MN APCD Presentation Slides with Agenda for February 8, 2016 (PDF)

Minnesota All Payer Claims Database
Public Use File Workgroup
Meeting 2: Initial Public Use Files
February 8, 2016
Kris Van Amber
Senior Management Consultant
Management Analysis &
Development
Minnesota Management & Budget
Linda Green
Vice President, Programs
Freedman HealthCare
Agenda
Welcome and Overview
40 minutes
Types of Public Use Files
PUF Support and Logistics
50 minutes
15 minutes
Public Comment
10 minutes
Next Steps and Adjourn
5 minutes
2
MEETING GOALS AND PROJECT OVERVIEW
3
Goals of Today’s Meeting
• Update the Workgroup on the trajectory of the Public Use File (PUF)
development process
• Gather feedback from the Workgroup on the development and release of the
initial PUFs in March 2016
• Discuss key considerations for PUF data release/access
• Share plans for presenting the PUFs to the user community and gathering
feedback
4
Overview of the MN APCD PUFs
• Legislature’s requirements for MN APCD Public Use Files:
•
•
•
•
Summary data must be made available to the public
Data about patients, providers and payers requires continuing protection
Provide initial PUFs by March 1, 2016; downloadable web-based files by June 30, 2019
Files should be available at no (or minimal cost) to the user
• Principles for PUF development
•
•
•
•
•
•
Iterative process that engages users at all stages
Incorporates feedback from stakeholders and users
Includes clear and explanatory documentation for data users
PUFs must be free or low-cost, available to anyone
Minimal data request process
Uses the MN APCD as the sole data source
5
PUF Development Process
• November 2015: PUF Workgroup Meeting
• Incorporate guidance and insight received
• January-February 2016: Develop framework for initial PUFs
• March 2016: Release initial PUFs and documentaiton
• Pre-release user webinar
• Post-release user survey and Q&A for feedback
• April 2016 and onwards: Future strategies to expand PUF
development and meet June 2019 directive
•
•
•
•
Incorporate guidance and insight received
Refresh data files
Create new stratifications
Align with ongoing research activities
6
Why are these files important?
• The PUFs demonstrate the value of the MN APCD to the members of the
public
• The PUFs contain the same information that the Minnesota Department of
Health uses to design, monitor, and evaluate new interventions and programs
to improve the health of Minnesotans.
• The PUFs offer a high-level look at trends
• The PUFs create “open data” by providing public access to state-produced
information
Discussion
What other points should we raise to the public about the value of these files?
7
Interviews with Minnesota Health Data Users
• Potential data uses
• Service utilization across care settings
• Understanding care provided to individuals over a long period of time
• Examining readmissions
• Advantages of summary tables
• Helps identify new research questions
• Is a validation/comparison tool for researchers’ own data
• Helps compare Minnesota’s population to the rest of the country
• Advantages of detailed summary files
• Allows researchers to conduct extensive data validation and nuanced analysis
• Researchers can use the same analytic approach with PUF data and their other datasets
• Provides patient-level information
• Desired characteristics of PUFs
•
•
•
•
•
•
As detailed as possible
Multiple years of
8 data to show trend
Type of service provided (x-ray, lab, examination, surgery)
Type of service setting (inpatient, outpatient, ambulatory surgical center )
Primary care vs. specialty provider
Tracking patients across different types of coverage (commercial, Medicare, public program)
8
Diverse Data Interests
• Different types of users with different needs
• Researchers want a more detailed file that is only useful to users with
extensive skill and training
• Other users want a file that provides useful answers without further analysis
• The aim: find a balance that starts out meeting some needs and move on
from there
9
SOURCE DATA
10
Data Parameters and Characteristics
of the Source Data
•
•
•
•
All Insured: Commercial, Medicaid, Medicare
Includes 2013 data but will be built out over time to add additional years
All settings of care
Paid claims
• Supports population analysis, trends, geographic variation
• Understanding the data
• Patient privacy protections
• Provider and payer information
• Data sources
11
FRAMEWORK FOR THE PUBLIC USE FILES
12
Characteristics of the PUFs
• Meaningful information that protects individuals’ privacy
Summary data only (not claims level)
Does not identify individual patients, payers, or providers
Uses age banding and zip code aggregation to group records together
Cell size suppression prevents re-identification (each cell in the file must include data for ≥11
unique individuals)
• Annual updates
• Freely accessible to the public
•
•
•
•
• Clear data documentation
• Topics include: protection of sensitive data, what is included in each file, and the most powerful
uses for the data
• User engagement to gather feedback
• Pre- and post-release user engagement and education (webinar, survey, Q&As with monthly
updates)
• Feedback and recommendations from stakeholders will be documented to inform future PUF
development
13
Plan for a Phased Approach to File Production
• Current thinking: Start with smaller summary files that provide helpful insights to a range of
users; then expand to more detailed summary files for more experienced researchers
• March 2016: Dates of service for 2013
• Summer 2016: Expand to include dates of service for 2013-2014 (with potential for partial
2015)
• Ongoing
•
•
•
•
Will refresh the initial March 2016 files by late summer, if possible
Beginning in 2017, will add new data when available
Will identify other themes to include in future PUFs
Medicare medical data for calendar year 2014 will be included in summer 2016 PUF release; anticipating
Medicare Part D 2014 data as well
Discussion
• Does the Workgroup agree that these plans are consistent with the guardrails?
