towards a value driven system
TRANSCRIPT
September 2009
Towards a Value Driven SystemMeasurement as a first step
2WisconsinSeptember 2009
-
Overview
• Goals of health care• Definitions –quality, cost,
efficiency, value • What can we measure now? • What COULD we measure in full
EHR environment in organized systems (?with right incentives)
• Moving forward incrementally
3WisconsinSeptember 2009
-
If you can’t determine (measure) where you are, where you have
been, and where you want to go- it is difficult to get there
4WisconsinSeptember 2009
-
What should a “value based” Health Care Do?
What should a “value based” Health Care Do?
A value-based health care system
20% of peoplegenerate
80% of costs
A: Keep or move population to healthy
or low
Healthy/Low Risk
At-Risk
HighRisk
Severe Disease
Current Health Care Spending
Active Disease
Source: HealthPartners
Ideal Spending
5WisconsinSeptember 2009
-
The Critical Aims of Health Care IOM
• Safe• Timely• Effective• Efficient• Equitable• Patient Centered
6WisconsinSeptember 2009
-
Definitions-towards the practical
• Value in health care: – A SUBJECTIVE assessment of the net
benefits (benefits-risks) of health care in relationships to the costs relative to other goods and services •Depends on who is doing the “valuing”
(patient, consumer, purchaser, “society”)•Based on variable amount of objective
information
• Value= (benefits-risks)/costs– Difficult to measure benefits, risks and
costs directly (at least right now)
7WisconsinSeptember 2009
-
Definitions-towards the practical
• Efficiency: – The physical relation between
resources used and health outcome…when the maximum set of possible improvements is obtained from a set of resource inputs (resources-cost). Palmer & Torgerson, BMJ 1999
• Measureable Efficiency: – An objective assessment of quality
relative to the costs (resources used and/or price paid)
8WisconsinSeptember 2009
-
Definitions
• Quality: – The degree to which health services for
individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge.
• Measureable quality: – reliable and valid assessment of the
structures, processes and outcomes related to provision of health care related to the aims of care that is safe, timely, effective, efficient, equitable and patient centered
9WisconsinSeptember 2009
-
Current to Ideal Quality & Cost Measurement
Useful Interventions
NOT Done
Under use
Useful Interventions
Done
Interventions of zero or negative
utility
Overuse
Ineffective
Interventions
Misuse
Cost of Misuse
Excess relative cost for useful care
Cost of Overuse
Cost of Avoidable
Consequences
COSTCOST
“Waste”
Total Use-Costs
QUALITYQUALITY
Episode based or Total Relative Resource Use or Costs for given patient (by disease, condition, episode or time) )
Measureable Outcomes- Clinical, Patient Experience Other
Structure
Process
Outcomes
Clinical-Pt
From high benefit to risk to low or no benefit to risk
Un Certain Utility
10WisconsinSeptember 2009
-
So where are we now?
• Quality measures– Underuse- ample number of measures– Overuse-misuse- a few useful measures
mostly for medications– Outcomes- best related to patient experience
(CAHPS), few clinical intermediate outcomes• Appropriateness (relative benefit-risk)
– virtually no measures• Cost-resources use
– Commercial (complex defined episode based-specific costs)
– NCQA RRU measures (time episode based total costs)
NO measures standardized For EMR use
11WisconsinSeptember 2009
-
So what to do now to “define” value?
• Use (where possible) NQF endorsed measures-preferably those used in PQRI and/or Hospital incentive program – Quality
• HEDIS-MD specified• Physician consortium for Performance
Improvement PCPI (selected)
– Cost-Resource use (NQF review late 2009)• HEDIS RRU (used at plan level 2007-09)• Episode of care –commercial, ABMS (2010)
12WisconsinSeptember 2009
-
Complex episode of care measures• Positives
– Widely used by health plans– Focused on specific procedure or claims
defined episode of care allows drill down • Negatives
– Nearly all products use some sophisticated (and non transparent) set of conventions to sort costs into discrete episodes and to risk adjust
– Need for large numbers to get valid results (400 plus episodes)
– Questionable use for public reporting given variation and difficulty in summarizing
– The more episodes created the lower the apparent cost
– Is not sensitive to the appropriateness of what was done
13WisconsinSeptember 2009
-
Population Resource Use Approach
• Positives– Publically available specifications with
public domain risk adjustment– Measures resource use of care for
groups of patients with index condition over time –sensitive to effects of in appropriate care
• Negatives– Does not look at disease or procedure
specific costs– Only limited drill down potential (best if
coupled with internal episode measures)
14WisconsinSeptember 2009
-
The Good NewsIt CAN be done
Quality and
Resource Use
HEDIS- 2007-08-09
15WisconsinSeptember 2009
-
NCQA Population RRU MeasuresNCQA Population RRU Measures• Total yearly relative
resource use over for people with– COPD (08)– Cardiac Conditions (08)– Hypertension,
uncomplicated (08)– Diabetes (07 & 08)– Asthma (07 & 08)
• Condition, episode delimited resource use for people with– Acute low back pain (07
