Ann Robinow - Defining Parameters/Cost Measurement: Condition Specific (PDF: 734KB/13 pages)

MN DEPARTMENT OF HEALTH
PROVIDER PEER GROUPING (PPG)
ADVISORY GROUP
DEFINING PARAMETERS: CONDITION SPECIFIC
COST MEASUREMENT: CONDITION SPECIFIC
ANN ROBINOW
MEETING 3: JULY 10, 2009
Introduction
Comments and changes to meeting summary?
Review of questions or comments since last meeting
Information on prevalence of conditions
Report Technical Advisory Group responses since
last meeting
Prevention Care status
Questions for Today’s Meeting:
Defining Parameters Section
How many and which specific medical conditions will be
measured?
MDH recommends Advisory Group identify 3 to 6 conditions
for first year implementation.
Quality data for all providers must be available for all
conditions selected. MN Community Measurement initiatives
& MN Quality Incentive Payment System identified
conditions will be most complete quality data sets available
in 2010.
Selected Conditions by Entity
Condition
Diabetes
Heart Failure
Low back pain
Coronary artery disease
Depression
Heart attack
Pneumonia
Asthma
Hypertension
Total knee
Adult preventive
Child preventive
Maternity
Prostate cancer
UTI
Hip fracture
Cholecystitis
COPD
MN QI
X
X
X
X
X
X
AQA
X
X
X
X
X
X
CMS
BTE
X
Baskets of Care
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Specific Medical Conditions: Survey
Results
Average Score (11 responses)
Lower = Higher Priority
Condition
Diabetes
Coronary artery disease
Maternity
Asthma
2.27
5.09
5.27
5.36
(1)
(2)
(3)
(4)
Pneumonia
Total Knee Replacement
5.82 (5-6)
5.82 (5-6)
Heart Failure
Heart attack
6.09 (7)
6.91 (8)
Depression
7.00 (9)
Low back pain
7.27 (10-11)
Preventive
Other: Annual Female
Screening
7.27 (10-11)
7.00 (one vote)
PPG Advisory Meeting Schedule
MEETING
DATE
TIME
Meeting 1
Thursday, June 11
Introduction/Background
Meeting 2
Friday, June 26
Defining Parameters
Meeting 3
Friday, July 10
Cost Measure for Conditions
Meeting 4
Friday, July 17
Quality Measures for Conditions
Cost Measure for Total Care
Meeting 5
Wednesday, July 22
Quality Measure for Total Care
Combining Cost & Quality
Meeting 6
Monday, July 27
Combining Cost & Quality
Meeting 7
Wednesday, September 2
Information Needs by Audience
Meeting 8
Friday, September 11
Revisit Outstanding Issues
Meeting 9
Wednesday, September 30
Final Review
Context For Today’s Discussion
High Level Steps in Peer Grouping
Questions for Today’s Meeting: Condition
Specific Cost Measurement
Cost & Risk Adjustment Questions:
1. What is cost?
2. What tool should be used to create episodes and
identify costs related to episodes?
3. How will costs be risk adjusted within episodes?
4. How should outliers be defined in condition specific peer
grouping and how will they be accounted for?
5. Should peer grouping be analyzed by payer type to risk
adjust for payer mix?
Questions for Today’s Meeting: Condition
Specific Cost Measurement
Provider Attribution Questions:
6. Who will be measured for condition specific?
7. How do we define the entity to be measured (unit of
analysis) for condition specific?
8. What is a “peer group” for condition specific
measures?
9. What recommendations/principles should be used to
determine patient attribution to providers?
10. How many providers should patients be attributed to
for condition specific peer grouping?
Condition Specific Cost Measurement
Technical Panel Recommendations
1.
2.
3.
4.
5.
6.
7.
Use commercial software; ETG most common.
Consider impact of other payment sources (e.g.: hospital dispro
share (DSH) or risk sharing); may be too challenging
Truncate or trim high end outliers; exclude low end outliers.
Risk adjustment for payer mix needs to be applied if not
adequately incorporated into grouper software.
Attribution to primary hospital under restricted conditions; concern
with ability to identify transfers.
Use RBRVS and DRG to measure resources
Feasible to combine all costs from all payers but separating out
price may be necessary due to payer mix and variation in price
within same payer.
How to Identify Related Costs to
Specific Conditions?
Definition of a Condition includes:
Clear beginning and ending points, can be time based
Which services included
Which services excluded
Most health plans use a software tool to define
Conditions: Episode Treatment Grouper (ETG) from
Ingenix Symmetry
Other programs: Medical Episode Group (MEG) from
Thomson Reuters, Marketbasket System from Cave Consulting
Group.
What Does Episode Software Do?
Selection of an
episode grouper
software tool will
address many of
today’s issues but
use of software
requires some
user input
decisions.
*From CMS Report on Evaluating Functionality of ETG & MEG, 8/2008
What is Cost?
Resource Use x Unit Price = Total Cost
Each component influences Total Cost so ability to compare components
separately is important.
MDH views its charge from the legislature as providing comparisons on variations
in actual cost, not just resource use.
