Table 2. Association of Star Rating and MAPD Contract Enrollment Before and After 2012 (Panel Data Hybrid Model)a.
| Enrollment in concurrent year (t) | Enrollment in subsequent year (t+1) | |||
|---|---|---|---|---|
| Estimate (95% CI) | p-value | Estimate (95% CI) | p-value | |
| Pre-bonus payment period (t = 2009 to 2011) | ||||
| Between effect: Star rating difference between contracts | 7,858.8 (-335.2, 16,052.7) | 0.060 | 9,280.1 (364.2, 18,195.9) | 0.041 |
| Within effect: Star rating change within a contract | 148.4 (-4,734.0, 5,030.9) | 0.952 | 692.5 (-2,109.4, 3,494.4) | 0.628 |
| Post-bonus payment period (t = 2012 to 2015) | ||||
| Between effect: Star rating difference between contracts | 8,322.2 (153.7, 16,490.7) | 0.046 | 10,849.9 (967.9, 20,731.8) | 0.031 |
| Within effect: Star rating change within a contract | 3,558.5 (-55.9, 7,172.9) | 0.054 | 11,337.1 (6,025.4, 16,648.8) | <0.001 |
| Pre-post combined period (t = 2009 to 2015) | ||||
| Between effect: Star rating difference between contracts | 8,088.5 (287.1, 15,889.8) | 0.042 | 9,749.6 (798.8, 18,700.4) | 0.033 |
| Within effect * Posta | 1,594.4 (-4962.6, 8,151.4) | 0.634 | 9,136.9 (2,106.7, 16,167.1) | 0.011 |
CI, confidence interval; MAPD, Medicare Advantage Prescription Drug plan.
aPost = 1 if year was from 2012 to 2015 and Post = 0 if year was from 2009 to 2011. Estimates were based on random effects models adjusted for contract-level repeated measures and controlling for contract types (health maintenance organization [HMO], point-of-service [POS], local provider organization [PPO], private fee-for-service [PFFS], or regional PPO), contract maturity (how many years the contract has been in existence, defined as the time elapsed between the year the contract became effective and the study year), and the lagged (t-1 year) variables of total number of plans in the contract, proportion of plans in the contract that offered Part D coverage, proportion of plans in the contract that were Special Needs Plans (i.e., those that limit membership to people with specific diseases or characteristics, and tailor their benefits, provider choices, and drug formularies to best meet the specific needs of the groups they serve), and the proportion of plans in the contract that were employer group health plans in the prior (t-1) year, and year dummy variables. “Within effect * Post” is the interaction term between “Within effect” and “Post.”