DRG Systeme in Europa - TU Berlin · different forms of hospital payment ... the NHIF top down microcosting x Finland 5 reference hospitals ... Change coding practice

2. Februar 2017 Krankenversicherung und Leistungsanbieter 1

Reinhard Busse, Prof. Dr. med. MPH FFPHFG Management im Gesundheitswesen, Technische Universität Berlin

(WHO Collaborating Centre for Health Systems Research and Management)&

European Observatory on Health Systems and Policies

DRG Systeme in Europa

Management im GesundheitswesenKrankenversicherung und Leistungsanbieter

Paymentmecha-

nism

Patient needs (risk

selection)

ActivityExpendi-

turecontrol

Technical efficiency

Trans-parency

QualityAdmini-strative

simplicityNumber of services/

case

Number of

cases

Fee-for-service + + + ― 0 0 0 ―

DRG based case payment

0 ― + 0 + + 0 ―

Global budget ― ― ― + 0 ― 0 +

Incentives ofdifferent forms of hospital payment

2Krankenversicherung und Leistungsanbieter2. Februar 2017

Paymentmecha-

nism

Patient needs (risk

selection)

ActivityExpendi-

turecontrol

Technical efficiency

Trans-parency

QualityAdmini-strative

simplicityNumber of services/

case

Number of

cases

Fee-for-service + + + ― 0 0 0 ―

DRG based case payment

0 ― + 0 + + 0 ―

Global budget ― ― ― + 0 ― 0 +

Incentives ofdifferent forms of hospital payment

European

countries 1990s/2000s

USA 1980s

“dumping” (avoidance), “creaming”

(selection) and “skimping” (undertreatment)

up/wrong-coding, gaming


Country Study Activity ALoS

US, 1983 US Congress - Office of

Technology Assessment, 1985

▼ ▼

Guterman et al., 1988 ▼ ▼

Davis and Rhodes, 1988 ▼ ▼

Kahn et al., 1990 ▼

Manton et al., 1993 ▼ ▼

Muller, 1993 ▼ ▼

Rosenberg and Browne, 2001 ▼ ▼

Empirical evidence (I): hospital activity and length-of-stay under DRGs

USA

1980s


Cf. Table 7.4 in Busse et al. 2011

European

countries

1990/ 2000s

Empirical evidence (II)

2. Februar 2017

Country Study Activity ALoS

Sweden,

early 1990s

Anell, 2005 ▲ ▼

Kastberg and Siverbo, 2007 ▲ ▼

Italy, 1995 Louis et al., 1999 ▼ ▼

Ettelt et al., 2006 ▲

Spain, 1996 Ellis/ Vidal-Fernández, 2007 ▲

Norway,

1997

Biørn et al., 2003 ▲

Kjerstad, 2003 ▲

Hagen et al., 2006 ▲

Magnussen et al., 2007 ▲

Austria, 1997 Theurl and Winner, 2007 ▼

Denmark, 2002 Street et al., 2007 ▲

Germany, 2003 Böcking et al., 2005 ▲ ▼

Schreyögg et al., 2005 ▼

Hensen et al., 2008 ▲ ▼

England,

2003/4

Farrar et al., 2007 ▲ ▼

Audit Commission, 2008 ▲ ▼

Farrar et al., 2009 ▲ ▼

France, 2004/5 Or, 2009 ▲ 5

Cf. Table 7.4

in Busse

et al. 2011

To get a common “currency” of hospital activity for

• transparency efficiency benchmarking &

performance measurement (protect/ improve quality),

• budget allocation (or division among providers),

• planning of capacities,

• payment ( efficiency)

Exact reasons, expectations and DRG usage differ

among countries – due to (de)centralisation, one

vs. multiple payers, public vs. mixed ownership.


Reasons for DRGs:

Timeline and purposes of introduction

7

Country

19

80

19

85

19

90

19

95

20

00

20

05

20

10

Original purpose Principal purpose in 2010

Austria Budgetary allocation Budgetary allocation, Planning

England Measuring hospital activity Payment

Estonia Payment Payment

FinlandMeasuring hospital activity, benchmarking

Planning, benchmarking, hospital billing

France Measuring hospital activity Payment

Germany Payment Payment

Ireland Budgetary allocation Budgetary allocation

Netherlands Payment Payment

Poland Payment Payment

Portugal Measuring hospital activity Budgetary allocation

Spain (Catalonia)

Payment Payment, benchmarking

Sweden PaymentMeasuring hospital activity, benchmarking

19

80

19

85

19

90

19

95

20

00

20

05

20

10

Introduction of DRGs

DRG-based hospital payment

Krankenversicherung und Leistungsanbieter2. Februar 2017

DRG-based case payments, DRG-based budget allocation

(possibly adjusted for outliers, quality etc.)

