Towards Chronic Care - Department of Health

Download Report

Transcript Towards Chronic Care - Department of Health

Options for Chronic Care
Under Activity Based Funding
IHPA Activity Based Funding Conference 13-16 May 2013
Steve Gillett
John Pilla
KPMG National Health Care Group
Purpose of Paper
ABF has largely focussed around products as defined by specific service events
(separations, occasions of service, attendances etc). This approach works reasonably well
for acute care, but less well for other types of care.
This is largely an artefact of history. It reflects how information was collected 50 years
ago when Fetter started developing DRGs
Preferred treatment models have changed with shorter episodes and greater use of
outpatient and out of hospital care. Increased emphasis on the patient journey and care
coordination, especially for chronic care. Emphasis on keeping people healthly and out of
hospital.
Steven Duckett talked about funding the episode and being more focussed on patient
outcomes. How can we fund care in ways that encourage this?
In this paper we start to explore funding models other than the traditional “casemix”
models and consider how they might coexist with the the existing IHPA model. We will
focus on the longest “episodes of care”, those associated with chronic conditions
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
The Rise and Rise of Chronic Conditions
1. Factors driving increases in chronic care requirements:Population aging (more elderly with chronic conditions) 
reduced mortality
2.Expect increased age specific prevalence
3. Prevalence = Existing Cases + Incidence – Deaths + Cure
4. Reductions in the age specific mortality rates:• Little recovery because of the definition of “chronic”
• Incidence (new cases)  uncertain?
5. Greater demand for services
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Current incentives are wrong
Under ABF the incentive is to admit multiple times.
Hospital admitted care can be the most appropriate care
setting for “acute episodes” of many chronic conditions.
However it is better to keep people in good health and
minimise “acute episodes”.
Under ABF there is no budget or incentive to spend to
keep people living with chronic care health problems and
out of hospital?
- GP fund holder arrangements
- Budgets for hospitals  Explored in this talk
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Objectives
Overall aim is to appropriate care in an appropriate stetting:1. Allow for substitution for less intensive care
2. Allow preventative measure to prevent severity increasing
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Desirable Requirements of a Funding Model
• Process must be compatible with the National pricing
framework.
• Be revenue neutral compared to episodic funding (unless
there is a specific incentive to move to one funding
model). Why?
• no cost data
• Based on feasibly collected data.
• Creates the right incentives and difficult to game
• Be able to be consistently reproduced
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Approaches to Funding Models
Model 1
Bundling services for the treatment of a specific condition
over a predefined period.
Model 2
Paying for the full health needs for an individual over a
predefined period.
Model 3
Offering incentive payments for reduction in hospitalisation
for patients in the target population.
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Australian Examples
• GP antenatal care for a period after the delivery
has been roled into the price for the delivery.
• Renal Dialysis in some jurisdictions has a mixed
model with an annual payment and a episode
based payment.
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Comparison of Approaches
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Condition Hierarchy ( Model 2:example using DxCG)
In capitation funding:
1. the concept of a
Principle Diagnosis
has no meaning as
this can change with
different episodes.
2. Don’t pay twice for
similar conditions
 which is more
important
Could AR-DRG
repetitive exclusion
used in defining
complications be used
for this?
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
One approach to setting Weights under Model 1
1.
Calculate total NWAU for each patient episode
normally and sum across all the health encounters
over the period.
2. Calculate a revise NWAU score and sum for all
patients. Revise by:1.
2.
3.
Remove all secondary diagnoses relevant to the condition .
Regroup and recalculate NWAU
Zero weight cases with a relevant principle diagnosis that :1. Do not use ICU
2. Do not have surgery
3. Other criteria ?
Sum across all episodes during the period
3.
Subtract the Revised NWAU from the actual NWAU
for each patient and calculate the difference
4. Find the average difference across all patients in the
group
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Constructing Regression Data for setting weights under Model 2
Hospital Annual ICD Summary for Mrs Jones
Year Admission
1
1
2
3
2
1
2
3
3
1
2
3
4
5
4
1
2
3
4
ICD Diags.
A,B,D,Y
A,C,D
A,B,C,F
A,C
A,C,H
A,B,D
A
A,C
A,C,D
A,B,C,F
A,E
A,B,D
A,D
A,B,G
A,B,C,F
WIES
W1
W2
W3
X1
X2
X3
Y1
Y2
Y3
Y4
Y5
Z1
Z2
Z3
Z4
Annual HCCs for Mrs Jones
Year
1
2
3
4
DCGs
DCG1,DCG2,DCG3,DCG4
DCG1,DCG2,DCG3,DCG6
DCG1,DCG2,DCG3,DCG4
DCG1,DCG2,DCG3,DCG4,DCG5
Hospital admissions Mrs Jones
Year
1
2
3
4
ICD Diags.
A,B,C,D,F
A,B,C,D,H
A,B,C,D,E,F
A,B,C,D,G,F
WIES
W=W1+W2+W3
X=X1+X2+X3
Y=Y1+Y2+Y3+Y4+Y5
Z=Z1+Z2+Z3+Z4
DxCG Sofware
ICD
A
B
C
E
F
G
H
DCG
DCG1
DCG2
DCG3
DCG3
DCG4
DCG5
DCG6
HCC
HCC1
HCC1
HCC2
HCC2
HCC3
HCC3
HCC4
Concurrent
Prospective
2
2
1
1
4
4
5
4
3
2
2
1
1
0
Annual HCCs for Mrs Jones
WIES
W
X
Y
Z
Year
1
2
3
4
DCGs
HCC1,HCC2,HCC3
HCC1,HCC2,HCC4
HCC1,HCC2,HCC3
HCC1,HCC2,HCC3
