minnesota ltss service access study: findings from years 1 and 2
DESCRIPTION
Minnesota LTSS Service Access Study: Findings from Years 1 and 2. Jessica Kasten and Rebecca Woodward August 14 th - 15 th 2014. Minnesota is National Leader in Publicly-funded LTSS. 1. - PowerPoint PPT PresentationTRANSCRIPT
© Truven Health Analytics Inc. All Rights Reserved. 1
Jessica Kasten and Rebecca WoodwardAugust 14th- 15th 2014
Minnesota LTSS Service Access Study: Findings from Years 1 and 2
© Truven Health Analytics Inc. All Rights Reserved. 2
Minnesota is National Leader in Publicly-funded LTSS
Ranked 1st in AARP Scorecard in overall performance across multiple dimensions, both in the 2011 and 2014 editions
Increased shares of people receiving LTSS in the community
Older Adults People with Disabilities
0%
20%
40%
60%
80%
100%
2000 2012
Ranked 3rd on Medicaid per-person spending specific to HCBS (2012)
1
© Truven Health Analytics Inc. All Rights Reserved. 3
Purpose of Study
Authorized by 2011 Legislature
If there were impacts of rate changes, how much?
What other factors were relevant to access?
How do findings support development of access measures for a DHS Dashboard?
The main purpose was to ascertain the extent to which provider rate changes affected recipients’ ability to access LTSS.
© Truven Health Analytics Inc. All Rights Reserved. 4
Study Period Timeline for Provider Rate Change Effects
© Truven Health Analytics Inc. All Rights Reserved. 5
3
1
2
Three Phases of Study
Close collaboration w
ith DH
S
Background and Selection of
Access Measures
Exploratory analysis of trends and encounter
data
Multivariate analysis
© Truven Health Analytics Inc. All Rights Reserved. 6© Truven Health Analytics Inc. All Rights Reserved. 6
Phase 1
Background on Service Access and Selection of Measures
© Truven Health Analytics Inc. All Rights Reserved. 7
Background on Service Access and Selection of Measures, 2012
Literature Review related to measurement of access in health care and LTSS
Review of how access to LTSS is assured in managed care
o Interviewed MN MCO key informants
Proposed several measure domains to explore in the quantitative analysis:
1. Comparison of services used to services authorized, with a significant discrepancy indicating an access constraint
2. Service utilization
3. Provider availability
Measures described in discussion of multivariate analysis
Truven Health gathered and synthesized background materials to inform the selection of LTSS service access metrics.
© Truven Health Analytics Inc. All Rights Reserved. 8
MCOs’ Perspectives on Access
Semi-structured telephone interviews (December 2012 - February 2013) using protocol approved by DHS
MCOs use numerous methods to assure access to LTSS
MCOs use several sources to assess their enrollees’ access to LTSS
MCOs generally did not think the rate changes affected access
Some did not think providers could sustain further cuts
Some thought the increase in PCA requirements adversely affected provider availability
© Truven Health Analytics Inc. All Rights Reserved. 9© Truven Health Analytics Inc. All Rights Reserved. 9
Phase 2
Exploratory Analysis of Trends and Encounter Data
© Truven Health Analytics Inc. All Rights Reserved. 10
Service Use Trends
Selected services based on multiple criteria (e.g. policy interest, adequate data, offered by multiple programs, etc.)
oPersonal Care Assistance (PCA)
oPrivate Duty Nursing (PDN)
oSkilled Nurse Visit (SNV)
oHomemaker
oConsumer Directed Community Supports (CDCS)
Examined number of recipients using the service and amount of service used over the study period
Examined by delivery system (FFS and managed care) Average number of people using the service increased both in FFS and
managed care for PCA, homemaker, and CDCS Trends not consistent between FFS and managed care for PDN or SNV
© Truven Health Analytics Inc. All Rights Reserved. 11
Encounter Data Review
Reviewed encounter claims for the 5 services included in the trends analysis
Reviewed most relevant claims fields with particular focus on units of service
Most important finding for multivariate analysis was the significant number of outliers in units of service for some services in some years
oAddressed by trimming the outliers to reasonable amounts based on DHS billing guidelines
© Truven Health Analytics Inc. All Rights Reserved. 12© Truven Health Analytics Inc. All Rights Reserved. 12
Phase 3
Multivariate Analysis
© Truven Health Analytics Inc. All Rights Reserved. 13
Multivariate Analysis Overview
Statistical study of 2 or more variables of interest at the same time
Include factors such as geographic area, age of recipient, level of likely LTSS need, etc.
