Population Management of Chronic Illness: Towards a Scalable Healthcare Infrastructure

1
Population Managementof Chronic IllnessTowards
a Scalable Healthcare Infrastructure
Bruce R. Schatz CANIS LaboratorySchool of
Library Information ScienceSchool of
Biomedical Health Information
Sciences University of Illinois at
Urbana-Champaign schatz_at_uiuc.edu ,
www.canis.uiuc.edu
Comprehensive Depression Center University of
Michigan Medical School Ann Arbor, January 3,
2002
2
Severe versus Average Health

• Depression Center for 35K visits per year
• At this Scale
• Multidisciplinary teams can treat patients
• Telephone questionnaires can follow-up
• State of Michigan has 1.5M at-risk persons
• At this Scale
• Need Healthcare Infrastructure for Population
  Monitoring

3
Outline of Talk

• The Promise (What) slides 4-11
• Population Monitoring of Average Health
• The Technology (How) slides 12-19
• Full-Spectrum Quality-of-Life Indicators
• The Plan (Here to There) slides 20-27
• Pilot Projects for Population Management

4
The Promise

• Population Monitoring
• of
• Average Health

5
The Problem of Chronic Illness

• Chronic Illness is the Economy!
• Acute can cure immediate symptom
• Chronic must manage over long time
• No Infrastructure for Chronic Healthcare
• twice a year community clinic
• twice a month alternative medicine
• twice a day self-care home monitors
• Most of Population has Chronic Illness
• Heart Diseases physical cause of death
• Affective Disorders mental burden of life
• Cancer, Arthritis, Asthma, Diabetes

6
What Works

• Multidisciplinary Teams treating Lifestyle
• Medicine physicians and nurses
• Health psychologists and social workers
• Decreases Readmissions for Heart Disease
• Why are these Teams effective?
• Treat all lifestyle factors (full-spectrum)
• Treat actual disease stage (dynamic)
• Treat actual patient status (adaptive)
• No Infrastructure for Chronic Healthcare
• Expert teams need expert training
• Doesnt scale to whole populations
• Cant reach underserved populations

7
Solution of Healthcare Infrastructure

• Specialty Center (100 at a time)
• Like Depression Center, use a team
• Treat each patient as an individual
• QoL Questionnaire (10K longitudinally)
• Assess Quality of Life with questions (SF-36)
• Patients administer, Physicians analyze
• Gross screening for immediate treatments
• At-Risk Population (1M continuously)
• Full range of stage and status
• Prevention requires early detection

8
What Scales

• Provider Pyramid
• Range of providers for range of needs
• More expert is more expensive
• Level of Service for Volumes of Persons
• Top (few severe) professionals (physicians)
• Middle screening and follow-ups
• Bottom (many average) amateurs (patients)
• Analogues from other Infrastructures
• Evolution of the Telephone (logical/physical)
• Medicine versus Health
• Railroads (physical) versus Banking (logical)

9
Population Management

• Strategy of Preventive Medicine (G. Rose)
• All Chronic Illness is Continuous
• To change Extreme, must change Average
• Infrastructure for Chronic Healthcare
• Must manage the Average (healthy)
• Now treat the Extreme (sick, severe)
• Decrease Average will Decrease Extreme
• Population versus Individual Management
• Population Management by Health Monitors
• Screen All the People All the Time
• Locate at-risk cohorts across population

10
Managed Expectations

• Quality of Life is the Goal
• Improve overall quality across spectrum
• Beyond simply damping down symptoms
• Many Features for Health Status
• in Canada R. Evans economic model
• in America Healthy People 2010
• Beyond Managed Care to Expectations
• Understand spectrum and make choices
• 80-year-olds are not 20-year-olds
• Empowering individuals at base of pyramid

11
Population Monitoring

• Possible to Monitor Whole Populations
• Daily Monitors, Full Spectrum of Features
• Relies on Internet to handle Questionnaires
• Cohort Clusters supplement Diagnoses
• Daily Feature Record for each Individual
• Detailed Records for whole Population
• Group Clusters of Similar Patients
• Cohort Clusters drive Treatments
• Treat by comparing Similar Cases
• Manage Expectations with Actual Cases
• Identify Risk based on Cohort Clusters

12
The Technology

• Full-Spectrum
• Quality-of-Life
• Indicators

13
Quality of Life Indicators

• General Purpose Instruments
• Paper-Based Assessment 30 questions
• Answerable by Patients across Populations
• Medical Outcomes Study (A. Tarlov)
• MOS produced general-purpose SF-36
• Specialty Practices in Big Cities
• Cure status for Acute condition
• Utility of QoL questionnaires
• Effective at gross screening
• VA study (3K) survival of heart surgery

14
Disease-Specific Questionnaires

• Specific Questions for Specific Disease
• 1000 QoL questionnaire instruments
• Paper-based, clinical trial screening
• Causal Model drives Questions
• KCCQ for Cardiomyopathy (CHF)
• Model based on fluid retention overload
• Majority of seniors with CHF dont have!
• Caring for Depression (K. Wells)
• MOS specific for Depression
• CES-D, Center Epidemiological Studies
• DIS, NIMH Diagnostic Interview Schedule

