Abstract Title

Evaluation of effect of adherence patterns on the sample size and power: A simulation study

Presenter Name

Surulivelrajan Mallayasamy

RAD Assignment Number

1917

Abstract

Purpose: Adherence to medication regimens is an important factor contributing to the success of a therapy both in clinical trials and practice. The objective of our study was to evaluate the effect of adherence patterns on the sample size and power of a clinical trial using population pharmacokinetic (PK)-pharmacodynamic (PD) model-based simulations linked to quantitative adherence models.

Methods: Longitudinal plasma concentration (PK) and pharmacological effect (PD) data were simulated in n=200 individuals per each group of test and standard-of-care (SOC) in each dataset. The population PK model used was a two compartment model with oral absorption. The PD model used was an indirect response inhibitory model. Two scenarios of PK behavior, A-short half-life (~12 hours) and B-long-half life (~35 hours) were simulated by altering the clearance parameter. Two scenarios of PD behavior, C-slower onset of effect (~4 weeks) and D- faster onset (~2 weeks) were simulated by altering the fractional turnover rate. Commonly seen drug PK-PD characteristics were generated by a combination of AC, AD, BC and BD scenarios. Non-Adherence, in terms of dose omissions (0-50%), was simulated as binary variable (missing a dose-0, taking a dose-1) using a discrete time first order Markov model. Test and SOC groups varied in their potency parameter in the PD model (EC50) such that test showed superior effect. Simulations were conducted using NONMEM software. The standard deviation (SD) of the effect at the 5th week of treatment was calculated from the simulated data and used for power and sample size calculations assuming various effect sizes.

Results: Increasing non-adherence increased the variability (SD) of outcome in the simulated trials. The drug feature of long-half life with faster onset (BD) was more tolerant to the effects of non-adherence on statistical power. The drug feature of short half-life with slower onset (AC) was the most affected type by non-adherence. The sample size requirements could double depending on the adherence level and effect size. For smaller effect sizes, non-adherence can cause a significant drop in power and require large sample sizes.

Conclusions: The effect of non-adherence on sample size and power is a function of drug PK-PD characteristics and effect size. Careful consideration of adherence patterns in clinical trial simulations could provide a valuable tool for designing successful trials.

Research Area

Other

Presentation Type

Poster

This document is currently not available here.

Share

COinS
 

Evaluation of effect of adherence patterns on the sample size and power: A simulation study

Purpose: Adherence to medication regimens is an important factor contributing to the success of a therapy both in clinical trials and practice. The objective of our study was to evaluate the effect of adherence patterns on the sample size and power of a clinical trial using population pharmacokinetic (PK)-pharmacodynamic (PD) model-based simulations linked to quantitative adherence models.

Methods: Longitudinal plasma concentration (PK) and pharmacological effect (PD) data were simulated in n=200 individuals per each group of test and standard-of-care (SOC) in each dataset. The population PK model used was a two compartment model with oral absorption. The PD model used was an indirect response inhibitory model. Two scenarios of PK behavior, A-short half-life (~12 hours) and B-long-half life (~35 hours) were simulated by altering the clearance parameter. Two scenarios of PD behavior, C-slower onset of effect (~4 weeks) and D- faster onset (~2 weeks) were simulated by altering the fractional turnover rate. Commonly seen drug PK-PD characteristics were generated by a combination of AC, AD, BC and BD scenarios. Non-Adherence, in terms of dose omissions (0-50%), was simulated as binary variable (missing a dose-0, taking a dose-1) using a discrete time first order Markov model. Test and SOC groups varied in their potency parameter in the PD model (EC50) such that test showed superior effect. Simulations were conducted using NONMEM software. The standard deviation (SD) of the effect at the 5th week of treatment was calculated from the simulated data and used for power and sample size calculations assuming various effect sizes.

Results: Increasing non-adherence increased the variability (SD) of outcome in the simulated trials. The drug feature of long-half life with faster onset (BD) was more tolerant to the effects of non-adherence on statistical power. The drug feature of short half-life with slower onset (AC) was the most affected type by non-adherence. The sample size requirements could double depending on the adherence level and effect size. For smaller effect sizes, non-adherence can cause a significant drop in power and require large sample sizes.

Conclusions: The effect of non-adherence on sample size and power is a function of drug PK-PD characteristics and effect size. Careful consideration of adherence patterns in clinical trial simulations could provide a valuable tool for designing successful trials.