Association of co-pay elimination with medication adherence and total cost.

OBJECTIVES: To determine whether elimination of co-pays for prescription drugs affects medication adherence and total health care spending. STUDY DESIGN: Retrospective comparative study. METHODS: We conducted a difference-in-differences comparison in the year before and after expansion of a Zero Dollar Co-pay (ZDC) prescription drug benefit in commercially insured Louisiana residents. Blue Cross and Blue Shield of Louisiana members with continuous disease management program enrollment were analyzed, of whom 6463 were enrolled in the ZDC program and 1821 were controls who were ineligible because their employers did not opt in. RESULTS: After ZDC expansion, medication adherence fell in the control group and rose in the ZDC group, with a relative increase of 2.1 percentage points (P = .002). Medical spending fell by $71 per member per month (PMPM) (P = .027) in the ZDC group relative to controls. Overall, there was no significant increase in the cost of drugs between treatment and controls. However, when drugs were further categorized, there was a significant increase of $8 PMPM for generic drugs and no significant difference for brand name drugs. Comparisons of medication adherence rates by household income showed the largest relative increase post ZDC expansion among low-income members. CONCLUSIONS: Elimination of co-pays for drugs indicated to treat chronic illnesses was associated with increases in medication adherence and reductions in overall spending of $63. Benefit designs that eliminate co-pays for patients with chronic illnesses may improve adherence and reduce the total cost of care.

Value-based insurance design in Louisiana: Blue Cross Blue Shield's Zero Dollar Co-pay program.

OBJECTIVES: To determine whether a program that eliminated pharmacy co-pays, the Blue Cross Blue Shield of Louisiana (BCBSLA) Zero Dollar Co-pay (ZDC) program, decreased health care spending. Previous studies have found that value-based insurance designs like the ZDC program have little or no impact on total health care spending. ZDC included an expansive set of medications related to 4 chronic diseases rather than a limited set of medications for 1 or 2 chronic diseases. Additionally, ZDC focused on the most at-risk patients. STUDY DESIGN: ZDC began in 2014 and enrolled patients over time based on (1) when a patient answered a call from a nurse care manager and (2) when a patient or their employer changed the benefit structure to meet the program criteria. During 2015 and 2016, 265 patients with at least 1 chronic condition (asthma, diabetes, hypertension, mental illness) enrolled in ZDC. METHODS: Observational study using within-patient variation and variation in patient enrollment month to identify the impact of the ZDC program on health spending measures. We used 100% BCBSLA claims data from January 2015 to June 2018. Monthly level event studies were used to test for differential spending patterns prior to ZDC enrollment. RESULTS: We found that total spending decreased by $205.9 (P = .049) per member per month, or approximately 18%. We saw a decrease in medical spending ($195.0; P = .023) but did not detect a change in pharmacy spending ($7.59; P = .752). We found no evidence of changes in spending patterns prior to ZDC enrollment. CONCLUSIONS: The ZDC program provides evidence that value-based insurance designs that incorporate a comprehensive set of medications and focus on populations with chronic disease can reduce spending.

Evaluating the performance of a predictive modeling approach to identifying members at high-risk of hospitalization.

Aims: To evaluate the risk-of-hospitalization (ROH) models developed at Blue Cross Blue Shield of Louisiana (BCBSLA) and compare this approach to the DxCG risk-score algorithms utilized by many health plans.Materials and Methods: Time zero for this study was December 31, 2016. BCBSLA members were eligible for study inclusion if they were fully insured; aged 80 years or younger; and had continuous enrollment starting on or before June 1, 2016, through time zero. Up to 2 years of historical claims data from time zero per patient was included for model development. Members were excluded if they had cancer, renal failure, or were admitted for hospice. The Blue Cross ROH models were developed using (1) regularized logistic regression and (2) random decision forests (a tree ensemble learning classification method). All models were generated using Scikit-learn: Machine Learning in Python. Prognostic capabilities of DxCG risk-score algorithms were compared to those of the Blue Cross models.Results: When stratifying by the top 0.1% of members with the highest ROH, the Blue Cross logistic regression model had the highest area under the receiving operator characteristics curve (0.862) based on the result of 10-fold cross-validation. The Blue Cross random decision forests model had the highest positive predictive value (49.0%) and positive likelihood ratio (61.4), but sensitivity, specificity, negative predictive values, and negative likelihood ratios were similar across all four models.Limitations: The Blue Cross ROH models were developed and evaluated using BCBSLA data, and predictive power may fluctuate if applied to other databases.Conclusions: The predictability of the Blue Cross models show how member-specific, regional data can be used to accurately identify patients with a high ROH, which may allow healthcare workers to intervene earlier and subsequently reduce the healthcare burden for patients and providers.