Excluded costs(e.g. for infrastructure; in U.S. also physician services)

Payments for non-patient care activities(e.g. teaching, research, emergency availability)

Payments for patients not classified into DRG system(e.g. outpatients, day cases, psychiatry, rehabilitation)

Other types of payments for DRG-classified patients(e.g. global budgets, fee-for-service)

Additional payments for specific activities for DRG-classified patients (e.g. expensive drugs, innovations),

possibly listed in DRG catalogues

For what types of activities? Scope of DRGs (I)

Krankenversicherung und Leistungsanbieter 82. Februar 2017

Psychiatry Day casesAcute

inpatient careOutpatient care Rehabilitation

Original DRG

systems

DRG system(included in orseparate fromoriginal DRGs)

DRG system(included in orseparate fromoriginalDRGs)

DRG system(identical ordifferent to

original DRGs)

DRG system(included in orseparate fromoriginal DRGs)

For what types of activities? Scope of DRGs (II)

Krankenversicherung und Leistungsanbieter 92. Februar 2017

Scope in the Netherlands: DBCs (diagnosis-treatment combinations); examples

Inpatient acute care incl. ICU

Ambulatory

specialist

care

Hospitalisation

Discharge

DBC 1

DBC 2

DBC 3

DBC 6

DBC 5

Ambulatory specialist

care

DBC 4


Patient classification system

Data collection

Price setting

Actual reimbursement

• Diagnoses• Procedures

• Severity• Frequency of revisions

• Demographic data

• Clinical data

• Cost data

• Sample size, regularity

• Cost weights

• Base rate(s)• Prices/ tariffs

• Average vs. “best”

• Volume limits

• Outliers• High cost cases

• Quality• Innovations

• Negotiations

Essential building blocks of DRG systems

Import 1

2

34


AP-DRG AR-DRG G-DRG GHM NordDRG HRG JGP LKF DBC

DRGs / DRG-like groups 679 665 1,200 2,297 794 1,389 518 979 ≈30,000

MDCs / Chapters 25 24 26 28 28 23 16 - -

Partitions 2 3 3 4 2 2* 2* 2* -

Basic characteristics of DRG-like PCS in Europe

Patient classification system

• Diagnoses

• Procedures

• Severity

• Frequency of revisions


Main questions relating to data collection

Clinical data

classification system for diagnoses and

classification system for procedures

Cost data

imported (not good but easy) or

collected within country (better but needs

standardised cost accounting)

Sample size

entire patient population or

a smaller sample

Many countries: clinical data = all patients;

cost data = hospital sample

with standardised cost accounting system

Data collection

• Demographic data• Clinical data

• Cost data



Diagnosis and procedure coding across Europe

Country Diagnosis Coding Procedure Coding

Austria ICD-10-AT Leistungskatalog

England ICD-10 OPCS - Office of Population Censuses and Surveys

Estonia ICD-10 NCSP - Nomesco Classification of Surgical Procedures

Finland ICD-10 NCSP - Nomesco Classification of Surgical Procedures

France ICD-10 CCAM - Classification Commune des Actes Médicaux

Germany ICD-10-GM OPS - Operationen- und Prozedurenschlüssel

Ireland ICD-10-AM ACHI - Australian Classification of Health Interventions

The Netherlands ICD-10 Elektronische DBC Typeringslijst

Poland ICD-10 ICD-9-CM

Portugal ICD-9-CM ICD-9-CM

Spain ICD-9-CM ICD-9-CM

Sweden ICD-10 NCSP - Nomesco Classification of Surgical Procedures

(almost)

standardisedno uniform standard available

14Krankenversicherung und Leistungsanbieter

Data collection


• Clinical data

• Cost data


2. Februar 2017

Number (share) of costdata collecting hospitals

Direct costallocation to patients

Data used for calculation of DRG weights

Austria20 reference hospitals (~8% of all hospitals)

grosscosting x

England all hospitals top down microcosting x

EstoniaAll hospitals contracted by

the NHIFtop down microcosting x

Finland5 reference hospitals

(~30% of specialised care)bottom up microcosting x

France99 hospitals (~ 13% of inpatient admissions)

mainly top down microcosting

x

Germany~250 hospitals

(~ 15% of all hospitals) mainly bottom up

microcostingx

Ireland

Imported DRG systems and weights (or with only minor modifications)Poland

Portugal

The Netherlandsunit costs: 15-25 hospitals

(~ 24% of all hospitals)bottom up microcosting x

Spain Imported DRG systems and weights

Sweden(~ 62% of inpatient

admissions)bottom up microcosting x

15

Collection of cost dataData collection


• Clinical data

• Cost data



“cost weight“ (varies by DRG)

