Modeling & Managing Complex Systems A Case Study of Healthcare Delivery

Copyright © 2008 Tennenbaum Institute. All rights reserved.

Knowledge and Skills for Enterprise Transformation.Knowledge and Skills for Enterprise Transformation.

Modeling & Managing Complex SystemsModeling & Managing Complex SystemsA Case Study of Healthcare DeliveryA Case Study of Healthcare Delivery

William B. RouseWilliam B. Rouse

Knowledge and Skills for Enterprise Transformation. 2

OverviewOverview• Complex Adaptive Systems Complex Adaptive Systems • Healthcare DeliveryHealthcare Delivery• Complexity of HealthcareComplexity of Healthcare• Case Study – Cost ControlCase Study – Cost Control• Healthcare EnterpriseHealthcare Enterprise• SummarySummary


Complex Adaptive SystemsComplex Adaptive Systems• They are They are nonlinear, dynamicnonlinear, dynamic and do not inherently reach fixed equilibrium points. and do not inherently reach fixed equilibrium points.

The resulting system behaviors may appear to be random or chaotic. The resulting system behaviors may appear to be random or chaotic.• They are composed of They are composed of independent agentsindependent agents whose behavior can be described as whose behavior can be described as

based on physical, psychological, or social rules, rather than being completely based on physical, psychological, or social rules, rather than being completely dictated by the dynamics of the system.dictated by the dynamics of the system.

• Agents' needs or desires, reflected in their rules, are not homogeneous and, Agents' needs or desires, reflected in their rules, are not homogeneous and, therefore, their therefore, their goals and behaviors are likely to conflictgoals and behaviors are likely to conflict -- these conflicts or -- these conflicts or competitions tend to lead agents to adapt to each other's behaviors.competitions tend to lead agents to adapt to each other's behaviors.

• Agents are Agents are intelligent, learnintelligent, learn as they experiment and gain experience, and as they experiment and gain experience, and change behaviors accordingly. Thus, overall systems behavior inherently change behaviors accordingly. Thus, overall systems behavior inherently changes over time.changes over time.

• Adaptation and learning tends to result in Adaptation and learning tends to result in self-organizingself-organizing and patterns of and patterns of behavior that emerge rather than being designed into the system. The nature of behavior that emerge rather than being designed into the system. The nature of such emergent behaviors may range from valuable innovations to unfortunate such emergent behaviors may range from valuable innovations to unfortunate accidents.accidents.

• There is There is no single point(s) of controlno single point(s) of control – systems behaviors are often – systems behaviors are often unpredictable and uncontrollable, and no one is "in charge." Consequently, the unpredictable and uncontrollable, and no one is "in charge." Consequently, the behaviors of complex adaptive systems usually can be influenced more than behaviors of complex adaptive systems usually can be influenced more than they can be controlled.they can be controlled.


HealthcareHealthcare

• Stakeholders & InterestsStakeholders & Interests• Example of Disease DetectionExample of Disease Detection• Network of NetworksNetwork of Networks


Stakeholders & InterestsStakeholders & InterestsStakeholderStakeholder Risk Mgt.Risk Mgt. PreventionPrevention DetectionDetection TreatmentTreatment

PublicPublic e.g., Buy Insurancee.g., Buy Insurance e.g., Stop e.g., Stop SmokingSmoking

e.g., Get e.g., Get ScreenedScreened

Delivery SysDelivery Sys PhysiciansPhysicians Physicians & Physicians & HospitalsHospitals

GovernmentGovernment Medicare, Medicare, Medicaid, Medicaid, CongressCongress

NIH, CDC, DoD, et al.NIH, CDC, DoD, et al.

Non-ProfitsNon-Profits American Cancer Society, American Heart American Cancer Society, American Heart Association, et al.Association, et al.

AcademiaAcademia Business SchoolsBusiness Schools Basic Science Basic Science DisciplinesDisciplines

Technology & Technology & Medical SchoolsMedical Schools

Medical Medical SchoolsSchools

BusinessBusiness Employers, Employers, Insurance Insurance

Companies, HMOsCompanies, HMOs

Guidant, Guidant, Medtronic, et al.Medtronic, et al.

Lilly, Merck, Lilly, Merck, Pfizer, et al.Pfizer, et al.


