10/21/2015

1

Stephen Duckett

@stephenjduckett

Presentation to

Tasmanian Economic Forum

October 2015

Improving efficiency in

hospital care

2

What are we trying to achieve?

Cost/outcome

Cost/output

Outcome/output

3

0

5

10

15

20

25

30

35

40

45Real growth

Growth if spending

a constant

proportion of GDP

Change in Australian governments’ expenditure 2003-2013

$ bn relative to CPI

Health expenditure is the fastest growing

segment of government expenditure

Welfare Health Education Defence Infra-

structure

Ageing,

comm &

disability

services

Govern-

ment

Other

4

Hospitals are the fastest growing segment of

health expenditure

-5

0

5

10

15

20

25

Hospitals Primarycare andmedicalservices

Other Pharmac-euticals

Privatehealth

insurance

Notfurther

specified

Real growth

Growth if spending

a constant

proportion of GDP

Change in Australian governments’ health expenditure 2003-2013

$ bn relative to CPI

10/21/2015

2

5

-$2,000

$0

$2,000

$4,000

$6,000

$8,000

$10,000

$12,000

$14,000

$16,000

Most of the spending increases is a result of

more services per person of a given age

Other

Health inflation

(beyond CPI)

Ageing

Population growth

Sources: AIHW; ABS; DOHA

Change in Australian governments’ health expenditure 2004-2011

2011$ m

Public

hospitals Medical

services

Medication

Private

hospitals

Research

Community

health

Dental

Other

6

Growth in spending is evenly balanced across

ages

Older population

$172

$108

$181

$250

$231

$187

$246

$246

$186

$438

$0

$500

$1,000

$1,500

$2,000

$2,500

$3,000

0s 10s 20s 30s 40s 50s 60s 70s 80splus

Ageing (more elderly

people)

More services (to people of the same age)

Why are we spending more per person? Contribution to increase in per-person costs, 1989 to 2010

Note: Less reliance ought to be placed on figures for 80+, as sample sizes are small and data categories change across surveys

Source: ABS Fiscal Incidence Studies (various years); ABS Cat. 3101.0 Table 59; Grattan analysis

Age group

7

Projected Commonwealth government grants for

public hospitals to states have been cut substantially

(sort of)

Source: 2014-15 Budget Overview p7 http://budget.gov.au/2014-15/content/overview/download/Budget_Overview.pdf

Grattan Health

Program on

hospital waste

10/21/2015

3

9

Waste abounds in hospitals: proportion of

beds with identified waste at audit

Note: Only one type of waste was recorded for each bed.

Resar, R. K., et al. (2011) Hospital inpatient waste identification tool, Institute

for Healthcare Improvement 10

There is significant within-state variation in

public hospital costs (2010-11 data)

Note: Some small hospitals (total admissions < 4,000 p.a.) not shown

$0 $500 $1,000 $1,500 $2,000 $2,500 $3,000

NSW

VIC

QLD

WA

SA

TAS

ACT

NT

Unexplained costs above the lowest level in each state ($ per admission) Hospitals with the lowest unexplained cost in each state = 0

Hospitals

Average level of unexplained costs

Avoidable costs unexplained costs above the average level

11

Avoidable costs add up to $1 billion a

year

$0

$50

$100

$150

$200

$250

$300

ACTNTTASSAWAVICNSWQLD

Avoidable cost by state, 2010-11

$ million

12

$2,000 $4,000 $6,000 $8,000 $10,000

ACT

TAS

QLD

WA

SA

NSW

VIC

Range Median

There is huge variation in the cost of

treatments, e.g. gall bladder removal …

Cost of laproscopic colecystectomy (gall bladder removal), unadjusted, 2010-11

$2,000 $4,000 $6,000 $8,000 $10,000

Note: H08B, the less complicated DRG category for the procedure

10/21/2015

4

13

$0 $10,000 $20,000 $30,000 $40,000

ACT

TAS

QLD

SA

WA

VIC

NSW

Range Median

Cost of hip replacement, unadjusted, 2010-11

$0 $10,000 $20,000 $30,000 $40,000

Note: I03B, the less complicated DRG category for the procedure

… and hip replacements

14

4,128

4,478

6,029

6,106

7,934

$0 $5,000 $10,000 $15,000

Admissions

Median cost

High volume hospitals

A

E

D

C

B

Costs aren’t driven by scale Cost of gall bladder removal, unadjusted, five high-volume hospitals, 2010-11

Note: H08B, the less complicated DRG category for the procedure

15

There are three steps to remove avoidable

costs, starting with setting the right price

Arbitrary

Pay for

costs caused

by patient

factors

Don’t pay for

avoidable

costs

Pay for care that works

Adjust for

adverse

events

Adjust for

readmissions

Pay for

pathways

Adjust for

outcomes

Their own

pervious

standard

What care

does cost More

research &

better data

needed Average cost

Current

system

Grattan

proposal

Below-

average cost

What care

should cost

What costs are funded?