WIES
W
X
Y
Z
Concurrent
6
7
6
6
Prospective
7
6
7
7
Mrs Jones Records for Weights Construction
NOTE: WIES
can be replaced
with NWAU
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Year
1
2
3
4
Dummy Variables with "Present" score
Age/Sex
Age/Sex
Age/Sex
Age/Sex
Group,
Group,
Group,
Group,
HCC1, HCC2, HCC3
HCC1,HCC2,HCC4
HCC1,HCC2,HCC3
HCC1,HCC2,HCC3
Dependant Variable
Concurrent Prospective
Model
Model
W
X
X
Y
Y
Z
Z
n.a.
Regression Model using HCCs for Weights under Model 2
WIES = ∑βi × AgeSexi + ∑γi × HCCi
Estimate
βi s and γi s - these become the weights
where
AgeSexi are dummy variables (0,1) for 18 age
sex cohorts
HCCi are dummy variables (0,1) for 184 HCCs
Note: requires large amounts of data
can give odd results like negative amounts
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Regression Model using HCCs
WIES = ∑βi × AgeSexi + ∑γi × HCCi
Estimate
βi s and γi s - these become the weights
where
AgeSexi are dummy variables (0,1) for 18 age
sex cohorts
HCCi are dummy variables (0,1) for 184 HCCs
Note: requires large amounts of data
can give odd results like negative amounts
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Example of running a version of Model 2 using regression derived weights
(Actual and predicted for chronic care patients who did not die during the period)
Health Service
Region 1
Region 2
Region 3
Region 4
Actual (12 Month*)
4.7
4.0
5.3
5.0
Concurrent
5.2
4.8
4.7
5.6
Use Ratio
0.92
0.83
1.13
0.89
Prospective
3.0
3.2
2.6
3.4
Close alignment with concurrent and Capitation
Lower prospective. Possibilities:1) Acute illness not picked up in age/sex adequately
2) Deaths not fully excluded (Only in hospital deaths)
3) Timing of 12 month period
4) Using trimmed data for weights (unlikely….similar when all results used)
5) Less resources actually required due to stablisation after and acute
period.
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Predictive Power of regression derived weights in the Model 2 example
(on a cohort of chronic care patients)
Concurrent Models (predict WIES in the year that the
diagnoses were reported)
1. HCCs only
2. HCCs and Age/Sex groups
Rsquare 60.3%
Rsquare 60.4%
Prospective Models (predict WIES for the year after the
diagnoses were reported)
1.
2.
3.
4.
HCCs only
HCCs and Age/Sex groups
Trimmed data HCCs only
Trimmed data HCCs + Age/Sex groups
Rsquare
Rsquare
Rsquare
Rsquare
16.2%
19.1%
59.4%
65.5%
Note: Higher RSquares expected on trimmed data (2.5% high and
low WIES useage).
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
A different approach:
John Hopkins Adjusted Care Groups (ACGs)
Not
Mutually
Exclusive
Diagnostic Codes (ICD-9-CM)
(n=14,000)
ADGs
(n=32)
1 ICD1 ADG
Age, Sex
Mutually
Exclusive
Adjusted Clinical Group (ACG)
(n=92)
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
1 Person1 ACG
Predictive Modelling using ACGs
Age
MultiMorbidity
Disease
Burden
(ACGs)
Gender
Risk Score
Selected
Medical
Conditions
(EDCs)
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Special Population
Markers
Pharmacy
Morbidity
(Rx-MGs)
Selected Prior
Use Measures
(optional)
Defining the Payment Model
Considerable debate about the parameters within the
funding model eg
1. Should it be based upon actual performance or based upon improved
performance or a combination of both.
2. Should it be based upon reaching a threshold or on a continuous scale or
a combination of both. If based upon a threshold how should the
thresholds be determined
3. Improved performance requires change. How can change be funded with
certainty. The CEO does not know if he will achieve target performance
until after the event.
4. Should indicative budgets be allocated at the start of the year assuming
each hospital receives a share for reaching targets. What happens to
“unallocated” money where targets are not reached.
Experience suggests that where hospitals do poorly
under the model they will lobby to change the rules
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
The HARP Experience in Victoria
Hospitals give a fixed amount of dollars for each enrolled
chronic care patient to “trial” new treatment modes
designed to keep people out of hospital. This was
additional to any ABF funding under existing
arrangements.
Analysis of enrolled patients showed:1. Patients were allocated to multiple programs at the same time (often in the same
hospital)
2. Patients were allocated to programs, subsequently discharged from the program and reenrolled some time later
3. Admissions through the emergency department went down but admissions from other
sources went up
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Developing a Safety Net For Chronic care
funding
Fixed episode
based
payment
Capitation
NWAU
End of year
Adjustment if
minimum
percentage is
not reached
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
Low
High
Percentage
Percentage
Casemix
Normal
ABFWIES
NWAU
Additional
episode
based NWAU
if high point is
exceeded *
* Simpler than an outlier model
Setting the percentage boundaries
1. Set the lower boundary by Policy (eg 25%)
2. Calculate the high boundary to be as close to revenue
neutral as possible
Capitation
NWAU
25%
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.
High
Percentage?
Normal ABF NWAU
Thank You
Contact:
email
[email protected]
phone (+61) 03 9288 6289
mobile (61+) 0407 72240
© 2010 KPMG, an Australian partnership and a member firm of the KPMG network of independent member firms
affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved.
The KPMG name, logo and "cutting through complexity" are registered trademarks or trademarks of KPMG International
Cooperative ("KPMG International"). Liability limited by a scheme approved under Professional Standards Legislation.