Disease DetectionDisease DetectionCostsCovered

PublicAwareness

PublicReadiness

ScreeningAvailable

ScreeningEffective

PublicCommunication

PublicEducation

PhysicianEducation

ConsumerAdvocacy

MedicalResearch

$$$ $ $Public, Delivery System, Government, Non-Profits, Academia, Business


Accreditation & Licensing• American Board of Medical Specialties• Accreditation Council for Graduate Medical Educ.• Accreditation Council for Continuing Medical Educ.• AOA Council on Postdoctoral Training• Federation of State Medical Boards• Joint Commission on Accreditation of Healthcare Org.• Liaison Committee on Medical Education

Examples of Other Stakeholders• American Assoc of Retired Persons• Leapfrog Purchasing Group• National Business Group on Health• Etc.

Networks of NetworksNetworks of Networks

Professional Associations• American Academy of Family Physicians• American Medical Association• American Osteopathic Association (AOA)• Council of Medical Specialty Societies• Etc.


ComplexityComplexity

• Market ComplexityMarket Complexity• Studying ComplexityStudying Complexity• Design ImplicationsDesign Implications• Management ImplicationsManagement Implications


Market ComplexityMarket Complexity• RetailRetail• TelecomTelecom• HealthcareHealthcare• Complexity ModelComplexity Model

– Network ModelNetwork Model– Calculating ComplexityCalculating Complexity– Complexity SurfaceComplexity Surface

• Complexity AssessmentComplexity Assessment


Food Services

Food &Drugs

GeneralMerchandisers

SpecialtyRetail

Consumers

FoodConsumerProducts

Electronics

Furniture

Household

Beverages

Apparel

Toys

Pharma

HomeEquipment

FoodProduction

Packaging

Transportation &Logistics

RetailRetail


Semiconductor& Other Electronics

Network &Infrastructure

Equipment

TelecommServiceProviders

Service &Billing

ConsumersComputerPeripherals

ComputerSoftware

ContentProviders

Government

TelecomTelecom


Semiconductor& Other Electronics

Supplier

Network &Infrastructure

EquipmentSupplier

TelecommServiceProviders

Service &Billing

Provider

ConsumersComputerPeripherals

Supplier

ComputerSoftwareSupplier

ContentProviders

Government

QualcommTexas Instruments

Intel

Nortel NetworksCisco

SiemensOpenwave

Alcatel-Lucent

PalmNokia

SamsungMotorola

Research in MotionNovatel Wireless

HP

MicrosoftSymbian

Research in Motion

Verizon WirelessT-Mobile

AT&TSprint-Nextel

FCCAmdocs

GoogleYahoo!CNN

Enterprises& End-Users


Pharmaceuticals HealthWholesalers

HealthProviders

Consumers

Pharmacy

MedicalEquipment

OtherEquipment

HealthInsurance

Government &Policy Makers

R&D Laboratories

HealthcareHealthcare


Network ModelNetwork Model1

2

3

Ni

1

2

3

Nij

1

2

3

Nijk

1

2

3

Nijkl

t01234

Ni = No. of 1st tier suppliers to ith product/service outletsNij = No. of 2nd tier suppliers to ijth Tier 1 supplier, e.g., OEMNijk = No. of 3rd tier suppliers to ijkth Tier 2 supplierNijkl = No. of 4th tier suppliers to ijklth Tier 3 suppliers


Calculating ComplexityCalculating Complexity

Σi = 1

Ni

- p (ni | t) log [p (ni | tm)] +

Σj = 1

Nij

- p (nj | ni t) log [p (nj | ni tm)] +

Σk = 1

Nijk

- p (nk | ni nj t) log [p (nk | ni nj tm)] +

Σl = 1

Nijkl

- p (nl | ni nj nk t) log [p (nl | ni nj nk tm)]

C = Σ ptmm = 1

T

{

}


Complexity SurfaceComplexity Surface

10 20 30 40 50 60 70 80 90 1001

3

50.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Complexity (Bits)

Breadth

Depth

Market Complexity


0

5

10

15

20

25

30

35

Aerospace Automotive Retail Healthcare Telecom

Co

mp

lexi

ty (

Bit

s)

Consumer Total

Complexity AssessmentComplexity Assessment


Design ImplicationsDesign Implications

• Two Design PrinciplesTwo Design Principles• Designed ComplexityDesigned Complexity• Enterprise AgilityEnterprise Agility• Enterprise ArchitecturesEnterprise Architectures


Two Design PrinciplesTwo Design Principles• The nature and extent of B2C service value

determines B2B service value, as well as the value of products and other value enablers.

