A Typology of Efficiency in Health Care Implications for Measurement

A Typology of Efficiency in Health Care: Implications for Measurement Paul G. Shekelle, M.D., Ph.D. December 4, 2006 Project Overview • AHRQ-funded project began in October 2005 • Three major tasks: – Create a typology of efficiency – Scan and review literature on efficiency – Evaluate existing measures of efficiency • Final report due February 2007 HusseyEfficiencyAMA-2 10/06/06 Overview of Talk • Highlight motivation for current work • Present RAND’s typology • Review existing measures • Examples HusseyEfficiencyAMA-3 10/06/06 Measuring Efficiency • Tremendous pressure exists from various stakeholders to measure “efficiency” – Concern about rising health care costs – Variability in intensity of resource use not associated with better processes and outcomes • Little is known about how well available metrics capture the quantities of interest – Considerable lack of common language, conceptual clarity • Little is known about the consequences (intended and unintended) of applying those metrics at different levels in the system • How is efficiency established in an environment with mixed payment methods? HusseyEfficiencyAMA-4 10/06/06 Many Fortune 50 Companies Are Demanding Cost and Quality Metrics on Physicians High/Low High/High Efficient Low/Low Low/High Effective HusseyEfficiencyAMA-5 10/06/06 Overview of Talk • Highlight motivation for current work • Present RAND’s typology • Review existing measures • Examples HusseyEfficiencyAMA-6 10/06/06 Efficiency Measures Typology Overview Typology is organized in three tiers 1. Perspective Who is asking what about whom, and why? 2. Output What is being produced? 3. Type of Efficiency What is the root cause of inefficiency? What are the inputs to output? HusseyEfficiencyAMA-7 10/06/06 Perspective • We identify several potential points of view: – Health care “firms” • Providers • Health plans • Purchasers – Individuals – Society HusseyEfficiencyAMA-8 10/06/06 Context Matters in Efficiency Measurement Stakeholder Primary Fiduciary Responsibility Typical Time Period As long as responsible for patient During hospital stay Physician Active patients in a panel Patients who are admitted Enrollees Hospital Health Plan Renewed annually Employer Employees and covered dependents All residents Length of employment Society Unlimited HusseyEfficiencyAMA-9 10/06/06 Output: What’s Being Produced? • We are interested in evaluating the efficiency with which particular health care products (outputs) are “manufactured” • Being explicit about the output is critical (and often not done) • We define two major categories of outputs: – Services – Health outcomes • Producers (firms) define outputs – Financial flows (what is being sold) influence definitions HusseyEfficiencyAMA-10 10/06/06 Examples of Outputs by Type Health Care Firm (Producer) Service Output Examples Health Outcome Output Examples Physician •Visits •Procedures •Diagnoses •Prescriptions •Preventable deaths •Quality adjusted life years Hospital •Discharges •Procedures •Inpatient days •Covered lives •Functional status •Preventable deaths •Preventable complications Health Plan •Quality adjusted life years HusseyEfficiencyAMA-11 10/06/06 Types of Efficiency • Within the context of perspective and outputs, we identify three major types of efficiency: – Technical Applies primarily to firms – Productive – Social • Social efficiency is more often the focus for society than for firms HusseyEfficiencyAMA-12 10/06/06 Technical Efficiency A firm achieves technical efficiency when it cannot produce the same output with any fewer inputs HusseyEfficiencyAMA-13 10/06/06 Productive Efficiency A firm achieves productive efficiency when it cannot produce the same output at a lower cost HusseyEfficiencyAMA-14 10/06/06 Technical and Productive Efficiency Measures Point to Different Root Causes of Efficiency Technical Efficiency Inputs are put to good use Productive Efficiency Inputs are put to good use + Best mix of inputs chosen + Lowest prices are paid HusseyEfficiencyAMA-15 10/06/06 Example: Technical vs. Productive Efficiency • Technical Efficiency – Hospital A has a good CPOE system and staff are able to use it well – Hospital B has a CPOE system but it is difficult to use; staff follow old order entry process, but now with the extra step of computer entry Hospital A has higher technical efficiency than Hospital B HusseyEfficiencyAMA-16 10/06/06 Example: Technical vs. Productive Efficiency • Productive Efficiency – Hospital A bought a CPOE system, Hospital B did not; Hospital A now turns around orders more quickly – Hospital A and Hospital C both bought a CPOE system, but Hospital A got a better deal Hospital A has higher productive efficiency than Hospitals B and C HusseyEfficiencyAMA-17 10/06/06 Social Efficiency • Social efficiency is achieved when no member of society can be made better off without making