HCUP/AHRQ updates

updated 6/19/2017

HCUP Offers Guidance for Using ICD-10-CM/PCS Data

A new ICD-10-CM/PCS Resources section on the HCUP User Support (HCUP-US) Web site provides information and educational materials for researchers using the Healthcare Cost and Utilization Project (HCUP) and other administrative databases.

The ICD-10-CM/PCS Resources section documents key differences in the structure of HCUP databases, provides general guidance and forewarning to users analyzing outcomes that may be affected by the transition to the ICD-10-CM/PCS coding system, and includes additional Web resources that may be useful to researchers.

If you have any questions or suggestions for additional Federal and State ICD-10-CM/PCS Web resources, please contact HCUP User Support.

2015 HCUP State Databases—Revised Structure and New Data Elements

The HCUP State Databases are annual, calendar-year files. The introduction of ICD-10-CM/PCS on October 1, 2015, means that the 2015 State Databases include a combination of codes:

  • Nine months of the data with ICD-9-CM codes (January 1, 2015, to September 30, 2015)
  • Three months of the data with ICD-10-CM/PCS (October 1, 2015, to December 31, 2015)

To alert users to this change in the data, the file structure of the 2015 HCUP State Databases stores the first three quarters of data with ICD-9-CM codes separately from the fourth quarter of data with ICD-10-CM/PCS codes.

In addition, the names of diagnosis- and procedure-related data elements under ICD-10-CM/PCS have been modified to clearly identify the new coding scheme. ICD-10-CM/PCS information on the 2015 Nationwide Databases will be available when the databases are released. Additional information on the 2015 State Databases is available in the ICD-10-CM/PCS Resources section. For questions, please contact HCUP User Support.

HCUP 1-hour webinars:

The Agency for Healthcare Research and Quality (AHRQ) is offering a two-part webinar series on the Healthcare Cost and Utilization Project (HCUP)’s research-enhancing databases, products, and tools.

September 6 – Overview of the HCUP Databases
Summary: This one-hour session will introduce health services and policy researchers to the HCUP databases and related resources through a general overview of HCUP’s databases.

September 13 – Overview of the HCUP Products and Tools
Summary: This one-hour session will introduce health services and policy researchers to the HCUP products and tools, with a particular emphasis on HCUPnet, the free online data query system.

> Information for both events will be available on the HCUP website in August on the HCUP Workshops and Webinars page.

HCUP Data Users’ Workshop:

On September 19, AHRQ will sponsor a full-day, intermediate-level HCUP data users’ workshop at its headquarters in Rockville, Maryland. This workshop will teach health services researchers and analysts how to use or improve their use of HCUP databases and products. Registration information and the workshop description will be available in August on the HCUP Workshops and Webinars page.

HCUP e-News

Summer 2017; Issue #51


Databases and Products

Recently Released: 2015 State Databases and Additional 2014 Databases

Since March 2017, the following State Databases have been released:

In addition, the Nationwide Readmissions Database (NRD) is now available for data years 2010–2012.

Complete listings of available databases by year can be found in the Database Catalog on the HCUP-US Web site.


Publication Spotlights

New HCUP Statistical Briefs Posted on HCUP-US

Since March 2017, the following HCUP Statistical Briefs have been released:

  • #222 Delivery Hospitalizations Involving Preeclampsia and Eclampsia, 2005–2014
  • #223 Surgeries in Hospital-Based Ambulatory Surgery and Hospital Inpatient Settings, 2014

To access these and other Statistical Briefs, please visit the Statistical Briefs page on the HCUP Reports page.

New HCUP Method Series Report Now Available

Since March 2017, the following Methods Series Report has been released:

Method Series Report #2017-02: Feasibility Report on Redesigning the Nationwide Emergency Department Sample (NEDS) discusses considerations for a redesign of the NEDS, evaluates the possible changes, summarizes comparative analyses of alternative designs to external data sources, and offers recommendations for redesigning the NEDS

This report and other HCUP reports can be found on the HCUP Reports page.

Publications Using HCUP Data

Feldman D, Swaminathan R, Geleris J, et al. Comparison of trends and in-hospital outcomes of concurrent carotid artery revascularization and coronary artery bypass graft surgery: the United States experience 2004 to 2012. JACC Cardiovasc Interv. 2017;10(3):286-98.

This study used the 2004–2012 National (Nationwide) Inpatient Sample (NIS) to examine trends in carotid endarterectomy and carotid artery stenting utilization when performed during the same hospitalization as a coronary artery bypass graft. In-hospital outcomes also were examined by patient age, sex, and carotid artery disease status. The article abstract is available via PubMed.

Whitman I, Agarwal V, Nah G, et al. Alcohol abuse and cardiac disease. J Am Coll Cardiol. 2017;69(1):13-24.

This study uses the 2005–2009 California State Inpatient Databases (SID), State Ambulatory Surgery and Services Databases (SASD), and State Emergency Department Databases (SEDD) to evaluate the relationship between alcohol abuse and atrial fibrillation, myocardial infarction, and congestive heart failure. The article abstract is available via PubMed.

 

To read additional recently published articles featuring HCUP data, please visit the Research Spotlights page on the HCUP-US Web site.


HCUP Q&A

Question:

It appears that HCUPnet has been redesigned. I have some questions regarding this newly redesigned site’s functionality.