• The magnitude of B2C complexity, relative to total market complexity, reflects market maturity– Both B2C and B2B complexity are expressed in terms

of information theory binary digits (bits),– B2B complexity is often increased, in turn increasing

total complexity, in order to reduce B2C complexity.

B2C = Business to ConsumerB2B = Business to Business


Designed ComplexityDesigned Complexity

• The enterprise as a system includes all stakeholder The enterprise as a system includes all stakeholder organizations whether they are your partners, collaborators, organizations whether they are your partners, collaborators, channels, competitor, or regulatorschannels, competitor, or regulators

• Increase complexity where you can best manage it, in order Increase complexity where you can best manage it, in order to decrease complexity to end users, i.e., patients and to decrease complexity to end users, i.e., patients and physiciansphysicians– You can manage design, development, manufacturing, sustainmentYou can manage design, development, manufacturing, sustainment– You cannot manage economies, markets, competitors, end usersYou cannot manage economies, markets, competitors, end users

• Support managing complexity by providing means to Support managing complexity by providing means to monitor and influence system state, performance, and monitor and influence system state, performance, and stakeholders – see Managing Complexity stakeholders – see Managing Complexity


Enterprise AgilityEnterprise Agility

• Optimization Within Design AssumptionsOptimization Within Design Assumptions– Objective is to balance performance and costsObjective is to balance performance and costs– Assure robust allocation of resourcesAssure robust allocation of resources

• Adaptation Beyond Design AssumptionsAdaptation Beyond Design Assumptions– Objective is to balance opportunities and risksObjective is to balance opportunities and risks– Assure ability to reallocate resourcesAssure ability to reallocate resources

• Tradeoffs Between Optimization vs. AdaptationTradeoffs Between Optimization vs. Adaptation– Options for contingencies enhance agilityOptions for contingencies enhance agility– Costs of options undermine optimalityCosts of options undermine optimality

• Architectures That Enable Ongoing Tradeoffs Architectures That Enable Ongoing Tradeoffs


Enterprise ArchitecturesEnterprise Architectures

Strategic Processes

Operational Processes

Information Systems

Information Technology


Management ImplicationsManagement Implications

• Value PhilosophyValue Philosophy

• Organizational BehaviorsOrganizational Behaviors

• Managing ComplexityManaging Complexity


Value PhilosophyValue Philosophy

• Value focuses on organizational outputs (or Value focuses on organizational outputs (or outcomes), rather than inputs.outcomes), rather than inputs.– Health states of patients vs. budgets of providersHealth states of patients vs. budgets of providers

• Value relates to benefits of outcomes, rather Value relates to benefits of outcomes, rather than outcomes themselves.than outcomes themselves.– Productivity improvements due to wellnessProductivity improvements due to wellness

• Value implies relevant, usable, and useful Value implies relevant, usable, and useful outcomes.outcomes.– Stakeholders have to understand and appreciateStakeholders have to understand and appreciate


Organizational BehaviorsOrganizational Behaviors

Traditional SystemTraditional System Complex Adaptive SystemComplex Adaptive System

RolesRoles ManagementManagement LeadershipLeadership

MethodsMethods Command & ControlCommand & Control Incentives & InhibitionsIncentives & Inhibitions

MeasurementMeasurement ActivitiesActivities OutcomesOutcomes

FocusFocus EfficiencyEfficiency AgilityAgility

RelationshipsRelationships ContractualContractual Personal CommitmentsPersonal Commitments

NetworkNetwork HierarchyHierarchy HeterarchyHeterarchy

DesignDesign Organizational DesignOrganizational Design Self OrganizationSelf Organization


Managing ComplexityManaging Complexity• System StateSystem State

– Current and projected value flowsCurrent and projected value flows– Current and projected problemsCurrent and projected problems