& 08)
• Cost Service Categories– Inpatient Facility– Surgery & Procedure – Evaluation &
Management– Pharmacy, ambulatory
• Utilization Service Categories– Inpatient Discharges– ED Discharges– MRI (low back pain only)
16WisconsinSeptember 2009
-
Features of NCQA RRU MeasurementFeatures of NCQA RRU Measurement• Transparent-anyone can “look inside”• Uses same population as HEDIS Quality Measures in
same disease-so can be reported together• Simple risk adjustment
– Age & Gender– Disease severity (Type 1 v Type 2 diabetes)– Presence of co-morbidities (yes-no)– Exclusions of other dominant conditions
• Active cancer, HIV/AIDS, ESRD, etc.– Member cost capped if exceeds specified amount – Adjusted for enrollment and pharmacy benefit status
(medical and pharmacy member months)• Standardized Fee Schedule—weighted resource use
– Cost differentiates between a more intense and less intense service—e.g., amputation & office visit
17WisconsinSeptember 2009
-
Resource Use and Quality ResultsResource Use and Quality Results
Plan
Diabetes Quality
Composite
Diabetes Medical Components Resource Use
Pharmacy
Resource Use
Combined Medical
Inpatient
Facility Eval & Mngmt
Surgery & Procedure
s
Plan A 1.06 1.14 1.32 1.00 0.89 1.14
Plan B 1.10 0.85 0.96 0.74 0.73 1.12
Plan C 1.10 0.80 0.84 0.79 0.71 1.16
Plan D 1.14 0.74 0.77 0.85 0.56 1.13
Plan E 0.97 0.73 0.79 0.76 0.54 1.19
Sample Diabetes Relative Resource Use in a Single State – HEDIS 2008:
Note: Same as1.00= average, Less than 1.00 = below average Greater than 1.00 = above average
18WisconsinSeptember 2009
-
Resource Use and Quality Results
19WisconsinSeptember 2009
-
Resource Use and Quality ResultsResource Use and Quality Results
Note: Plan results are based on national relative indices for quality and RRU.
AB
CD
E
High Quality, Low Use High Quality, High Use
Low Quality, Low Use Low Quality, High Use0.70
0.80
0.90
1.00
1.10
1.20
1.30
Qua
lity
0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50Relative Resource Use
Source: HEDIS 2008, with national RRU.Exclusions: 1. Eligible population in plan < 400. 2. Outliers for Total Medical RRU components.
Commerical HMOs in One StateHEDIS 2008 Quality & RRU -- Diabetes
20WisconsinSeptember 2009
-
And now the bad newsResource Use Measures
• Even though transparent-resource use measures are complex
• Due to inherent wide variation of costs in patient groups, need very large sample sizes (>400) at least for public reporting
• Data incompleteness and lack of standardization is rampant– Bundling of services– Missing data (losing data in reporting
cost-resource use makes you look better)
21WisconsinSeptember 2009
-
More bad newsComparison with Quality
• No consistent relationship between resource use-cost and quality– Weak correlations
• Higher quality with higher pharmacy and outpt E&M• Lower quality with higher in-patient and surgery-
procedures
• Too few or too weak quality measures to link in most diseases (diabetes, CV, asthma, hypertension exceptions)
• Most KEY quality measures still require manual chart review (paper AND some EMR’s)-so nearly impossible to measure at individual MD level
22WisconsinSeptember 2009
-
But a bright futureSome promising developments
• Creation of standardization of measures in EMR and EHR environments– Limitations of EMR measurement
(physician versus patient focus)
• Measures building on EHR capabilities• Measurement of Appropriateness-
overuse (relative risk- benefit) measures-towards a real VALUE measurement
23WisconsinSeptember 2009
-
Standardization of measures in EHRs
• Stampede of activity related to ARRA and 35 plus Billion– Multiple standard setting organizations
racing to create protocols for specifying and embedding measures, extracting data and reporting in EMR-EHRs
– Definition of “meaningful use” for 2011, 2013, and 2015
– Likely to have profound effects on EMR’s, EHRs and measures (project to create standard specifications for EMR measures launched)
24WisconsinSeptember 2009
-
New Measures for EHRs
25WisconsinSeptember 2009
-
Advances using E-data
• Measures linked to guidelines and clinical decision support (concurrent measurement)– Adherence to guidelines-exceptions and rapid
cycle learning– Overall CV risk reduction (verses single
measures)– Appropriateness- relative risk benefit (more to
come)
• Measures of outcomes of clinical care– Reduction of CV risk (Archimedes)-real time and
in follow-up– Functional status over time– Intensification of treatment and exceptions
• Coordination of care
26WisconsinSeptember 2009
-
Advances in other realms
• Patient experience of care– More sophisticated surveys (medical
home related)– More sophisticated data collection
•Email-webased-time of encounter
– Incorporation into quality improvement- payment driven
27WisconsinSeptember 2009
-
Major GAPAbility to Measure Relative Benefit and Risk Directly
AHRQ-NQF-NCQA-PCPI Conference
28WisconsinSeptember 2009
-
Relative Benefit Risk (?and cost) • There are virtually NO measures of
quality related to many of the big ticket items driving cost
• Areas identified by NPP as critical– Surgical procedures (by pass surgery)– Other procedures (endoscopy)– Diagnostic imaging (PET, MRI, CT scans)– Screening (excess pap smears,
mammograms etc)
29WisconsinSeptember 2009
-
Why this gap?