“base rate“ or adjustment

Relative weight(e.g. Germany)

1.0€ 3000 (+/-)

(varies slightly by state)

Raw tariff(e.g. France)

€ 30001.0 (+/-)

(varies by region and hospital)

Raw tariff(e.g. England)

£ 30001.0 – 1.32

(varies by hospital)

Score (e.g. Austria) 130 points € 30

X

X

X

Price setting

• Cost weights

• Base rate(s)

• Prices/ tariffs


How to calculate costs and set prices fairly (I)


X

• Based on good quality data

(not possible if cost weights imported)

• “Cost weights x base rate”

vs. “Tariff + adjustment” vs. Scores (see below)

• Average costs vs. “best practice” (for few HRGs in England)

17

Country Monetary conversion/

adjustment factors

Applicability of conversion

rate / adjustment factors

Austria (Implicit) Point value Depending on state

England Market forces factor Hospital-specific

Estonia Base rate Nationwide

Finland Base rate Hospital-specific

France (1) Regional adjustment

(2) Transition coefficient (until

2012)

(1) Region-specific

(2) Hospital-specific

Germany Base rate State-wide

Ireland Base rates (1) Specific to one of four hospital

peer groups

(2) Hospital-specific

Netherlands Direct (no conversion) Not applicable

Poland Point value Nationwide

Portugal Base rate Hospital peer group

Spain (Catalonia) (1) Direct (no conversion)

(2) Base rate

(1) Not applicable

(2) Region-wide (CMS-DRGs)

Sweden Base rate County-specific

How to calculate costs and set prices fairly (II)

Price setting

• Cost weights

• Base rate(s)

• Prices/ tariffs



18

How many patient records did we use?

Busse R et al. BMJ 2013Krankenversicherung und Leistungsanbieter2. Februar 2017

19

How many DRGs exist per episode?


20

How wide is the price variation?


21

Size of bubble:

number of DRGs

Range:

DRG weights

(index case = 1)

14x

28x

2x

2.5x 2.5x

1x

1.1x 1.5x

60x

30x

1.5x

1.7x

5x

3.4x

16 10

14

10

10 10

Busse R et al. BMJ 2013

2. Februar 2017 Krankenversicherung und Leistungsanbieter

Krankenversicherung und Leistungsanbieter 22

To become even more specific, definition of AMI index case and 6 other vignettes

1 & 2:

not invasive

(1 = death)

3 - 6:

invasive

(3 & 5 = BMS,

4 & 6 = DES,

6 = death)

Quentin W et al. Eur Heart J 2013

2. Februar 2017

23

€ 2601 € 4533

€ 2189 € 1837

€ 2926 € 7933

AMI: relative DRG payments I (index case = 1)

16% receive stents

57% receive stents

Qu

enti

n W

et

al. E

ur

Hea

rt J

20

13


25

Stroke episode:index case & six case vignettes

Peltola M & Quentin W Cerebrovasc Dis 2013


26

Stroke episode: classification criteria

Candidates:

• Age

• Primary diagnosis (stroke or infarction vs. bleeding …)

• Secondary diagnoses/ comorbidities/ complications

• Procedures (e.g. systemic thromboloysis)

• Death/ time of death

• Length-of-stay

• Use of stroke unit


2. Februar 2017 27

Stroke episode: classification criteria

ENG

EST

GER

Krankenversicherung und Leistungsanbieter

28

Stroke episode: resulting reimbursement

Thrombolysis

& >7 daysStroke

unit

Day care

Secondary

diagnoses

Thrombo-

lysis

Secondary

diagnoses

Death

Peltola M & Quentin W Cerebrovasc Dis 2013



Incentives of DRG-based

hospital payment

Strategies of hospitals

1. Reduce costs per

patient

a) Reduce length of stay

optimize internal care pathways

inappropriate early discharge (‘bloody discharge’)

b) Reduce intensity of provided services

avoid delivering unnecessary services

withhold necessary services (‘skimping/undertreatment’)

c) Select patients

specialize in treating patients for which the hospital has a competitive

advantage

select low-cost patients within DRGs (‘cream-skimming’)