• System PerformanceSystem Performance– Current and projected value, costs & value/costCurrent and projected value, costs & value/cost– Current and projected options for contingenciesCurrent and projected options for contingencies

• System StakeholdersSystem Stakeholders– Involvement of each stakeholder groupInvolvement of each stakeholder group– Performance of each stakeholder groupPerformance of each stakeholder group

• Information SystemsInformation Systems– Measurement, modeling & display of system stateMeasurement, modeling & display of system state– Agile “What If?” experimentation & adaptationAgile “What If?” experimentation & adaptation


Case StudyCase Study

• MotivationMotivation– Tripled %GDP for Tripled %GDP for realreal healthcare costs for 1965-2005 healthcare costs for 1965-2005– 50% of cost growth attributable to technological innovation50% of cost growth attributable to technological innovation– ““Running on Empty”Running on Empty”

• ApproachApproach– Growth ModelGrowth Model– Learning ModelLearning Model– Process ModelProcess Model

• ConclusionsConclusions


ApproachApproach

• Overall PhenomenonOverall Phenomenon• Model No. 1 & ResultsModel No. 1 & Results• Model No. 2 & ResultsModel No. 2 & Results• Model No. 3 & ResultsModel No. 3 & Results• ImplicationsImplications


TechnologyInnovation

IncreasedEfficiency

IncreasedEffectiveness

DecreasedRisk

IncreasedUse

DecreasedCost/Use

IncreasedExpenditures

LongerLife

ImprovedCare


Model No. 1 -- GrowthModel No. 1 -- Growth

Cost (1 – α) Use (1 + β) = Total (1 + δ)

where α = Annual rate of cost reduction β = Annual rate of usage growth δ = Annual allowable total growth

α = (β – δ)/(β + 1)


Results No. 1Results No. 1

TreatmentTreatment Annual Rate of Usage Annual Rate of Usage GrowthGrowth

Minimum Annual Rate Minimum Annual Rate of Cost Reductionof Cost Reduction

AngiographyAngiography 10%10% 9%9%

AngioplastyAngioplasty 15%15% 13%13%

DialysisDialysis 12%12% 11%11%

Hip ReplacementHip Replacement 10%10% 9%9%

Knee ReplacementKnee Replacement 11%11% 10%10%


LimitationsLimitations

• Model Provides No Mechanism for Model Provides No Mechanism for Achieving Cost ReductionsAchieving Cost Reductions

• Model Does Not Differentiate Elements of Model Does Not Differentiate Elements of Healthcare Delivery ProcessHealthcare Delivery Process


Model No. 2 -- LearningModel No. 2 -- Learning

Cost (t=T) = Cost (t=0) No. Uses (t=T)-Rate

Percent Cost Per Use for Percent Cost Per Use for Each Doubling of UsesEach Doubling of Uses

Rate for Rate for Learning ModelLearning Model

70%70% 0.5150.515

80%80% 0.3220.322

90%90% 0.1520.152

Note: This is a well-developed concept in a wide range of production processes.


Cost Reductions @ 5% GrowthCost Reductions @ 5% Growth

Cost Per Use for Three Learning Rates

-

20

40

60

80

100

100

110

122

134

148

163

180

198

218

241

265

293

323

356

392

Uses

Co

sts

Pe

r U

se

70%

80%

90%


Expenditure Growth @ 5%Expenditure Growth @ 5%

Growth of Expenditures @ 5%

-

5,000

10,000

15,000

20,000

25,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Year

To

tal

Co

sts


Cost Reductions @ 10% GrowthCost Reductions @ 10% Growth

Cost Per Use for Three Learning Rates

-

20

40

60

80

100

100

121

146

177

214

259

314

380

459

556

673

814

985

1192

1442

Uses

Co

sts

Pe

r U

se

70%

80%

90%


Expenditure Growth @ 10%Expenditure Growth @ 10%

Growth of Expenditures @ 10%

-

10,000

20,000

30,000

40,000

50,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Year

To

tal

Co

sts



Results at 30 YearsResults at 30 Years

RateRate No. of Uses No. of Uses Cost/UseCost/Use Total ExpendituresTotal Expenditures

