The $3.84 Billion Question
Analyzing Generic Drug Adoption in Medicare Part D
A Data Analytics Portfolio Project | February 2026
$3.84B
Potential annual Medicare savings from just 25% brand to generic conversion in 2023
Analyzing Generic Drug Adoption in Medicare Part D
A Data Analytics Portfolio Project | February 2026
Potential annual Medicare savings from just 25% brand to generic conversion in 2023
In 2023, Medicare Part D spent $57.3 billion on medications that had cheaper, medically identical alternatives sitting on pharmacy shelves. My analysis found that if just 25% of those brand name prescriptions had been switched to their FDA-approved generic equivalents, Medicare could have saved $3.84 billion in a single year.
Medicare Part D, which began providing prescription drug coverage in 2006, serves approximately 54.8 million Medicare beneficiaries as of 2025.1 When the FDA approves a generic drug as "therapeutically equivalent" to a brand name drug, these products are pharmaceutically equivalent and bioequivalent. They deliver identical amounts of active ingredient into the bloodstream at the same rate.3 Generics typically cost 80-85% less than their brand name counterparts,6 and from 2009 to 2018, the U.S. healthcare system saved nearly $2 trillion through generic drug use.7
As a pharmacy clerk, I've watched patients struggle with medication costs while standing just feet away from cheaper alternatives that would work exactly the same way. This portfolio project emerged from that experience, using data analytics to quantify a problem I've seen firsthand: why do Medicare beneficiaries continue receiving expensive brand name drugs when identical generic options exist?
This analysis examines 2023 Medicare Part D prescriber data, a critical year to understand because it represents the landscape just before major policy reforms took effect. The Inflation Reduction Act's redesign of Medicare Part D introduced transformative changes in 2024-2025, including elimination of the coverage gap, implementation of a $2,000 annual out of pocket cap (indexed to $2,100 for 2026), and the Medicare Drug Price Negotiation Program.101112 What you're seeing here is a historical analysis: what the generic adoption opportunity looked like before these structural reforms fundamentally altered the program.
Using two CMS datasets containing prescriber level and drug level data, my analysis identified 454 medications with both brand and generic versions available in 2023. The findings reveal not just the scale of the savings opportunity, but where it's concentrated, and the answers are surprising.
Not all generic drugs are created equal in the eyes of the FDA. When a generic drug receives an "A" rating in the FDA's Orange Book (officially called "Approved Drug Products with Therapeutic Equivalence Evaluations"), it means the generic is deemed therapeutically equivalent to the brand name reference drug and can be safely substituted.5 This rating isn't given lightly: bioequivalence is established when the 90% confidence intervals for the generic to brand ratios of both maximum drug concentration (Cmax) and area under the plasma concentration time curve (AUC) lie within 80-125%.4
In simpler terms: FDA-approved therapeutically equivalent generics deliver the same amount of active ingredient into your bloodstream at the same rate as the brand name drug.3 Same clinical outcome, same safety profile, same effectiveness. The only differences are typically inactive ingredients (like fillers or dyes) and the price.
Here's where it gets interesting: generic drugs typically cost 80-85% less than their brand name counterparts.6 That's not a small discount. It's a massive price differential for medications the FDA has certified as medically identical. My analysis found that in 2023, this price gap created a $57.3 billion opportunity across just 454 drugs where both brand and generic versions existed.
When Medicare pays for a brand name drug instead of its generic equivalent, the program spends more money for the same health outcome. Unlike individual patients who might have copay structures that mask the true cost difference, Medicare bears the full price differential. Over time, these choices compound: small decisions by individual prescribers aggregate into billions of dollars in excess spending across the program.
