Hospital Alternative Reimbursement Models, and DRGs

Transcription

1 Hospital Alternative Reimbursement Models, and DRGs

2 Topics 1 Alternative Reimbursement Models Fixed Fee options 2 Diagnosis Related Groups and Case Mix Risks, Rationale and Incentives 3 Clinical Coding Technical Considerations

3 Objectives To provide a quick overview and history To go into some detail as to the inner workings To discuss some technical detail with respect to application in our environment Present both a Funder and Provider view

4 1 Alternative Reimbursement Models, and Fixed Fee Options

5 Alternative Reimbursement Models Why are ARMs such a big issue? FFS incentives the more you do, the more you get paid Third party payer incentives costs incurred one step removed from activities Cost plus business models incentives to invest in latest and greatest, and make return

6 ARMs Increasing Risk Transfer Capitation Global Fees Passes on admit rate risk, many different approaches, many complexities it is possible, but hasn t been done en masse in SA Brings other costs into the fixed fee, could be done in graduation, consider correlations of costs, and variable charging Fixed Fees Incentive to manage all hospital costs within an admission, incentives are to creep classifications Per Diems Incentive to manage intraday costs, incentive to increase LOS and possibly LOA Fee for Service Incentive to increase all aspects of utilisation, degree depends on the detailed fee and billing structures

7 Fixed Fee Options Bottom up approach Define a basket of costs for a given procedure Usually Clinically substantiated Could only do for small list of commodity procedures Not useful for comprehensive ARM Top down data approach Look for clusters in the data similar clinical categories and similar resource use Leads invariably to DRGs

8 Overview of DRGs Essentially a clustering of clinically similar admissions, organized into groups that are expected to have similar resource usage Inputs include diagnosis codes, procedure codes, birth weights, ages, gender, death Objectives of a good grouper include A manageable list of groups (hundreds, not thousands) Should make clinical sense Variables required should be commonly available Groups should contain admissions with similar expected resource use These objectives trade off against each other

9 Overview of DRGs Early development used LOS as resource measure Typical structure Admissions classified into a small number of Major Diagnosis Categories (say 25 or so) Then each of these is divided into smaller more specific groups (say in total) Groups are then further subdivided into narrower groups, usually based on complications or co-morbidity. There seems to be the most divergence in methods at this stage.

10 Example of DRG hierarchy and partitioning process Case mix definition by DRG; Fetter et al

11 Structure examples Base DRG (top 10) Total Caesarean Delivery 4.00% Complex Cataract Procedures, Sameday 3.25% Dental Extractions and Restorations 3.15% Other Major Affective Disorders 2.04% Colonoscopy 1.88% Vaginal Delivery 1.81% Antenatal and Other Obstetric Admission 1.70% Spinal Disorders 1.70% Colonoscopy W Major Diagnosis 1.63% Cystourethroscopy, Sameday 1.54% Complication % Admissions W/O CC 77.7% W CC 19.1% W MCC 3.2% DRG - example Avg Cost Kidney and Urinary Tract Infections (females) W/O CC Kidney and Urinary Tract Infections (females) W CC Kidney and Urinary Tract Infections (females) W MCC

12 History of DRGs in the USA In 1983 US Congress enacted a DRG based PPS for all Medicare patients First large scale in New Jersey in 1970 s 1987 New York State legislates DRG based PPS for all non Medicare patients leads to AP-DRGs Developed in Yale late 1960 s by Prof Fetter for HCFA Intention was to develop framework for monitoring utilisation of services in a hospital setting Mid 1980 s HCFA commissions Yale to refine grouper, leads to RDRGs Further refinements over time led to APR- DRGs, S-DRGs, with latest CMS version being MS-DRGs

13 History of DRGs in Europe * Methods to determine reimbursement rates for diagnosis related groups (DRG): A comparison of nine European countries, March 2006

14 History of DRGs in South Africa M developed and distributed the IR-DRG grouper to private sector stakeholders, based on the unique coding structures chosen 3M updated and revised the grouper until 2003 when they pulled out of our market No further development was done and no new groupers developed, until Discovery started modifying the IR grouper to allow for new codes introduced since 2003 as well as making improvements in other areas. This process stopped when discussions began regarding changing national coding schemas

15 Case Mix Using DRGs we can explicitly calculate differences in cost arising from differences in case mix Were there more heart bypasses relative to tonsillectomies this year than last year? Were there more complicated tonsillectomies relative to uncomplicated tonsillectomies this year than last year? Using the average cost per DRG in a chosen base year we can calculate a case mix index over time and between hospitals For example