Main focus was rate effects (FYs 2008-12)
Explored same set of services from Phase 2, except for CDCS
o CDCS presented methodological challenges
Included large number of State data sources
Added Rural Urban Commuting Area (RUCA) classification of geographic areas
What has been the impact of rate changes, relative to other potential correlates, on access to LTSS in Minnesota?
© Truven Health Analytics Inc. All Rights Reserved. 14
Multivariate Outcome Variables
Multivariate Model
Access Measure as Dependent Variable
Access Measure Description
Service Authorized Amount vs. Used
Measure 1 (FFS Only) % difference between authorized and used amounts of LTSS Service X, with access constraint defined as a discrepancy of >15%
Utilization Measure 2 Out of those eligible, use or non-use of LTSS Service X within a given yearly quarter
Measure 3 Out of service users, amount (units) of LTSS Service X used within a given yearly quarter
Provider Availability
Measures 4a and 4b Number of enrolled LTSS providers per county (4a) and participating LTSS providers per county (4b)
Measures 5 Ratio of unique recipients to unique participating LTSS providers
© Truven Health Analytics Inc. All Rights Reserved. 15
Explanatory Variables
Zip code characteristics (e.g. RUCA)
Provider rate changes Recipient
characteristics that vary over
time (e.g. age)
Recipient characteristics
that do not vary over time
(e.g. gender, race)
© Truven Health Analytics Inc. All Rights Reserved. 16
Measure 1 Results: Discrepancy Between Authorized and Used Amounts of Service
© Truven Health Analytics Inc. All Rights Reserved. 17
Measure 2 Results: Use vs. Non-Use of Service
© Truven Health Analytics Inc. All Rights Reserved. 18
Measure 3 Results: Amount of Service Used
© Truven Health Analytics Inc. All Rights Reserved. 19
Measure 4a Results: Enrolled Provider Counts
© Truven Health Analytics Inc. All Rights Reserved. 20
Measure 4b Results: Participating Provider Counts
© Truven Health Analytics Inc. All Rights Reserved. 21
Measure 5 Results: Ratio of Unique Recipients to Unique Participating Providers
© Truven Health Analytics Inc. All Rights Reserved. 22
Multivariate Summary
Designed and analyzed access measures tailored to available data and DHS’ interests
Novel approach with few, if any, precedents
Most of the measures showed some rate change effects with Measure 3 (amounts of service used) showing the largest effects
Provider availability measures showed the least rate change effects
PCA appears to be the service, of the four examined, most greatly affected by the rate changes
Other factors such as age, level of LTSS need, and geographic area had much larger influence than the rate changes on access in Measures 1 and 2, but comparable or smaller-sized effects in Measure 3
Enrollment in managed care often has a larger effect on access measures as compared to the effects of other factors
© Truven Health Analytics Inc. All Rights Reserved. 23
Study Limitations
Main focus and charge were to determine whether there were rate change effects
Not able to explore whether other statistical approaches might explain the access measures better (i.e. better “fit” to data)
With no available control group, an observational study like this shows associations, not causation
Difficult to control for policy or programmatic changes (e.g. PCA reform)
Likely other factors we have neither identified nor controlled for Presence of an informal caregiver Level of LTSS need for people without assessments
© Truven Health Analytics Inc. All Rights Reserved. 24
Next Steps
Development of technical appendix
Consider which measures best lend themselves to Dashboard metrics and what the most useful “drill-down” variables should be
Age group Geographic location (RUCA, county, other) Program (waiver, home care)
Develop Dashboard and test measures
© Truven Health Analytics Inc. All Rights Reserved. 25
COMMENTS AND QUESTIONS
© Truven Health Analytics Inc. All Rights Reserved. 26
More than Data. Answers.
Jessica [email protected] 547-4379
Rebecca [email protected] 254-5353