This analysis examines 2023 data, not because it's the most current, but because it captures a critical moment in Medicare Part D history. In January 2023, the Inflation Reduction Act's $35-per-month insulin cost cap took effect for Medicare Part D enrollees,8 representing the federal government's first major intervention to directly limit out of pocket costs for a specific drug category. Research had shown that 1.5 million Medicare beneficiaries using insulin would have saved $734 million in Part D costs if the cap had been in effect in 2020.9
Between 2023 (when this data was collected) and 2026 (when this analysis is being conducted), Medicare Part D underwent transformative structural changes under the IRA:101112
The insulin market saw particularly dramatic shifts after 2023. Following Medicare's $35 cap, major manufacturers (Eli Lilly, Novo Nordisk, Sanofi) announced voluntary $35 cap programs for uninsured and commercially insured patients in 2023-2024.23 By 2025-2026, the Medicare Drug Price Negotiation Program's Selected Drug Subsidy provisions began affecting insulin pricing dynamics.12 These changes fundamentally altered the cost sharing landscape that existed when this data was collected.
This project should be understood as historical analysis: what Medicare Part D spending and savings opportunities looked like before these major policy changes took effect. The brand versus generic utilization patterns identified here may have shifted as the IRA provisions were implemented. However, understanding the prereform baseline is critical for evaluating whether policy changes actually moved the needle on generic adoption, and whether the core problem of prescribers choosing expensive brands over identical generics has been addressed or merely masked by cost caps and price negotiations.
This project utilized two publicly available Centers for Medicare & Medicaid Services (CMS) datasets from 2023: the Medicare Part D Prescriber by Drug and Geography dataset (containing prescriber level information on drug spending, claims, and beneficiary counts) and the Medicare Part D Spending by Drug dataset (containing drug level aggregate spending data). Together, these datasets provide comprehensive coverage of Medicare Part D prescription activity across all 50 states plus U.S. territories.
My analysis followed a systematic approach across four phases:
Phase 1: Drug Identification — I identified all drugs with both brand and generic versions available in 2023 and classified drug types based on naming patterns (insulin products, drugs ending in -prazole, -statin, etc.). All generic equivalents were verified against FDA Orange Book therapeutic equivalence standards.35
Phase 2: Utilization Analysis — I calculated brand utilization rates (the percentage of prescriptions filled with brand name versions when generics existed) for each drug, computed total Medicare Part D spending on drugs with generic alternatives, and analyzed geographic variation in generic adoption rates across all states and territories.
Phase 3: Savings Calculation — I modeled potential savings at various brand to generic conversion rates (10%, 25%, 50%, 75%, 100%) using the formula: Savings = Brand Spending × Conversion Rate × Generic Discount Rate. The calculations used conservative estimates based on documented generic cost differentials.6
Phase 4: Data Validation — I verified mathematical consistency across all calculations, cross-referenced totals against CMS published spending figures, and ensured all outputs maintained referential integrity.
All data processing was conducted using Python, with analysis organized into seven separate scripts for transparency and reproducibility. The processed data produced seven Tableau-ready CSV files covering top savings opportunities, conversion scenarios, state level analysis, and drug specific breakdowns. Tableau was used for all visualization and dashboard creation.
In 2023, Medicare Part D's total gross spending across all drugs reached approximately $275.8 billion. My analysis found that $57.3 billion of this total (20.8%) was spent on 454 drugs that had both brand and generic versions available. This means more than one in every five dollars Medicare spent on prescription drugs went to medications where therapeutically equivalent, significantly cheaper alternatives existed.
If just 25% of brand name prescriptions for these 454 drugs had been converted to their generic equivalents in 2023, Medicare could have saved $3.84 billion. This is a conservative scenario, not wholesale replacement of all brand name drugs, but a modest shift toward greater generic adoption. It assumes that three out of every four brand prescriptions would continue unchanged, with only one in four switching to the generic version.
The savings potential scales significantly at higher conversion rates. My analysis modeled five scenarios:
Even modest increases in generic adoption (shifting just one in ten brand prescriptions to generic) would generate over $1.5 billion in annual Medicare savings.
These aren't theoretical savings from unproven interventions or experimental drugs. This is money Medicare could have saved in 2023 by using medications the FDA has already certified as therapeutically equivalent to the brand name versions.35 The infrastructure exists, the medications are approved and available, and the clinical outcomes are identical. The question isn't whether generic substitution works. It's why it isn't happening more often.
The $3.84 billion savings opportunity wasn't evenly distributed across all 454 drugs. My analysis found that the top 20 drugs by savings potential accounted for a disproportionate share of the opportunity. What's striking isn't just the concentration. It's which drugs dominated the list.
