1. Efficiently Measuring Efficiency: Is Judgment the Correct Path
Howard Beckman, MD
Medical Director
RIPA
Rochester, NY

2. Cost Efficiency: Competition and Judgment
Urgent need for Cost Efficiency
Current in vogue models involve public reporting, tiering, limiting panels
Inherent in these models are competition and judgment

3. Cost Efficiency: Competition and Judgment
The core measurement for cost efficiency is the efficiency index – a comparison of one practitioner’s case-mix adjusted costs to peers
Based on responsible or total costs
Can be age and sex adjusted
Cost variables such as facility or pharmacy costs can be flattened to focus attention on what can be changed

4. Efficiency Indexes – Pros and Cons
+ Based on comparisons with peers or benchmarked group
+ Comparisons encourage a response
+ Case mix adjusted through episode grouping
software
– Judgmental - evaluates doctor, not behaviors
– Reductionistic – assumes generally + or -
– Limited actionability – costly to get to action
- Not adequately severity adjusted

5. Internal Medicine and Family Practice Number of Measures A Doctor is 25% Above or 25% Below Peers In Specialty +5 +4 +3 +2 +1 -1 -2 -3 -4 -5
ripa reported Feb 2006

6. Methodological Problems with Efficiency Indexes
Practitioners are generally efficient at some things but not others (82% in the middle)
Few distinctly better (11%) and worse (8%) overall physicians
Focusing on the practitioner creates defensiveness, humiliation and the creation of committed enemies

7. Methodological Problems with Efficiency Indexes
Setting targets incents selecting most easily treated patients/discharging recalcitrant ones
Hitting target may involve incremental treatment that causes more harm than good
Those who do the best were doing the best BEFORE incentives put in place

8. Conclusions
For cost efficiency, the physician is not the most effective unit of analysis
Efficiency indexes are too indirect and personal to be actionable
Fear as the motivational tool does not promote collaboration
Focusing on appropriately selected behaviors is a more logical strategy

9. The Next Generation: Adding Overuse and Underuse Measures to the Quality Paradigm

14. What Is Needed
By condition, find the local-regional variation in specific services
Understand if the variation represents overuse or misuse – have the quality conversation
Find overuse-misuse reduction opportunities for a whole specialty = find best practices
Create a series of measures based on reducing overuse or underuse offering interventions based on best practices
Reduce costs while improving quality – not by chance, but by DESIGN
The challenge was to find a general, systematic way of finding specific overuse/misuse issues from our data. In underuse there is one way to do things right (e.g. in diabetes). For overuse, there are lots of different ways to spend money. So we realized we had to find what was happening locally. We expect this might vary from region to region, and we are undertaking a study with RAND to see to what extent that is true.

15. MPPT™ Analysis of Hypertension
MPPT™ Analysis of Hypertension* (ETG 0281, Benign HTN w/o comorbidity, among 260 internists)
We would like to present analyses of three conditions (based on Episode Treatment Groups®) for a hypothetical HMO with 500,000 members. The analyses are modified versions of actual results obtained by FMA.
In the first analysis, hypertension care among internists, each quintile of physicians consists over 50 internists in a hypothetical network and would represent over 2000 episodes of care. An entire quintile’s data is aggregated to discern the differences in practice pattern between lower and higher cost physicians.
After removing services that are special cases, the only significant cost drivers are lab test, office visits and pharmacy. The underlying tables enable drilldown to specific services.
Now imagine we took an average line for each of these three cost centers and found out how much each quintile was above or below that line. (Out actual process is more complex and works at the individual service level.)
Episode Treatment Groups® and ETG® are trademarks of Symmetry Health Data Systems, an Ingenix company.
* Provided by FMA Hypothetical Costs for Illustration Only

16. Cost Variation – All in Pharmacy
Comparing each cost to the average, service by service, gives the variation in practice patterns.
Note that for this group of internists and this condition, lab tests and office visits make almost no difference at all.
If everyone prescribed like Quintile 1, the savings opportunity would be over $2,000,000 per year for our hypothetical 500,000 member plan.
Opportunity: Over $2,000,000 per year
Provided by FMA Hypothetical Costs for Illustration Only

17. Pharmacy Analysis: Best Practice is Quintile 1
All five quintiles have about the same total utilization. Quintile 1 uses the recommended medications (thiazides, ACE-I, and beta-blockers) at the highest rate. ACE-Is are better proven than ARBs and tolerated by > 85% of patients, but ARB use increases by quintile. CCB/ACE-I combinations are much more costly than a CCB plus an ACE-I. Two simple rules – use ACE-Is and ARBs in a ratio of 85% to 15%, and use CCB separately from ACE-I – would save over $900,000 per year in this hypothetical population.

18. Removing Benign Skin Growths Office visits and procedures drive costs
Only office visits and procedures make a difference. The highest group had more pathology costs.
Opportunity: $1.5 Million for
a 500,000 member HMO
Provided by FMA Hypothetical Costs for Illustration Only

19. Drilling Down on Procedures
Drilling down on procedures, we see the highest group uses more excisions $130 - $150) while lowest uses lower cost procedures $50-80). In the fourth group there is a question about over coding certain auxiliary procedures (closure in layers). There would be about 50,000 episodes of benign skin lesion removal per year in this system. Estimated savings opportunity: $1.5 million.
Provided by FMA Hypothetical Costs for Illustration Only

20. Creating a Blueprint for Change Provided by Focused Medical Analytics

21. Conclusions
Focus on reducing overuse instead of relying on efficiency indexes
Find specific action items to improve value
Direct attention to meaningful action items to engage practitioners as partners
Change physician behavior through incentives, avoid “punishing bad docs” as primary motivational strategy