1
people have died from curable diseases
since this page started loading...
💀

Methodology, Parameters, and Calculations

Keywords

health economics methodology, clinical trial cost analysis, medical research ROI, cost-benefit analysis healthcare, sensitivity analysis, Monte Carlo simulation, DALY calculation, pragmatic clinical trials

Overview

This appendix documents all 10 parameters used in the analysis, organized by type:

  • External sources (peer-reviewed): 8
  • Calculated values: 2
  • Core definitions: 0

Quick Navigation

Calculated Values (2 parameters) • External Data Sources (8 parameters)

Calculated Values

Parameters derived from mathematical formulas and economic models.

Drug Cost Increase: Pre-1962 to Current: 105x

Drug development cost increase from pre-1962 to current ($24.7M → $2.6B = 105×)

Inputs:

\[ \begin{gathered} k_{cost,pre62} \\ = \frac{Cost_{dev,curr}}{Cost_{pre62,24}} \\ = \frac{\$2.6B}{\$24.7M} \\ = 105 \end{gathered} \]

Methodology:80

✓ High confidence

Sensitivity Analysis

Sensitivity Indices for Drug Cost Increase: Pre-1962 to Current

Regression-based sensitivity showing which inputs explain the most variance in the output.

Input Parameter Sensitivity Coefficient Interpretation
Pharma Drug Development Cost Current 1.3110 Strong driver
Pre 1962 Drug Development Cost 2024 Usd -0.3181 Moderate driver

Interpretation: Standardized coefficients show the change in output (in SD units) per 1 SD change in input. Values near ±1 indicate strong influence; values exceeding ±1 may occur with correlated inputs.

Monte Carlo Distribution

Monte Carlo Distribution: Drug Cost Increase: Pre-1962 to Current (10,000 simulations)

Monte Carlo Distribution: Drug Cost Increase: Pre-1962 to Current (10,000 simulations)

Simulation Results Summary: Drug Cost Increase: Pre-1962 to Current

Statistic Value
Baseline (deterministic) 105x
Mean (expected value) 104x
Median (50th percentile) 104x
Standard Deviation 9.03x
90% Range (5th-95th percentile) [90.6x, 119x]

The histogram shows the distribution of Drug Cost Increase: Pre-1962 to Current across 10,000 Monte Carlo simulations. The CDF (right) shows the probability of the outcome exceeding any given value, which is useful for risk assessment.

Exceedance Probability

Probability of Exceeding Threshold: Drug Cost Increase: Pre-1962 to Current

Probability of Exceeding Threshold: Drug Cost Increase: Pre-1962 to Current

This exceedance probability chart shows the likelihood that Drug Cost Increase: Pre-1962 to Current will exceed any given threshold. Higher curves indicate more favorable outcomes with greater certainty.

RECOVERY Trial Cost Reduction Factor: 82x

Cost reduction factor demonstrated by RECOVERY trial ($41K traditional / $500 RECOVERY = 82x)

Inputs:

\[ \begin{gathered} k_{RECOVERY} \\ = \frac{Cost_{P3,pt}}{Cost_{RECOVERY,pt}} \\ = \frac{\$41K}{\$500} \\ = 82 \end{gathered} \]

Methodology:66

✓ High confidence

Sensitivity Analysis

Sensitivity Indices for RECOVERY Trial Cost Reduction Factor

Regression-based sensitivity showing which inputs explain the most variance in the output.

Input Parameter Sensitivity Coefficient Interpretation
Recovery Trial Cost Per Patient -2.4783 Strong driver
Traditional Phase3 Cost Per Patient 2.4635 Strong driver

Interpretation: Standardized coefficients show the change in output (in SD units) per 1 SD change in input. Values near ±1 indicate strong influence; values exceeding ±1 may occur with correlated inputs.

Monte Carlo Distribution

Monte Carlo Distribution: RECOVERY Trial Cost Reduction Factor (10,000 simulations)

Monte Carlo Distribution: RECOVERY Trial Cost Reduction Factor (10,000 simulations)

Simulation Results Summary: RECOVERY Trial Cost Reduction Factor

Statistic Value
Baseline (deterministic) 82x
Mean (expected value) 71.2x
Median (50th percentile) 72.4x
Standard Deviation 15.3x
90% Range (5th-95th percentile) [50x, 94.1x]

The histogram shows the distribution of RECOVERY Trial Cost Reduction Factor across 10,000 Monte Carlo simulations. The CDF (right) shows the probability of the outcome exceeding any given value, which is useful for risk assessment.

Exceedance Probability

Probability of Exceeding Threshold: RECOVERY Trial Cost Reduction Factor

Probability of Exceeding Threshold: RECOVERY Trial Cost Reduction Factor

This exceedance probability chart shows the likelihood that RECOVERY Trial Cost Reduction Factor will exceed any given threshold. Higher curves indicate more favorable outcomes with greater certainty.

