The Impact of an Innovative HRA-based Wellness & Prevention - - PowerPoint PPT Presentation

The Impact of an Innovative HRA-based Wellness & Prevention Demonstration on Claims-Based Health Outcomes Session: New Insights in Medicare Spending and Outcomes Academy Health, June 24 2013, Baltimore Betty Tao Fout, PhD and Daniel Weinberg, MBA, Ph.D. IMPAQ International

Presenter Disclosures

2

Betty Fout, Ph.D. IMPAQ International

The research in this article was supported by the Centers for Medicare & Medicaid Services under Contract Number HHSM-500-2006-00007I-07 with IMPAQ International, LLC. The views expressed in this article are those of the author and do not necessarily reflect the views of the U.S. Department of Health and Human Services, the Centers for Medicare & Medicaid Services or IMPAQ International.

Senior Risk Reduction Demonstration(SRRD)

• Three-year CMS program to reduce risk and manage

health using health risk assessments (HRAs) and tailored feedback

• Two Vendors (A and B) with private sector HRA + risk

reduction program experience

• Target population
• FFS Medicare beneficiaries enrolled in Part A and B
• Between the ages of 67 and 74 at program onset
• Not enrolled in hospice or have ESRD
• Original entitlement due to age, not disability

3

Study Design Overview

4

Intervention Group Administrative control group (no contact) Submit HRA No response

• r opt-out

Standard Treatment Enhanced Treatment Generic mailing RANDOMIZE Eligible Medicare FFS Beneficiaries RANDOMIZE

For each vendor:

• National sample
• Two local samples

Vendors’ Risk Reduction Programs (vary by individual risk and treatment arm)

Study Design Overview, Continued

• National sample and two local samples per vendor
• Local samples - partnered with regional Area Disability

and Resource Centers (ADRCs)

• Vendors used proprietary risk algorithms to triage

participants and tailor intervention

• Demonstration period: May 2, 2009 to April 30, 2012

(three years)

• Vendor A participated in the first two years only
• Vendor B participated for all three years

5

SRRD Evaluation

• Implementation analyses: site visits to vendors and

interviews

• Impact analyses
• Link study beneficiaries to their Medicare claims history
• Baseline period – January 1, 2008 to April 30, 2009
• Examined the impact of SRRD on certain claims-based
• utcomes (purpose of this paper)
• Examine the impact of SRRD on HRA risk variables
• Examine whether the program was “budget neutral”

6

Scope of Interim Claims-based Impact Analysis

• “Cast a wide net” to detect impact
• A number of outcome measures from various domains
• A number of SRRD subgroups
• “Initial” evaluation
• After the first year of data available
• Will inform final analysis which will include three years
• f outcome data

7

Methodology

• Given the RCT design, simply comparing means

across treatment and control groups should yield unbiased impact estimates

• Regression approach was chosen to
• Get more precise impact estimates
• Control for still existing differences
• Robustness checks
• Dealing with extreme values
• Estimating impacts using alternative methods

8

Methodology – cont.

• Treatment – Control groups of primary interest:
• Standard Intervention (Arm1) to HRA-only (Arm 3)
• Enhanced Intervention (Arm 2) to HRA-only (Arm 3)
• Baseline period: January 1, 2008 to April 30, 2009
• Outcome period: May 1, 2009 to April 30, 2010
• Regression approach controls for baseline

differences

• Age, Gender, Race, Dual Eligibility
• Baseline value of the dependent variable
• Baseline value of total Medicare payments

9

Participant Counts by Treatment Arm

10

Vendor A Vendor B Year 1 Year 2 Re-Enrolleesa Year 1 Year 2 Re-Enrolleesa Arm 1 2,489 1,833 1,365 2,458 1,792 1,299 Arm 2 2,495 1,797 1,326 2,459 1,722 1,232 Arm 3 2,489 2,203 1,742 2,458 2,115 1,623 All Arms 7,473 5,833 4,433 7,375 5,629 4,154

a Re-enrollees returned HRAs in both Year 1 and Year 2.

Baseline Characteristics - Vendor A National

Arm 2 Arm 3 Arm 2 – Arm 3 Medicare Expenditures Total Medicare Payments $7,966 $7,433 $533 Inpatient Payment $3,040 $2,663 $376 Outpatient Payment $1,374 $1,362 $12 Carrier Payment $2,634 $2,600 $34 ED Payment $114 $106 $8 Medicare Use Any Inpatient 16.3% 16.4%

• 0.07%

Inpatient Length of Stay 1.48 1.19 0.29* Any Outpatient 75.9% 76.2%

• 0.26%

Outpatient Days 4.8 4.8 0.0 Carrier Days 21.4 20.9 0.5 ED Visits 0.47 0.43 0.04 Health Status HCC Risk Score (2009) 0.85 0.85 0.00 HRA Overall Risk Variable % High Risk 7.6% 7.3% 0.34% % Medium Risk 27.4% 28.6%

• 1.18%

% Low Risk 65.0% 64.1% 0.85%

11

Statistical significance is shown at the 1% (***) , 5% (**) and 10% (*) levels.