5%5% 412412 $48/use$48/use $19,874$19,874

10%10% 15861586 $24/use$24/use $38,256$38,256

Note: Uses = 100, Cost/Use = $100, Total = $10,000 at Year =1


LimitationsLimitations

• Model Exhibits Impressive Cost Model Exhibits Impressive Cost Reductions Due to Production Learning, Reductions Due to Production Learning, But It Does Not Suggest Where & How But It Does Not Suggest Where & How This Learning HappensThis Learning Happens

• Model Does Not Reflect the Process Model Does Not Reflect the Process Whereby Healthcare Is DeliveredWhereby Healthcare Is Delivered


Labor Labor Labor

Technology Technology Technology

Tre

atm

ent

Dia

gnos

is

Rec

over

y

Det

ectio

n

Technology

Labor

Model No. 3 -- ProcessModel No. 3 -- Process


Expanded Cost ModelExpanded Cost ModelCost (t) = Cost of Labor (t) + Cost of Tech. (t)

CTOT (t) = CPUL (t) NU (t) + CPUT (t) NU (t)

CPUL (t) = CPUL (1) NU (t)-RateL

CPUT (t) = CPUT (1) NU (t)-RateT

NU (t) = NU(1) (1+β) t-1



Required Efficiency (% Cost Per Use Per Doubling)

0%

10%

20%

30%

40%

50%

60%

70%

80%

5% 10% 15%

Technology Use Growth Rate

La

bo

r C

os

t P

er

Us

e

GDP = 0%

GDP = 2%

GDP = 4%


ImplicationsImplications

• In order to limit the growth of total healthcare In order to limit the growth of total healthcare spending to the growth of the GDP, some spending to the growth of the GDP, some combination of three things is neededcombination of three things is needed– Limiting the growth of technology useLimiting the growth of technology use– Limiting the cost of technology useLimiting the cost of technology use– Decreasing the cost of labor associated with useDecreasing the cost of labor associated with use

• Overall, savings due to learning are the key to Overall, savings due to learning are the key to affordabilityaffordability– Learning rates of < 70% are very difficult to achieveLearning rates of < 70% are very difficult to achieve


Sources of LearningSources of Learning• Labor efficiency, changes in personnel mix, Labor efficiency, changes in personnel mix,

standardization, specialization, methods improvements, standardization, specialization, methods improvements, better use of equipment, changes in the resource mix, better use of equipment, changes in the resource mix, product and service redesign, and shared best practices.product and service redesign, and shared best practices.– Less Labor Per UseLess Labor Per Use: Achievable via individual learning and : Achievable via individual learning and

productivity enhancements.productivity enhancements.– Less Expensive Hours Per UseLess Expensive Hours Per Use: Achievable by, for example, : Achievable by, for example,

substituting assistant physicians or nurse practitioners for substituting assistant physicians or nurse practitioners for physicians. In this case, the experts can be used as physicians. In this case, the experts can be used as orchestrators of cadres of much less expensive clinicians.orchestrators of cadres of much less expensive clinicians.

– No Labor Per UseNo Labor Per Use: Often technology enabled. For example, : Often technology enabled. For example, web-based scheduling and account management can enable web-based scheduling and account management can enable patients to substitute their labor for that of providers, as has patients to substitute their labor for that of providers, as has been experienced in the airline, banking, and retail industries.been experienced in the airline, banking, and retail industries.


““Bending the Curve”Bending the Curve”1.1. Producing and Using Better InformationProducing and Using Better Information

1.1. Promoting Health Information TechnologyPromoting Health Information Technology2.2. Center for Medical Effectiveness & Health Care Decision MakingCenter for Medical Effectiveness & Health Care Decision Making3.3. Patient Shared Decision MakingPatient Shared Decision Making

2.2. Promoting Health and Disease PreventionPromoting Health and Disease Prevention1.1. Public Health: Reducing TobaccoPublic Health: Reducing Tobacco2.2. Public Health: Reducing ObesityPublic Health: Reducing Obesity3.3. Positive Incentives for HealthPositive Incentives for Health

3.3. Aligning Incentives with Quality and EfficiencyAligning Incentives with Quality and Efficiency1.1. Hospital Pay-for-PerformanceHospital Pay-for-Performance2.2. Episode-of-Care PaymentEpisode-of-Care Payment3.3. Strengthening Primary Care and Care CoordinationStrengthening Primary Care and Care Coordination4.4. Limit Federal Tax Exemptions for Premium ContributionsLimit Federal Tax Exemptions for Premium Contributions

