Research on collaborative innovation optimization strategies for digitally enabled higher education ecosystems.

Digitally enabled higher education involves the in-depth use of new-generation digital technology, which has subverted and innovated the traditional teaching mode, driven the development of high-quality teaching and learning, and improved teachers' teaching experience, and increased efficiency. Based on ecosystem theory, this paper constructs a higher education ecosystem with the government, enterprises, and universities as the core participating subjects. It considers the participating subjects' effort level and the ecosystem's overall benefits under the three scenarios of noncooperative research and development (R&D), cost sharing, and cooperative R&D. The results show that (1) the service innovation effort level of the three parties increases with increasing human resource level and technology maturity, and the government's benefit decreases with increasing cost of fulfilling social responsibility. (2) The government's cost subsidies to universities and enterprises can enhance the service innovation level of both parties and increase the optimal returns of the three parties and the ecosystem as a whole. (3) In the cooperative R&D game scenario, the effort level of the three parties and the total ecosystem returns are greater than those in the noncollaborative R&D scenario, and after determining the subsidy coefficients of the government, Pareto optimality of the three parties and the ecosystem as a whole can be achieved. The conclusions of this study can aid in understanding the dynamic evolution mechanism of digitally enabled higher education and provide a realistic decision-making reference for higher education ecosystem managers.

PERSPECTIVE: A Path to Value-Based Insurance Design for Mental Health Services.

BACKGROUND: Aligning cost of mental health care with expected clinical and functional benefits of that care would incentivize the delivery of high value treatments and services. In turn, ineffective or untested care could still be offered but at costs high enough to offset the delivery of high value care. AIMS: The authors comment on Benson and Fendrick's paper on Value-Based Insurance Design (VBID) for mental health in the September 2023 special issue of this journal. The authors also present a preliminary framework of key ingredients needed to consider VBID for mental health treatments and services. METHODS: The authors briefly review current and past efforts to contain costs and improve quality of mental health care, which include (for example) use of carve-out and carve-in programs, evaluation of cost sharing models, impact of accountable care organizations, and studying other benefit designs and impact of federal and state policies. RESULTS: Using PTSD as an example, key ingredients of VBID for mental health services were identified and include the following: tools for case identification and monitoring progress over time at the population level; specific treatments and services with evidence of clinical effectiveness, cost-effectiveness, and health equity; and an approach to document the specific treatment or service was delivered (versus another treatment or service that may lack evidence). DISCUSSION: The inability to afford mental health care is a top barrier to treatment seeking. People who do elect to spend time and money on mental health care are further disadvantaged by accessing care that is not well regulated and the quality at best is questionable. VBID could be an important lever for increasing access to and use of high value mental health care. Partnerships among the research, practice, and policy communities can help ensure research solutions meet needs of these two communities. IMPLICATIONS FOR HEALTH CARE: VBID holds promise to make high value mental health care more affordable while discouraging low value treatments and services. IMPLICATIONS FOR HEALTH POLICIES: While evidence gaps remain, these gaps can be filled concurrently with pursuit of VBID for mental health services. IMPLICATIONS FOR FUTURE RESEARCH: This paper identifies important research opportunities to help make VBID a reality for mental health care.

Health Plan Switching and Health Care Utilization: A Randomized Clinical Trial.

Importance: While the Patient Protection and Affordable Care Act (ACA) helped make health insurance premiums more affordable with premium tax credits, ACA marketplace enrollees continue to face barriers to care. Objective: To investigate the effect of informational emails on plan switching and health care utilization. Design, Setting, and Participants: This randomized clinical trial was conducted during the 2021 special enrollment period in California's Affordable Care Act marketplace among households that reported receiving unemployment insurance and were enrolled in non-silver-tier plans. The trial targeted 42 470 households that became temporarily eligible for cost-sharing reduction (CSR) silver plans that covered 94% of medical costs (CSR silver 94 plans) as a result of the 2021 American Rescue Plan Act. Intervention: Households were randomized to either a no-email control group or to a treatment group receiving 2 informational emails encouraging households to switch to CSR plans. Main Outcomes and Measures: The primary outcome was the switch rate to a CSR silver plan by July 31, 2021. Secondary outcomes include various measures of health care utilization in the second half of 2021 (July 1, 2021, to December 31, 2021). Health care utilization was measured by rates of practitioner visits, emergency department visits, hospitalizations, and prescription fills. Results: Of the 42 470 households (head of household mean [SD], age, 41.4 [13.3] years; 51.7% male), 10 650 (25.1%) were in the control group and 31 820 (74.9%) were in the treatment group. The emails led to a statistically significant 3.1-percentage point (95% CI, 2.6-3.6 percentage points) increase in CSR silver 94 enrollment (a 74.8% relative increase) by July 31, 2021, and a 1.3-percentage point (95% CI, 0.2-2.4 percentage points) increase (a 2.3% relative increase) in practitioner visits by December 31, 2021. The emails had no detectable effect on prescription fills, emergency department visits, or hospitalizations. Conclusions and Relevance: The results of this randomized clinical trial provide experimental evidence that, with access to more affordable health care, individuals are more likely to visit practitioners. Trial Registration: ClinicalTrials.gov Identifier: NCT05891418.

