The Real Drivers of Quality: The View from an Accountable Care Organization

The Real Drivers of Quality: The View

from an Accountable Care Organization

Brent C. James, M.D., M.Stat.

Executive Director, Institute for

Health Care Delivery Research

Intermountain Healthcare

Salt Lake City, Utah, USA

Annenberg Hall, Harvard University, Cambridge, Massachusetts

Wednesday, 19 August 2009 -- 10:45a - 11:15a

1.

Well-documented, massive, variation in

practices

(beyond the level where it is even remotely possible that all

patients are receiving good care)

2.

High rates of inappropriate care

3.

Unacceptable rates of preventable care-

associated patient injury and death

4.

A striking inability to "do what we know works"

5.

Huge amounts of waste and spiraling prices,

that limit access

(46.6 million uninsured Americans, and still

climbing)

(1)

Continued reliance on the "craft of medicine"

(clinicians as stand-alone experts)

runs up against

(2)

Clinical uncertainty

in the context of

(3)

Payment that encourages utilization

The craft of medicine

(each physician an expert)

placing her patient's health care needs before any

other end or goal,

An individual physician

drawing on extensive clinical knowledge gained

through formal education and experience

Can craft

a unique diagnostic and treatment regimen

customized for that particular patient.

This approach will produce the best

result possible for each patient.

Clinical uncertainty

Enthusiam for unproven methods ...

The maxim, "If it might work, try it" ...

Quality means "spare no expense" ...

1.

Lack of valid clinical knowledge

regarding best treatment

(poor evidence)

2.

Exponentially increasing new medical knowledge

(doubling time has decreased to ~8 years; at current rates, a clinician will need to learn, unlearn, then relearn half of their medical knowledge base 5 times during a typical career)

3.

Continued reliance on subjective judgment

4.

Limitations of the expert mind when making complex

decisions

Miller, 1956: The magic number 7, plus or minus 2: some limits on our capacity for processing information Eddy: "The complexity of modern medicine exceeds the capacity of the unaided human mind"

A care delivery environment where

1.

Every patient encounter contributes to reliable

knowledge regarding best care

2.

Every clinician is in formal life-long training

- organizational resources to track key research developments

- regularly detail front-line clinicians in "state of the art" care high-priority conditions

- "best care" integrated into clinical workflows, so that evidence-based best practice is the lowest energy state, default action

3.

Any time any clinician says "in my experience," they

mean "in my measured experience"

- patients first seek a personal counselor and advocate -- usually a physician or nurse - they translate their experience -- including clinical outcomes -- through that relationship

- signficiant health care encounters nearly always involve a series of choices among competing options (far beyond the initial choice of a physician or a hospital)

- true definition of transparency (from 2 separate IOM committees): all involved (clinical counselors, as well as patients) have sufficient (not necessarily perfect) information to make good choices at each decision point

4.

Information infrastructure helps organize and prioritize

complex decisions

- decision support: "sharpen the image" by making complex decisions more transparent and more tightly focused)

A key, proven tool: mass customization

Shared baselines

(a form of Lean Production)

-A multidisciplinary team of health professionals:

1.

Select a high priority care process

2.

Generate an evidence-based "best practice" guideline

3.

Blend the guideline into the flow of clinical work

staffing

training

supplies

physical layout

educational materials

measurement / information flow

4.

Use the guideline as a shared baseline, with clinicians

free to vary based on individual patient needs

5.

Measure, learn from, and (over time) eliminate variation

arising from professionals; retain variation arising

from patients

("mass customization")

Are practice protocols

A case management system

(to control cost outcomes)

a research system

(for continuous learning,

to improve medical outcomes)

--Physicians

It is more important that you do it the

same

than that you do it

"right"

When you "do it the same:"

error rates fall

costs fall

you can apply the scientific method to

systematically improve

-- staff is more efficient; you more are efficient

-- less complexity = fewer mistakes = better outcomes

-- regardless of where you start,

you will end up with best demonstrated care practices

(Truth is found more often from mistakes than from confusion ...)

Francis Bacon (1561 - 1626)

Quality as a core business strategy

Integrated clinical / operations

management structure

1998:

(an outcomes tracking system)

Integrated management information systems

1997:

cost structure vs. net income

(mediated by payment mechanisms)

integrated facility / medical expense budgets

Integrated

(aligned)

incentives

1999:

Full roll-out and administrative integration

2000:

(strategic)

Key process analysis

(very strong)

Shared vision

- what is Intermountain's mission;

- how do we add value to the communities we serve?

1996:

Managing clinical knowledge

1.

Generate initial evidence-based best practice guideline

(flowchart)

2.

Blend the guideline into clinical workflow

(clinical flow sheets, standing order sets, etc.)

3.

Design outcomes tracking reports

(using electronic data warehouse)

4.

Design and coordinate decision support

(electronic medical record)

5.

Design patient and professional education materials

Initial development phase

6.

Keep the Care Process Model current

(research pipeline; protocol

variations; outcomes; improvement suggestions)

7.

Academic detail front-line teams

(Clinical Learning Days)

8.

Run the referral clinic

(last step in treatment cascade)

9.

