Excess volume and moderate quality of inpatient care following DRG implementation in Germany Reinhard Busse, Prof. Dr. med. MPH FFPH Dept. Health Care Management, Technische Universität Berlin, Germany (WHO Collaborating Centre for Health Systems Research and Management) & European Observatory on Health Systems and Policies 1
Of course, since its inception in 23, Germany s DRG system (G-DRGs) has been constantly improved and is rightly considered a world leader 2
annual revisions have improved the cost-predictive value of the country-wide weights and make up-coding harder - Early years: Major revisions to increase precision Later years: development has stabilized Year 23 24 25 26 28 21 212 214 DRGs total 664 824 878 954 1137 12 1193 1196 Base-DRGs 411 471 614 578 64 69 595 588 236 454 353 318 293 29 287 4 5 7 8 9 9 9 9 Inpatient DRGs total 664 824 878 952 1132 1195 1189 1191 - valuated 642 86 845 912 189 1154 1149 1148 - unvaluated 22 18 33 4 43 41 4 43 Day care DRGs total 2 5 5 5 5 - valuated 1 1 1 1 2 - unvaluated 1 4 4 4 3 R2 all cases.4556.5577.6388.685.729.7443.754.7671 R2 inlier.6211.722.7796.7884.8166.843.844.8533 Unsplit Severity levels 3
using more classification criteria than in many other countries ENG EST GER 4
while base rates are used to accomodate regional differences Relative cost weight Patient characteristics Gender, Age, Diagnoses, Severity Base rate X = Hospital individual until 29; Uniform state-wide from 21 Treatment options Procedures, Technologies, G-DRG payment Intensity Payment example: Normal birth without cc in Berlin in 21 Relative cost weight.541 X Base rate Berlin 2927.5 = Payment 1584 5
and LOS adjustments and supplementary fees individualize payment to avoid skimping/ creaming and to incentivize innovations Relative cost weight + Supplementary fees X Base rate + LOS adjustment Year For services not (yet) included in DRG cost weights G-DRG payment = 23 24 25 26 28 21 212 214.1229.71.1148.27.1257.63.1265.7.1168.97.1373.76.1465.34.1464.14 Supplementary fees 26 71 83 115 143 15 159 - valuated 1 35 41 64 81 82 95 - unvaluated 25 36 42 51 62 64 64 Range of cost weights: min.-max. (rounded) 6
resulting in costs per case which are moderate in international comparison US$ PPP Bypass 4 35 3 25 2 15 1 5 SLV GER ISR KOR POR SWE AUS CAN FRA Source: Koechlin, F., L. Lorenzoni and P. Schreyer (21), Comparing Price Levels of Hospital Services Across Countries: Results of Pilot Study, OECD Health Working, Papers, No. 53, OECD Publishing. http://dx.doi.org/1.1787/5km91p4f3rzw-en FIN NOR USA
resulting in costs per case which are moderate in international comparison US$ PPP Hip Implantation 2 18 16 14 12 1 8 6 4 2 SLV ISR GER KOR FIN POR FRA SWE CAN Source: Koechlin, F., L. Lorenzoni and P. Schreyer (21), Comparing Price Levels of Hospital Services Across Countries: Results of Pilot Study, OECD Health Working, Papers, No. 53, OECD Publishing. http://dx.doi.org/1.1787/5km91p4f3rzw-en AUS USA
which, next to the DRGs per se, is also attributed to the hospital budgets - Sickness funds negotiate activity based DRG budgets every year with every hospital listed in the regional hospital plans Casemix X Base rate - + Surcharges = Hospital budget Budget over-run adjustment (hospital pays back): - - + Supplementary fees 65 % (standard DRGs), 25 % (drugs, medical, polytrauma and burns DRGs), Negotiations for certain DRGs (those that are difficult to predict) Budget under-run adjustment (hospital receives compensation) : - 2% (standard DRGs) 9
BUT overall hospital expenditure has increased noticeably 1,45 2.9% of GDP +41% 1,4 1,35 +32% Increase in cases 1,3 +27% 1,25 GDP Costs/ hospital case Hospital expenditure 1,2 1,15 1,1 1,5 1 22 23 24 25 26 27 28 29 21 211 212 213 2.6% 1
3 Acute care hospital discharges per 1 25 +12% (+1.7%/year for DRG-reimbursed cases) 2-4% 15-16% -29% 1 5 2 Austria Belgium Denmark France Germany Italy Netherlands Norway Spain Sweden Switzerland United Kingdom EU members before May 24 25 21 11
7 Acute care hospital beds per 1 6-12% 5 4-17% 3 2 2-31% -19% 25 21 Austria Belgium Denmark France Germany Italy Netherlands Norway Spain Sweden Switzerland United Kingdom EU members before May 24
Growth in number of cases (27-212) was very different on regional level Average annual change of case numbers 13
Is differing demand responsible for this? Supply Prices Change of case numbers in hospitals Morbidity change leads to change in number of cases Morbidity Proximity to death: More admissions to hospital in order to safe lives Mortality Change of Population: More citizens more cases Population Remaining impact of age Age (Proxy for other determinants ) Getting neutralized, e.g. changed preferences or lifestyles of population Not measurable, not age-related determinants Capacities 14
Demand-side changes (morbidity, mortality, population size) only partially explain increasing case numbers Relative der numbers Fallzahlen RelativeVeränderung change of case Veränderung Mortalität Change due todurch mortality Veränderung Bevölkerungszahl Change due todurch population Fallzahlveränderung insgesamt Change of case numbers Veränderung Morbidität Change due todurch morbidity Einflüsse zusammen Total demand factors 15% 1% 5% % -5% -1% -15% These are the persons born at the end of WW II 5 1 15 2 25 3 35 4 45 5 55 6 65 7 75 8 85 9 Age groups 15