16 Case Mix over time Base Year Admissions Average Cost Case Weight DRG A 2,365 13, DRG B , DRG C , DRG D 5,961 9, Overall 8,815 11, Case weights set so that case mix index equals 1 in base year

17 Case Mix over time Base Year Admissions Average Cost Case Weight DRG A 2,365 13, DRG B , DRG C , DRG D 5,961 9, Overall 8,815 11, Next Year Admissions Average Cost Case Weight DRG A 3,406 14, DRG B , DRG C , DRG D 6,438 10, Overall 10,438 12, Increase 12.0% 1.8%

18 Case Mix over time Base Year Admissions Average Cost Case Weight DRG A 2,365 13, DRG B , DRG C , DRG D 5,961 9, Overall 8,815 11, Same case weights used in next year Next Year Admissions Average Cost Case Weight DRG A 3,406 14, DRG B , DRG C , DRG D 6,438 10, Overall 10,438 12, Increase 12.0% 1.8%

19 Case Mix over time Base Year Admissions Average Cost Case Weight DRG A (27%) 2,365 13, DRG B (4%) , DRG C (1%) , DRG D (68%) 5,961 9, Overall 8,815 11, Next Year Admissions Average Cost Case Weight DRG A (33%) 3,406 14, DRG B (5%) , DRG C (1%) , DRG D (62%) 6,438 10, Overall 10,438 12, Increase 12.0% 1.8% Change in mix of admission types changes case mix index. 1.8% of 12% increase explained by case mix changes

20 Case Mix by hospital Hospital 1 Hospital 2 Overall Admits Average Cost Admits Average Cost Admits Average Cost Case Weight 2,365 13,598 4,973 12,911 7,338 13, , , , , , , ,961 9,805 3,201 9,953 9,162 9, ,815 11,386 8,715 12,311 17,530 11, Average Cost difference 8.13% Hospital 1 Case weight Hospital 2 Case weight Case weight difference 9.0% Residual Avg Cost difference -0.8% Hospital 2 has a higher average cost per admission

21 Case Mix by hospital Hospital 1 Hospital 2 Overall Admits Average Cost Admits Average Cost Admits Average Cost Case Weight 2,365 13,598 4,973 12,911 7,338 13, , , , , , , ,961 9,805 3,201 9,953 9,162 9, ,815 11,386 8,715 12,311 17,530 11, Average Cost difference 8.13% Hospital 1 Case weight Hospital 2 Case weight Case weight difference 9.0% Residual Avg Cost difference -0.8% But Hospital 2 also has a higher case mix

22 Case Mix by hospital Hospital 1 Hospital 2 Overall Admits Average Cost Admits Average Cost Admits Average Cost Case Weight 2,365 13,598 4,973 12,911 7,338 13, , , , , , , ,961 9,805 3,201 9,953 9,162 9, ,815 11,386 8,715 12,311 17,530 11, Average Cost difference 8.13% Hospital 1 Case weight Hospital 2 Case weight Case weight difference 9.0% Residual Avg Cost difference -0.8% On a case mix adjusted basis, Hospital 2 is therefore more efficient!

23 Applications cost efficiency measurement, attribution The hospital cost per admission Tariff Pharmacy Alternative reimbursement Price Utilisation Consumables Ethicals Ward Theatre Equipment Price Mix and utilisation Length of stay Mix and utilisation Level of acuity

24 2 Risks, Rationale and Incentives

25 Risk Transfer ARMs are often referred to as a Risk Transfer What s actually transferred is Risk and the Incentive to manage costs Risks (Incentives) transferred include: Random variability (in cost, not frequency of admissions) Environmental and business risks Selection and moral hazard risk, to some extent General budgeting and expense overrun risks New technology Regulatory risk Changes in clinical practice leading to increased LOS or LOA Changes in clinical practice leading to increased utilisation of non tariff items

26 Rationale and Incentives Agreement on a fixed fee for hospital admissions with appropriate risk adjustment means hospitals have an incentive to reduce all input costs related to admissions Well, sort of

27 Rationale and Incentives It depends on what share of the hospitals revenue falls under these arrangements, and to what extent they are of the same form It is unlikely that hospitals can expend resources to manage (i.e. maximise profit) a portion of their business as FFS and a portion as Fixed Fee While it may be possible to bill according to each funder public or private (e.g. by applying negotiated billing rules for example) managing completely different incentives is asking a lot