Manage specialist care managers

Maintenance phase

Core work group

(knowledge expert)

responsibility

Computerized decision support

Start with a practice guideline

Implement it on a clinical information system

transparent to users

automatically pulls all necessary clinical information

codifies process knowledge through expert system support

decision support (advisories, alerts) as opposed to mandates

or cookbooks

Span the continuum of care

(a longitudinal data

repository with all inpatient and outpatient information)

Use it for both management

(cost control)

and improvement

(learning, research)

Antibiotic process management

1988

Total patients

25,288

Antibiotic patients

8,051

(31.8%)

21,898

11,624

(53.1%)

Overall LOS

Antibiotic LOS

5.11

7.5

4.50

7.3

Antibiotic cost/patient

$122.66

$ 78.37

Pharmacy costs

$3,979,561

Antibiotic costs

$ 987,547

(24.8%)

$7,185,817

$ 924,876

(12.9%)

1994

higher severity of illness (case mix index) no change in patterns of bacterial resistance

more wide-spectrum antibiotic use higher per unit pharmacy costs

(reduced total tonnage, better crop rotation)

Mortality

3.65%

2.65%

Evans et al. A computer-assisted management program for antibiotics and other antiinfective agents. New Engl J Med 1998; 338(4):232-8 (Jan 22).

ADEs at LDS Hospital

89

90

91

92

93

94

95

96

97

98

99(3)

Year

0

100

200

300

400

500

600

0

100

200

300

400

500

600

Elements of coordinated care

1.

A primary care practice

2.

A network of support: referral specialists and hospitals

3.

Embedded care management nurses

4.

"Shared baseline" evidence-based best practice

support

5.

Communication system

(electronic medical record)

that

- spans the continuum of care ("lifetime" longitudinal patient record, including personal health record) - implements and deploys "shared baseline" evidence-based best practice protocols

Problems and chronic conditions

Medication profile

Preventive care summary

Pertinent labs

Pertinent exams

Passive reminders

organized by illness

General

patient

status

information

Disease

specific

information

21% 31% 26% 39%

1 year

2 years

0%

10%

20%

30%

40%

50%

Percent readmissions

0%

10%

20%

30%

40%

50%

Control

Care management

Complex diabetes patients - hospitalization rates

0

1

2

3

4

Years

0.7

0.75

0.8

0.85

0.9

0.95

1

Proportion alive

Control

Care management

Complex diabetes patients - mortality rates

Physician productivity

(WRVUs - work relative value units)

0

100

200

300

400

Average WRVUs

Physicians with embedded care management support were

significantly (8%) more productive than controls

16.6 15.4 14.5 14.4 13.8 13.1 12.3 12.7 _12.1 10.8 9.5 9.2 9. 4 8.6 8.2 8 8. 3 7.9 7.5 7.7 7.8 7.6 7.6 7.6 8.3 7.8 7.3 7.9 _{7.7 7.7 7.}7 8 8.6 Jun 99S ep_Dec Mar 00 JunSepDec Mar 01 Jun_Sep_Dec Mar 0 2 JunSepDec Mar 03JunSe p Dec Mar 04 JunSe p Dec Mar 05 JunSepDec Ma r 06JunSepDec Mar 0 7 Jun 0 5 10 15 20

% diabetic patients with HgA1c > 9

0 5 10 15 20

Poor HbA1c control

31.5 32.8 34.5 34.1 34 .9 36.4 38. 6 37.4 38.4 41.9 45.7 45.5 45.3 48.1 51.1 50.8 48.8 51.3 53. 8 53.2 _52.5 54. 6 56 54.8 55 .8 57.2 59.1 57.3 57.3 57.9 59.5 57.8 56.1 Jun 99S ep_Dec Mar 00 JunSepDec Mar 01 Jun_Sep_Dec Mar 0 2 JunSepDec Mar 03JunSe p Dec Mar 04 JunSe p Dec Mar 05 JunSepDec Ma r 06JunSepDec Mar 0 7 Jun 0 25 50 75

% diabetic patients with HgA1c < 7

0 25 50 75

Excellent HbA1c control

39.4 40.5 41 .2 41.8 42.5 42.9 43.9 44.4 44 .2 44.1 44 .9 46 47 .1 48. 7 49.3 50.3 50. 7 51 52.2 54.4 55.3 56.7 58.7 59.8 60.8 61 .2 62.6 63. 5 63.7 63.6 63.5 Dec99_Mar 00JunSe p Dec Mar 01 Jun_Sep_Dec Mar 0 2 JunSepDec Mar 03JunSe p Dec Mar 04 JunSe p Dec Mar 05 JunSepDec Ma r 06JunSepDec Mar 0 7 Jun 0 25 50 75

% diabetic patients with LDL < 100

0 25 50 75

Excellent lipid control

The healing professions are changing

From

craft-based practice

individual physicians, working alone

handcraft a customized solution for each patient

based on a core ethical commitment to the patient and

vast personal knowledge gained from training and experience

To

profession-based practice

groups of peers, treating similar patients in a shared setting

plan coordinated care delivery processes

which individual clinicians adapt to specific patient needs

(e.g., standing order sets) (housestaff ::= apprentices)

early experience shows

less expensive less complex

better patient outcomes

(facility can staff, train, supply an organize to a single core process) (which means fewer mistakes and dropped handoffs, less conflict)

Why "profession-based" practice?

1.

It produces better outcomes for our patients

2.

It eliminates waste, reduces costs, and

increases available resources for patient care

3.

It puts the caring professions back in control

of care delivery

4.

It is the foundation for useful shared electronic

Only one pertinent question:

Assume that front-line clinicians are

-

as smart you are

-

as dedicated to patients as you are

-

as hard-working as you are

-

as motivated as you are

- are the only ones with fundamental knowledge

of how the front-line process actually works

;

But they usually don't control the systems that set

the context within which they work ...

How will your proposed intervention