Quantifying the Risk of Disability and Death using Medical Claims Data US patent 7,249,040 and paten - PowerPoint PPT Presentation

Loading...

PPT – Quantifying the Risk of Disability and Death using Medical Claims Data US patent 7,249,040 and paten PowerPoint presentation | free to download - id: 3f732-ODNiZ



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Quantifying the Risk of Disability and Death using Medical Claims Data US patent 7,249,040 and paten

Description:

Clinical profiling for disease management ... Cardiovascular. Disease ... No Cardiovascular. 12. Bayes' Theorem. Example for Cardiovascular Disease and LTD ... – PowerPoint PPT presentation

Number of Views:61
Avg rating:3.0/5.0
Slides: 30
Provided by: gregb84
Learn more at: http://www.ehcca.com
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Quantifying the Risk of Disability and Death using Medical Claims Data US patent 7,249,040 and paten


1
Quantifying the Risk of Disability and
Death using Medical Claims Data (US patent
7,249,040 and patents pending) Presented
atThe National Predictive Modeling
SummitDecember 13, 2007
  • Greg Binns, PhD

2
Overview
  • Challenges
  • Enhanced Risk Selection
  • Loss Ratio Analysis and Model Validation
  • Clinical profiling for disease management
  • Summary and discussion

3
Challenges for Disability and Life
  • Risk selection
  • Manual rateslittle discrimination
  • Experiencelittle credibility and the Lexian PDF
    implies credibility worse than you thought
  • Competitionwild variability in pricing the same
    case
  • Risk management
  • Pricing multiple lines
  • Clinical profiling for disease management
  • Solutionuse clinical information from medical
    claims for more accurate forecasts, pricing and
    DM

4
Enhanced Risk Selection
5
StrategyWinning by Changing the Rules
  • Using better information (all medical claims and
    diagnoses)
  • Forecasting claim cost more accurately using
    proprietary Clinical/Statistical Models
  • Modifying the distribution systemreview all
    groups in medical plan or TPA then quote on
    groups with the greatest profit potential

6
More Accurate Risk SelectionAll Lines, All
Groups
CUSTOMIZED CLINICAL STRUCTURE by Line
DECISION SUPPORT STOP LOSS
CAPCOST and FIRST DOLLAR
INTEGRATED PERSON AND GROUP DATA
RAW DATA
CLINICAL/ STATISTICAL ANALYTICS
SPECIFIC AGGREGATE STOP LOSS
LIFE
STD
Underwriting/ Decision Support Products
LTD
Data and Process Steps
7
Paradigm Shift
  • Evaluate risk and target favorable groups using
    Clinical/Statistical Models
  • Provide more accurate pricing of Disability and
    Life for medical customers
  • Lower loss ratio and its variability
  • Cross-sell with first dollar or medical stop loss
    coverage
  • Lower future riskstarget high risk Disability
    and Life employees for disease management

8
Evaluating Disability and Life Risk
  • Medical claims and eligibility data required for
    cases to be underwritten
  • Medical and Disability or Life claims do not need
    to be linked at the person or group level for
    model developmentkey breakthrough
  • Different Clinical/Statistical Models required
    for different insurance products
  • Compare clinical risk to demographic and
    experienceClinical/Demographic Ratio

9
Clinical/Statistical Models
  • Benefits of medical underwriting without the cost
    or intrusion
  • Far greater range in the person-level estimate of
    incidence rates and severity
  • Direct estimate of future riskforward looking
  • Clinical profile for disease management

10
Chain of Events for Disability/DeathNo Clinical
Condition Condition Develops Diagnosis
and Treatment (usually) Disability/Death
Probability LTD Given Lung CA .09
Probability Lung CA .0001
Female 45-49 Unknown Clinical Conditions
Female 45-49 Develops Lung Cancer
Female 45-49 Disabled with Lung Cancer
Probability of Disability.004
11
Need Probability of LTD claim, Given Medical
Condition (e.g., Cardiovascular)Probability
(LTD claimCardiovascular) Prob(LTD claim
Cardiovascular)/ Prob(Cardiovascular)
No LTD No Cardiovascular
Cardiovascular Disease
LTD Claims
LTD Cardiovascular Overlap
12
Bayes Theorem Example for Cardiovascular
Disease and LTD
  • Probability of (LTD Claim Cardiovascular)
  • Prob(Cardiovascular LTD Claim) Prob(LTD
    Claim)/
  • Prob(Cardiovascular)

13
Life Clinical vs. Demo ModelsGroup Forecast
Accuracy Improves due to Increased Precision at
Person Level
14
LTD Clinical vs. Demo ModelsGroup Forecast
Accuracy Improves due to Increased Precision at
Person Level
15
Preliminary Validation for Life
  • Groups with High Clinical/Demo Ratios have much
    higher actual death rates than Low Clinical/Demo
    Groups but similar Demographic Risk

16
Preliminary Validation for LTD
  • Groups with High Clinical/Demo Ratios have 55
    Greater Experience/Manual Ratio than Low
    Clinical/Demo Groups

17
Group Level Clinical Risk About 1/3
Groups 10 Over, 1/3
Groups 10 Under Demo Average
18
Potential Profit Impact
  • Based on one clients LTD data for cases under
    1,000 lives
  • Avoiding the worst 5 of cases would result in
    increasing margins from 14 to 36
  • Assume avoid ½ of bad groups, margin becomes 25
    or 11 increase
  • Profit improved by targeting groups with low
    clinical risk compared to demographic risk
  • 21 groups have clinical/demo ratiolt.80
  • 10 premium reduction gives clinical loss
    ratio(.8/.9)(current loss ratio).89 or lower
    of current

19
Potential Profit Impact (cont.)
  • Life validation
  • Groups with Clinical/Demo Risklt2.0
  • 75 claims
  • 86 premium
  • Implies 13 reduction in current Loss Ratio
    (.75 claims)/(.86 premium)
    (current LR) 87 current LR
  • Groups with Clinical/Demo Riskgt2.0
  • 25 claims
  • 14 employees or premium in those groups
  • Implies 79 increase over current Loss Ratio
  • 10 reduction in loss ratio is target for LTD and
    Life

20
Pricing Strategy
  • Current manual and process flow remain as
    foundation for underwriting
  • Blend Clinical/Demographic Ratio into pricing
    using credibility theoryinclude experience if
    reasonable credibility
  • Pricing considerations
  • Price sensitivity and persistency rates
  • Competitors and their strategy
  • New vs. renewal for ancillary linesnote all
    groups are medical renewals due to data
    requirements
  • Discounts for multiple ancillary lines

21
Example Pricing Grid LTD Clinical/Demo Ratio
22
Pricing ConsiderationsCorrelations between Lines
23
Clinical Profiling for DM
24
Example Clinical Profile for Disease
ManagementSummary
25
Example Clinical Profile for Disease
ManagementMental Disorders
26
Example Clinical Profile for Disease
ManagementMusculoskeletal Disorders
27
Summary
  • Medical plans will have huge competitive
    advantage
  • Superior risk selection using medical claims
  • Cross-sell with medical or stop loss coverage
  • Cash flowLife and LTD premium about 3 medical
  • High persistency rates favor incumbentchange is
    slower than anticipated but inevitable
  • Future risk mitigation through disease management

28
Additional Topics
  • Privacy issues
  • Individuals
  • Groups
  • Modeling other lines
  • Other?

29
Thanks
  • Greg Binns, PhD
  • greg.binns_at_TruRisk.com
  • 847.295.2891 phone
  • 847.295.2892 fax
About PowerShow.com