28 Rationale and Incentives It also depends on the nature and term of the contract(s) If the contract is short term and based on transparent FFS pricing, realizing all efficiencies in year one means next years negotiation will be tough (for the hospital) and the return on all effort expended will be yielded in year one only A longer term formula driven contract allows a more amicable spread of returns and still gives certainty to the funder Continuous pursuit of efficiencies will plateau at some point ARMs can also promote hospital specialisation, with increasing focus on those DRGs that are profitable

29 Rationale and Incentives For the Funder (Medical Scheme / State) : Gives certainty as to the amount of each hospital claim Very valuable for small funds (i.e. private sector schemes) as the risk of large claims is transferred to the hospital Negotiation then relates to a fee schedule (in theory) which makes matters simpler particularly if there is some industry consistency Aligns the financial incentives of funders and hospitals to manage costs downward Move away from FFS incentives to increases costs to model that incentivizes lower costs

30 Rationale and Incentives For the Hospital Incentives are aligned between funder and shareholders DRGs better define care provided as products instead of components Pursuits for efficiency yield profitability Savings non tariff costs drop to the bottom line Savings in tariff items yield additional profit, after allowing for overhead cost components and volume changes

31 Rationale and Incentives Fixed fees in other countries have been driven mostly by Government who set the price, and most often in environments where the Doctors are employed or allowed to be employed. In a voluntary ARM environment, the business case should prevail must be a win-win scenario

32 Rationale and Incentives What about quality incentives? There is a risk that utilisation is managed to the point of impinging on quality Which aspects of quality; There are many Headline quality measures are death and complication rates The issue extends to perceived quality by the patient, and the Doctor Commodity items used may not be quite so comparable (masks, gloves, etc) What about incentives for the Doctor? They are primary decision makers about resource use case by case And are not involved in the alignment of incentives

33 3 Clinical Coding and Technical Considerations

34 Clinical Coding This is the heart and soul of the DRG Main inputs are Diagnosis and Procedure codes, usually based on Discharge Codes usually agreed between Hospital and Administrator / MCO Coding is not perfect by any means, and there is an element of coding maturity over time Our combination of CPT and ICD came about in mid 1990 s from meetings between RAMS and HASA - brought about by the newly formed PHISC a unique hybrid of coding

35 Clinical Coding Contentious coding is dealt with through an interactive process 1. Identify DRG outliers Supported by random audits of coding, on both sides 4. Rerun DRGs 3. Funder and hospital agree on new coding 2. Investigate coding Takes time to complete data reconciliation and coding agreements

36 Clinical Coding Small changes in coding can have significant impacts (changes to coding systems/habits, REF coding, and chronic coding) Coding of chronic diseases has implications for co-morbidity measurement, but a lack of standards in chronic coding makes this difficult Some immeasurables and resultant implicit assumptions include BMI Family history Smoking status Income / socio economics

37 Applications - Reimbursement Actual practical method could be a fee schedule by DRG, or a case weights schedule to be applied to overall average cost per admission Case weights based on the average cost per admission in each DRG group, in the base period; could also use trimmed data Base periods are usually a given year Case weight shapes over time respond to trends in technology, utilisation changes, as well as regulatory and industry changes

38 Other Applications Applications other then reimbursement include Analysis of drivers of expenditure increases Identify what proportion of increased costs are due to changes in case mix Enables apples with apples comparisons between hospitals Manageable number of groups allows statistical balance between reliability and ability to interpret results in discussions between funders and providers Can be used as a resource planning tool Quality of care analysis and monitoring on a case mix adjusted basis

39 Technical issues include Data reconciliation, consistency of definitions, Statistical fit of the DRG grouper to the data The role and methods of trimming Coding standards Other risk factors (input cost differences, supply side issues, type of hospital) Benefit design / entitlement affects demand

40 Benefit Design and Managed Care Benefit design has an impact on the case weights and the measure of case mix changes Limits, Sublimits, Deductibles, and Co-payments all affect the cost per admission Consistency in benefits helps matters Over time changes in benefit design from one year to the next can be problematic Within the Funder s options member movements between options can be problematic Similar issues with the application of managed care any changes in the application of managed care that affect the cost per admission can make assignment of increases difficult

41 Other Risk Factors The addition of other risk factors could improve predictability But the usual balance between credible cell size exists Any additional factors should be considered carefully Geographical area where prices or costs differ by region modifications must be made (in the US hospitals fees are adjusted to take into account differences in nursing salaries