2. Increase revenue per

patient

a) Change coding practice

improve coding of diagnoses and procedures

fraudulent reclassification of patients, e.g. by adding inexistent

secondary diagnoses (‘up-coding’)

b) Change practice patterns

provide services that lead to reclassification of patients into higher

paying DRGs (‘gaming/overtreatment’)

3. Increase number of

patients

a) Change admission rules

reduce waiting list

admit patients for unnecessary services (‘supplier-induced demand’)

b) Improve reputation of hospital

improve quality of services

focus efforts exclusively on measurable areas

Being aware of incentives and hospital strategies

How European DRG systems reduce unintended behaviour: 1. long- and short-stay adjustments

LOS

Revenues

Deductions(per day)

Surcharges(per day)

Short-stay outliers

Long-stay outliers

Inliers

Lower LOSthreshold

Upper LOSthreshold


• Volume limits



• Negotiations


How European DRG systems reduce unintended behaviour: 2. Fee-for-service-type additional payments


• Volume limits



• Negotiations

England France Germany Nether-lands

Payments per hospital stay

One One One Several possible

Payments for specific high-cost services

Unbundled HRGs for e.g.:• Chemotherapy•Radiotherapy•Renal dialysis•Diagnostic imaging•High-cost drugs

Séances GHM for e.g.:• Chemotherapy•Radiotherapy•Renal dialysis

Additional payments:• ICU• Emergency care• High-cost drugs

Supplementary payments for e.g.:• Chemotherapy•Radiotherapy•Renal dialysis•Diagnostic imaging•High-cost drugs

No

Innovation-related add’lpayments

Yes Yes Yes Yes (for drugs)


Type of

adjustmentMechanism Examples

Hospital

based

DRG/

disease

based

Patient

based

Payment for an individual patient is

adjusted upwards or downwards by a

certain amount

No payment is made for a case

Certain readmissions within 30 days are not paid

separately but as part of the original admission (e.g., in

England and Germany)

Complications (that is, certain conditions that were not

present upon admission) cannot be used to classify

patients into DRGs that are weighted more heavily (e.g.,

in the United States)

How DRG systems reduce unintendedbehaviour: 3. adjustments for quality


Type of


Hospital

based

DRG/

disease

based

Payment for all patients with a certain

DRG (or a disease entity) is adjusted

upwards or downwards by a certain

percentage

DRG payment is not based on average

costs but is awarded to those hospitals

delivering ‘good quality’

Insurers negotiate with hospitals that DRG payment is

higher/lower if certain quality standards are met/not met

(e.g., in Germany and the Netherlands)

DRG payment for all hospitals is based on ‘best practice’;

that is, costs incurred by efficient, high-quality hospitals

(e.g., in England)

Patient

based



certain amount









Type of


Hospital

based

Payment for entire hospital activity is


certain percentage

Hospital receives an additional

payment unrelated to activity

Predefined quality results are met/not met (e.g., in

England)

Overall readmission rate is below/above average or

below/above agreed target (e.g., in the United States)

Hospitals install new quality improvement measures (e.g.,

in France)

DRG/

disease

based

Payment for all patients with a certain

DRG (or a disease entity) is adjusted

upwards or downwards by a certain

percentage

DRG payment is not based on average

costs but is awarded to those hospitals

delivering ‘good quality’

Insurers negotiate with hospitals that DRG payment is

higher/lower if certain quality standards are met/not met

(e.g., in Germany and the Netherlands)

DRG payment for all hospitals is based on ‘best practice’;

that is, costs incurred by efficient, high-quality hospitals

(e.g., in England)

Patient

based



certain amount









4. Frequent revisions of PCS and payment rates


Country PCS Payment rate

Frequency of updates Time-lag to data Frequency of updates Time-lag to data

Austria Annual 2–4 years 4–5 years 2–4 years

England Annual Minor revisions annually; irregular

overhauls about every 5–6 years

Annual 3 years (but adjusted

for inflation)

Estonia Irregular (first

update after 7 years)

