asthma.

The second major change was the shift to a plurality rule in assigning ZipTowns to HSAs, so that all towns are assigned to an HSA with none remaining contested. ZipTowns are towns or groups of towns with over 1,000 residents and relatively contained ZIP code boundaries. ZipTowns were assigned to hospital service areas as follows:

1.  If the plurality of a ZipTown’s discharges were from a Vermont hospital or Dartmouth Hitchcock Medical Center (DHMC) in Lebanon, New Hampshire, the ZipTown was assigned to that hospital’s service area.

2.  If the plurality of a ZipTown’s discharges were from a non-Vermont hospital other than DHMC, the ZipTown was assigned to the HSA of the Vermont hospital (or DHMC) with the next highest number of discharges. The HSAs continue to include two multi-hospital areas: the White River Junction HSA (Dartmouth Hitchcock Medical Center, Mt. Ascutney Hospital, and the Veterans Administration Hospital) and the Brattleboro HSA (Brattleboro Memorial Hospital and Grace Cottage Hospital). These HSAs include multiple hospitals because resident hospitalizations are split among the hospitals with no clear majority of inpatient discharges concentrated at any one hospital.

In addition to these definitional changes, HSA names were changed. To distinguish these new HSAs from previous ones, HSAs were given the name of the most populous town in their area.

Return to Top