4.4. Correcting Price Signals in the Health Care MarketCorrecting Price Signals in the Health Care Market1.1. Reset Benchmark Rates for Medicare Advantage PlansReset Benchmark Rates for Medicare Advantage Plans2.2. Competitive BiddingCompetitive Bidding3.3. Negotiated Prescription Drug pricesNegotiated Prescription Drug prices4.4. All-Payer Provider Payment Methods and RateAll-Payer Provider Payment Methods and Rate5.5. Limit Payment Rate Updates in High-Cost AreasLimit Payment Rate Updates in High-Cost Areas


InterpretationInterpretation

• Seven of these recommendations would Seven of these recommendations would tend to reduce use ratestend to reduce use rates– 1.2, 1.3, 2.1, 2.2, 2.3, 3.1, 3.2 1.2, 1.3, 2.1, 2.2, 2.3, 3.1, 3.2

• Nine of these recommendations focus Nine of these recommendations focus on reducing costs byon reducing costs by– Increased efficiency (1.1, 1.2, 3.3)Increased efficiency (1.1, 1.2, 3.3)– Increased market-based competition (3.4, Increased market-based competition (3.4,

4.1-4.5)4.1-4.5)


ConclusionsConclusions• Successful Technology Innovation Leads to Successful Technology Innovation Leads to

Growing Markets and Increased RevenuesGrowing Markets and Increased Revenues– Revenue to Innovator Revenue to Innovator Cost to Payer Cost to Payer

• Such Growth Is Viewed More Favorably When Such Growth Is Viewed More Favorably When Individuals Pay, Rather Than Third PartiesIndividuals Pay, Rather Than Third Parties– Possible Market-Based Mechanisms?Possible Market-Based Mechanisms?

• Increasing System Efficiency Is Needed to Assure Increasing System Efficiency Is Needed to Assure the Affordability of Technology Innovationthe Affordability of Technology Innovation– Required Improvements Are Very Substantial Required Improvements Are Very Substantial


Healthcare EnterpriseHealthcare Enterprise

• Entities & ContextEntities & Context

• Enterprise ArchitectureEnterprise Architecture

• Implications for ChangeImplications for Change

• Hierarchical NetworkHierarchical Network


Entities & ContextsEntities & Contexts

Entities Entities Interacting Interacting

Context of Context of InteractionsInteractions

PeoplePeople ProcessesProcesses

ProcessesProcesses OrganizationsOrganizations

OrganizationsOrganizations SocietySociety


Clinical Practices(People)

Economic Model &Incentive Structure

Healthcare Ecosystem(Society)

Human Productivity &Healthcare Costs

Delivery Operations(Processes)

Patient Care &Health Outcomes

Care Capabilities &Health Information

System Structure(Organizations)

Economic Returns &Performance Information

Competitive Positions &Economic Investments


Implications for ChangeImplications for Change

• Value can be increased in two waysValue can be increased in two ways– Increase the health outcomesIncrease the health outcomes– Decrease the cost of outcomesDecrease the cost of outcomes

• Increasing value provided by Clinical PracticesIncreasing value provided by Clinical Practices– Constrained by the nature of Delivery OperationsConstrained by the nature of Delivery Operations

• Increasing value provided by Delivery OperationsIncreasing value provided by Delivery Operations– Constrained by System StructureConstrained by System Structure

• Increased value provided by System StructureIncreased value provided by System Structure– Constrained by the EcosystemConstrained by the Ecosystem


Hierarchical NetworkHierarchical Network

Intra-Level Information Flow & Incentives

Inter-Level Information Flow

& Incentives

Delivery Operations(Processes)

Clinical Practices(People)

Healthcare Ecosystem(Society)

System Structure(Organizations)


SummarySummary

• Complex Adaptive Systems Complex Adaptive Systems • Healthcare DeliveryHealthcare Delivery• Complexity of HealthcareComplexity of Healthcare• Case Study – Cost ControlCase Study – Cost Control• Healthcare EnterpriseHealthcare Enterprise


Documents

Modeling & Managing Complex Systems A Case Study of Healthcare Delivery