David Haggstrom Regenstrief Conference Slides

Experience and Potential for Transformational Change in

Cancer Care Delivery

David A. Haggstrom, MD, MAS

Regenstrief Conference

October 3, 2007

Tranformational change• “Together we can transform the American health care

system into one that provides the highest quality of care for all of its citizens.”– Carolyn Clancy, AHRQ

• Connect the System. Through electronic health records, standards, and information exchange.

• Measure and Publish Quality. Work with doctors and hospitals to define benchmarks for what constitutes quality care.

• Measure and Publish Price. Agreement is needed on what procedures are covered in each "episode of care."

• Create Positive Incentives. Reward those who offer & those who purchase high-quality, competitively priced health care.

RAND quality report

• Individuals receive 55% of recommended care

• “A key component of any solution…is the routine availability of information on performance at all levels. Making such information available will require a major overhaul of our current health information systems, with a focus on automating the entry and retrieval of key data for clinical decision making and for the measurement and reporting of quality.”

McGlynn E et al., NEJM, 2003.

Cancer policy statementsInstitute of Medicine, April 1999:

“A cancer data system is needed that can provide quality benchmarks for use by systems of care: hospitals, provider groups, and managed care systems”

National Cancer Institute, cancer.gov, 2002:“Work is underway to make cancer a working model for quality of care research and the translation of this research into practice”

Regenstrief Cancer Care Engineering, May 2007:“Develop an hierarchical system-based approach to improving cancer care: establishing an Indiana prototype for colorectal cancer care”

National Initiative for Cancer Care Quality– Development of quality measures:

• Literature review of existing indicators, guidelines, RCTs• Subsequent expert review (clinical, health services researchers,

nurses, patients)

– Method: patient survey & medical record review• 47% of eligible patients enrolled

– Proportion of patients who received recommended care• 86% of breast cancer patients

– Adherence less than 85% for 18/36 breast cancer measures• 78% of colorectal cancer patients

– Adherence less than 85% for 14/25 colorectal ca measures

– Colorectal cancer post-treatment surveillance: 50%– Appropriate referral: 13-59%– Respect for patient preferences: 57-71%

Schneider E et al., Jnl Clin Onc, 2005.

Outline• How do we measure quality?• With what data do we measure quality?• What technologies can transform our use of quality data?

– health information exchange– clinical decision support– personal health records

• How do we study technology implementation?• Who is accountable for quality?

– provider-level– system-level– patient-level

Data

Quality measureTransformative

technology Provider

Patient

System

Health care quality and Overutilization

Unmeasured quality gap

Good Quality Poor Quality

Good Quality

UnderuseOveruse

• Application of medical interventions without known medical benefit, or worse, with known lack of benefit

• Overtreated: Why too much medicine is making us sicker and poorer, Shannon Brownlee, 2007

• Cancer examples:• bone marrow transplant for

breast cancer• PSA screening• surveillance testing



Health care quality and Uncertainty

Uncertainty

Certainty

• Clinical uncertainty• Risk adjustment

• Decision-making uncertainty• Patient preferences &

values vary• Measure “informed” or

“good decisions” in concordance with patient preferences, Mulley et al.

Other dimensions of quality

Patient satisfaction Patient-provider communication

Quality-of-life Symptom control

Importance of measurement

Act Plan

Study Do

PDSA cyclesPDSA cycles

Act Plan

Study Do

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

General population Cancer

population Longitudinal care

How do we measurecancer care quality?

Recurrence

Surveillance

General population SEER

cancer registries

Medicare

With what data do we measure quality?United States

Haggstrom DA et al., Cancer. 2005.

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

General population

Medicare/Medicaid

With what data do we measure quality?Michigan

State cancer registry

Bradley CJ et al., Cancer, 2005.

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

General population

With what data do we measure quality?Iowa

Private claimsState

cancer registry

Doebbeling BN et al., Med Care,1999.

