Table 3. Logistic Regression Results for Voluntary Disenrollment Rates: Calendar Years (CYs) 1999-2001.
| Independent Variable | Dependent Variable1 | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| 1999 N=111,775 Pseudo R2 = 0.10 |
2000 N=66,0192 Pseudo R2 = 0.13 |
2001 N=131,679 Pseudo R2 = 0.11 |
||||
|
| ||||||
| Odds Ratio | P-Value | Odds Ratio | P-Value | Odds Ratio | P-Value | |
| Beneficiary Choice Descriptor | ||||||
| Plan Origination: Fee-for-Service3 | — | — | — | — | — | — |
| Plan Origination: GHO | 0.865 | 0.000 | 0.965 | 0.244 | 0.929 | 0.000 |
| Plan Origination: NewlyEligible for Medicare Part B in Current CY | 0.758 | 0.000 | 0.666 | 0.000 | 0.838 | 0.001 |
| Plan Origination: Newly Eligible for Medicare Part B in Previous CY | 0.878 | 0.000 | 0.844 | 0.000 | 0.907 | 0.000 |
| Number of Days Enrolled in GHO Prior to Disenrollment (Ranked from 1-5)4 | 0.678 | 0.000 | 0.819 | 0.000 | 0.681 | 0.000 |
| Previously Enrolled in Same GHO | 1.212 | 0.000 | 1.280 | 0.000 | 1.085 | 0.001 |
| First-Ever GHO Enrollment | 0.504 | 0.000 | 0.676 | 0.000 | 0.415 | 0.000 |
| Beneficiary Characteristic | ||||||
| White, Non-Hispanic3 | — | — | — | — | — | — |
| Black, Non-Hispanic | 0.987 | 0.675 | 1.113 | 0.007 | 0.995 | 0.854 |
| Hispanic, Any Race | 1.034 | 0.069 | 1.052 | 0.021 | 1.101 | 0.000 |
| 64 Years<3 | — | — | — | — | — | — |
| 65-69 Years | 1.117 | 0.002 | 0.964 | 0.450 | 1.125 | 0.000 |
| 70-74 Years | 1.090 | 0.021 | 0.914 | 0.069 | 1.087 | 0.005 |
| 75-79 Years | 1.096 | 0.019 | 0.908 | 0.062 | 1.034 | 0.280 |
| 80 Years+ | 1.091 | 0.027 | 0.978 | 0.669 | 0.970 | 0.337 |
| Eligible for Medicaid | 1.472 | 0.000 | 1.922 | 0.000 | 1.709 | 0.000 |
| Eligible for Hospice Benefits | 3.739 | 0.000 | 2.807 | 0.000 | 2.355 | 0.000 |
| Female | 0.966 | 0.062 | 0.994 | 0.007 | 0.930 | 0.000 |
| MA (Plan) Descriptor | ||||||
| For-Profit Plan | 1.458 | 0.000 | 1.396 | 0.000 | 1.809 | 0.000 |
| MA Plan's Enrollment (Continuous Variable) | 1.000 | 0.008 | 1.000 | 0.000 | 1.000 | 0.000 |
| MA Plan's Experience in County (Ranked from 1-4)5 | 1.011 | 0.367 | 1.421 | 0.000 | 1.332 | 0.000 |
| MA Plan's County Penetration Rate (Continuous Variable)6 | 0.924 | 0.000 | 0.936 | 0.000 | 0.973 | 0.000 |
| Market (County) Descriptor | ||||||
| MA Plan Availability in County (Number of MA Contracts Operating in County) | 0.976 | 0.000 | 0.998 | 0.796 | 1.037 | 0.000 |
| Aggregate MA Plan County Penetration Rate (Continuous Variable)6 | 1.017 | 0.000 | 1.003 | 0.133 | 0.992 | 0.000 |
| County Had at Least 1 End-of-Year MA Withdrawal or Service Area Reduction | 0.981 | 0.363 | 0.801 | 0.000 | 1.026 | 0.128 |
Group Health Organization (GHO) enrollment spell ended in a voluntary disenrollment.
The observations included in the regression equation for 2000 are much fewer than in 1999 and 2001 due to a much greater number of missing values for the variable Aggregate Medicare Advantage (MA) Plan County Penetration Rate in 2000. Observations with a missing value for any of the regression variables were eliminated from the estimations.
Comparison group.
1=less than 90 days; 2=90 to less than 180 days; 3=180 to less than 365 days; 4=1 year to less than 5 years; 5=5 years or longer.
Length of time MA contract had been operating in the county as of December 31 of the CY: 1 = less than 1 year; 2 = 1 to less than 5 years; 3 = 5 to less than 9 years; 4 = 9 or more years.
An individual MA Plan's County Penetration Rate = The number of MA enrollees in the MA plan's contract and county divided by the total number of Medicare beneficiaries residing in the county; Aggregate MA Plan County Penetration Rate = The total number of MA enrollees in the county in all MA plans divided by the total number of Medicare beneficiaries residing in the county.
SOURCE: Laschober, M., Mathematica Policy Research: Calculations based on constructed GHO Beneficiary and Spell Level files.