Baseline Characteristics - Vendor B National

Arm 2 Arm 3 Arm 2 – Arm 3 Medicare Expenditures Total Medicare Payments $7,310 $7,801

• 490

Inpatient Payment $2,685 $3,071

• 386

Outpatient Payment $1,351 $1,327 24 Carrier Payment $2,645 $2,674

• 29

ED Payment $97 $92 6 Medicare Use Any Inpatient 17.0% 18.7%

• 1.8%

Inpatient Length of Stay 1.2 1.3

• 0.2

Any Outpatient 77.5% 76.8% 0.7% Outpatient Days 4.9 4.9 0.0 Carrier Days 20.65 21.07

• 0.41

ED Visits 0.4 0.4 0.0 Health Status HCC Risk Score (2009) 0.85 0.84 0.00 HRA Overall Risk Variable % High/Medium Risk* 29.1% 30.1%

• 1.1%

% Low Risk 70.9% 69.9% 1.1%

12

*Vendor B’s overall risk variable did not allow for distinguishing between high and medium risk categorizations.

National Arm 2 Impacts – Vendor A

Impact Arm 2 mean Arm 3 mean % Impact Medicare Payment Total Medicare Payments

• 958**

5,807 6,765

• 14.2%

Inpatient Payment

• 491

2,090 2,581

• 19.0%

Outpatient Payment

• 280***

915 1,196

• 23.4%

Carrier Payment

• 200*

2,037 2,237

• 8.9%

ED Payment 2 87 85 1.9% Medicare Use Any Inpatient

• 2.1%**

12.7% 14.8%

• 14.4%

Inpatient Days

• 0.3*

0.9 1.2

• 24.6%

Any Outpatient

• 2.2%*

70.5% 72.6%

• 3.0%

Outpatient Days

• 0.2*

3.6 3.8

• 6.5%

Carrier Days

• 0.6

16.3 16.9

• 3.5%

ED Visits 0.0 0.3 0.3 0.6% Preventive Screening Colorectal Cancer 1.7% 24.3% 22.7% 7.3% Breast Cancer

• 0.5%

60.3% 60.7%

• 0.8%

Cardiovascular 0.4% 66.0% 65.5% 0.7%

13

Statistical significance is shown at the 1% (***) , 5% (**) and 10% (*) levels.

National Arm 2 Impacts – Vendor B

14

Impact Arm 2 mean Arm 3 mean % Impact Medicare Payment Total Medicare Payments 519 7,034 6,516 8.0% Inpatient Payment 405 2,835 2,430 16.7% Outpatient Payment 34 1,177 1,143 3.0% Carrier Payment 54 2,248 2,194 2.5% ED Payment

• 0.5

84.8 85.3

• 0.6%

Medicare Use Any Inpatient

• 0.2%

14.3% 14.5%

• 1.1%

Inpatient Length of Stay 0.2 1.3 1.1 17.1% Any Outpatient 1.4% 73.5% 72.0% 2.0% Outpatient Days

• 0.1

3.7 3.8

• 2.7%

Carrier Days 0.4 17.1 16.7 2.2% ED Visits 0.0 0.3 0.3 1.9% Preventive Screening Colorectal Cancer 0.6% 23.4% 22.9% 2.5% Breast Cancer 0.0% 60.7% 60.6% 0.0% Cardiovascular 1.0% 66.4% 65.4% 1.5%

Findings

• First year findings show that Vendor A’s Enhanced National

program “reduces” next year’s total Medicare expenditures by $958 or 14.2%

• Almost half of “savings” come from inpatient expenditures. Other

half come from outpatient and carrier expenditures

• Estimates of utilization impacts are consistent
• Significant reduction in any hospital stay (14.4%) and number of inpatient

days (24.6%)

• Modest reduction in any outpatient admission (3.0%) and number of
• utpatient days (6.5%)
• No impacts on ED visits or preventive screening use
• No impacts detected for Vendor B’s Enhanced National program

15

Interpretation/Conclusions

• Vendor A’s Enhanced National program impacts too high?

Too soon? No early increase in utilization.

• Robustness Checks
• Outliers (not an issue)
• Alternative estimators (similar results)
• Groups driving impacts (high-risk groups)
• “Placebo” effects? (YES!!)
• There appears to be an impact of just filling out an HRA

(Arm 3) – increase in utilization, at least in the first year

• Being in Arm 2 brings back that utilization to where it was
• Thus, there actually is no impact once we take “placebo”

effect into account

16

Lessons Learned

• Program/Evaluation Design:
• Just filling out a Health Risk Appraisal may alter health

care seeking behavior

• Oversampling high risk populations may be important

(dual eligibles, minorities, etc.)

• Program Design:
• Customizing by risk level (impactability) may be critical
• Reaching out to hard-to-reach may be critical
• Analyze additional years of data – forthcoming

final impact analyses

17

Further Information

• Interim evaluation report is available at:

http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and- Reports/Reports/Downloads/SRRDEvaluation_FirstYearReport_03212012_IMPAQ.pdf

• For more information, contact Dr. Pauline Karikari-Martin

(CMS project officer) at:

pauline.karikarimartin@cms.hhs.gov

18