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

With what data do we measure quality?Indiana

‘Connect the System’

Cancer population Longitudinal care

IN State cancer registry

SPINVA cancer

registryLINK

VA-INPC

longitudinal care• Medicare• Medicaid• Private claims• Regenstrief EMR• VA admin. claims• VA EMR

New possibilities

with EMR• Test results

– Laboratory

– Radiology•Free text query

Add new quality measures in EMR

• Satisfaction• Communication• Quality-of-life• Symptoms

• Proportion of patients who underwent screening – primary care• Adequate lymph node retrieval & evaluation - surgery• Proportion receiving radiation therapy/chemotherapy - oncology• Proportion who underwent follow-up of abnormal test – who?• Proportion receiving postoperative surveillance – who is responsible?

VA-INPCCancer

population Longitudinal data

Primary care – quality measure

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance


Cancer population Longitudinal data

VA/INPC

Surgery treatment – quality measure

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

• Proportion of patients who underwent screening – primary care• Adequate lymph node retrieval & evaluation - surgery• Receipt of radiation therapy/chemotherapy - oncology• Proportion who underwent follow-up of abnormal test – who?• Proportion receiving postoperative surveillance – who is responsible?


VA/INPC

Oncology treatment – quality measure

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

Transformative technology #1:Health information exchange

Primary care Surgery Oncology

timely quality performance reports

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

Transformative technology #2:Clinical decision support

High riskAve risk / Follow-up of abnormal tests Surveillance colonoscopy

ONCWATCH REMINDERS

real-time quality performance

Cancer screening

Control

hospitals

β site

No CDS

Active system re-design

5 more hospitals

Passive dissemination5 regional hospitals

Learning system

1. CLINICAL PROCESSES

a. Cancer screening

b. Diagnostic colonoscopy

c. Surveillance colonoscopy

2. Organizational surveys• Culture• Teamwork• Leadership

New CDS

Implementation Plan

Qualitative data–Workflow–Usability

Control

hospitals

Implementation researchof health information technology

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


High riskLow risk / Follow-up of abnormal tests Surveillance colonoscopy

Physician

specialtyGeneral internist +/- gastroenterologist +/- surgeon

real-time quality performance

Cancer screening

REMINDERS

• Proportion who underwent follow-up of abnormal test

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

Uncertain accountability

Primary careRadiologist

Proceduralist

Who is responsible?

Information exchange

• Proportion receiving postoperative surveillance

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


Primary careSurgery

Oncology

Who is responsible?

Information exchange

Approaches touncertain accountability

Promoting continuity infragmented health care system

1. System approach:• Assign provider accountability to larger organizations/units

than single provider

2. Patient-centered approach• Share accountability through personal health records

‘Extended medical staff’definition

• Data source: Assigned physicians & patients using Medicare claims (2002-2004)

• Physicians extended medical staff– Inpatient MDs: assign MD to hospital where they

provided care to most inpatients– Outpatient MDs: assigned MD to hospital where

admitted most patients• Patients extended medical staff

– Inpatients: assigned based on plurality of discharges over specific period

– Outpatients: assigned to physician (primary or specialty care) who provided most of their care in outpatient setting

• then assigned to physicians’ primary hospital

Fisher E et al., Health Affairs,2006

Concentration of patients among extended medical staff

% of physician inpatient work at primary hospital

90%

% of patient received services from extended medical staff

73%

Patient panel size Individual providers Extended medical staff

0-49 50% <1%

50-499 48% 2%

>500 2% 98%

Advantages of ‘extended medical staff’ as locus of accountability

‘Positive incentives’• Performance measurement

– focus on longitudinal experience of patient & address fragmentation of care by pooling accountability

• Local organizational accountability– influence resource distribution, IF measured on BOTH

quality & cost– larger organizations have capacity to invest in

improving quality & lowering costs

Patient-centered approachPromoting continuity in

fragmented cancer care system

Survivorship Care Plan

1.Cancer type, treatment received, & potential toxicities

2.Tailored information about timing & content of recommended follow-up

3.Recommendations regarding preventive practices & how to maintain health & well-being

4.Availability of psychosocial services in community

Personal health record

1.Diagnosis, treatment received, & potential side-effects

2.Tailored information about timing & content of recommended follow-up

3.Recommendations regarding preventive practices & how to maintain health & well-being

