Pharmaceutical Evidence PlatformAccelerating Drug Approvals with
Causal Inference
Variacle integrates pharmaceutical evidence from multi-center clinical databases to dramatically reduce patient recruitment burden and accelerate the path to drug approval with statistical guarantees.
Example clinical trial: Bisoprolol vs Metoprolol
Total Sample Size Impact
ORIGINAL SAMPLE SIZE
131
65 + 66 patients
→
POST-VARIACLE SAMPLE SIZE
330
65 + 265 patients
This means ~2.5 times more recruitments for free while increasing to 90% success probability, all from integrating evidence.
Study Design
A randomized, open-label study comparing Bisoprolol with Metoprolol Succinate sustained-release on heart rate and blood pressure in hypertensive patients (CREATIVE Trial - NCT01508325).
- •Exposure: Beta-blocker type (Bisoprolol vs Metoprolol)
- •Duration: 12-week treatment period
- •Sample: Bisoprolol (n₁=65), Metoprolol (n₀=66)
Primary Outcome
Change from Baseline in Mean Ambulatory Diastolic Blood Pressure (DBP) in the last 4 hours after 12-week treatment measured via 24-hour ambulatory blood pressure monitoring (ABPM).
Sample Size Analysis
Pre-Variacle Distribution
Distribution of mean estimates with SEM = σ/√n for each group.
Bisoprolol: Mean = -4.45 mmHg, SEM = 1.33 (SD = 10.74 for n=65)
Metoprolol: Mean = -3.39 mmHg, SEM = 1.38 (SD = 11.23 for n=66)
Bisoprolol (n=65)
Mean: -4.45
SD: 1.33
Sample Size: 65
Metoprolol (n=66)
Mean: -3.39
SD: 1.38
Sample Size: 66
Variacle's - Evidence Integration
Using the eICU Collaborative Research Database
Data Source
eICU Collaborative Research Database - A freely available multi-center database for critical care research containing over 139,000 critical care patient records from 208 hospitals.
Distribution Comparison
Pre-Variacle vs Post-Variacle
Pre-Variacle: Traditional Approach
Bisoprolol (n=65)
Mean: -4.45
SD: 1.33
Metoprolol (n=66)
Mean: -3.39
SD: 1.38
Post-Variacle: With eICU Evidence Integration
Bisoprolol (refined)
Mean: -4.45
SD: 1.33
Metoprolol (refined)
Mean: -3.08
SD: 0.69
Understanding Statistical Power
Statistical power is the probability that a hypothesis test will correctly detect a true effect when one exists. It represents the probability of rejecting the null hypothesis when it is false (avoiding a Type II error).
Given a desired power level and significance level, we can calculate the required sample size of a control group using the following formula. For a fixed sample size n₁ in the treatment group and control group sample size n₀ with significance level α, power (1-β), common standard deviation σ, and effect size δ:
Sample Size for Control Group (n₀):
Applied Example: CHD Risk Reduction
According to meta-analysis by Law et al. ("Use of blood pressure lowering drugs in the prevention of cardiovascular disease: meta-analysis of 147 randomised trials in the context of expectations from prospective epidemiological studies"), a diastolic blood pressure reduction of δ = 5 mmHg corresponds to approximately a 5% reduction in coronary heart disease (CHD) events.
FALSE DISCOVERY RATE
5%
α = 0.05
DESIRED POWER
90%
Power = 0.90
NEEDED METOPROLOL SAMPLE SIZE
234
n₀ required (90% power)
Variacle's Achievement
By refining the Metoprolol distribution using eICU evidence, Variacle achieves an effective sample size of 265:
neff = σ² / σ²Variacle = 265
Total Sample Size Impact
ORIGINAL SAMPLE SIZE
131
65 + 66 patients
→
POST-VARIACLE SAMPLE SIZE
330
65 + 265 patients
This means ~2.5 times more recruitments for free while increasing to 90% success probability, all from integrating evidence.