Association of State Insulin Out-of-Pocket Caps With Insulin Cost-Sharing and Use Among Commercially Insured Patients With Diabetes : A Pre-Post Study With a Control Group.

BACKGROUND: Twenty-five states have implemented insulin out-of-pocket (OOP) cost caps, but their effectiveness is uncertain. OBJECTIVE: To examine the effect of state insulin OOP caps on insulin use and OOP costs among commercially insured persons with diabetes. DESIGN: Pre-post study with control group. SETTING: Eight states implementing insulin OOP caps of $25 to $30, $50, or $100 in January 2021, and 17 control states. PARTICIPANTS: Commercially insured persons with diabetes and insulin users younger than 65 years. Subgroups of particular interest included members from states with insulin OOP caps of $25 to $30, enrollees with health savings accounts (HSAs) that require high insulin OOP payments, and lower-income members. MEASUREMENTS: Mean monthly 30-day insulin fills and OOP costs. RESULTS: State insulin caps were not associated with changes in insulin use in the overall population (relative change in fills per month, 1.8% [95% CI, -3.2% to 6.9%]). Insulin users in intervention states saw a 17.4% (CI, -23.9% to -10.9%) relative reduction in insulin OOP costs, largely driven by reductions among HSA enrollees; there was no difference in OOP costs among nonaccount plan members. More generous ($25 to $30) state insulin OOP caps were associated with insulin OOP cost reductions of 40.0% (CI, -62.5% to -17.6%), again primarily driven by a larger reduction in the subgroup with HSA plans. LIMITATIONS: Single national insurer; 9-month follow-up. CONCLUSION: Insulin OOP caps were associated with reduced insulin OOP costs but no overall increases in insulin use. A proposed national insulin cap of $35 for commercially insured persons might lead to meaningful insulin OOP savings but have a limited effect on insulin use. PRIMARY FUNDING SOURCE: Centers for Disease Control and Prevention and National Institute of Diabetes and Digestive and Kidney Diseases.

Prescription Drug Dispensing and Patient Costs After Implementation of a No Behavioral Health Cost-Sharing Law.

Importance: On January 1, 2022, New Mexico implemented a No Behavioral Cost-Sharing (NCS) law that eliminated cost-sharing for mental health and substance use disorder (MH/SUD) treatments in plans regulated by the state, potentially reducing a barrier to treatment for MH/SUDs among the commercially insured; however, the outcomes of the law are unknown. Objective: To assess the association of implementation of the NCS with out-of-pocket spending for prescription for drugs primarily used to treat MH/SUDs and monthly volume of dispensed drugs. Design, Settings, and Participants: This retrospective cohort study used a difference-in-differences research design to examine trends in outcomes for New Mexico state employees, a population affected by the NCS, compared with federal employees in New Mexico who were unaffected by NCS. Data were collected on prescription drugs for MH/SUDs dispensed per month between January 2021 and June 2022 for New Mexico patients with a New Mexico state employee health plan and New Mexico patients with a federal employee health plan. Data analysis occurred from December 2022 to January 2024. Exposure: Enrollment in a state employee health plan or federal health plan. Main Outcomes and Measures: The primary outcomes were mean patient out-of-pocket spending per dispensed MH/SUD prescription and the monthly volume of dispensed MH/SUD prescriptions per 1000 employees. A difference-in-differences estimation approach was used. Results: The implementation of the NCS law was associated with a mean (SE) $6.37 ($0.30) reduction (corresponding to an 85.6% decrease) in mean out-of-pocket spending per dispensed MH/SUD medication (95% CI, -$7.00 to -$5.75). The association of implementation of NCS with the volume of prescriptions dispensed was not statistically significant. Conclusions and Relevance: These findings suggest that the implementation of the New Mexico NCS law was successful in lowering out-of-pocket spending on prescription medications for MH/SUDs, but that there was no association of NCS with the volume of medications dispensed in the first 6 months after implementation. A key challenge is to identify policies that protect from high out-of-pocket spending while also promoting access to needed care.