What

standard

should

hospitals

reach?

16

Medical workforce supply and demand – three

scenarios

80,000

85,000

90,000

95,000

100,000

105,000

110,000

115,000

Comparison Productivity gain Low demand

Supply Demand

Number of doctors 2025

10/21/2015

5

17

Job satisfaction improved with more complex

roles

Hospital specialists' overall job satisfaction by responses to the statement "I often undertake tasks that could be done by somebody less qualified

than me"

0%

20%

40%

60%

80%

100%

Strongly

disagree

(n=137)

Disagree

(n=811)

Neutral

(n=487)

Agree

(n=1040)

Strongly

agree

(n=427)

Satisfaction:

Very dissatisfied

Not sure

Moderately

satisfied

Very satisfied

Moderately

dissatisfied

18

There was very strong agreement with a wide range of

substitution options

Respondents were asked to what extent they agreed that the following shifts of workload would

reduce the cost without reducing quality and safety

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Interns to ENs

ENs to cleaners

Residents to clerical workers

ENs to clerical workers

RNs to personal care assistants

Specialists to RNs

RNs to clerical workers

Specialists to physician assistants

Residents to physician assistants

Residents to RNs

Interns to nurse practitioners

Interns to RNs

RNs to ENs

Residents to nurse practitioners

ENs to personal care assistants

Specialists to nurse practitioners

Occupational therapists to allied health assistants

Physiotherapists to physiotherapy assistants

Strongly agree Agree Neither Disagree Strongly disagree

5

19

Pharmacy staff mix and automation,

European hospitals, 2010

0

5

10

15

20

0

1

2

1–49 50–99 100–199 200–299 300–399 400–599 600–799 800–999 1000–1499 1500–2000 >2000

Hospital size (beds)

Pharmacists

Staff / 100 beds

Assistants

Per cent hospitals with stock dispensing robots (dots)

Frontini, R., Miharija-Gala, T. and Sykora, J. (2012) 'EAHP survey 2010 on hospital pharmacy in Europe: Part 1. General frame and staffing', European Journal of Hospital Pharmacy: Science and

Practice, 19(4), p 385-387 (and Part 2)

20

Substitution opportunities reported in

Victoria

0%

20%

40%

60%

80%

Nutrition OT Orthotics Physio Podiatry Social work

Speech path

Acute

Sub-acute

Percentage of time that allied health practitioners spend completing assistant-attributable tasks.

10/21/2015

6

21

Different barriers as a proportion of all responses (top barrier)

Professional culture and industrial relations were seen

as the biggest barriers to substitution across all fields …

7

0% 10% 20% 30% 40%

Blank

Implementation or transition costs

Substitute workforce availability

Management capacity

Other

Substitute workforce capacity/quality

Tradition

Registration restrictions

Industrial relations

Professional culture

22

Opportunity 1: Nursing assistants

• Identify roles currently

performed by registered

nurses that can be

performed by nursing

assistants

• Nurse assistants are

away of recruiting from a

different pool (non-school

leavers, returning to

workforce)

Registered nurses

Nursing assistants

23

Opportunity 2: Specialist roles for nurses

• International and Australian

evidence show that nurses can do

low-risk, high-volume procedures

well and safely

• We propose specialist nurse

endoscopy and nurse sedationist

roles

• Nurse endoscopists would only

provide less-complex endoscopies

(no biopsies or other interventions)

• Nurse sedationists would only work

in low-risk age groups and cases

Registered nurses

RN advanced roles

24

Opportunity 3: Allied health assistants

Physio, OT

Allied health assistants

(Physio, OT)

0% 20% 40% 60% 80%100%

Interns to ENs

ENs to cleaners

Residents to clerical…

ENs to clerical workers

RNs to personal care…

Specialists to RNs

RNs to clerical workers

Specialists to…

Residents to…

Residents to RNs

Interns to nurse…

Interns to RNs

RNs to ENs

Residents to nurse…

ENs to personal care…

Specialists to nurse…

Occupational…

Physiotherapists to…

Strongly agree

10/21/2015

7

25

GP

Practice

nurse

Physician

assistant

Opportunity 4: Physician assistants

Narrow scope

of practice

Wide scope

of practice

Autonomy

Delegation

GP or other medical practitioner

Physician assistant

26

Transition grants to facilitate change

1. Hospitals sign up to

• embed new roles in their workforce

• agree to a target ratio of existing

and new (not simply adding)

2. Hospitals get a time-limited grant to

fund a proportion of the cost of

introducing new workforce. It tapers

over time.