• For these interventions, it is difficult to judge quality without dealing with the relative benefit and risks of applying procedure to a given set of patients – Example of problem
•Excellent “quality” score -but patients had a low probability of getting any benefit from the procedure
•Fair quality score – but all patients really had high probability of benefiting from the procedure
30WisconsinSeptember 2009
-
Barriers
•Technical-scientific– Weakness of evidence and analysis of
evidence (including efficacy and effectiveness)
– Paucity of consensus, evidenced based guidelines
– Need for clinically rich data to determine appropriateness
– Range in expert consensus based on specialty and weighting of evidence
31WisconsinSeptember 2009
-
Barriers
• Political-cultural – Lack of funding for development (evidence,
analysis or application)– Overuse encouraged by fee for service
reimbursement (one persons overuse is another’s income)
– Belief that doing something is better than nothing
– Public belief in technical solutions to health problems
– Legal concerns (populations versus individual outcomes)
– Strong push within organized medicine to use predicted clinical effectiveness ONLY for decision support or feedback
32WisconsinSeptember 2009
-
Essential
Relative Risk Benefit 10 8 6 4 2 0
SmallBenefit To Risk
Desirable Optional Harmful
Characterizing Relative Benefit and Risk
Net harm
Highbenef
it to risk
Appropriate Uncertain In appropriate
Overuse
33WisconsinSeptember 2009
-
Low Hanging Fruit
• Overuse: Application of an otherwise effective intervention to patients for whom the benefits are small or negative relative to the risks AND/OR costs
• “Currently definable overuse”: overuse in groups of patients that can be defined using currently available data (claims, age, gender etc)
34WisconsinSeptember 2009
-
Priorities for Overuse• Interventions with high variation and high
cost (NPP defined areas) – Imaging, Procedures Screening
• Look where existing criteria or guidelines have been developed indicating lack of benefit– US Preventive Services Task Force-”D”
recommendations– American College of Cardiology-American College of
Radiology
• Look for specific areas where “overuse” could be defined by existing claims or demographic data (duplication of services, hospital readmissions, ambulatory sensitive readmissions)
35WisconsinSeptember 2009
-
Currently definable overuse projects(initial focus of PCPI and NCQA work) • Imaging
– Cardiac– Sinusitis
• Procedures – GI (endoscopy) – Angioplasty, Angiography, By-Pass Surgery– Induction of labor– Low back pain (surgery, injection)
• Screening (over screening)• Other related measures under development
(possibly waste-overuse)– Readmissions – Ambulatory sensitive or avoidable admissions
36WisconsinSeptember 2009
-
BUT- reducing overuse using existing data
is only a first step
Next: Begin research and collaboration
with others to create framework to address full range of benefit risk
measurement
37WisconsinSeptember 2009
-
Uses of relative benefit- risk “scores”
• Use in Clinical Decision Support– Real time “academic detailing” to reduce low
benefit use (MGH)– Peer review of clinicians who do a high risk,
low benefit procedure – Comparison to actual outcomes (learning)
• Use as Performance Measures– Overall performance based feedback to
individual clinicians (7.5 versus 4.5) – Review by group leaders or by Boards (MOC) – Substitute for utilization review– Linked to payment
38WisconsinSeptember 2009
-
Work to date on Relative Risk Benefit
• RAND studies (Appropriateness)• American College of Cardiology
– Have developed Criteria for CV PET scanning, Percutaneous Coronary Interventions, others
• American College of Radiology– “Relative Clinical Utility” criteria for >30
imaging procedures– Mass General has developed computer based
ordering decision support (Radiology Order Entry)
• Very little beyond these –and above were done either for research or within the professional organizations for QI only purposes
39WisconsinSeptember 2009
-
Conclusion: Critical area to pursue
BUTDifficult and expensive to create
measures
Requires full electronic health record system
Likely to be controversial
40WisconsinSeptember 2009
-
Path Beyond • Need for further technical –
methodological work– How best to create criteria and measures of
relative risk benefit ? – How to build interface with
• Effectiveness research • Computer modeling of effectiveness• Clinical decision support (patient and provider)
• Focus would be from start on development of benefit-risk measures for use in EHRs since detailed clinical data in electronic formats needed to – To identify eligible populations – To calculate benefit/risk scores– To link to concurrent clinical decision support
41WisconsinSeptember 2009
-
Sooo- in Conclusion• We are making progress-but a lot
slower than we would like to do in creating evidence for VALUE determinations
• Limited at present by data, data collection, payment system and a lot more
• Some help is on the way-but uncertain how much or when
• So-ON WISCONSIN (except when playing Penn State)