Evolution analysis of low-carbon cooperation of service providers based on Moran process in cloud manufacturing.

Low-carbon cooperation among cloud manufacturing service providers is one way to achieve carbon peak and neutrality. Such cooperation is related to the benefits to service providers adopting low-carbon strategies and stochastic factors such as government low-carbon policies, providers' environmental awareness, and demanders' low-carbon preferences. Focusing on the evolutionary process of service providers' low-carbon strategy selection under uncertain factors, a stochastic evolutionary game model is constructed based on the Moran process, and the equilibrium conditions for low-carbon cooperation among providers are analyzed under benefit-dominated and stochastic factor-dominated situations. Through numerical simulation, the effects of the cloud platform's cost-sharing coefficient for low-carbon investment, matching growth rate, carbon trading price, and group size on providers' low-carbon strategy evolution are analyzed. The research results show that increasing the cloud platform's low-carbon cost-sharing, carbon trading price, and group size can promote low-carbon cooperation among service providers. With greater low-carbon investment costs and greater stochastic factor interference, the providers' enthusiasm for low-carbon cooperation decreases. This study fills the research gap in the low-carbon cooperation evolution of cloud manufacturing providers based on the stochastic evolutionary game and provides decision-making suggestions for governments and cloud platforms to encourage provider participation in low-carbon cooperation and for providers to adopt low-carbon strategies.

Effects of nurse visit copayment on primary care use: Do low-income households pay the price?

Nurses are increasingly providing primary care, yet the literature on cost-sharing has paid little attention to nurse visits. We employ a staggered difference-in-differences design to examine the effects of adopting a 10-euro copayment for nurse visits on the use of public primary care among Finnish adults. We find that the copayment reduced nurse visits by 9%-10% during a one-year follow-up. There is heterogeneity by income in absolute terms, but not in relative terms. The spillover effects on general practitioner (GP) use are negative but small, with varying statistical significance. We also analyze the subsequent nationwide abolition of the copayment. However, we refrain from drawing causal conclusions from this due to the lack of credibility in the parallel trends assumption. Overall, our analysis suggests that moderate copayments can create a greater barrier to access for low-income individuals. We also provide an example of using a pre-analysis plan for retrospective observational data.

Cost sharing for oral lenvatinib among commercially insured patients.

OBJECTIVE: To use a nationwide pharmaceutical claims database to evaluate cost-sharing trends for commercially insured patients with cancer who were prescribed lenvatinib (Lenvima). STUDY DESIGN: IBM MarketScan databases were used to evaluate lenvatinib costs for patients with employer-based commercial insurance, and for patients 65 years and older, Medicare claims for fee-for-service plans. METHODS: Patients were included if they had least 1 outpatient pharmaceutical claim for lenvatinib paid on a noncapitated basis from 2015 to 2019. Median and IQR costs were estimated and inflation adjusted to 2019 US$ for 30-day supplies and reported as total, insurance liability, coordination of benefits, and out-of-pocket costs. RESULTS: A total of 685 patients had at least 1 pharmaceutical claim for lenvatinib, which included patients with thyroid (n = 251; 36.6%), renal cell (n = 202; 29.5%), hepatocellular (n = 160; 23.4%), and endometrial (n = 48; 7.0%) cancer. The median (IQR) number of prescriptions per patient was 3 (2-7), and the median (IQR) total days of supply was 90 (45-210) days. The median (IQR) 30-day cost of lenvatinib was $17,253 ($15,597-$18,120). Median (IQR) 30-day insurance liability was $16,847 ($15,000-$17,981). Median (IQR) 30-day coordination of benefits was $0 ($0-$0). Median (IQR) 30-day patient out-of-pocket cost was $32 ($0-$100). However, the maximum 30-day out-of-pocket cost in our patient cohort was $12,538. CONCLUSIONS: In this cohort, insurance was liable for the majority of total lenvatinib drug costs, and 75% of patients paid $100 or less per month out of pocket. This information can be used by care teams to counsel insured patients. Health systems and drug manufacturers must identify patients with high out-of-pocket costs and provide convenient access to financial assistance programs so that patients are not forced to forgo the benefits of these drugs due to financial barriers. Value-based payment models and drug pricing reform are also needed to address underlying drivers of high drug costs.

