Page History

Date

Thursday, March 23, 2023, at 1:00 pm ET

Recording

3-23-2023 Townhall Meeting Recording

Passcode: +BWWl99!

Agenda

1. Welcome from CMS
2. Quality Conference - Coming Soon!
3. Monthly Satisfaction Survey
4. Databricks (DBX) Updates and Training Videos
5. Data Updates
6. Introducing the Portal
7. Training Updates and Enhancements
8. Q&A

CMS Quality Conference 2023 - Building Resilient Communities

• The conference will be held virtually on Monday, May 1 through Wednesday, May 3, 2023.

• “Come to the Data” will be presented by Ken Howard the Director of the Division of Engineering and Architecture.

• The exact day and time coming soon.

• Register here: https://www.cmsqualcon.com/en

Databricks Update

Databricks Notebook Solution (Phase One) 

• What will be in scope for the first release?
  • All native notebook capabilities and library installations.
  • Languages: Python, R, Scala, Structured Query Language (SQL)
  • Workbench and Centralized Data Repository (CDR) Data
  • Two different clusters- R/Machine Learning (ML), Python/SQL
  • Users can share their notebooks.
  • Users can visualize their results.
• Training Hub 
  • Example: How to connect to GitHub.
  • Users can provide any ideas for future curriculum.
• Support
  • #help-service-center-sos Slack channel
  • Email servicecentersos@cms.hhs.gov 
• When will we be onboarding users? 
  • April 

Databricks & SAS Integration (Phase Two) 

• What is included in this release?
  • Ability to connect SAS Programs to Databricks (DBX) SQL for isolated, scalable compute.
  • SAS Viya connection to external systems and databases
  • SQL endpoints
• When will this be released?
  • Q2
• What training will be offered?
  • Example: Converting from HiveQL to American National Standards Institute (ANSI) SQL.

What are the benefits of Databricks?

• More powerful compute platform

  • Faster query performance

  • ML capabilities

  • Data modeling and data visualization

• More stable environment

  • Auto-scales

• Supports multiple languages.

Databricks Training Videos

• Reference the recording for the following two training videos (soon to be released on the Training Hub):

  • Using Databricks Notebooks

  • How to Import a Repository into Databricks.

Data Updates

Medicare Severity Diagnosis Related Groups (MSDRG) Availability 

• MSDRG data is now available in the CDR.

  • MSDRG is a system of classifying a Medicare patient’s hospital stay into various groups in order to facilitate payment of services.

• In the nch_part_a schema, you will find three tables containing all MSDRG codes produced by the CMS MSDRG Grouper software, as well as the Part A Input data:

  • claims_msdrg_output

  • claims_msdrg_output_diagnosiscodeinfo

  • claims_msdrg_output_procedurecodeinfo

• Codes began refreshing in the CDR on claims records process week 01/06/2023 with Version 40 of the software.

• Work is underway to perform a historical backfill of MSDRG codes.

National Claims History (NCH) Claims Data Duplication Issue

• We recently reported a claims issue affecting Part A and B Claims with the specific NCH weekly processing date of 10/11/2022.
  • There are separate claims (different claim_sk values and beneficiary information and other claim field values) that have the same nch_clm_sgmt_link_num and nch_daily_proc_dt.
  • Some users concatenate these two values to create a unique claim identifier. 
• Resolution: 
  • During this month’s maintenance window, we wrote the corrected nch_part_a claims header tables to production.
  • The the appropriate claims in the Part A schema now an NCH daily processing date of 10/10/2022.
• Next Steps: 
  • We will move forward with correcting the NCH_PART_B Hive tables and writing them into production at the beginning of next month during the next maintenance window.

Hospital Quality Reporting (HQR) Data Issue

• The HQR team reported an issue with the quality_measures_hqr_data_form schema this past December.
• The years 2022 and 2023 numerator, denominator, and score fields were supposed to be the values for the whole year (Q1 – Q4), but instead they were the values submitted for Q4.
• Resolution: The HQR team reported this issue as resolved as of 3/13. An updated data dictionary can be found on the CDR Data Catalog.

Introducing the Portal

A single place for everything Data & Analytics!

• Single sign-on (SSO) Experience to Data and Analytics (D&A) Tools

• My Access experience

• The new home for the CCSQ Public Content & Training Hub!

• Data sharing capabilities

• User collaboration

My Data Use Agreement (DUA) Experience – MVP!

• Privileged Users will be able to upload an organization’s DUA to the D&A Portal.

• Automatically applying coarse-grain access controls to organizations when accessing the CDR.

• DUA Upload History by organization with the ability to view the original DUA.

We Need You!

• Look for announcements to participate in future User Discovery Interviews, and Usability Testing opportunities.

• Contact the CCSQ D&A team for more info. 

Training Updates

Confluence Clean Up Efforts

• Audit of Confluence content

• Remove outdated content.

• Language review

• Data Catalog

Training Hub

• Micro Training Videos

• Knowledge Base Articles

• Reference the recording for a walk through of the Training Hub- Archive.

Q&A

1. During Phase One for Databricks, will Zeppelin be decommissioned?

   A - The end goal is to sunset Zeppelin, but it will not be as quick as once we release Databricks. 

2. Is Databricks taking the place of SAS Viya? If not, will we be able to continue to use SAS Viya and HiveQL code language?

   A - Databricks is not replacing Viya. In the Phase Two release, users will be able to connect Viya or other tools like Redshift to Databricks. Training will be provided on training to ANSI SQL since Databricks supports this instead of HiveQL. 

3. Will a recording be available for us to review on our own later?

   A -  All micro-training videos shown during the Townhalls will be made available out on Confluence for everyone to reference.

4. Do we need separate Healthcare Quality Information Systems (HCQIS) Access Roles and Profiles (HARP) user role for Databricks?

   A -  Yes. Onboarding information will be sent in an email that includes your onboard date, training material, and support. 
   

5. Does Databricks have access to existing workbench (S3)?

   A -  It will. 
6. Can users install and load R packages by themselves?
   A - Yes. 

7. Will it be a requirement to eventually use DataBricks or can we continue working in SAS Viya if we prefer?

   A - No, users can continue using SAS Viya. The endpoint will be changed. Users will be using the SQL warehouse for Databricks.

8. Will the python notebook have ML functionality in a future release?

   A - Yes. 

9. Do the Databricks website trainings require us paying to access them? A lot of sites like SAS Viya and Tableau require payments to access trainings.

   A - No. The Databricks company has many free video trainings users can leverage. 
   

10. What specific HARP role do we need to request?

    A - Onboarding instructions containing HARP role information will be shared shortly. 

11. Is there an upper limit for how many jobs one user can submit simultaneously? Is there an upper limit for memory usage or time running?

    A - Databricks has can auto-scale. Once it hits a certain memory usage, the job will be killed automatically. 

12. Will MSDRG be add to Complete Statistical Analytic Table (CSAT)?

    A - The D&A Team left in the claim_sk join field to all the tables to make it easier for end users to join with other data sets. MSDRG data is at the claim level, so it’s only the claim_sk from the claim_sgmt_line_num = 1. Assuming that users are working at the claim level in CSAT, users should be able to join to the claims_msdrg_output table on claim_sk. 

13. Also, does the claim id issue relate to CSAT?

    A - The duplication isn't affecting CSAT at this time. 

14. When will the Training Hub be available? 

    A - The Training Hub- Archive is already available today.