Strategies for Medicaid Care Management Programs September 23, 2008 The 2 nd National Predictive Modeling Summit Linda Shields, RN, BSN, Senior Associate

1 Mercer Predictive Modeling Objectives & Techniques  Identify members that are projected to be high cost in the future for additional interventions, in an effort to reduce their future expenditures  Stratify members by their projected health care needs to be able to determine the appropriate intervention  Identify members that are currently inexpensive and are at the early stages of a disease onset, that would have not been identified by more traditional risk adjustment techniques  The Adjusted Clinical Groups (ACGs) and Diagnostic Cost Groups (DCGs) risk adjustment system have both developed predictive modeling components that are included in their risk adjustment models  Mercer has recently completed several projects that utilized the ACG system to evaluate the efficiency of managed care organizations (MCOs) and Fee for Service populations

2 Mercer Medicaid Case Study  A review of a State’s Fee-for-Service Medicaid population was performed using the ACG model to better understand the underlying population and identify care management opportunities  The ACG system offers multiple measures that can be used to identify subsets of members that would benefit the most from a care management program. These measures include: –Predictive Modeling Score –93 Mutually Exclusive Risk Groups –6 Resource Utilization Bands (RUBs) –Chronic Condition Markers –Co-morbidities –Hospital Dominant Conditions

3 Mercer Predictive Modeling  The PM score represents the probability that an individual will be in the top 5% most expensive members the following year  PM scores range from 0 to 1  A PM score of 0.95 indicates that there is a 95% chance that a member will be among the top 5% most expensive members the next year  Members with a PM score of 0.9 or higher will likely be very expensive the next year, but this score will identify a small number of members  Selecting a lower PM score will identify more members, however some of these members will have lower costs in the following year

4 Mercer Year 1 PM Score High Risk Members: (PM score of 0.6 or higher) Year 2 Utilization Low PM Score in Year 1High PM Score in Year 1 Chronic Condition Total Members Total $ PMPM Inpatient $ PMPM ER $ PMPM Inpatient Days 1,000 PY ER Visits 1,000 PY Total Members Total $ PMPM Inpatient $ PMPM ER $ PMPM Inpatient Days 1,000 PY ER Visits 1,000 PY Arthritis75$584$82$ $1497-$15-1,000 Asthma674$375$80$ $5,066$1,055$769,7312,622 Back Pain366$441$110$266251,20412$1,890$593$572,6562,754 CHF30$1,695$774$135, $2,788$1,555$6317,4551,488 COPD107$642$189$272,0631,18220$1,908$590$364,6081,468 Depression272$809$199$331,1691,49131$1,577$565$575,6922,465 Diabetes192$622$103$237931,0198$2,054$483$406,3081,385 Hyper- lipidemia 185$408$86$ $3,393$1,595$10012,7664,851 Hypertension214$484$153$ $1,946$1,087$775,4403,360 Ischemic HD66$902$265$181, $956$26$381051,579 Renal Failure4$ $2,665$568$503,3101,241 None7,010$255$76$ $1,939$674$203, Total 9,195$318$88$ $2,368$728$516,1232,011

5 Mercer Risk Groups and RUBs  Another alternative is to look at a member’s RUB group assignment  The distribution of members across the 93 risk groups can also be used to evaluate the health status of the members and identify members for care management programs  This comparison can be simplified by looking at the distribution of members across the six Resource Utilization Bands (RUBs)  RUBs group ACGs with similar expected costs

6 Mercer Year 1 RUB Assignment Year 2 Utilization Chronic Condition Non User RUB Administrative RUBLow RUBMedium RUBHigh RUB Very High RUB Arthritis--$270$485$789$1,064 Asthma--$178$329$575$3,279 Back Pain-$31$232$406$620$1,641 CHF---$1,192$1,756$2,994 COPD--$30$488$897$1,285 Depression--$742$663$841$1,759 Diabetes--$663$581$746$1,137 Hyperlipidemia--$169$422$409$1,293 Hypertension--$176$395$554$2,092 Ischemia HD-$946$412$1,299 Renal Failure--$1,300-$2,265- None$199$94$174$402$397$1,093