Three insulin products claimed the top three positions:
Combined, these three insulin formulations alone represented over $1.2 billion in potential savings from just a 25% conversion rate. And their brand utilization rates (between 93% and 99%) revealed that despite generic availability, prescribers were overwhelmingly choosing brand name products.
The top 20 list extended beyond insulin to include other high-cost, high-volume medications where generics existed but weren't being prescribed despite FDA therapeutic equivalence certification.35
The insulin numbers demand closer examination. My analysis identified multiple insulin formulations in the top savings opportunities, all sharing one characteristic: brand utilization rates between 93% and 99%. This means that even when generic insulin versions were available and FDA-approved as therapeutically equivalent,35 prescribers were choosing brand name products in nearly every case.
Here's what makes this particularly notable: January 2023 marked the implementation of the Inflation Reduction Act's $35-per-month insulin cost cap for Medicare Part D enrollees.8 This policy capped what beneficiaries paid out of pocket, making insulin more affordable for patients. But it didn't address which version of insulin was being prescribed (brand or generic). While patients were protected from high costs, Medicare was still bearing the full price differential between expensive brand name insulins and their therapeutically equivalent generic alternatives.
The insulin landscape continued shifting throughout 2023 and beyond. In March 2023, manufacturer Eli Lilly announced plans to cap insulin prices at $35,14 with other manufacturers following suit.23 These were voluntary programs rather than systematic shifts toward generic prescribing. By 2025-2026, the Medicare Drug Price Negotiation Program's provisions began affecting insulin pricing dynamics,12 but the fundamental question remained: when generics exist, why weren't they being prescribed?
The insulin pattern illustrates a crucial insight: patient cost caps and manufacturer pricing agreements, while beneficial for beneficiary affordability, don't necessarily change prescribing behavior. The 93-99% brand utilization rates in 2023 suggest that even as policy interventions focused on lowering patient costs, the underlying practice of prescribing expensive brands over identical generics persisted. This represents a different kind of savings opportunity, one focused on system level efficiency rather than patient out of pocket costs.
One plausible explanation for high brand utilization could have been regional variation. Perhaps certain states or regions had prescribing cultures that favored brand name drugs, while others embraced generics more readily. My analysis examined generic adoption rates across all 50 states plus U.S. territories to test this hypothesis.
What I found was striking uniformity. Generic adoption rates across states ranged from 88.3% to 91.6%, a span of just 3.3 percentage points. This isn't the kind of variation that explains a multi-billion-dollar savings opportunity. Whether you looked at California or Wyoming, New York or Mississippi, the pattern was essentially the same: generics were widely adopted overall, but certain specific drugs showed persistently high brand utilization regardless of geography.
The narrow geographic variation has important implications: the savings opportunity isn't primarily driven by regional prescribing differences or state level policy variations. Instead, the concentration of potential savings in specific high-cost drugs (particularly the insulin products with 93-99% brand utilization) suggests that the problem lies with particular medications rather than geographic patterns. An intervention to increase generic adoption in one state would likely face the same barriers as an intervention in any other state.
This finding suggests that interventions to increase generic adoption would likely need to be drug specific rather than region specific. Rather than targeting states or regions with low generic adoption rates, efforts should focus on changing prescriber behavior for particular high-cost medications where generic alternatives exist but aren't being prescribed. The insulin case exemplifies this: the problem wasn't isolated to certain geographic areas. It was systemic across the entire Medicare Part D program.
The data reveals what happened in 2023 (high brand utilization despite generic availability) but not why. Understanding the barriers to generic adoption requires examining the complex ecosystem of prescribers, patients, pharmacists, regulators, and payers. Research has identified several documented barriers that help explain the patterns observed in this analysis.
Research has shown that barriers to generic drug use include negative perceptions about generic medicines, concerns about efficacy and safety, and preference for familiar brand names.15 Studies of specific patient populations have documented concerns that generic substitution might lead to breakthrough symptoms or side effects,16 even though the FDA maintains that therapeutically equivalent generics are just as safe and effective as brand name drugs.17 These perceptions affect both prescribers and patients: a prescriber may default to the brand name product they learned in medical school, while patients may request brand names they've seen advertised or used previously.