4.Availability of psychosocial services in community

RiskAssessment

Screening Diagnosis Treatment Surveillance

Provider -Oncologist-Radiationtherapist-Surgeon

-PrimaryCare-Radiologist-Proceduralist

-PrimaryCare

-PrimaryCare-Oncologist-Surgeon-Radiologist-Proceduralist

PHR

Transformative technology #3:Personal health records

Palliativecare

Patient/

caregiver

Longitudinal care

• Current functions:– Patient education– Self-management tools– Pharmacy refills

• Functions in near future:– on-line appointments– patient/provider messaging

• Pilot program:– full access to VA electronic health record

Study/project designs

Appropriate care

Clinical processes

Quality

personal health record

no personal health record

Patient perceptionsquality of cancer care

• Population: population-based survey of 1,067 patients with colorectal cancer nine months after diagnosis

• Mean problem score (in descending order):– Health information 48– Treatment information 32– Psychosocial care 32– Coordination of care 21– Access to care 12

Ayanian J et al., Jnl Clin Onc, 2005.

Racial/ethnic differencespatient perceptions of quality of care

Health care quality and AccessUnmeasured quality gap


Population with access

to health care system

• “Chemotherapy cost not covered for illegal immigrants because does not qualify as emergency”,

New York Times,

September 23, 2007

No

access

Transformationtechnology & implementation

Transformative technologiesInformation exchange

timely performance feedbackClinical decision support

real-time peformance feedbackPersonal health records

patient-centered decision support

Implementation science Better understand adoption of technology with multiple methods & designs

Transformationaccountability

Solitary episodes of care delivery accountability for multiple episodes of care

shared among multiple providers• Large organizations have capacity to act

give health information and recommendations to patient in personal health record

• Patient is willing to act – patient health is at stake

Transformationwhat we measure is what we change (at least on purpose)

Leverage ‘system connectedness’

Measure quality in many dimensionsUnderuseOveruseAccessPatient experience

Act Plan

Study Do

Act Plan

Study Do

Indy quality performance measure the right things

at the right time

Act Plan

Study Do

Act Plan

Study Do

Thank you

Health care quality and Overutilization



Good Quality

UnderuseOveruse

• Application of medical interventions without known medical benefit, or worse, with known lack of benefit

• “Avoiding the unintended consequences of growth in medical care: how might more be worse?” 1999, JAMA, Fisher E, Welch G

• Overtreated: Why too much medicine is making us sicker and poorer, Shannon Brownlee, 2007

• Cancer examples:• bone marrow transplant for breast

cancer• PSA screening• surveillance testing

Health care quality and AccessUnmeasured quality gap


Population with access

to health care system

• African Americans more frequently than whites

• lost medical insurance coverage after cancer diagnosis

• denied coverage after changing jobs

• reached their insurance spending limits

• “CMS - chemotherapy cost not covered for illegal immigrants because does not qualify as emergency”,

New York Times,September 23, 2007

No

access

Implementation research

• Measuring context…

• Organizational surveys:– Teamwork– Leadership– Culture

• Qualitative methods

• Human factors engineering

What is measured targets what is changed

Approporiate clinical use Overuse Risk-adjusted use Preference-concordant use Patient experience Patient symptoms Access

• Proportion of care delivered outside the VA for CRC care? Denise Hynes?

Importance of measurement

Act Plan

Study Do

PDSA cyclesPDSA cycles

XX

Act Plan

Study Do

XX

Subspecialty care

Primary care


Poor QualityGood Quality

Shared care

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


Oncwatch

High riskLow risk / Follow-up of abnormal tests Surveillance colonoscopy

REMINDERS

Physician

specialtyGeneral internist +/- gastroenterologist +/- surgeon

real-time feedback

Cancer screening

Sample sizes

Patient population size

Assessment of providers as individuals

Assessment of providers as group (extended medical staff)

0-49 50% <1%

50-499 48% 2%

>500 2% 98%

Cancer examples of use of non-evidence-based interventions

Bone marrow transplant for breast cancer

PSA Surveillance tests among cancer

survivors

Atlanta measures

• More detail about measures???