Regional hot spots ( net, i.e. adjusted for all demand-side factors) Hamburg Average annual change of case numbers Ingolstadt 16
Summary demand side Variation of the population age structure seems to have the biggest effect on the demand side resulting in a slight increase of numbers. Age is after controlling for all other demand factors associated with case numbers but not linearly increasing (maybe due to agerelated treatment options) and distribution of age-effect between supply and demand remains unclear. Regional distribution shows that, after controlling for other demand factors, increasing case numbers are clustered within regions Relatively and absolutely strong growing MDCs have a below average influence on demand. 17
Relative change of case numbers Looking at the supply-side: incentives work changes of cost weights correlate with case numbers Size of effect: +.2% more cases per +1% increase in DRG weight (i.e. +1.9%/ year instead +1.7%) doubling of case increase for DRG weight increase of +8% Relative change of cost weights 18
Relative change of case numbers and hospitals which decrease LOS most increase case numbers most (or vice versa?) Relative change of ALOS 19
Summary - supply side A 1 % change of DRG-weights leads to an attributable.2 % change of case numbers hospitals respond on price changes which is a usual economic reaction A statement on the medical appropriateness of increasing numbers is not possible (yet) Infrastructural overcapacities might have a huge influence on increasing case numbers Effect of other revenue components (e.g. supplementary fees) can t be estimated First evidence on supply behaviour of hospitals in the DRG-era in Germany, more research necessary 2
Implications and considerations for service steering Constant and representative group for calculation of cost weights More diagnosis (instead of procedure) orientation for DRGs Mandatory and interdisciplinary medical second opinion for selected diagnoses More weight on reimbursement components for availability of e.g. emergency capacities Reform of hospital planning and investment financing mechanisms Redefinition of budget overrun regulation More collaboration between ambulatory and hospital sector overcome borders More evaluation and scientific competition better data 21
What about quality? Cost and survival of AMI patients in 1 European hospitals 1.85.9.95 cost 1.5 1.1 C+ C -.1 Q -.5 survival FINLAND GERMANY SWEDEN.5 Q+.1 FRANCE SPAIN Based on Häkkinen et al. 22 Health Policy 214
Staffing makes a big difference (shown here for AMI patients admitted in Berlin 24-27 within/ outside core hours and with/without physician-escorted EMS) + 6% + 4% + 9% Maier et al., Clin Res Cardiol 21 23
1.5 1.1 1.15 Cost and survival of stroke patients in 94 European hospitals.85.9.95 1 cost C+ C -.1 Q -.5 survival FINLAND GERNANY SWEDEN 24.5 Q +.1 FRANCE SPAIN Based on Häkkinen et al. Health Policy 214
no stroke unit (734 hospitals) stroke unit before 27 (26 hospitals) 25 Nimptsch & Mansky, Int J Stroke 213 The wide variation in quality can be explained by the availability of stroke units
Implications and considerations for quality improvement Concentrate services (e.g. AMI is treated in 14 hospitals, with 8 of them having fewer of one patient per week) Better measurement of indication quality More specification of structural quality (e.g. stroke is treated in 13 hospitals, of which only 35% have a stroke unit) Broader publication of quality data Quality agreements with hospitals during budget negotiation or hospital planning Pay-for-performance 26
Linking quality to DRGs our proposals for Germany Patient-based Indication Quality Structural Quality Process Quality Outcome No payment if no indication DRG/ disease-based (all cases Hospital-based (all with same DRG/ diagnosis) cases within hospital) Deduction per DRG and share on case without indication; no payment if minimum quantity for specific treatments is not reached Unverified procedure codes are neglected by grouping algorithm Certain DRGs are not billable if specific structures are not in place (e.g. Stroke Unit) Hospital acquired infections are neglected by grouping algorithm Best practice DRG-weights if costs for better quality are proven to be higher Payment rules for unplanned Surcharge for significant above readmissions average quality Reporting of No payment if quality data is Deductions if quality data is not not available available for numerous cases Quality Budget deduction if structural deviation from hospital plan Base rate deduction if quality data is wrong or incomplete 27
G-DRG-based hospital payment: Conclusion Strengths Weaknesses Transparency and improved documentation No system to reward/penalize hospitals for quality Fair (uniform) reimbursement Minimal (only state-based) adjustment for different input prices Precision of DRG system Increasing complexity with number of DRGs Precision of cost weight calculation Uniform accounting system but no full sample of hospitals Transparent methodology of developing and updating the system Weak instruments to manage hospital volumes 28
www.eurodrg.eu www.mig.tu-berlin.de 8 June 215 DRGs in Europe - towards transparency, efficiency, and quality? 29