1–2 years Annual 1–2 years

Finland Annual 1 year Annual 0–1 year

France Annual 1 year Annual 2 years

Germany Annual 2 years Annual 2 years

Ireland Every 4 years Not applicable (imported AR-DRGs) Annual (linked to

Australian updates)

1–2 years

Netherlands Irregular Not standardized Annual or when

considered necessary

2 years, or based on

negotiations

Poland Irregular – planned

twice per year

1 year Annual update only of

base rate

1 year

Portugal Irregular Not applicable (imported AP-DRGs) Irregular 2–3 years

Spain

(Catalonia)

Biennial Not applicable (imported

3-year-old CMS-DRGs)

Annual 2–3 years

Sweden Annual 1–2 years Annual 2 years

2007

2.1

1.3

2.5

2006

0.00.4

1.5

-2.1-1.8

0.3

1.1

2009

1.41.5

2.62.9

2008

1.1

2.3

3.8

2010

-2.0

A-Segment A-Segment corrected for budget reductions B-Segment 2005 B-Segment 2008 B-Segment 2009

Free prices – to be negotiated between insurers and providers

Share %

B-segment

Main

additions

to

B-segment

Knee Arthritis

Cataract

Hip Arthritis

Slipped Disc

Diabetes

Pregnancy Birth

Cataract

Pacemaker

Meniscus

Breast Cancer

Cardio

Stroke

Follow-up Cardio

Skin Cancer

10% 19% 34%

Fixed prices – set by health regulator

Price development hospital DBCs 2006-2010 (%, nominal)

Sources: Marktscan Medisch Specialistische Zorg 2011; Nza. Onderhandelen over ziekenhuiszorg; Vektis 2009

Prices for list B–DBCs in the Netherlandscan be negotiated Actual

reimbursement

• Volume limits

• Outliers

• High cost cases

• Quality

• Innovations

• Negotiations


Health care reform succeeded in lowering prices, but it did not curb volume growth

10

-0.54)

09

0.3

08

5.5

07

1.3

06

1.0

05

-0.9

04

0.1

03

2.1

02

7.4

01

11.0

00

5.3

99

5.6

Volume growth (%)

Generic

Inflation2)

10

4.0

09

6.4

08

5.4

07

4.3

06

3.2

05

4.6

04

6.4

03

4.1

02

5.4

01

2.8

00

0.1

99

0.2

• But since the health

care reform volume

growth accelerated

• Today’s challenge:

volume growth

reduction without the

waiting lists of the

1990s

• Health care reform

(competition) has

indeed led to lower

prices (driven by B-

segment)

The 2005-2006 Reform Paradigm• Volume growth is a fact of life: ageing,

innovation• More efficiency is needed to deal with

volume growth• Competition will lead to more efficiency

and lower prices

Strict

budgeting

Wait list

reduction

Health care reform:

competitions

09 10

3.5

6.8

08

10.9

04

6.5

03

6.2

02

12.9

07

5.6

06

4.1

05

3.7

01

13.8

00

5.4

99

5.8

Price increase (%)Total growth in hospital expenditures (%)1)

1) Hospital expenditure include day and/or night cost and include specialist health care (4) Estimate based on “Marktscan Medisch specialistische zorg2011”2) Consumer Price Index CBSSources: CBS Statline (Zorgrekeningen; expenditures at current and constant cost); RIVM Performance Of Dutch Health Care 2010; Stijging Zorgkostenontrafeld; VGE; Marktscan Medisch specialistische zorg 2011; BoStrategy& analysis


• Volume limits

• Outliers

• High cost cases

• Quality

• Innovations

• Negotiations


Conclusions

• DRG-based hospital payment is the main method of provider payment in Europe, but systems vary across countries– Different patient classification systems

– DRG-based budget allocation vs. case-payment

– Regional/local adjustment of cost weights/conversion rates

• To address potential unintended consequences, countries– implemented DRG systems in a step-wise manner

– operate DRG-based payment together with other payment mechanisms

– refine patient classification systems continously (increase number of groups)

– place a comparatively high weight on procedures

– base payment rates on actual average (or best-practice) costs

– reimburse outliers and and high cost services separately

– update both patient classification and payment rates regularly

• If done right (which is complex), DRGs can contribute to increased transparency and efficiency – and possibly quality


2. Februar 2017 37Krankenversicherung und Leistungsanbieter

Documents

DRG Systeme in Europa - TU Berlin · different forms of hospital payment ... the NHIF top down microcosting x Finland 5 reference hospitals ... Change coding practice