Type 2 diabetes (T2D) is prevalent and costly in the United States. In 2021, more than 38 million people, 11.65% of the population, in the United States had T2D.¹ The national healthcare system and patients with T2D were facing significant economic burden.² The estimated direct medical cost and health productivity losses in 2022 were $307 billion and $106 billion, respectively.³

Excessive diabetes healthcare costs are often associated with medication nonadherence.⁴ Fukuda et al⁴ highlighted that among newly diagnosed patients with T2D, medication adherence was associated with significantly higher cumulative healthcare costs, ranging from $937 to $44 673.⁵ More so, better medication adherence is typically associated with healthcare cost savings. For instance, Polonsky et al⁵ found that annual medical spending per patient with T2D decreased by $4413 for adherent patients—medication possession ratio was 80% or higher. This reduction increased to $5170 for individuals aged 65 years or older. Similarly, Axon and colleagues reported that adherent patients incurred 14.7% lower total healthcare expenditures compared with nonadherent patients.⁶ Furthermore, several other diabetes studies have reported that medication adherence at the population level can lead to healthcare cost savings amounting to billions of dollars, further highlighting the critical role of medication adherence in reducing healthcare expenditures.^7,8

Cost-associated medication nonadherence may be overcome by value-based insurance designs (VBIDs). In a study investigating the predictors of medication adherence, the authors found that medication adherence was negatively related to cost sharing, for every $1 decrease in cost-sharing, the proportion of days covered would increase by 1.1 days. In addition, they reported that better medication adherence was associated with fewer healthcare utilizations and lower total healthcare expenditures.⁹ Similarly, another study observed that a VBID offering copayment reductions and eliminations for certain cholesterol-lowering medications increased medication adherence by 2.8%.¹⁰ Comparable findings have been reported by many chronic conditions.^11–14 Furthermore, VBIDs designed to reduce cost-sharing can result in significant healthcare cost savings.^15–18

In 2020, Blue Cross and Blue Shield of Louisiana (BCBSLA) implemented a Zero-Dollar Co-pay (ZDC) program, aimed at alleviating financial burdens, addressing cost-related medication nonadherence, reducing healthcare utilization, and improving clinical outcomes, such as Hemoglobin A1c for eligible beneficiaries. The program operated by eliminating all copayments for antidiabetic medications, which were primarily generic drugs. The ZDC program launched in 2020 expanded the eligible cohort based on the previous iterations in 2014 and 2018 by applying less strict eligibility criteria.^18,19 This study aims to evaluate whether the ZDC program can reduce total healthcare expenditures among BCBSLA fully insured members with T2D. It contributes to the existing literature by examining the effects of a copayment elimination program, whereas prior studies have more commonly assessed copayment reduction strategies. By focusing on the complete removal of copayments, this study provides new evidence on the potential of financial incentives to influence patient behavior. Unlike interventions that incorporate educational or counseling components, the ZDC program was implemented as a standalone financial intervention, enabling a clearer assessment of its independent impact. To enhance the validity of our findings, we applied a 2-way fixed effects difference-in-differences design with balanced baseline characteristics between the intervention and comparison groups, allowing for a more robust causal interpretation.

We identified 4530 members in the ASO group as the control group and 3035 members in the ZDC group. After the weighting process, the 2 groups have comparable baseline characteristics (Table 1). The average ages were 52.9 years old (standard deviation [SD] = 11.27) and 48.75 years old (SD = 12.44) for the control and the ZDC group members, respectively. Among the control group, 55.61% of the members were female, compared with 57.43% in the ZDC group. Additionally, 80.86% of the control group and 77.07% of the ZDC group resided in urban areas. Hypertension was the most prevalent chronic condition in both groups, affecting 73.86% of the control group and 51.30% of the ZDC group. On average, members of the ZDC group were younger and exhibited a better health profile, as indicated by the prevalence of chronic conditions and the average diabetes complications severity index scores.