7 Mercer Chronic Condition Markers & Co- Morbidities  The ACG grouper also identifies members with chronic conditions that are amenable to care management interventions  These chronic condition markers can be used to evaluate the prevalence of chronic conditions within a population  The cost and complexity of caring for a patient with any of these chronic conditions will be affected by the number of co-morbidities that each member has, which will impact their health status  Members with multiple chronic conditions would have a marker for each condition  To avoid counting a member in multiple disease categories, a chronic condition hierarchy was used to assign each member to 1 chronic disease category  The hierarchy that was used to assign members is as follows: — Renal Failure, CHF, COPD, Ischemic HD, Depression, Asthma, Diabetes, Hyperlipidemia, Hypertension, Arthritis, and Low Back Pain

8 Mercer Year 1 Number of Chronic Conditions Year 2 Utilization # of Chronic Conditions # of Members Total $ PMPM Inpatient $ PMPM ER $ PMPM Inpatient Days 1,000 PY ER Visits 1,000 PY 07,034$260$77$ ,456$505$123$ $734$209$281,4591, $866$215$ , $1,041$275$372,1141, $1,387$348$333,6451, $1,546$474$373,5871,304 74$2,166$735$4310,9571,304 81$1,717-$69-2,000 91$ $3,324$1,223-11,000-

9 Mercer Hospital Dominant Conditions  A hospital dominant condition is a diagnosis that has a high probability of requiring the member to be hospitalized in the following year  The higher the number of hospital dominant conditions a member has, the greater their health care needs will be in the following year  The following chart relates a member’s Year 1 number of hospital dominant conditions to their Year 2 expenditures  Members with 1 or more hospital dominant conditions were significantly more expensive the following year

10 Mercer Year 1 Hospital Dominant Conditions Year 2 Utilization # of Chronic Conditions # of Members Total $ PMPM Inpatient $ PMPMER $ PMPM Inpatient Days 1,000 PY ER Visits 1,000 PY 08,960$315$86$ $1,004$237$351,3951, $1,790$709$665,5772, $2,874$1,406$4415,6291,984 45$1,810$1,120$785,0911,455 52$3,493$1,005$1215,4002, $6,690$4,102$3157,0001,000

11 Mercer Combined Risk Index  The combination of PM score, RUB group, number of chronic conditions, and number of hospital dominant conditions can be used to identify a subset of members that will be high cost in the following year  Within each chronic condition category the Combined Risk Index identifies a cohort of significantly more expensive members  Parameters of the Combined Risk Index can vary to identify more members, which will result in less separation between the high and low risk group, or identify a smaller subset that will have greater separation

12 Mercer Year 1 Combined Risk Index Year 2 Health Care Utilization Low PM Score in Year 1High PM Score in Year 1 Chronic Condition Total Members Total $ PMPM Inpatient $ PMPM ER $ PMPM Inpatient Days 1,000 PY ER Visits 1,000 PY Total Members Total $ PMPM Inpatient $ PMPM ER $ PMPM Inpatient Days 1,000 PY ER Visits 1,000 PY Arthritis68$561$59$ $960$223$172, Asthma643$341$73$ $2,788$581$484,6981,735 Back Pain353$397$109$266351,18425$1,732$351$431,4312,215 CHF17$1,372$627$64, $2,563$1,322$4612,8071,238 COPD80$519$139$161, $1,422$455$493,8602,070 Depression248$721$143$309311,40655$1,624$647$564,7552,408 Diabetes178$624$112$248591,02122$1,080$161$262,1031,128 Hyper- lipidemia 171$390$89$ $1,246$411$422,9132,155 Hypertension200$401$90$ $1,795$1,087$375,9431,886 Ischemic HD44$640$186$ $1,265$285$302,7241,215 Renal Failure2$ $2,322$494$432,8801,080 None6,955$252$75$ $1,023$333$242,0901,287 Total 8,959$297$81$ $1,621$508$393,8691,699

