Title: Two decades of casemix - great ideas crippled by tunnel vision
1Two decades of casemix - great ideas crippled by
tunnel vision
- Professor Kathy Eagar
- Centre for Health Service Development
- University of Wollongong
- 2008 Casemix Conference, Adelaide, 16-19 November
2008
2Overview
- Personal reflections on 20 years involvement in
casemix development in Australia - Where we started
- What was promised
- What was achieved
- Where we are in 2008
- Where we need to be
3What I have learned in the last 30 years
- No matter what I look at, I find significant
variation between hospitals and between community
health and care providers - costs
- quality
- health outcomes
- Most of which cant be explained
- were different, our patients are sicker,
our clients have more needs
4TYPES OF VARIATION
- 1 Variation due to differences in the ways that
health services treat patients - 2 Variation due to differences in the kinds of
patients treated
5Casemix and Variation
- We must be able to control for one type of cause
of variation in order to understand the other - Casemix classifications have the potential to
help to control for variations between patients - By controlling for variations between patients
we produce information which can potentially help
to understand the differences between providers - costs, quality, outcomes
6A definition of casemix
- The mix of cases
- The classification of patient episodes based on
those patient attributes that best explain the
cost of care (cost drivers) - A generic term
- The Diagnosis Related Group classification is one
casemix system - But there are lots of others
- Casemix is not a synonym for DRGs
7Known cost drivers in healthcare (1992)
- acute inpatients
- rehabilitation
- psychiatry
- palliative care
- emergency
- neonatology
- diagnosis, age, procedure (as a proxy for
diagnosis) - functional impairment, ability to manage
Activities of Daily Living (ADLs) - ADL function, symptom severity, social and
economic circumstances, aggression - pain, symptoms, carer support, ADL function
- urgency, symptoms
- birth weight
8An observation
- Those factors that drive costs are also measures
of need for health care as well as predictors
of health outcomes - acute inpatients - diagnosis, age, procedure
- rehabilitation - functional impairment, ability
to manage Activities of Daily Living (ADLs) - mental health - ADL function, symptom severity,
social and economic circumstances - palliative care - pain, symptoms, carer support,
ADL function - emergency - urgency
- neonatology - birth weight
9Implication
- If developed well, casemix classifications can be
used to measure need for health care (at both the
individual and population level) and measure both
service quality and patient health outcomes - casemix-adjusted health outcomes
- The possibilities are endless!
10A potted history of casemix in Australia
11A bit of history
- 1983 - US Prospective Payment System based on
DRGs - 1987-1993 First Australian casemix developments
by the Commonwealth - George Palmer, Don Hindle, Richard Scotton,
Stephen Duckett and others - 1988 - first Australian Casemix Conference
12The original thinking behind using DRGs
13More history
- 1988 - Casemix Development Program (CDP) included
in the 1988-1993 Medicare Agreement (5m) - The CDP became the Hospital Information and
Performance Information Program in the 2003
Australian Health Care Agreements - Casemix Development Program Goal
- development of a hospital payment system which
takes into account casemix - adjusted output - Target date for new system - 30 June 1993
14Casemix Development Program Priorities
- Patient abstracting and coding
- Patient costing
- Quality assurance
- Education
- Information systems
- Payment system design
15Casemix systems
- 1992 - First Australian DRG system (AN-DRG) based
on input from the Australian Clinical Casemix
Committee - 1998 - AR-DRGs introduced, with progressive
versions since
16Payment system design
- 1990 Scotton and Owens Case payment in
Australian hospitals issues and options - preferred option - Commonwealth fund hospitals
directly at the marginal cost of the DRG, with
states and territories funding the remaining
costs
17Patient abstracting and coding (1992)
- The DRG classification system should be used for
the classification of acute patient episodes of
care. It should not be used for the
classification of all patients who are treated in
an "acute" hospital. - Recommendation
- That there be three acuity levels (care types)
termed acute, sub-acute and non-acute
18Patient abstracting and coding
- That the AN-DRG classification be used to
classify acute episodes of care. - That the DRG classification system not be applied
to sub-acute and non-acute episodes of care. - That psychiatric episodes of care, regardless of
acuity, be excluded from the DRG classification
system
19Problems with DRG-centred models
- DRGs don't work for many case types
- mental health
- rehabilitation
- chronic illness
- intensive care
- DRGs not sufficiently refined
- multiple conditions
- principal diagnosis
203 Care Types
- Diagnosis-related care (acute)
- Function-related care (sub-acute) and
- Supportive care (non-acute, including residential
care and community substitutes) - Mental Health
- Closer to function-related than diagnosis-related
care
21Other casemix classifications would be required
- National acute care ambulatory classification
system/s - emergency departments
- outpatients
- community health
- National sub-acute classification system
- National supportive care (non-acute)
classification system - National mental health classification system
22An episode of care
- A phase of treatment defined by either
- the setting (inpatient outpatient, community)
- the goal (oncology palliative care etc)
- May be treatment for a whole Episode of Illness
or just one part of it - With the capacity to measure the outcomes of each