Barriers can occur at multiple points, including state laws on generic substitution, factors related to generic availability, and consumer and prescriber behavior.18 Some states require pharmacists to substitute generics unless the prescriber specifies otherwise, while others take a more permissive approach, allowing but not requiring substitution.19 Additionally, some states require patient consent before generic substitution.
Prescribers can specify "dispense as written" (DAW) or "brand necessary," which overrides generic substitution laws. Research on specialty drugs has found that DAW specifications can be significantly higher than the general prescription population. One study of generic imatinib found 24% of prescriptions were marked DAW, compared to just 3.5% for medications with available generics overall.20
PBMs, which manage prescription drug benefits on behalf of insurers and employers, create formularies and negotiate rebates that can create counterintuitive barriers to lower-cost medicines.21 The complex incentive structures can sometimes favor higher-cost drugs when rebates and other considerations are factored in. While the details of PBM negotiations are often proprietary, the structure can create situations where a higher-priced brand name drug might be preferred on a formulary despite a generic alternative being available.
These barriers don't operate in isolation. They interact and compound. A prescriber concerned about patient preferences might write "dispense as written" to avoid pharmacy substitution. State laws that require patient consent add friction to generic substitution. PBM formularies that favor certain brands create economic incentives that run counter to generic adoption. The result is a system where the path of least resistance often leads to brand name prescriptions, even when therapeutically equivalent generics exist.
While the 2023 landscape has now been transformed by IRA provisions, this historical analysis offers insights that remain relevant as Medicare Part D continues to evolve. The patterns identified here (drug specific savings concentration, persistent high brand utilization despite generic availability, and minimal geographic variation) suggest several considerations for different stakeholders.
The narrow geographic variation in generic adoption (88.3-91.6%) coupled with savings highly concentrated in specific drugs like insulin suggests that interventions should be drug specific rather than region specific. Research has shown that state level generic substitution laws affect adoption rates,24 yet the 2023 data reveals that even with varying state laws, high brand utilization persisted for certain medications across all geographies.
As the Medicare Drug Price Negotiation Program takes effect and Part D cost sharing structures continue to change, monitoring whether brand versus generic utilization patterns shift will be important for understanding policy effectiveness. The insulin case study from 2023 (where 93-99% brand utilization persisted even after the $35 cost cap took effect8) demonstrates that patient facing affordability measures don't automatically change prescribing behavior or system level costs.
The analysis underscores several opportunities for prescribers and pharmacists:
Generic-first prescribing: When clinically appropriate, defaulting to generic prescriptions for therapeutically equivalent drugs35 can generate substantial system savings without compromising patient outcomes.
Patient education: Pharmacists and prescribers explaining that FDA-rated "A" therapeutic equivalents deliver the same clinical outcomes317 can help address perception barriers that discourage generic adoption.1516
Systematic review: Periodically reviewing established patients to identify opportunities for cost-effective generic switches, particularly for high-cost medications like insulin where brand utilization was found to be 93-99% in 2023.
This analysis points to several areas warranting further investigation:
Prescriber decision-making: Qualitative research exploring why prescribers continue brand name prescribing despite generic availability could reveal intervention points not captured in administrative data.
Post-IRA impact assessment: Comparing 2023 brand utilization patterns to 2025-2026 data would assess whether IRA provisions changed prescribing behavior or merely shifted who pays. The question isn't just whether patients pay less, but whether the system is using lower-cost therapeutically equivalent alternatives.
Insulin-specific deep dive: The 93-99% brand utilization rates for insulin products in 2023 deserve targeted investigation. Understanding whether subsequent policy changes (manufacturer voluntary caps,23 negotiation program provisions,12 and Part D restructuring10) have affected this pattern would provide valuable insights into policy effectiveness.
Like all analyses, this project has important limitations that should be understood when interpreting the findings:
This analysis examines 2023 data. The Medicare Part D landscape has fundamentally changed due to IRA provisions implemented in 2024-2026, including the coverage gap elimination, $2,000 out of pocket cap, and Medicare Drug Price Negotiation Program.101112 Current savings opportunities and brand utilization patterns may differ substantially from what existed in 2023.