3. At mid-point, hospitals must prove they

are halfway to meeting the final staffing

ratio. The second half of the grant is

conditional upon demonstrating

progress.

4. The transition grant finishes at end of

workforce introduction period, by which

time new workforce roles embedded.

5. Savings will be passed back to the

system by a fall in the cost of care (and

prices).

Temporary subsidy for new roles

Change in workforce mix

Phase out subsidy with

system savings

27

What are we trying to achieve?

Cost/outcome

Cost/output

Outcome/output

28

Most variation analyses look at geographic

variation and find large disparities …

-125% -100% -75% -50% -25% 0% 25% 50% 75% 100% 125% 150%

MLA procedure rate: difference from national average

Cholecystectomy

Colectomy

Hip replacement

Lumpectomy

Mastectomy

Knee replacement

CABG

Open prostatectomy

Closed prostatectomy

Tonsillectomy

Appendectomy

Hysterectomy

Source: Grattan Institute

10/21/2015

8

29

… but that doesn’t tell you much

Procedure

rates

Highest

Lowest

… is it due to different case-mix?

… is it due to severity of condition?

… is it due to patient preferences?

… is it due to lack of certainty about effectiveness?

… what does regional analysis mean anyway?

High could be bad but …

Low could be bad but…

There’s little clarity about when variation is legitimate

That’s made it difficult to develop effective policy

30

Increasing certainty that variation can

identify inappropriate care

Variation in rates suggests inappropriate care

31

Increasing certainty that variation can

identify inappropriate care

Variation in rates suggests inappropriate care

On average this treatment should not be provided routinely for this class of patient

32

Increasing certainty that variation can

identify inappropriate care

Variation in rates suggests inappropriate care

On average this treatment should not be provided routinely for this class of patient

High level evidence that on average this treatment should not be provided for this class of patient

10/21/2015

9

33

Increasing certainty that variation can

identify inappropriate care

Variation in rates suggests inappropriate care

On average this treatment should not be provided routinely for this class of patient

High level evidence that on average this treatment should not be provided for this class of patient

Definitive advice that this treatment should not be provided in this patient

34

We analyse 5 ‘do-not-dos’ and 3 ‘do-not-do

routinely’ treatments from NICE, MSAC and Prasad

Do-not-dos:

• Vertebroplasty for osteoporotic vertebral fractures

• Arthroscopic lavage or debridement for OA of the

knee

• Laparoscopic uterine nerve ablation for chronic pelvic pain

• Removing healthy ovaries during a hysterectomy

• HBOT for a range of conditions (inc.

osteomyelitis, cancer, and non-diabetic wounds

and ulcers)

Do-not-do routinely:

• Fundoplication for gastro-intestinal reflux

• Episiotomy for spontaneous vaginal births

• Amniotomy to augment a normal delivery

Patients with ‘legitimating’ diagnoses are excluded

35

79

302

813

4659

0 1000 2000 3000 4000 5000

Nerve ablation

Ovary removal

Vertebroplasty

Knee arthroscopy

Hyperbaric oxygen

Do-not-do procedures, Australia, 2010-11

35

A large proportion of relevant patients have

do-not-dos

0% 10% 20% 30% 40% 50% 60%

Ovary removal

Hyperbaric oxygen

Vertebroplasty

Knee arthroscopy

Nerve ablation

36

-

20

40

60

80

100

120

-

1

2

3

4

5

NSW VIC QLD WA SA TAS NT

-

5

10

15

20

NSW VIC QLD WA SA TAS NT

Hyperbaric oxygen

chamber for various

conditions

Vertebroplasty for

compression fractures

Knee arthroscopy for

osteoarthritis

Removal of healthy ovaries

during hysterectomy

Nerve ablation

for pelvic pain

Rate

pe

r 1

00

0

‘DN

D’ p

atie

nts

Rate

pe

r 1

00

0

‘DN

D’ p

atie

nts

0

10

20

30

40

50

60

NSW VIC QLD WA SA TAS NT

0

20

40

60

80

100

120

NSW VIC QLD WA SA TAS NT

Rates of do-not-dos vary across states

10/21/2015

10

37

0%

20%

40%

60%

80%

Hospital

Average

Hyp

erb

aric

Art

hro

scop

y

Vert

eb

rop

lasty

Ova

ry r

em

ova

l

Nerv

e a

bla

tion

0%

10%

20%

30%

40%

50%

Amniotomy Episiotomy Fundoplication

Proportion of relevant patients getting do-not-do procedure Proportion of relevant patients getting do-not-do routlinely procedure