The optimal co-insurance rate for outpatient drug expenses of Iranian health insured based on the data mining method.

OBJECTIVE: A more equal allocation of healthcare funds for patients who must pay high costs of care ensures the welfare of society. This study aimed to estimate the optimal co-insurance for outpatient drug costs for health insurance. SETTING: The research population includes outpatient prescription claims made by the Health Insurance Organization that outpatient prescriptions in a timely manner in 2016, 2017, 2018, and 2019 were utilized to calculate the optimal co-insurance. The study population was representative of the research sample. DESIGN: At the secondary level of care, 11 features of outpatient claims were studied cross-sectionally and retrospectively using data mining. Optimal co-insurance was estimated using Westerhut and Folmer's utility model. PARTICIPANTS: One hundred ninety-three thousand five hundred fifty-two individuals were created from 21 776 350 outpatient claims of health insurance. Because of cost-sharing, insured individuals in a low-income subsidy plan and those with refractory diseases were excluded. RESULTS: Insureds were divided into three classes of low, middle, and high risk based on IQR and were separated to three clusters using the silhouette coefficient. For the first, second, and third clusters of the low-risk class, the optimal co-insurance estimates are 0.81, 0.76, and 0.84, respectively. It was equal to one for all middle-class clusters and 0.38, 0.45, and 0.42, respectively, for the high-risk class. The insurer's expenses were altered by $3,130,463, $3,451,194, and $ 1,069,859 profit for the first, second, and third clusters, respectively, when the optimal co-insurance strategy is used for the low-risk class. For middle risks, it was US$29,239,815, US$13,863,810, and US$ 14,573,432 while for high risks, US$4,722,099, US$ 6,339,317, and US$19,627,062, respectively. CONCLUSIONS: These findings can improve vulnerable populations' access to costly medications, reduce resource waste, and help insurers distribute funds more efficiently.

Acute Diabetes Complications After Transition to a Value-Based Medication Benefit.

Importance: The association of value-based medication benefits with diabetes health outcomes is uncertain. Objective: To assess the association of a preventive drug list (PDL) value-based medication benefit with acute, preventable diabetes complications. Design, Setting, and Participants: This cohort study used a controlled interrupted time series design and analyzed data from a large, national, commercial health plan from January 1, 2004, through June 30, 2017, for patients with diabetes aged 12 to 64 years enrolled through employers that adopted PDLs (intervention group) and matched and weighted members with diabetes whose employers did not adopt PDLs (control group). All participants were continuously enrolled and analyzed for 1 year before and after the index date. Subgroup analysis assessed patients with diabetes living in lower-income and higher-income neighborhoods. Data analysis was performed between August 19, 2020, and December 1, 2023. Exposure: At the index date, intervention group members experienced employer-mandated enrollment in a PDL benefit that was added to their follow-up year health plan. This benefit reduced out-of-pocket costs for common cardiometabolic drugs, including noninsulin antidiabetic agents and insulin. Matched control group members continued to have cardiometabolic medications subject to deductibles or co-payments at follow-up. Main Outcomes and Measures: The primary outcome was acute, preventable diabetes complications (eg, bacterial infections, neurovascular events, acute coronary disease, and diabetic ketoacidosis) measured as complication days per 1000 members per year. Intermediate measures included the proportion of days covered by and higher use (mean of 1 or more 30-day fills per month) of antidiabetic agents. Results: The study 10 588 patients in the intervention group (55.2% male; mean [SD] age, 51.1 [10.1] years) and 690 075 patients in the control group (55.2% male; mean [SD] age, 51.1 [10.1] years) after matching and weighting. From baseline to follow-up, the proportion of days covered by noninsulin antidiabetic agents increased by 4.7% (95% CI, 3.2%-6.2%) in the PDL group and by 7.3% (95% CI, 5.1%-9.5%) among PDL members from lower-income areas compared with controls. Higher use of noninsulin antidiabetic agents increased by 11.3% (95% CI, 8.2%-14.5%) in the PDL group and by 15.2% (95% CI, 10.6%-19.8%) among members of the PDL group from lower-income areas compared with controls. The PDL group experienced an 8.4% relative reduction in complication days (95% CI, -13.9% to -2.8%; absolute reduction, -20.2 [95% CI, -34.3 to -6.2] per 1000 members per year) compared with controls from baseline to follow-up, while PDL members residing in lower-income areas had a 10.2% relative reduction (95% CI, -17.4% to -3.0%; absolute, -26.1 [95% CI, -45.8 to -6.5] per 1000 members per year). Conclusions and Relevance: In this cohort study, acute, preventable diabetes complication days decreased by 8.4% in the overall PDL group and by 10.2% among PDL members from lower-income areas compared with the control group. The results may support a strategy of incentivizing adoption of targeted cost-sharing reductions among commercially insured patients with diabetes and lower income to enhance health outcomes.