External Data Sources

Parameters sourced from peer-reviewed publications, institutional databases, and authoritative reports.

CPI Multiplier: 1980 to 2024: 3.8:1

CPI inflation multiplier from 1980 to 2024 (280.48% cumulative inflation)

Source:9

Uncertainty Range

Technical: 95% CI: [3.75:1, 3.85:1] • Distribution: Normal

What this means: We’re quite confident in this estimate. The true value likely falls between 3.75:1 and 3.85:1 (±1%). This represents a narrow range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The normal distribution means values cluster around the center with equal chances of being higher or lower.

Input Distribution

Probability Distribution: CPI Multiplier: 1980 to 2024

Probability Distribution: CPI Multiplier: 1980 to 2024

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence

Average Annual New Drug Approvals Globally: 50 drugs/year

Average annual new drug approvals globally

Source:13

Uncertainty Range

Technical: 95% CI: [45 drugs/year, 60 drugs/year] • Distribution: Lognormal

What this means: This estimate has moderate uncertainty. The true value likely falls between 45 drugs/year and 60 drugs/year (±15%). This represents a reasonable range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Average Annual New Drug Approvals Globally

Probability Distribution: Average Annual New Drug Approvals Globally

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence

Pharma Drug Development Cost (Current System): $2.60B

Average cost to develop one drug in current system

Source:68

Uncertainty Range

Technical: 95% CI: [$1.50B, $4B] • Distribution: Lognormal (SE: $500M)

What this means: There’s significant uncertainty here. The true value likely falls between $1.50B and $4B (±48%). This represents a wide range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Pharma Drug Development Cost (Current System)

Probability Distribution: Pharma Drug Development Cost (Current System)

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence • 📊 Peer-reviewed

Pragmatic Trial Median Cost per Patient (PMC Review): $97

Median cost per patient in embedded pragmatic clinical trials (systematic review of 64 trials). IQR: $19-$478 (2015 USD).

Source:74

Uncertainty Range

Technical: 95% CI: [$19, $478] • Distribution: Lognormal

What this means: This estimate is highly uncertain. The true value likely falls between $19 and $478 (±237%). This represents a very wide range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Pragmatic Trial Median Cost per Patient (PMC Review)

Probability Distribution: Pragmatic Trial Median Cost per Patient (PMC Review)

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence

Pre-1962 Drug Development Cost (1980 Dollars): $6.50M

Average drug development cost before 1962 FDA efficacy regulations, adjusted to 1980 dollars (Baily 1972)

Source:80

Uncertainty Range

Technical: 95% CI: [$5.20M, $7.80M] • Distribution: Lognormal

What this means: This estimate has moderate uncertainty. The true value likely falls between $5.20M and $7.80M (±20%). This represents a reasonable range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Pre-1962 Drug Development Cost (1980 Dollars)

Probability Distribution: Pre-1962 Drug Development Cost (1980 Dollars)

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence • 📊 Peer-reviewed

Pre-1962 Drug Development Cost (2024 Dollars): $24.7M

Pre-1962 drug development cost adjusted to 2024 dollars ($6.5M × 3.80 = $24.7M, CPI-adjusted from Baily 1972)

Source:80

Uncertainty Range

Technical: 95% CI: [$19.5M, $30M] • Distribution: Lognormal

What this means: This estimate has moderate uncertainty. The true value likely falls between $19.5M and $30M (±21%). This represents a reasonable range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Pre-1962 Drug Development Cost (2024 Dollars)

Probability Distribution: Pre-1962 Drug Development Cost (2024 Dollars)

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence • 📊 Peer-reviewed

Recovery Trial Cost per Patient: $500

RECOVERY trial cost per patient. Note: RECOVERY was an outlier - hospital-based during COVID emergency, minimal extra procedures, existing NHS infrastructure, streamlined consent. Replicating this globally will be harder.

Source:83

Uncertainty Range

Technical: 95% CI: [$400, $2.50K] • Distribution: Lognormal

What this means: This estimate is highly uncertain. The true value likely falls between $400 and $2.50K (±210%). This represents a very wide range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Recovery Trial Cost per Patient

Probability Distribution: Recovery Trial Cost per Patient

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence

Phase 3 Cost per Patient: $41K

Phase 3 cost per patient (median from FDA study)

Source:101

Uncertainty Range

Technical: 95% CI: [$20K, $120K] • Distribution: Lognormal

What this means: This estimate is highly uncertain. The true value likely falls between $20K and $120K (±122%). This represents a very wide range that our Monte Carlo simulations account for when calculating overall uncertainty in the results.

The lognormal distribution means values can’t go negative and have a longer tail toward higher values (common for costs and populations).

Input Distribution

Probability Distribution: Phase 3 Cost per Patient

Probability Distribution: Phase 3 Cost per Patient

This chart shows the assumed probability distribution for this parameter. The shaded region represents the 95% confidence interval where we expect the true value to fall.

✓ High confidence