While FDA-rated therapeutic equivalence means generic drugs deliver the same clinical outcomes,3517 individual patients may have legitimate reasons for continuing brand name medications. These include different inactive ingredients that may affect tolerability in sensitive individuals, established dosing with narrow therapeutic index (NTI) drugs where prescribers prefer consistency, and prior authorization or step therapy requirements that may complicate switching. The administrative data shows what was prescribed but not why.
The savings calculations assume a straightforward cost differential between brand and generic drugs based on documented price gaps.6 In practice, Medicare Part D pricing involves complex negotiations, manufacturer rebates, and risk corridors that may affect actual realized savings. The modeled savings represent potential system level cost reductions, not guaranteed budget impacts.
The data reveals prescribing patterns but not the clinical reasoning behind them. High brand utilization could reflect various factors: prescriber preference based on clinical experience, patient request, prior authorization requirements favoring certain products, or clinical considerations not captured in administrative data. The analysis identifies the pattern but cannot definitively explain all contributing factors.
This analysis of 2023 Medicare Part D data reveals that $57.3 billion was spent on 454 drugs with generic alternatives available, and that a 25% shift from brand to generic could have saved Medicare $3.84 billion in a single year. The opportunity was particularly concentrated in insulin products, where 93-99% brand utilization persisted despite generic availability and despite the $35 insulin cost cap that took effect in January 2023.8
These findings represent a historical snapshot of Medicare Part D immediately before major structural reforms took effect. The IRA's Part D redesign, out of pocket spending caps, and Medicare Drug Price Negotiation Program have fundamentally altered the landscape since 2023.101112 However, this analysis demonstrates the scale of the generic adoption opportunity that existed in the prereform period and raises important questions about prescribing patterns, policy effectiveness, and system level cost management.
As Medicare continues to evolve, the core insight remains relevant: when therapeutically equivalent generic drugs exist,35 ensuring their adoption represents one of the most straightforward pathways to controlling prescription drug spending while maintaining patient access to effective medications. The question this analysis raises isn't whether generic substitution works. The FDA has already certified these drugs as therapeutically equivalent.17 The question is what barriers prevent it from happening more systematically, and whether policy reforms address prescribing behavior or simply shift who bears the cost.
This analysis used two publicly available Centers for Medicare & Medicaid Services (CMS) datasets from 2023:
Both datasets are available through the CMS data portal and represent the most comprehensive publicly available information on Medicare Part D prescription activity.
All data processing was conducted in Python using pandas for data manipulation, with analysis organized into seven separate scripts:
This modular approach allowed for transparent verification of each analysis phase and ensured reproducibility of results.
All generic equivalents were verified against FDA Orange Book standards for therapeutic equivalence.35 Only drugs with FDA "A" ratings (indicating pharmaceutical equivalence, bioequivalence, and safe substitutability) were included in the analysis. The bioequivalence standard requires that 90% confidence intervals for generic to brand ratios of Cmax and AUC fall within 80-125%.4
Potential savings were calculated using the formula:
Savings = Brand Spending × Conversion Rate × Generic Discount Rate
Where:
This approach models potential system level savings, not guaranteed budget impacts, as actual Medicare Part D pricing involves manufacturer rebates and risk corridors not fully captured in the public datasets.
All calculations underwent multiple validation checks:
The analysis produced seven Tableau-ready CSV files:
tableau_top20_drugs.csv - Top 20 drugs by savings potentialtableau_all_drugs.csv - All 454 drugs with categorizationstableau_savings_scenarios.csv - Savings at multiple conversion ratestableau_state_analysis.csv - State-level generic adoption ratestableau_brand_vs_generic_costs.csv - Cost per claim comparisonstableau_summary_stats.csv - Key statistics for dashboard KPIstableau_insulin_analysis.csv - Insulin-specific deep diveThe CMS public datasets provide comprehensive coverage but have some limitations:
These limitations are inherent to working with publicly available administrative data and affect what questions can be definitively answered.
All factual claims in this analysis are supported by the following sources:
Centers for Medicare & Medicaid Services (CMS) Public Datasets (2023):
Available at: https://data.cms.gov/