Figures 1–3 display the event studies and exhibit the trend of the treatment effect of the ZDC program for each month during the follow-up period. The event studies suggest a positive effect of the ZDC program on reducing total and medical spending during the follow-up period. Although estimates are accompanied by wide 95% confidence intervals, several post-intervention event-time estimates are statistically significant, indicating the reductions in healthcare utilization and supporting the potential cost-saving benefits of the program. In addition, given the lack of evident nonparallel trends during the baseline period, we did not conduct formal parallel pretrend test.

Table 2 summarizes the treatment effect and the heterogeneity of the treatment effect across the overall population and subgroups. The ZDC program was able to reduce total healthcare and medical spending among the ZDC population. Specifically, the program was associated with a $121.76 reduction in total healthcare spending per member per month (PMPM) (P = .002) and a $131.50 reduction in medical spending PMPM (P < .001). However, the ZDC pharmacy benefit had no significant effect on pharmacy spending ($9.74; P = .54).

The subgroup analysis of the heterogeneity treatment effect of the ZDC program revealed that the complex users benefited the most from the ZDC program (see Appendix Tables 1-3 for balancing tables and Appendix Figs. 1-9 for corresponding event studies in Supplemental Materials found at https://doi.org/10.1016/j.jval.2025.07.017). Among control group members, 2199 were pre-ZDC users, 711 were pre-ZDC nonusers, and 1620 were complex users. In the ZDC group, 1746 were pre-ZDC users, 579 were pre-ZDC nonusers, and 710 were complex users. For complex users, the ZDC program was associated with a $283.44 (P = .006) reduction in total healthcare spending PMPM and a $385.45 (P < .001) reduction in the medical spending PMPM. In addition, the ZDC program was also associated with a $102.01 (P = .022) increase in pharmacy spending PMPM. No significant findings were observed within the pre-ZDC users and the pre-ZDC nonusers groups.

The ZDC program was effective in reducing total healthcare expenditures, with reductions in medical expenditures serving as the primary driving factor. Among all ZDC-eligible T2D beneficiaries, receiving the ZDC pharmacy benefit was associated with a statistically significant reduction in medical expenditures and a nonsignificant increase in pharmacy expenditures, with the greatest improvements observed in the complex users group. The findings of this study suggest that the impact of VBIDs on total healthcare spending can be achieved through financial incentives and improved medication adherence. The results of our study are consistent with the findings from the early iterations of the BCBSLA’s ZDC program. The 2014 iteration of BCBSLA’s ZDC program, which was combined with disease management and health coaches, indicated a $205.90 decrease PMPM in total healthcare spending.¹⁸ Moreover, the 2018 expansion of the ZDC program dropped the requirement of health coaches and still yielded similar findings, albeit at a lower magnitude. Among these 2018 ZDC beneficiaries, the ZDC program reduced medical costs by $71 PMPM and led to a $8 PMPM increase in generic drug use.¹⁹

Our work also aligns with previous literature. For instance, Maeng et al²¹ observed that a ZDC program offered to Geisinger Health System employees was associated with a $144 reduction in total healthcare spending PMPM. Additionally, another value-based patient support program was linked to a significantly lower all-cause medical cost. Patients who received the program benefit experienced a 29.2% reduction in all-cause medical costs compared with those who did not.²² Another important aspect of the relationship between VBIDs and healthcare costs is that VBIDs that reduce cost sharing typically lead to increased pharmacy spending because patients are encouraged to take more medications. However, this increase in pharmacy costs is often offset by savings in medical spending.^7,19,22–24