There are outliers with troubling patterns of

care

38

Information gap 1:

What not to do

0% 10% 20% 30% 40%

Consumer involvement documented

Setting identified

Users identified

Endorsed by other agencies

Replicable description of review

Recommendations linked to evidence

Professionals involved identified

Development process described

Quality indicators for Australian clinical practice guidelines, 2005-2013

Source: National Health and Medical Research

Council

0

0.2

0.4

0.6

0.8

1

1.2

1994199920042009

0

2

4

6

8

10

1994199920042009

Articles (million)

Articles (thousand)

• There is a huge volume of evidence

• Guidance focuses on what to do, is of variable quality, is

inconsistent & hard to use

• 50+ organisations work on disinvestment and their

approaches are largely uncoordinated and inconsistent

All

Systematic reviews

PubMed artices, 1994-2013

39

Information gap 2:

Who’s doing what

0%

20%

40%

60%

80%

Worse About the same Better or much better

Overall quality of health care Safe and skilled workforce

Responding to health care incidents

Proportion of board members Victorian LHNs, views on own network relative to average Victorian network

Notes: n = 233, 70% response rate, 96% of networks included

Source: Bismark et al (2013)

40

Hospital Name – 2010-11

Multiples of

national

rate DND/Rs

Relevant

patient

group Do-not-dos

HBOT DNDs - - - - - -

Removal of healthy ovaries 13.0 8 183

Vertebroplasty for CFs 0.0 0 31

Knee arthroscopy for OA 0.5 2 95

Nerve ablation for pelvic pain 0.6 1 75

Do-not-do routinely

Fundoplication for GORD 0.6 3 366

Episiotomy 2.9 211 1507

Amniotomy 0.4 26 1912

- - not in comparator group

Over benchmark

Less than 10% under benchmark

Recommendation 2:

ACSQHC report to all providers & funders

10/21/2015

11

41

Accountability gap

Recommendation 3: clinical reviews with

consequences

Identify outliers

Inform outliers that they are being

closely monitored

No further action State to initiate

external clinical review

Are they still

outliers after one

year?

Yes No

Does clinical

review support

practices? No further action Set clear targets for

improvement

No Yes

Are targets met?

No further action

No Yes

Financial and/or

governance sanctions

42

Recommendation 4:

Improve variation measurement

• Find more do-not-dos elsewhere (e.g.

Cochrane) and add more do-not-do

routinely treatments

• Link patient separations to

• analyse treatments that should not

be given first-line

• adjust for readmissions

• allow better adjustments for

morbidity

00%

10%

20%

30%

40%

50%

Linked imaging and pathology results

MBS and PBS diagnosis data

Linked records from MBS, PBS, hospitals

Hospital records (used in this report)

Measurable with:

Many more NICE do-not-dos can be measured with data we already collect

• Link to PBS and MBS data to acute data to allow measurement of more do-not-dos

(e.g. primary care do-not-dos, polypharmacy, patients not getting routine first-line

drug therapies)

• Pilot morbidity database for GP care in a few PHNs – collect data as part of MBS

billing

43

Some of our choices

• How much ‘benefit of doubt’ to give?

oIs a ‘Do Not Do’ a ‘Never Do’?

• Who should initiate investigation for potentially

inappropriate care?

• Is it OK for private hospital to be the focus (vs

surgeon)

• When should private insurers be able to deny

payment?

oWhen ACSQHC makes a determination?

oWhen clinical review makes a determination?

oWhen hospital fails to respond to external review?

44

What hospitals might do:

Table the Grattan report for discussion with the relevant

clinical governance group:

• Are any of the DNDs or DNDRs are an issue in the hospital?

• There are other issues we didn’t look at which are prominent in the

public debate (e.g. diagnostic test use). Are they relevant?

• How robust are your clinical governance processes?

• Is appropriateness of care being systematically monitored?

• What are the accountability mechanisms for clinical

choices?

• NB: I don’t think there are big savings for hospitals here

• NB: I do think this will be an increasing clinical governance

issue

10/21/2015

12

45

What Tasmania could do:

• Initiate clinical reviews where appropriate

for both the Do Not Dos and the Do Not Do

Routinelys

• Ask ACSQHC to develop a guideline

program about clinically ineffective care

• Ask ACSQHC to use link other data sets to

investigate other DNDs which can't be

identified with the Hospital Morbidity data

set

• Maybe start looking at cost-ineffective

care?

46

Some suggestions

• Drive (technical) efficiency through

• benchmarking

• tight(er) pricing (change range)

• workforce reform (change mean)

• Drive allocative efficiency through

• benchmarking

• P4P/nP4nP

• accountability for ineffective care

stephen.duckett@grattan.edu.au