|
|
|
|
|
Arlene S. Ash, PhD, School of Medicine, Dept. of Medicine, Section of General Internal Medicine, Boston University, Health Care Research Unit, 720 Harrison Ave. #1108, Boston, MA 02118, 617 638 7518, aash@bu.edu, Marilyn Schlein Kramer, MBA, DxCG, Inc., 25 Kingston St. Suite 200, Boston, MA 02111, Randall Ellis, PhD, Economics Department, Boston University, 270 Bay State Road, Boston, MA 02215, and Yang Zhao, PhD, Global Health Outcomes Research, Eli Lilly and Company, Lilly Corporate Center, DC 1833, Indianapolis, IN 46285.
Background: Starting in 2006, the Center for Medicare and Medicaid Services (CMS) will subsidize private, competitive pharmaceutical insurance plans for voluntary purchase by Medicare beneficiaries. The standard benefit is complex and CMS must certify that alternative benefits are “actuarially equivalent.” To “level the playing field” across plans that attract enrollees with different levels of need, CMS’s subsidy, forecast to be 76% of all costs, is mandated to be risk adjusted. Because the Medicare drug benefit is new, neither actual levels of drug spending, nor the detailed information required to accurately predict that spending, are available in CMS’s transaction files.
Methods: We use Medstat’s MarketScan Medicare Supplemental & COB Research Data from 1998 through 2000 (n~700,000) to examine the distribution of “year-2” drug costs and the explanatory power of various “year-1” predictors of these costs. We examine the implications of our findings for CMS’ obligations for implementing risk adjustment.
Results: Annual drug costs averaged $1472 and are highly skewed. Age and sex alone or with year-1 diagnoses are far less predictive than knowing the pattern of year-1 drug utilization (R2s increase from 1% to 45%). Even with the proposed 76% subsidy, the voluntary drug benefit is a poor actuarial bet for many people with very low year-1 drug utilization.
Conclusions: Medicare drug spending is highly variable and (given the right data) highly predictable. The fact that the government has far less powerful information for predicting costs than private insurers and individual beneficiaries poses serious challenges for successfully implementing the new Medicare drug benefit.
Learning Objectives:
Keywords: HCFA, Healthcare Costs
Presenting author's disclosure statement:
I have a significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.
Relationship: I have an equity interest in DxCG, Inc. whose models are used in this research. I also receive royalties and salary from DxCG, Inc.