Comprehensive Medicare Medical Scenario Modeling & Plan Cost Comparison

Study Overview

AISmartBids by One & Done is an smart Medicare optimization platform that identifies significant inefficiencies in current Medicare plan enrollment. Based on a large-scale simulation of over 1 million realistic enrollee scenarios using data aligned with Centers for Medicare & Medicaid Services (CMS), the platform found that most seniors are enrolled in suboptimal plans, costing them an average of $4,300 annually.

The analysis shows that 77% of enrollees could save over $1,000 per year, while 95% of standalone Prescription Drug Plan (PDP) users would benefit financially and have better coverage by switching to Medicare Advantage Prescription Drug (MAPD) plans.

If applied nationally, optimized plan matching could generate up to $137 billion in annual out-of-pocket beneficiary savings across Medicare Advantage and Part D.

The study highlights a systemic issue: broker-driven plan recommendations are often influenced by commission structures, leading to biased guidance that may not prioritize the consumer's best financial or healthcare outcomes. This problem is intensifying as insurers reduce commissions on more cost-effective plans.

AISmartBids by One & Done addresses this gap by removing sales incentives and delivering unbiased, data-driven recommendations. The platform evaluates plans based on each enrollee's actual and projected healthcare usage, presenting a comprehensive “all-in” cost view, which includes premiums, deductibles, co-pays, and ancillary benefits like dental, vision, and hearing.

Additionally, the platform provides ongoing value through automatic annual re-shopping, ensuring users remain in the most cost-effective and appropriate plan as market conditions evolve.

Methodology Summary

The Medicare test scenario generation methodology models realistic healthcare and insurance situations for 1,170 synthetic users, each assigned 100 unique scenarios—yielding a total of 1,170,000 simulations. Income levels start at a minimum of $10,000, with a low-income bias applied to specific user types (e.g., all D-SNP users and 25% of PDP-only users have incomes below the 2025 Federal Poverty Level threshold of $21,150). Medicare Part A premiums are set to $0 for over 99% of users, while Part B and Part D premiums are determined based on income brackets published by the Centers for Medicare & Medicaid Services (CMS) and adjusted gross income averages from IRS data.

Prescription drug needs are assigned based on usage tiers (low, medium, high) and are randomly selected from a list of 24 of the most common medications for older adults. Medical usage profiles reflect a range of realistic healthcare interactions, including hospitalizations and outpatient care.

Cost estimates for services and premiums incorporate both fixed and randomized components to mirror real-world variation. Scenario logic also introduces weighted preferences for Special Needs Plans (SNP, ~20%), standalone Part D (PDP-only, ~30%), and out-of-network coverage (~20%). These inputs, combined with randomized variation and statistical weighting, ensure a diverse and representative dataset that reflects realistic user behavior and cost outcomes.

The sections below detail the methodology, input assumptions, weighting strategies, and data sources used in this study.

Constants & Randomization

• Number of Scenarios: 1,170,000
  • There are 1,170 users and each user has 100 scenarios.
• Number of Synthetic Users: 1,170
  • Patient names and street addresses are sourced from the users.csv file in the Synthea dataset (limited to Massachusetts). To ensure geographic diversity, each record is supplemented with a randomly selected city, county, and state from the USGS GNIS PopulatedPlaces_National.txt file. All locations are validated against CMS public data to confirm available Medicare plans.
• Randomization: Scenario variability is introduced using documented data and techniques designed to reflect realistic, real-world fluctuations.
  • Offsets: A uniform random offset is applied to base values. Descriptions such as "±X%" or "±Y" indicate a value selected randomly within that range above or below the base.
  • Weighting: Where applicable—premiums, income levels, filing statuses—values are selected based on weighted probabilities derived from real-world data.
  • Randomness: Drug lists and individual scenario configurations are randomized to ensure diversity across the dataset.
• Special Needs Plans (SNPs): Includes I-SNP (institutional long-term care), D-SNP (dual eligibility), C-SNP (chronic conditions), and PACE (Programs of All-inclusive Care).
  • I-SNPs are assigned a higher weight than suggested by existing documentation to provide greater visibility for PACE plans, which tend to be underrepresented without this adjustment.
  • Approximately 20% of scenarios are designated as SNPs (±2%), distributed as: 80% D-SNP, 15% I-SNP (including PACE), and 5% C-SNP.
  • Sources: Kaiser Family Foundation (KFF); Better Medicare Alliance
• PDP Only:
  • KFF data projects approximately 34% of beneficiaries will choose standalone Part D PDPs through 2024-2025, declining to ~30% by 2026.
  • The system applies a ~30% preference for standalone PDPs (±5%). Without this adjustment, PDPs would rarely appear in results since they are typically not the most cost-effective option.
• Out-of-Network (OON): Estimated at ~20% of users (±2%). Since most people favor HMO plans that do not permit OON coverage, this estimate will be refined as additional data becomes available.

Premiums & Income

See our Medicare Information page to learn more about Medicare parts and premiums.