13 Mercer Care Management Applications  Risk scores can be used to identify members with high predicted concurrent and prospective scores. These members can be expected to be high-cost now and into the future  ACG and RUB groups can be used to identify members with multiple significant health problems  Predicted modeling scores identify members who are predicted to be high- cost in the annual time period following the risk assignment period  EDC groups can be used to identify members with chronic conditions that will likely need services in the future  Hospital dominant conditions identify members, who will likely require hospitalizations in the near future  Combinations of these factors can be used to create a Care Management Profile which identifies members who will likely have high health care utilization in the future  Helps to identify specific patients at risk and to develop appropriate interventions to both improve clinical outcomes and potentially avoid or decrease future utilization patterns and costs

14 Mercer Care Management Profile Examples Profile AreaCase 1Case 2 Age4740 GenderMaleFemale Risk Score Predictive Modeling Score0.93 Hospital Dominant Conditions22 FrailtyNoYes ArthritisNo AsthmaYesNo Congestive Heart FailureYes Chronic Renal FailureNoYes Congestive Obstructive Pulmonary DiseaseNo DepressionNoYes DiabetesYesNo HyperlipidemiaYesNo HypertensionYes Ischemic Heart DiseaseYesNo Low Back PainNo

15 Mercer Factors to Consider When Selecting Disease Category  Prevalence rates of disease conditions  Service utilization levels and costs associated with each condition  Existence of evidence-based treatment guidelines  Generally recognizable problems in therapy documented in the literature or large variation in practice  Large number of patients exists whose therapy could be improved  Preventable acute events  The potential of cost savings within a relatively short period  The ability of behavior change to impact the disease conditions

16 Mercer Considerations when Choosing a Care Management Program  Each program may be used by itself or in combination with any other  Individual components within each program should be selected for use based upon program goals and available resources  The largest opportunities to achieve substantial and early cost savings lie in decreasing ER usage, inpatient admissions, readmissions or length of hospital stays  Care improvements exist in implementing strategies that decrease member disease burden, elicit member behavior change and support compliance with evidence-based guidelines

17 Mercer Top 10 Disease Conditions Identified As Most Prevalent in Year 2 (Members with a Risk Score of >.60)  Low Back Pain  Asthma  Hypertension  Hyperlipidemia  Depression  Arthritis  Diabetes  Ischemic Heart Disease  Congestive Obstructive Pulmonary Disease  Congestive Heart Failure  Chronic Renal Failure

18 Mercer Disease Focus: Why Asthma?  Clinical Guidelines – Nationally Recognized & Accepted – Readily Available  Volume – Largest # Members – Greatest %  Dollars – Total PMPM approx. $600  Impactable – ER Usage – Avoid Triggers – Medication Management  Short Term Return – Manage Costs – Improve Outcomes

19 Mercer Member Complexity When considering Care Management strategies it is essential to understand clinical relationships, interactions and frequency of conditions within the targeted population.

20 Mercer Managing Comorbidities

21 Mercer Health Risk Assessment Self Care Mailers Strategies for Managing Increasing Member Complexity Case Management Disease Management Self Management Training Population Health Management Targeted Risk Assessment High Disease Burden Single High Impact Disease Users Users & Non-Users Predictive Modeling Decision Support Nurse Advice Line Population Segment Multiple Chronic Conditions High Cost/High Use Low Level Use for Minor Conditions & Potential for Risk Factors Unknow n Risk Factors

22 Mercer What is Disease Management? “Disease Management is a system of coordinated health care interventions and communications for populations with conditions for which patient self-care efforts are significant.” –-Disease Management Association of America (DMAA)