• What does the Workgroup think the next refresh should focus on: more years or more
stratifications?
• Are there other things we should keep in mind?
14
TYPES OF FILES
15
March 2016 PUFs – Probable Structure/Format
• Spreadsheet or csv (depending on total size)
• One year of data
• Patient residence by county and 5-digit ZIP code
• Age bands at 5 year intervals where appropriate
• Each row will show a count of patients
• Each file contains an “all other” category to help users understand what
proportion of the total data is included in the PUF
Discussion
• Have we interpreted the guardrails too narrowly?
16
Initial File Type: Diagnosis Summary File
•
Shows the frequency and distribution of principal diagnoses that Minnesotans received
during an inpatient hospital stay or other settings
•
What can users learn from this file?
•
•
•
•
Disease incidence – condition categories from actual claims for hospital patients
Disease prevalence – prevalence of conditions by geography and demographics in hospital stays
Hospitalizations due to chronic conditions – prevalence of chronic diseases
Access to care
17
Initial File Type: Utilization Summary File
• Shows the types of settings in which Minnesotans receive care
• Care settings include inpatient hospital stays, outpatient visits, ambulatory
surgical center visits, physician office visits, and Emergency Department visits
• What can users learn from this file?
• Utilization – variation in service utilization by geography, demographics, and place of
service
• Access to care
• In potential future files, health system cost trends – total cost of care by service
setting
18
Initial File Type: Service Line Summary File
• Information on the frequency and distribution of health care services provided to
Minnesotans
• One file for medical claims
• Would show:
• Service Code (CPT/HCPCS/ICD)
• Counts of number of patients
• File permits customized queries
• Most likely user is a researcher with reasonably extensive experience with data
query tools and deep knowledge of drug and procedure codes, code groupings,
and terminology
• What can users learn from this file?
• Prevention – rates of screening and prevention services delivered to Minnesotans
• Access to care
19
Questions
• Do these sets of files represent a good starting point for the development of
PUFs, while staying within guard rails?
• Are there other “middle ground” PUF topics that should be in the pipeline?
• Are there any other population or service characteristics that would enrich
these PUFs?
20
PUF SUPPORT AND LOGISTICS
21
PUF Data Documentation
• Data documentation for the 2016 PUFs will include:
•
•
•
•
Explanation of data included in the APCD
Challenges to summarizing APCD data
Disclaimers on what is and is not included in the PUFs
Overview of how the PUFs were created
• Data source
• File size and data aggregation
• Protection against re-identification
• Name and brief description of the data elements included in each summary file
• Summary of the PUF data access process
Discussion
• What is the expected audience for the data documentation? How can we best frame the
documentation for this audience?
• Are the topics listed above consistent with users’ expectations? What else might users need
to know about the data?
22
How can users access the PUFs?
• The PUFs are considered public data according to MN state law.
• Simple data request process to access the PUFs
• Users will request access via email
• Users will be provided simple training materials that explain how the data should be
best used
• Depending on file size, users will receive the PUFs via CD or other removable media in
compressed format
• Having users’ name and contact information will allow MDH to help provide
technical assistance and gather feedback from users
• User feedback will inform the direction of future PUFs
• The first set of PUFs will be available on March 1, 2016 with a refresh in late
summer, if possible
23
PUF User Engagement
• User engagement strategy for the PUFs is based on similar efforts in other states
• Pre-release: Data User Webinar (late February)
•
•
•
•
Introduce the contents, format, and potential uses of the initial PUFs
Clarify limitations of the PUFs
Review the data documentation
Explain the data access process
• Post-release: Web-based survey for users
• Online survey via Survey Monkey
• Designed to gather user feedback on their experiences with requesting, obtaining, and using the files
• Allows users to submit questions and/or recommendations
• Post-release: Monthly Q&A document
• Will answer questions received from users via the survey tool or sent directly to MDH
• Published and updated monthly from March-June 2016
Discussion
• Would users be comfortable with receiving outreach for their feedback?
• At what point should users be solicited for feedback?
24
Summary: Anticipated Next Steps
• February 2016: Finalize framework for initial PUFs with 2013 data
• March 2016: Release initial PUFs
• Pre-release user webinar
• Post-release user survey and Q&A for feedback
• April 2016 and onwards: Develop future strategies to expand PUF
development and meet June 2019 directive
• Incorporate guidance and insight received
• Summer 2016: Release more detailed PUF with additional years of data
• Fall 2016: Update March 2016 files
Discussion
• Com m ents or questions?
25
PUBLIC COMMENT
26
NEXT STEPS AND ADJOURN
Please send any additional comments or feedback to
[email protected] by Monday, February 15, 2016
Workgroup members are invited to join the PUF Data User webinar in late
February (more details to be provided).
27