Control

sites

α/β site

Non-OncWatch

Active implementation

5 more sites

Passive diffusion5 regional sites

Learning system

1. CLINICAL PROCESSES

a. Cancer screening

b. Diagnostic colonoscopy

c. Surveillance colonoscopy

2. Organizational surveys• Culture• Teamwork• Leadership

OncWatch

Implementation Plan

•Qualitative data–Workflow–Usability

Control

sites

• If a man will begin with certainties, he will end in doubts; but if he will be content to begin with doubts, he will end in certainties. -Francis Bacon (1561-1626),_

Cancer quality/performance measures

• Measuring the Quality of Breast Cancer Care in Women– Evidence Report/Technology Assessment No. 105. (University of

Ottawa Evidence-based Practice Center) AHRQ. October 2004.

• Cancer Care Quality Measures: Diagnosis and Treatment of Colorectal Cancer

– (Duke Evidence-based Practice Center) AHRQ. May 2006.

• Process measures• Outcome measures

Survivorship care plan

Survivorship care planPersonal health record

Types of uncertainty

ERROR: Measure quality of physician performance only by measuring clinical processes.

UNINTENDED CONSEQUENCE: Patients making informed and shared decisions with their physicians not to undergo screening or treatment are labeled as receiving poor quality care. Sick patients who would not benefit from aggressive screening or treatment may also be labeled as receiving poor quality care.

RESEARCH NEED: In areas of clinical uncertainty, measure quality by measuring the presence of “good decisions”, not only what decision is made. Take into account patient illness or preferences when measuring quality, otherwise, the measures may create incentives for inappropriate or unwanted clinical care.

• “The only man who behaves sensibly is my tailor; he takes my measurements anew every time he sees me, while all the rest go on with their old measurements and expect me to fit them”– George Bernard Shaw

General population

Medicaid

How do we measure performance?Michigan

State cancer registry

Bradley CJ, Gardiner J, Given CW, Roberts C. Cancer, Medicaid enrollment, and survival disparities. Cancer. 2005 Apr 15;103(8):1712-8.

Bradley CJ, Principal Investigator, “In-Depth Examination of Disparities in Cancer Outcomes.” $1,630,646. Funding Source: National Cancer Institute. 2004-2008.

Risk Assessm

ent

Diagnosis

Cancer Treatm

ent

Screening

Palliative Care

Recurrence

Surveillance

Health care quality improvement or implementation research

Poor

Quality

Health

Care

Good QualityPoor QualityGood Quality

Measurement of Cancer Care Measurement of Cancer Care QualityQuality

OrganizationOrganizationss

ProvidersProviders

Pathway 1:Pathway 1:IMPLEMENTATIOIMPLEMENTATIO

NN

Quality improvement framework

GoodGood PoorPoor

•PatientsPatients•Referring Referring CliniciansClinicians•PurchasersPurchasers INCENTIVE TO INCENTIVE TO

CHANGECHANGE

Pathway 2:Pathway 2:SELECTIONSELECTION

REPORTREPORTSS

PPUUBBLLIICC

Berwick D, Institute for Healthcare Improvement, Berwick D, Institute for Healthcare Improvement, Medical CareMedical Care



ProvidersProviders


NN


GoodGood PoorPoor


CHANGECHANGE


REPORTREPORTSS

PPUUBBLLIICC




ProvidersProviders


NN


GoodGood PoorPoor



ProvidersProviders


NN


GoodGood PoorPoor


CHANGECHANGE


REPORTREPORTSS

PPUUBBLLIICC



• Cheesy?

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

• When Mrs. Hewitt sued Habana over her mother’s death, for example, she found that its owners and managers had spread control of Habana among 15 companies and five layers of firms.

• As a result, Mrs. Hewitt’s lawyer, like many others confronting privately owned homes, has been unable to establish definitively who was responsible for her mother’s care.

• New York Times, September 23, 2007

• The limits of health services research:– Public health

– Chemotherapy for immigrants not covered

IOM: Crossing the Quality Chasm

• Effective

• Patient-centered

• Timely

• Efficient

• Equitable

• Safe

Tranformational change• “Together we can transform the American health care system into

one that provides the highest quality of care for all of its citizens.”– Carolyn Clancy, AHRQ

• Connect the System. Every medical provider will have some system for electronic health records. Standards need to be set so all health information systems can quickly and securely communicate and exchange data.

• Measure and Publish Quality. Every case, every procedure has an outcome. Some are better than others. To measure quality, we must work with doctors and hospitals to define benchmarks for what constitutes quality care.