15
Health Status Indicators

• General-Purpose for Social Correlations
• Whitehall study (M. Marmot)
• 12K civil servants in England
• SF-36 longitudinal screening (8K)
• Health status inverse of Socioeconomic
• Special-Purpose for Treatment Outcomes
• Depression Center Outreach (M-DOCC)
• IVR (Interactive Voice Response)
• Brief CDS (21 questions) plus SF-12
• Treatment Outcomes and Screening

16
Depression Screening

• MOS Depression Study (Rand/UCLA)
• 2K patients out of 22K in MOS
• In specialty practices Boston, Chicago, LA
• 5 longitudinal assessments over 4 years
• Every 6 months for 2 years then at 4 years
• Details of the Screening
• 2 stages of screening with CES-D and DIS
• Screen for MDD (major depressive disorder)
• 2nd for chronic dp (dysthymic disorder)
• Telephone follow-up for COD interview

17
Beyond Screening

• Why are Some People Healthy? (R. Evans)
• Major categories are disease, health care,
  health function, genetic endowment, physical
  environment, social environment, individual
  response, behavior, well-being, prosperity.
• Healthy People 2010
• 467 objectives in 28 focus areas
• www.health.gov/healthypeople
• Measure Full-Spectrum Health Status
• Detailed QoL in each detailed category

18
Full-spectrum Dry-runs

• Our first dry-run
• 500 questions from 20 QoL questionnaires
• Use Evans categories with 2 more levels
• Needed more Breadth especially Depth
• Collection Software by Medical Scholars
• Plans for next dry-run
• Multiple categorization for different views
• Encode nurses at Carle and at Barnes (Rich)
• For Depression, Encode the Center!

19
Computer-based Questionnaires

• Treat actual disease stage (dynamic)
• Computer assessment handles full-spectrum
• Database of all questions (500K)
• Individual session asks only 30 questions
• Tree-walking Categories by Breadth-First
• Treat actual patient status (adaptive)
• MOS knows this the problem (McHorney)
• GRE as the paradigm
• Session answers determine questions
• Historical answers determine questions

20
The Plan

• Pilot Projects
• for
• Population Monitoring

21
Population Management

• Possible to Monitor Whole Populations
• Daily Monitors, Full Spectrum of Features
• Internet Software handles Questionnaires
• Cohort Clusters supplement Diagnoses
• Daily Feature Record for each Individual
• Detailed Databases for whole Population
• Analyze Clusters of Similar Patients
• Cohort Switching drive Treatments
• Manage Expectations with Actual Cases
• Improve Health by Switching Cohorts

22
Peer-Peer Computations

• Local Interaction
• Your PC does small computations
• e.g. screensaver for SETI
• Global Merging
• Partition computation into small parts
• Each local forms part of global whole
• Large-Scale Distribution
• 3M users of SETI_at_Home
• Public Health applications already 1M users!

23
Peer-Peer for Medicine

• Intel Philanthropic P2P Program
• www.intel.com/cure
• Evolved engine from SETI
• United Devices commercial software
• 1M volunteers for Cancer computation
• Cancer Research Project (Oxford University)
• Partitioned Screening of Molecules
• Data-centered driven by Indexing needs
• Health monitors feasible for Individuals
• at Scale of whole Populations!

24
Getting from Here to There

• Develop Full-spectrum Questionnaire
• Merge existing Quality of Life instruments
• Encode knowledge from Medical Professionals
• Develop Dynamic Adaptive Administration
• Software to handle Interactive Sessions
• Software to build Individual History
• Software to build Population Database
• Deploy to test Population (30-50 persons)
• Develop Cohort Similarity Clustering
• Algorithms for Statistical Feature Matching
• Lifestyle Coaching via Cohort Switching

25
Healthcare Infrastructure

• Scalable Pilot Project
• 3000-5000 patients across ranges for 3-5 years
• Full-spectrum depth-first for Depression
• Provider Pyramid across County from Center
• Towards Ordinary Medicine
• Handle 1M persons for clinical trial
• Push out from M-CARE, Ford/GM
• All of Michigan, clusters not categories
• Automated questionnaires and data analysis
• Affective computing for Affective disorder

26
Ordinary Medicine

• Centralized Medicine does not Scale
• Distributed Healthcare does Scale
• Pilot is thousands of persons (1K)
• Customary to push down to Individual
• MOS to screen single person (1)
• Revolutionary to push up to Population
• IHM to screen millions of persons (1M)

27
Further Reading

• Richard Berlin and Bruce Schatz
• Population Monitoring of Quality of Life for
  Congestive Heart Failure, Congestive Heart
  Failure, 7(1)13-21 (Jan/Feb 2001).
• G. Rose, The Strategy of Preventive Medicine
• (Oxford University Press, 1992).
• K. Wells, R. Strum, C. Sherbourne, L. Meredith,
  Caring for Depression
• (Harvard University Press, 1996).
• R. Evans, M. Barer, T. Marmor (eds),
  Why are some People Healthy and Others Not? The
  Determinants of Health of Populations
• (New York Aldine de Gruyter, 1990).