23 Mercer Typical Disease Management Programs  Asthma  Chronic Obstructive Pulmonary Disease  Congestive Heart Failure  Ischemic Heart Disease  Diabetes  Depression  Anxiety  Hypertension  Hyperlipidemia

24 Mercer Disease Management Components for Success  Decreasing treatment variability  Closing the gap between current treatment patterns and optimal treatment guidelines  Provider adherence to nationally accepted guidelines  Clinical pathways available to direct interventions  Appropriate adjustments are made to guidelines to account for multiple co-morbid conditions or unique member situations  Guidelines, translated into layman’s language, are shared with members as a means of supporting self-care behaviors  Member & Provider Buy In

25 Mercer What is Case Management? “Case management is a collaborative process of assessment, planning, facilitation and advocacy for options and services to meet an individual's health needs through communication and available resources to promote quality cost-effective outcomes.” –-Case Management Society of America (CMSA)

26 Mercer Typical Cases Managed  Terminally Ill (Cancers)  Major Trauma (Accidents, Loss of Limb, Traumatic Brain Injury)  Physical Disability (Quadriplegia, Spina Bifida)  Fatal Conditions (HIV/AIDS)  Sudden Event (MI, Stroke)  Chronic Conditions (CHF, Asthma, Diabetes)  High Risk (Pregnancies, Preemies)  Complex Cases (Comorbidities, Psycho/Social/Economic Issues)  Transplants (Organ, Skin, Corneal)

27 Mercer Case Management Success  Decreased Utilization  Improved Clinical Conditions  Provider & Member Buy In  Collaboration Across Disciplines  Financial Savings primarily achieved through coordination of interventions among complex care providers & benefit management

28 Mercer Key Principles: Total Health Management  Address entire health care continuum  Everyone in Population  Emphasize Long-Term Behavioral Change & Risk Modification  Data Driven Programs  Not limited to single disease condition

29 Mercer Health Care Continuum

30 Mercer Behavioral Modification

31 Mercer Stages of Change CDC–Strategy of Change

32 Mercer

33 Mercer Impact of Risk Factors  Those with Lifestyle Risk Factors cost 10% - 70% more than those not at risk  Managing risk factors can: — Decrease the disease burden to the individual — Improve quality outcomes — Decrease the consumption of costly resources

34 Mercer Methodology: Managing Risk Factors

35 Mercer Member’s Involvement & Buy In Necessary  Active participation  Understand the importance of compliance with the treatment plan  Understand their condition  Identify and avoid trigger points  Reduce Risk Factors  Utilize tools and self-help materials provided to assist in taking an active role in self-care

36 Mercer Medicaid Specific Barriers to Care  Transportation  Language  Literacy Level  Medical Literacy  Knowledge Gaps  Economic Issues  Lack of Technology  Demographics/Locating the Member  Provider Reimbursement

37 Mercer Recommendations: Option #1 Disease Management Program

38 Mercer Option #2: Proactive Care Management Program  Traditional health care management focused on treating existing illness or disease. Proactive Care Management focuses interventions along the health care continuum from optimal health to illness  Options include building a program, contracting with a vendor to provide a program or a combination of building, and outsourcing/assembly  Program strives to proactively teach self-help behaviors that promote health, decrease development of risk factors, avoid behaviors that trigger acute events and help avoid disease development or to slow disease progression  For proactive care management programs to be successful, a careful analysis of the required skills and resources must occur  Due to the focus on prevention, behavioral change, and compliance with evidence-based guidelines additional resources not currently in place may be required

39 Mercer Indicators of Success  HEDIS &/or HEDIS-like Scores  Client Specific Goals  Enrollment  Satisfaction – Member – Provider  Utilization of Resources – ER – Inpatient – Rx

40 Mercer Currently In Progress  Care Management Program Gap Analysis  Systems Review  Evidence-based practice guidelines  Provider Education  Review practice models  Analysis of Routine reporting/feedback loop  ER Strategy