The cost savings associated with ZDC program were also highly consistent with our ZDC medication adherence findings. In our analysis of the ZDC program’s impact on medication adherence, we examined the same cohort of BCBSLA fully insured members with T2D who were eligible for the 2020 iteration of the program. We found that the ZDC pharmacy benefit was associated with a 4.4 percentage point increase (P < .001) in the proportion of days covered (PDC) for all antidiabetic medications, and a 5.4 percentage point increase (P < .001) in PDC specifically for ZDC-eligible antidiabetic medications²⁵ We observed heterogeneous treatment effects of the ZDC program across different user groups. The greatest improvement in medication adherence was found among the complex users, with increases of 9.1 percentage points for all antidiabetic medications and 10.0 percentage points for ZDC-eligible medications. In contrast, no significant improvement in PDC was observed among the pre-ZDC nonusers.²⁵ Notably, our cost analysis also suggested that the increase in pharmacy expenditure was offset by the cost savings in medical spending, underscoring the potential cost-effectiveness of the ZDC program. The relationship between cost-related nonadherence and healthcare spending is well documented in other studies.⁴ Iuga and McGuire stated that poor medication adherence will lead to poor health outcomes and increased healthcare utilization, which, in turn, drove up healthcare spending and out-of-pocket cost.²² Additionally, our findings on pharmacy expenditures are consistent with our analysis of monthly drug counts. The ZDC pharmacy benefit was associated with a 0.9 increase in the monthly count of all antidiabetic medications (P < .001) and a 0.074 increase in the monthly count of ZDC-eligible antidiabetic medications (P < .001). However, it is important to note that the observed—though nonsignificant—increase in overall pharmacy spending was not limited to antidiabetic medications.⁴ In summary, although causality cannot be definitively established, the medical cost savings observed in this study may be partially attributed to improved medication adherence and changes in drug use patterns resulting from the ZDC program. We will further examine the socioeconomic impact on the uptake of the ZDC pharmacy benefit. First, BCBSLA eliminated copayments for a wide range of antidiabetic medications and other preventive medications, such as antihypertensive medications. A similar study that evaluated the impact of preventive drug list, which removed copayments and deductibles for a wide range of medications, on patients’ out-of-pocket cost also observed significant reductions in the out-of-pocket cost and increase in medication utilization.²⁶ Second, in 2020, the average copayment for first-tier medications was approximately $11 per drug per year,²⁷ whereas the average annual income per person in Louisiana was $30 117²⁸—notably below the national average. This financial landscape suggests that even relatively modest out-of-pocket costs may pose a barrier to medication adherence for lower-income individuals. As such, eliminating these copays through the ZDC program could have provided a strong financial incentive for beneficiaries to initiate or maintain use of essential diabetes medications. This may partly explain the rapid onset of the program’s impact, as illustrated in Figure 1, in which improvements in total healthcare savings appear soon after implementation. The socioeconomic context in Louisiana likely heightened the responsiveness to cost-sharing reductions, amplifying the program’s early effects.

Our study has several limitations. First, sicker BCBSLA beneficiaries were excluded because of prior participation in the 2014 and 2018 ZDC programs, which may have reduced the statistical power to detect the program’s effect of the 2020 iteration. Additionally, the ZDC pharmacy benefit was not randomly assigned, making it difficult to control for unobservable factors that could influence the results. Furthermore, because our study period coincided with the COVID-19 pandemic, we recognize that the healthcare system and patient behaviors were significantly disrupted during this time. Although we adjusted for COVID-19 diagnoses in our regression models, this approach may not fully capture the broader and more complex effects of the pandemic—such as changes in healthcare utilization patterns, access to care, or medication-taking behaviors. As a result, residual confounding related to the pandemic may remain and could potentially bias our estimates. Moreover, the majority of our study population was of working age, whereas other studies may focus on an older population, limiting the generalizability of our findings. Our data also lacked information on race and socioeconomic status, restricting our ability to explore health disparities among different subgroups. In addition, because of data limitations, we were unable to disaggregate medical expenditures, healthcare utilization, or cost savings by specific disease categories, nor could we break down pharmacy expenditures by drug class or individual medications. This restricts our ability to determine whether the observed reductions in spending were driven by changes in diabetes-related care or other healthcare services. Finally, the follow-up period was limited to 18 months, which may not fully capture the long-term effects of the ZDC program.

The ZDC program was associated with a significant reduction in total healthcare expenditures among ZDC-eligible members with T2D, primarily driven by decreased medical spending while maintaining stable pharmacy costs. These findings emphasize the effectiveness VBIDs in addressing cost-related medication nonadherence, ultimately leading to lower healthcare utilization and improved financial sustainability. This study highlights the potential of VBIDs as a strategic approach to optimizing healthcare spending while enhancing patient outcomes.

Supplementary data associated with this article can be found in the online version at https://doi.org/10.1016/j.jval.2025.07.017.