another member worse off – Giving more resources to one person implies that those resources have been taken away from someone else – Appeal of “waste” is the notion that those resources do not benefit anyone currently HusseyEfficiencyAMA-18 10/06/06 What should be our third tier? • The concepts of technical v. productive efficiency resonates well with economists, but resonates less well with others • We are exploring an alternative third tier that looks at the types of inputs rather than the technical v. productive concept • In the alternative version, the approach to measuring the input used will affect the conclusions that can be drawn about how to improve efficiency • Inputs could be characterized as costs, resource counts, costs using standardized prices, etc. HusseyEfficiencyAMA-19 10/06/06 What about Quality? • What role should the quality of the output play in a measure of efficiency? • Some have proposed that any efficiency measure must include a measure of quality • We favor keeping efficiency separate from quality: – Inputs for certain health care processes share conceptual and measurement features – Metric to measure the quality of the output can vary greatly – the example of surgery – Common use of these terms in the US separate efficiency and quality HusseyEfficiencyAMA-20 10/06/06 RAND’s Efficiency Typology Society Perspective Health Care Firms Health Plans Providers Purchasers Individuals Output Services Health Outcomes Type Technical Productive Social HusseyEfficiencyAMA-21 10/06/06 Overview of Talk • Highlight motivation for current work • Present RAND’s typology • Review existing measures • Examples HusseyEfficiencyAMA-22 10/06/06 Hospital Efficiency Dominates Peer-Reviewed Literature Hospital Physician Health plan Nurse Medicare Area Other Focus/unit 0 20 40 60 80 100 Number of articles HusseyEfficiencyAMA-23 10/06/06 Econometric Analyses Dominate Measures in Peer-Reviewed Literature Type of article 0 50 100 Number of articles 150 200 Review/meta-analysis Descriptive Method development Econometric analysis HusseyEfficiencyAMA-24 10/06/06 Typical Measure from Peer-Reviewed Literature • Cit = f (Yit , Pit , β) + ui + vit • C is total costs • Y is outputs – Hospital discharges and outpatient visits • P is inputs – Capital costs and wages • Estimated using stochastic frontier analysis HusseyEfficiencyAMA-25 10/06/06 Different Worlds of Efficiency Measures • There is an almost total separation between the published studies of health care efficiency and the use of efficiency measures by providers, payers, and purchasers • Measures in use generally developed by vendors HusseyEfficiencyAMA-26 10/06/06 Vendor-Developed Measures • Episode-based: ETGs, MEGs, CRGs – Claims grouped into episodes and attributed to physicians – Measure is cost per episode (productive efficiency) – Also can look at resource use per episode (technical efficiency) HusseyEfficiencyAMA-27 10/06/06 Vendor-Developed Measures • Population-based: ACGs, DxCGs – Patient populations weighted by morbidity burden – Measure is cost per risk-adjusted patient per year (productive efficiency) – Also can look at resource use (technical efficiency) HusseyEfficiencyAMA-28 10/06/06 Conclusions from Review of Measures • Total disconnect between efficiency measurement by academics and vendors – Less consensus efficiency measures than quality measures • Little analysis of scientific soundness of measures • Almost all measures use services as outputs HusseyEfficiencyAMA-29 10/06/06 Overview of Talk • Highlight motivation for current work • Present RAND’s typology • Review existing measures • Examples HusseyEfficiencyAMA-30 10/06/06 Efficiency of Lasik Surgeons MD1 Input (per procedure) RN Labor RN Cost Anesthesia Anesthesia cost Total input cost MD2 MD3 1 hour $40/hour 40cc $0.10/cc $44 45 minutes $40/hour 40cc $0.10/cc $34 45 minutes $40/hour 40cc $0.05/cc $32 Output (total) Lasik Procedures Visual Functioning Patient Experience 8/day +10 points 89 8/day +10 points 89 8/day +10 points 89 MD2 and MD3 more technically efficiency than MD1 MD3 also more productively efficient than MD1 and MD2 HusseyEfficiencyAMA-31 10/06/06 Conclusion • Disconnect between academic world and vendors on efficiency measurement • Not the same level of consensus as seen on quality measures – Limited understanding of economics by noneconomists – Lack of research on scientific soundness of measures – Lack of actionable measures HusseyEfficiencyAMA-32 10/06/06 Some Challenges Ahead • Important to be explicit about the perspective, output, and type of efficiency – Not currently done systematically • Important to develop measures to fill gaps – Account for quality and outcomes of care – Social efficiency • Important to evaluate efficiency measures for scientific soundness, usability, etc. • We need agreement on the role the quality of the output should play in a measure of efficiency HusseyEfficiencyAMA-33 10/06/06