Site of Ambulance Origination and Billing for Out-of-Network Services.

Importance: The No Surprises Act implemented in 2022 aims to protect patients from surprise out-of-network (OON) bills, but it does not include ground ambulance services. Understanding ground ambulance OON and balance billing patterns from previous years could guide legislation aimed to protect patients following ground ambulance use. Objective: To characterize OON billing from ground ambulance services by evaluating whether OON billing risk differs by the site of ambulance origination (home, hospital, nonhospital medical facility, or scene of incident). Design, Setting, and Participants: Cross-sectional study of the Merative MarketScan dataset between January 1, 2015, and December 31, 2020, using claims-based data from employer-based private health insurance plans in the US. Participants included patients who utilized ground ambulances during the study period. Data were analyzed from June to December 2023. Exposure: Medical encounter requiring ground ambulance transportation. Main Outcomes and Measures: Ground ambulance OON billing prevalence was calcuated. Linear probability models adjusted for state-level mixed effects were fit to evaluate OON billing probability across ambulance origins. Secondary outcomes included the allowed payment, patient cost-sharing amounts, and potential balance bills for OON ambulances. Results: Among 2 031 937 ground ambulance services (1 375 977 unique patients) meeting the inclusion and exclusion criteria, 1 072 791 (52.8%) rides transported men, and the mean (SD) patient age was 41 (18) years. Of all services, 1 113 676 (54.8%) were billed OON. OON billing probabilities for ambulances originating from home or scene were higher by 12.0 percentage points (PP) (95% CI, 11.8-12.2 PP; P < .001 for home; 95% CI, 11.7-12.2 PP; P < .001 for scene) vs those originating from hospitals. Mean (SD) total financial burden, including cost-sharing and potential balance bills per ambulance service, was $434.70 ($415.99) per service billed OON vs $132.21 ($244.92) per service billed in-network. Conclusions and Relevance: In this cross-sectional study of over 2 million ground ambulance services, ambulances originating from home, the scene of an incident, and nonhospital medical facilities were more likely to result in OON bills. Legislation is needed to protect patients from surprise billing following use of ground ambulances, more than half of which resulted in OON billing. Future legislation should at minimum offer protections for these subsets of patients often calling for an ambulance in urgent or emergent situations.

Why Cost Sharing on Its Own Will Not Fix Health Care Costs.

This Viewpoint describes issues with cost sharing for health care costs and suggests improvements to current cost sharing systems.

Estimating The Impact Of Out-Of-Pocket Cost Changes On Abandonment Of HIV Pre-Exposure Prophylaxis.

Oral HIV pre-exposure prophylaxis (PrEP) is highly effective for preventing HIV. Several different developments in the US either threaten to increase or promise to decrease PrEP out-of-pocket costs and access in the coming years. In a sample of 58,529 people with a new insurer-approved PrEP prescription, we estimated risk-adjusted percentages of patients who abandoned (did not fill) their initial prescription across six out-of-pocket cost categories. We then simulated the percentage of patients who would abandon PrEP under hypothetical changes to out-of-pocket costs, ranging from $0 to more than $500. PrEP abandonment rates of 5.5 percent at $0 rose to 42.6 percent at more than $500; even a small increase from $0 to $10 doubled the rate of abandonment. Conversely, abandonment rates that were 48.0 percent with out-of-pocket costs of more than $500 dropped to 7.3 percent when those costs were cut to $0. HIV diagnoses were two to three times higher among patients who abandoned PrEP prescriptions than among those who filled them. These results imply that recent legal challenges to the provision of PrEP with no cost sharing could substantially increase PrEP abandonment and HIV rates, upending progress on the HIV/AIDS epidemic.

Contraceptive Use Among Traditional Medicare And Medicare Advantage Enrollees.