Resource use measures inputs to provide service --- a measure of
utilization.
Different systems to measure types of services
Physician: Resource Based Relative Value Unit (RBRVS)
Hospital: DRG for inpatient, APC for outpatient
Ancillary & Pharmacy: no commonly accepted method or no method
HealthPartners has developed a system to reprice all services (HPRRVs)
Price is the unit price or contracted rate for service.
Cost Measurement Options
How should Cost Outliers be accounted for in
Condition Specific Peer Grouping?
Goal: Represent with the greatest confidence possible, a
provider’s consistent average cost per episode.
Outliers are removed in order to not overstate or
understate a provider’s average cost data due to a few
unique cases.
Outliers are unique cases that are either unusually more
expensive than the “norm” or unusually less expensive than
the “norm”.
Often, outliers are trauma cases or other extremely
complicated cases (e.g. organ transplant, premature
triplets) that require an intensive level of resources.
Outlier Adjustment Options for Condition
Specific Cost Measurement
Link Outliers to Population Size (can be applied to all three options)
Pros: Limits impact of statistical variations by defining based on standard deviations specific to population.
Cons: Need to establish multiple formulas for outlier adjustment.
Condition Peer Grouping Types of Risk
Adjustments for Cost : Severity of Illness
Differences in severity of illness & co-morbidity of patients, even
with the same condition, can require more resources and impact
cost.
Ingenix Symmetry Episode Treatment Grouper (ETG) is most
accepted and widely used by plans in MN. Technical Panel
recommends.
Software tools make adjustments for severity of illness by
grouping patients with same condition together.
ETGs consider age, complications, presence of surgery,
presence of co-morbidities in grouping some conditions.
A second level of risk adjustment can be added to further
adjust for population within the same condition. HealthPartners
applies second level of risk adjustment.
Condition Peer Grouping Types of Risk
Adjustments for Cost : Demographics
Differences in demographic characteristics such as age, gender,
race/culture of patients, even with the same condition can
influence level of resources needed and impact cost.
ETGs includes demographic characteristics.
Data limitations to demographic characteristics: no
information on race and ethnicity.
Adjustments for geographic/zip code residence are not
accounted for.
Condition Peer Grouping Types of Risk
Adjustments for Cost : Payer Mix
Differences in payer mix of providers treating patients with same
condition will have different contracted prices for same services and
impact average total actual cost.
Government payers generally pay less than commercial.
Provider groups have differing negotiation leverage with
commercial payers.
Geographic location of provider can influence payer mix.
Uncompensated care is not included in total cost.
Tech Panel: Grouper software often “calibrate” separately
for payer type. If not, need to risk adjust.
Payer Mix Adjustment Options
Specific Condition Measurement
Condition
Who to Measure
Unit of Analysis
Peer Group
Diabetes
Primary Care
Endocrinologist
Clinic site (when possible)
All measured
providers
Pneumonia
Hospital
Individual Hospital
All measured hospitals
Heart Failure
Primary Care
Cardiology
Clinic site (when possible)
All measured
providers
Heart Failure
Hospital
Individual Hospital
All measured hospitals
Total Knee
Replacement
Orthopedic
Clinic site (when possible)
All measured
providers
Total Knee
Replacement
Hospital
Individual Hospital
All measured hospitals
Coronary Artery
Primary Care
Cardiology
Clinic site (when possible)
All measured
providers
Asthma
Primary Care
Pediatrician
Pulmonologist
Allergist
Clinic site (when possible)
All measured
providers
How will a Patient’s Cost be Attributed for
Condition Peer Grouping?
Physician Attribution Options for Condition
Specific Cost Measurement
Summary Recommendations
Options
Issue
Episode Software
Recommendation
ETG
Cost
Actual
Reprice
Actual & Reprice
Payer Mix Adjustment
No adjustments
Compare by payer
Normalize payer mix
Outlier Adjustment
Remove
Truncate
Trim
(specific to population size)
Attribution to one or
many physician
providers?
Single
Multiple
Attribution to primary or
multiple hospitals?
Primary w/in time frame
All hospitals admitted
PPG Advisory Meeting Schedule
MEETING
DATE
TIME
Meeting 1
Thursday, June 11
Introduction/Background
Meeting 2
Friday, June 26
Defining Parameters
Meeting 3
Friday, July 10
Cost Measure for Conditions
Meeting 4
Friday, July 17
Quality Measures for Conditions
Cost Measure for Total Care
Meeting 5
Wednesday, July 22
Quality Measure for Total Care
Combining Cost & Quality
Meeting 6
Monday, July 27
Combining Cost & Quality
Meeting 7
Wednesday, September 2
Information Needs by Audience
Meeting 8
Friday, September 11
Revisit Outstanding Issues
Meeting 9
Wednesday, September 30
Final Review
Preview of Next Meeting :
Quality Measures Condition Specific
Cost Measurement for Total Care
What quality measures should be included in physician
peer grouping for Condition Specific?
What quality measures should be included in hospital
peer grouping for Condition Specific?
Should analysis of Condition Specific include a
composite quality measure?
What types of risk adjustment should be performed
for Condition Specific quality measures?