  • Will the new HCUPnet site have all the functionality of the archived site?
  • How do I select patient and hospital characteristics? Outcomes and measures? Previously, these options were displayed on one screen during the step-by-step query process on the archived site.
  • For one of my queries, I am querying several ICD-9-CM diagnosis codes. I am interested in both mean and median costs and charges, but median does not appear to be an available option. Is that intentional?
  • In the same query of multiple ICD-9-CM diagnosis codes, I noticed that standard errors are not present. Is that intentional?
  • I recently purchased the Nationwide Readmissions Database (NRD) and am using HCUPnet as a point of comparison for my results. My analysis is looking at readmission rates for certain diagnosis-related groups (DRGs). Although my counts/rates are close, they do not match those from HCUPnet—do you know why that might be?

Answers:

Will the new HCUPnet site have all the functionality of the archived site?

Yes, the new, redesigned HCUPnet site will have all the functionality of the older, archived site. AHRQ is in the process of transferring functionality to the new site and will have most searches available by June 15, when the archived site is retired.

AHRQ will add the remaining searches and some additional enhancements in successive updates through the summer. Current functionality lets you query inpatient (including readmissions), emergency department, ambulatory surgery, and community (county) statistics, however not all search options are possible yet. Searches for inpatient stays can be conducted using ICD-9-CM codes, Clinical Classifications Software (CCS), DRG, major-diagnostic category (MDC), and service line classifications; searches for other data types may be available only by CCS or by CCS and ICD-9-CM codes. The two-way tables option was just added to the redesigned site last month.

It is important to note that the newly redesigned HCUPnet site will have some additional functionality that the archived site does not offer currently. For instance, users now can select an option that displays related diagnoses and procedures for conditions of interest.

Users are encouraged to explore HCUPnet regularly to determine what has been added over the course of the summer. For additional information or questions related to HCUPnet, please contact HCUP User Support.

How do I select patient and hospital characteristics? Outcomes and measures? Previously, these options were displayed on one screen during the step-by-step query process on the archived site.

To obtain this information you will need to follow a two-step process.

First, from the HCUPnet homepage, create a new analysis in which you will indicate the type of information in which you are interested (i.e., hospital inpatient, emergency department, or ambulatory surgery data) as well as whether you are interested in specific diagnoses or procedures or all patients in general.

Once the results are displayed from your initial analysis on the HCUPnet homepage, you can refine the search by using the navigation bar on the left-hand side to make changes to your analysis. “Outcomes and Measures” display the same options available to users on the archived as well as some newly added options such as readmissions. Although previously combined, “Patient Characteristics” and “Hospital Characteristics” now are two separate options available to users, each of which includes the same options as the archived site with new options to be added in the future.

In addition to selecting outcomes and measures or patient/hospital characteristics, you can change your setting of care, the geographic setting, or the year of your analysis as well as make changes to the way diagnoses or procedures are classified or displayed.

To add an option to your query, click on the adjacent plus sign and then submit your request. The screen then will regenerate your results.

For one of my queries, I am querying several ICD-9-CM diagnosis codes. I am interested in both mean and median costs and charges, but median does not seem to be an available option. Is that intentional?

If you selected the “Combine All Codes” option when entering your diagnoses or procedures of choice, then yes, median charges and costs are not available measures. You cannot obtain median values when all codes are combined because HCUPnet is unable to calculate medians by reaggregating the medians for each individual ICD-9-CM code. Median charges and costs statistics are available, however, if you modify your search so that each ICD-9-CM code is queried separately. Another option is to use a pre-aggregated classification of diagnoses, such as CCS or DRG or alternatively, purchase the full HCUP database.

For additional information on purchasing HCUP databases, please contact the HCUP Central Distributor.

In the same query of multiple ICD-9-CM diagnosis codes, I noticed that standard errors are not present. Is that intentional?

Like the option for median costs or charges, HCUPnet is unable to calculate standard errors when the option “Combine All Codes” is selected. This is a limitation of the way that the data are aggregated for the Web site. To obtain this information, you would need to purchase the full HCUP database.

It is also important to note that in general, standard errors are not present in HCUPnet for State-level statistics because they are often not applicable. A standard error is a measure of the precision of a statistic. It reflects the amount that a sample statistic’s value would fluctuate if a large number of samples were to be drawn using the same sampling design. All the HCUP Nationwide Databases are samples and because of this design methodology, standard errors are provided in HCUPnet given that all nationwide statistics are weighted to be representative of the entire U.S. population. The HCUP State Databases, however, are not samples, they capture the universe of inpatient and outpatient discharge records in participating HCUP States.

I recently purchased the Nationwide Readmissions Database (NRD) and am using HCUPnet as a point of comparison for my results. My analysis is looking at readmission rates for certain diagnosis-related groups (DRGs). Although my counts/rates are close, they do not match those from HCUPnet—do you know why that might be?

The discrepancy you describe most likely is the result of differing methodology between your analysis and HCUPnet. In your specific analysis, you are using the NRD data element DRG, which takes into consideration whether a condition was present on admission (POA).

National statistics in HCUPnet based on DRGs were created without consideration of POA (HCUP data element DRG_NoPOA). That decision was made because POA information is not uniformly available from all participating HCUP Partners.

Additional discrepancies may be the result of other varying methodology such as inclusion/exclusion criteria or the definition of an index stay. Additional information on the methodology used to define readmissions in HCUPnet is available at https://hcupnet.ahrq.gov/#method-readmissions or in the Introduction to the NRD.



The HCUP e-News is available online at the HCUP-US Web site and distributed via the HCUP Mailing List.

For more information and questions regarding the HCUP e-News, please email hcup@ahrq.gov