Medicare is the primary source of health insurance coverage for reproductive-age people with Social Security Disability Insurance. However, Medicare does not require contraceptive coverage for pregnancy prevention, and little is known about contraceptive use in traditional Medicare and Medicare Advantage. We analyzed Medicare and Optum data to assess variations in contraceptive use and methods used by traditional Medicare and Medicare Advantage enrollees, as well as among enrollees with and without noncontraceptive clinical indications. Clinically indicated contraceptives are used for reasons other than pregnancy prevention, including menstrual regulation or to treat acne, menorrhagia, and endometriosis. Contraceptive use was higher among Medicare Advantage enrollees than traditional Medicare enrollees, but use in both populations was low compared with contraceptive use among Medicaid enrollees. We found significant variation by Medicare type with respect to contraceptive methods used. Relative to traditional Medicare, the probability of long-acting reversible contraception was more than three times higher in Medicare Advantage, and the probability of tubal sterilization was more than ten times higher. Overall, Medicare enrollees with noncontraceptive clinical indications had twice the probability of contraceptive use as those without them. Medicare coverage of all contraceptive methods without cost sharing would help address financial barriers to contraceptives and support the reproductive autonomy of disabled enrollees.

The Impact of Prescription Drug Coverage on Disparities in Adherence and Medication Use: A Systematic Review.

Prescription drug cost-sharing is a barrier to medication adherence, particularly for low-income and minority populations. In this systematic review, we examined the impact of prescription drug cost-sharing and policies to reduce cost-sharing on racial/ethnic and income disparities in medication utilization. We screened 2,145 titles and abstracts and identified 19 peer-reviewed papers that examined the interaction between cost-sharing and racial/ethnic and income disparities in medication adherence or utilization. We found weak but inconsistent evidence that lower cost-sharing is associated with reduced disparities in adherence and utilization, but studies consistently found that significant disparities remained even after adjusting for differences in cost-sharing across individuals. Study designs varied in their ability to measure the causal effect of policy or cost-sharing changes on disparities, and a wide range of policies were examined across studies. Further research is needed to identify the types of policies that are best suited to reduce disparities in medication adherence.

The Reality of Cost Sharing in Japan.

Financial pressure on younger generation is mounting in Japan, a super-ageing society with staggering economy. The revision on the co-insurance rate for 70-74 with "Standard" category was implemented to mitigate such pressure, seeking better balance across generations in sharing the burden of healthcare cost. It raised the rate from 10% to 20% over the period of five years from 2014 to 2018. This report examined how it changed the share of cost sharing (cost sharing as percentage to total healthcare expenditure), among the 70-74 with "Standard" category in Citizens Health Insurance programme in 44 prefectures. It specifically focused on change in the population's actual share of cost sharing (ASCS) that better reflect the genuine amount of payment actually made by the patients themselves. The average ASCS increased from 7.28% (2013) to 10.78% (2019), resulting wider gap from the statutory planned share of cost sharing (i.e., the statutory co-insurance rate of 10% in 2013, and 20% in 2019). Also found was increased variance among prefectural ASCS, which may suggest a possibility of un-designed effect by the revision, of encouraging a move towards ability and willingness to pay. In terms of cost containment effect, Japan needs to consider various non-conventional options, including review of the current use of healthcare resources. First and foremost, however, the true state of cost sharing should be recognized in terms of ASCS and shared more widely as a reality. Such effort is essential in discussion of how to keep embracing the country's life line, UHC.

Why higher copayments for opioids did not reduce use among Medicare beneficiaries.

To examine whether higher cost-sharing deterred prescription opioid use. Medicare Part D claims from 2007 to 2016 for a 20% random sample of Medicare enrollees. We obtain estimates of the effect of cost-sharing on prescription opioid use using ordinary least squares and instrumental variables methods. In both, we exploit the variation (change) in cost-sharing within plans over time for a sample of beneficiaries who remain in the same plan. Focusing on changes in cost-sharing within a plan for a constant sample of beneficiaries mitigates potential bias from plan selection and using a constant set of weights derived from use in year (t) eliminates changes in the cost-sharing indexes due to (endogenous) consumer choice in year (t+1). Part D plans adopted benefit changes designed to reduce opioid use, including moving opioids to higher cost-sharing tiers. Increasing plan copayments for hydrocodone or oxycodone was associated with reductions in plan-paid claims and offsetting increases in cash claims. Widespread availability of low-cost generics combined with the anti-clawback provision in Part D mediated the effect of higher cost sharing to curb opioid use. As plans moved generic opioids to higher cost-sharing tiers, beneficiaries simply paid cash prices and aggregate use remained largely unchanged. The anti-clawback provision in Part D, intended to protect beneficiaries from price gouging, limited plans' ability to constrain opioid use through typical demand-side measures such as increased cost-sharing.