episode of care - And link episodes of care to form episodes of
illness (patient journeys) and measures
outcomes at that level
23(No Transcript)
24(No Transcript)
25(No Transcript)
26(No Transcript)
27(No Transcript)
28(No Transcript)
29Some underpinning ideas
- Casemix classifications for each episode of care
- The information technology to bundle multiple
episodes of care into an episode of illness - The capacity to develop smart funding models
based on an explicit understanding of who would
carry what risks - A recognition of new models of care, especially
more definitive care in the community
30Casemix matrix - horizontal and vertical
approaches
31Casemix development studies
- Urgency, Disposition and Age Groups (UDAGs)
- Australian National Sub-Acute and Non-Acute
Patient classification (AN-SNAP) - Mental Health Classification and Service Costs
study (MH-CASC) - Smaller or state specific studies such as
- Home Care, Community Health, Outpatients
- Internationally - disease staging, disease
clustering
32Great ideas
- Casemix ideas in the 1990s offered so much
promise. Not only paying for healthcare but also - Measuring population need
- Measuring outputs
- A meaningful way of measuring health outcomes and
- Assessing value for money
33Some real progress
34In 1992 there was no agreement about
- the definition of a hospital
- the definition of the boundary between
non-inpatient and inpatient care - how to calculate hospital length of stay
- the classification of boarders
- the classification of newborns
- the recording of activity occurring in Emergency
Departments - the definition of a Principal Diagnosis
35The AR-DRG system
- Recognised internationally as one of the best
- A basic information building block in various
payment and funding systems
36Use of other classifications byStates and
Territories
- Use of the Australian National Sub-Acute and
Non-Acute Patient (AN-SNAP) classification in
some states - Version 2 released in 2007, 10 years after
Version 1 - Use of outpatient clinic and emergency department
classifications in some states
37Some lost opportunities
38The way some people still think
39Theres more to health care than acute inpatient
care!
- But, 20 years on, the only classification that is
recognised and supported at a national level is
the DRG system - With virtually nothing known at the national
level about other forms of care, their costs or
their outcomes
40DoHA website 2008
- The term Casemix refers to the type or mix of
patients treated by a hospital or unit. Casemix
based funding is one of the key funding models
currently used in Australian health care services
for reimbursement of the cost of patient care.
Casemix was developed for in-patient use, but
some states and territories have developed
Casemix funding models for outpatient and
ambulatory care settings.
41Casemix classifications on the DoHA website 2008
- Australian Refined-DRGs
- Historical information about Australian
National-DRGs
42The use of DRGs
- DRGs have gradually been re-defined as a tool for
payment and funding - used to inform the funding of hospitals, not
clinical units within hospitals - with few exceptions, in funding models that lack
the incentives to systematically drive change at
the clinical level - Little systematic use for measuring (improving)
health outcomes or quality - with little clinical involvement or interest in
them any more
43Outpatients and community health
- Most development (by States) has focused on
classifying occasions of service based on the mix
of services (clinic names) rather than the mix of
cases - Little or no progress in defining episodes of
care beyond inpatient. Instead, the level of
classification is the attendance/occasion of
service. So - cant measure outcomes
- cant get the risk sharing right
44Little evidence of smart purchasing
452008 and beyond
46Multiple national reviews and strategies under
development
- To redesign the Australian health system
- National Health and Hospitals Reform Commission
- National primary care strategy
- National prevention taskforce
- National performance indicators
- If casemix is such a great idea, why isnt it
central to national reform thinking?
47National report cards
- The Commonwealth has been clear that it wants to
move to national public reporting of hospital
performance - Resistance from both states and territories and
clinical associations is based on concern that
the results will be misleading - If casemix is the way of adjusting for variations
in the mix of patients, why is everyone so
worried? - And, if the existing DRG system isnt good enough
for the job, what have we done wrong over the
last 20 years?
48New models of care
- Acute inpatient classification has been
progressively refined over 20 years (at a high
cost), but... - New models of care have been developing elsewhere
- in the community
- And we have almost no systems in place to
classify these new models or to routinely assess
the impact or value for money of substitutable
models - Developing these systems is 20 years overdue
49Principal Diagnosis
- The concept of a Principal Diagnosis is
increasing problematic - Changing patterns of morbidity
- more people with multiple chronic diseases and
- Changing models of care
- less people admitted to hospital with a clear
principal diagnosis - Time to revisit ideas such as disease staging and
disease clusters?
50A better future
- Casemix ideas are terrific its time that we
start using them to help with the real
challenges - A realisation that services of equal cost are not
of equal value (and that services of equal value
are not of equal cost) - Shift the concerns
- from cost to value for money,
- from outputs to outcomes.
51What measure of success?
- Well know the answer to 3 questions
- 1 Are we doing the right thing?
- 2 Are we doing it right?
- 3 Are we getting the right result?