• Measure and Publish Price. Price information is useless unless cost is calculated for identical services. Agreement is needed on what procedures are covered in each "episode of care."

• Create Positive Incentives. All parties—providers, patients, insurance plans, and payers—must be subject to contractual arrangements that reward those who offer and those who purchase high-quality, competitively priced health care.

• “Changing what is possible”

MedicareMedicaid

Private

Insurance Uninsured

Indiana Network for Patient Care

Regenstrief Medical Record System (RMRS)

General population

Cancer population

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

VA

INPC

StateRegistry

SPIN

Risk Assessm

ent

TimelyDiagnos

is

Cancer Treatme

nt

Screening

Surveillance for

Recurrence

Genomics

Risk of Cancer Death

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


population (xx%)

SEER(20% US)

StateRegistries(Indiana)

VA

Cancer registriesD

A

T

A

S

O

U

R

C

E

Primary care practice

• VA• IU-MG

VA Medicare Medicaidprivate

insurance Uninsured

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

Cancer population

SEER(20% US)


VA

Cancer registriesD

A

T

A

S

O

U

R

C

E

VA Medicare Medicaidprivate

insurance Uninsured

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

Cancer population

SEER(20% US)


VA

Cancer registriesD

A

T

A

S

O

U

R

C

E

SPIN

RiskAssessment


Patient/caregiver

Patient/caregiver

Provider ProviderProviderProvider Provider

Follow-up of abnormal tests may be improved by coordination

Survivorship care may be improved by coordination

Patient/caregiver

Patient/caregiver

Patient/caregiver

Patient/caregiver

PHR

Transformative technology #3:Personal health records

Patient/caregiver

Patient/caregiver

Palliativecare

Patient/

caregiver

Longitudinal care

VA

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


population

(xx%)

RE

S

E

A

R

C

H

QU

E

S

T

I

O

N

Genomic risk

Health Services Research

SEER(20% US)


VA

Cancer registries

Fragmented health care delivery = fragmented data

age

insurance coverage

location65

VA MedicareMedicaid

private

insuranceUninsured

SEER

SEER-Medicare

MedicareWellpoint

UnitedHealth Medicaid

medical record

VA INPC


age

insurance coverage

location65

VA MedicareMedicaid

private

insuranceUninsured

SEER

SEER-Medicare


age

insurance coverage

location65

VA MedicareMedicaid

private

insuranceUninsured

SEER

SEER-Medicare

RiskAssessment


PatientPatient


Follow-up of abnormal FOBTs may be improved by coordination


PatientPatient Patient/caregiver

Patient/caregiver

PatientPatient

Coordination of careacross the continuum

PHR

65

insurance coverage

location

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


population

(xx%)

RE

S

E

A

R

C

H

QU

E

S

T

I

O

N

Genomic risk

Health Services Research

SEER(20% US)


VA

Cancer registries

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


population

SEER(20% US)


VA

Cancer registriesD

A

T

A

S

O

U

R

C

E


Health Care


Certainty

Clinical uncertainty

Health Care Quality and Uncertainty

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


populationSEER MedicareLongitudinal careState

cancer registry

Medicaid

How do we measure performance?

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

General population State

cancer registry

Private claims



Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance


population Longitudinal data

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

General population

MedicareState

cancer registry

How do we measure performance?Virginia?

Penberthy L, McClish D, Manning C, Retchin S, Smith T. The added value of claims for cancer surveillance: results of varying case definitions. Med Care. 2005 Jul;43(7):705-12.

Risk Assessm

ent

Diagnosis

Cancer Treatme

nt

Screening

Recurrence

Surveillance

General population



RiskAssessment


Patient/caregiver

Patient/caregiver


Follow-up of abnormal FOBTs may be improved by coordination


Patient/caregiver

Patient/caregiver

Patient/caregiver

Patient/caregiver

Patient/caregiver

Patient/caregiver

Coordination of careacross the continuum

PHR

David Haggstrom Regenstrief Conference Slides

Health & Medicine

quality care

colorectal cancer care

quality of care research

highest quality of care

cancer care deliverydavid

quality benchmarks

reporting of quality

cancer examples