• Medicare Part A Premiums: Set by CMS based on qualifying work quarters. Since 99% of individuals have 40 or more qualifying quarters, a $0 Part A premium is assigned to all users.
  • Sources: CMS Fact Sheet; Medicare Costs
• Medicare Part B and Part D Premiums: Calculated using each individual's adjusted gross income and tax filing status from two years prior.
  • Sources: CMS Fact Sheet; Medicare Costs
• FPL Threshold: $21,150 — used to determine eligibility for Dual Eligibility and Extra Help.
  • Source: U.S. Department of Health and Human Services (2-person household, 2025)
• Minimum Income: $10,000 — low enough to qualify for subsidies, with a floor to prevent unrealistically low income generation.
• Filing Status: Based on IRS statistical weights for the 65+ age group.

  Status	Weight
  Individual	47%
  Joint	51%
  Separate	2%

• Adjusted Gross Income: Based on IRS statistical data for the 65+ age group.
  • 90% assigned income < $212K (includes low-income scenarios)
  • 4% between $212K - $266K
  • 4% between $266K - $334K
  • 1% between $334K - $750K
  • 1% above $750K
• Low Income Bias: All D-SNP users and 25% of PDP-only users are assigned an income ≤ FPL threshold, triggering subsidies like Dual Eligibility and Extra Help.
  • Sources: Kaiser Family Foundation (KFF); IRS

Medical Usage

Each scenario is assigned one of three usage tiers (low, medium, high), used to calculate past and projected costs using fixed or randomized cost estimates.

• Usage Scenarios: Past and future costs are aligned by usage tier to ensure meaningful comparisons. If a low-usage profile is selected for future costs, it is paired with a low-usage profile for past costs. This answers: "If you had this plan last year under similar usage, what would your costs have been?" Matching usage levels avoids unrealistic comparisons such as high past usage paired with low future usage, which could artificially inflate perceived savings.

Scenario Data by Usage Tier

• Common Drugs: The system uses the 24 most commonly prescribed medications for adults aged 65+, using average dosage amounts. Compiled from FDA and American Society of Health-System Pharmacists (ASHP) data.

  Drug Name (Generic)
  amlodipine 5 MG Oral Tablet
  24 HR nifedipine 30 MG Extended Release Oral Tablet
  prednisone 50 MG Oral Tablet
  montelukast 10 MG Oral Tablet
  omeprazole 40 MG Delayed Release Oral Capsule
  albuterol 0.09 MG/ACTUAT Metered Dose Inhaler (200 actuations)
  citalopram 30 MG Oral Capsule
  clopidogrel 75 MG Oral Tablet
  fluoxetine 20 MG Oral Capsule
  gabapentin 800 MG Oral Tablet
  hydrochlorothiazide 25 MG Oral Tablet
  sertraline 100 MG Oral Tablet
  simvastatin 20 MG Oral Tablet
  furosemide 40 MG Oral Tablet
  lisinopril 20 MG Oral Tablet
  esomeprazole 40 MG Delayed Release Oral Capsule
  duloxetine 40 MG Delayed Release Oral Capsule
  atorvastatin 40 MG Oral Tablet
  tamsulosin hydrochloride 0.4 MG Oral Capsule
  24 HR metoprolol succinate 100 MG Extended Release Oral Tablet
  levothyroxine sodium 0.1 MG Oral Tablet
  alendronic acid 10 MG Oral Tablet
  losartan potassium 50 MG Oral Tablet
  24 HR bupropion hydrochloride 300 MG Extended Release Oral Tablet

• Drugs by Usage Tier:
  • Low: 2 prescriptions/year — Typically healthy individuals with few or no chronic conditions.
  • Medium: 5 prescriptions/year — Individuals managing one or more stable chronic conditions (e.g., high blood pressure, high cholesterol, arthritis).
  • High: 10 prescriptions/year — Individuals with multiple chronic conditions (e.g., diabetes, heart disease, COPD, mental health disorders).
  • Sources: NIH; Georgetown University; CDC NCHS; Kaiser Family Foundation (KFF)
• Drug Costs: Calculated by AISmartBids using national averages and the coverage levels provided by individual plans.

Costs

This table summarizes typical medical expenses without insurance coverage, serving as the foundation for the cost estimates used by AISmartBids.

• Past Cost Calculations: Each user's cost profile is generated using a unique set of parameters to reflect realistic variation. Scenario-specific usage levels (low, medium, or high) are randomly assigned and determine the number of visits, procedures, and other medical interactions.
  • Primary Doctor: Scenario PCP visits x average visit cost.
  • Specialists: Scenario specialist visits x average visit cost.
  • Urgent Care: Scenario urgent care visits x average visit cost.
  • ER Visits: Scenario ER visits x a random amount ±$50 of average ER cost.
  • Hospital: Combined short and long hospital visits x a random amount ±$200 of average per-day cost.
  • Dental: Cost of one annual dental exam + scenario dental procedures x a random amount ±$150 of average procedure cost.
  • Vision: Cost of one annual eye exam + scenario glasses or contacts x a random amount ±$25 of average cost.
  • Hearing: Scenario hearing aids x a random amount ±$300 of average hearing aid cost.
  • Health Premiums: Random amount ±$100 of average premium cost.
  • Drug Premiums: Random amount ±$150 of average drug premium cost.
  • Prescriptions: Random amount ±$50 of average drug cost for Medium usage; half that for Low; double for High.
  • Other Costs: Set to $0 for testing purposes (users may enter miscellaneous costs in the live tool).
• Future Cost Calculations: The tool evaluates each scenario by applying its costs to every health plan available in the user's area—revealing the actual cost per plan and highlighting the difference between those costs and the average scenario cost.