Faculty Mentor
Karl E. Peace
Location
Russell Union Ballroom
Type of Research
On-going
Session Format
Poster Presentation
College
Jiann-Ping Hsu College of Public Health
Department
Department of Biostatistics, Epidemiology, and Environmental Health Sciences
Abstract
The LARGO trial established rasagiline (1 mg/day) as an effective adjunct to levodopa in Parkinson’s disease (PD) patients with motor fluctuations, demonstrating clinically meaningful reductions in daily “off-time” (poor or absent motor function) and improvements in Unified Parkinson’s Disease Rating Scale (UPDRS)/Clinical Global Improvement (CGI) endpoints compared with placebo. Building on this evidence base, we propose a design strategy that re-uses LARGO-like assumptions while improving statistical efficiency for future PD evaluations of rasagiline versus placebo. Methodologically, we leverage key results emphasizing the Balaam 4-sequence, 2-period design (AA, AB, BA, BB). Under the standard crossover model with direct and carryover effects, the Balaam direct-effect contrast is constructed so that period effects and carryover effects cancel algebraically, yielding a direct treatment difference estimator that is not conditional on carryover, unlike the classic Grizzle 2×2 estimator, where the direct effect is biased unless differential carryover is absent. The ASA Biopharmaceutical Section Proceedings discussion by Peace further highlights that incomplete “responders-only crossover” structures (e.g., White-type designs) can reintroduce carryover contamination and reduce efficiency due to response-dependent second-period sample sizes. Using LARGO-motivated parameters for change in daily off-time (rasagiline ≈ −1.2 hours; placebo ≈ −1.0 hours; SD ≈ 2; NI margin example −0.5), we outline a simulation framework comparing a parallel non-inferiority analysis to a Balaam-style crossover embedded in a randomized-withdrawal design to exploit within-subject contrasts while respecting ethical continuity of benefit. These design principles motivate a PD trial blueprint that maintains LARGO’s clinical endpoint while reducing sensitivity to carryover and improving inferential efficiency for rasagiline versus placebo. The anticipated contribution is a PD-focused trial strategy that strengthens the interpretability of treatment comparisons by design, separating direct effects from carryover sensitivity while transparently characterizing efficiency trade-offs for realistic PD heterogeneity.
Program Description
.
DOI
10.20429/GS4.2026.024
Start Date
4-23-2026 2:00 PM
End Date
4-23-2026 4:00 PM
Recommended Citation
Biswas, Purbasha; Kersey, Jing X.; and Peace, Karl E., "A Balaam-Modified Approach to LARGO-Like Parkinson’s Trials: Efficient Estimation of Rasagiline Effects Under Motor Fluctuations" (2026). GS4 Student Scholars Symposium. 219.
https://digitalcommons.georgiasouthern.edu/research_symposium/2026/2026/219
A Balaam-Modified Approach to LARGO-Like Parkinson’s Trials: Efficient Estimation of Rasagiline Effects Under Motor Fluctuations
Russell Union Ballroom
The LARGO trial established rasagiline (1 mg/day) as an effective adjunct to levodopa in Parkinson’s disease (PD) patients with motor fluctuations, demonstrating clinically meaningful reductions in daily “off-time” (poor or absent motor function) and improvements in Unified Parkinson’s Disease Rating Scale (UPDRS)/Clinical Global Improvement (CGI) endpoints compared with placebo. Building on this evidence base, we propose a design strategy that re-uses LARGO-like assumptions while improving statistical efficiency for future PD evaluations of rasagiline versus placebo. Methodologically, we leverage key results emphasizing the Balaam 4-sequence, 2-period design (AA, AB, BA, BB). Under the standard crossover model with direct and carryover effects, the Balaam direct-effect contrast is constructed so that period effects and carryover effects cancel algebraically, yielding a direct treatment difference estimator that is not conditional on carryover, unlike the classic Grizzle 2×2 estimator, where the direct effect is biased unless differential carryover is absent. The ASA Biopharmaceutical Section Proceedings discussion by Peace further highlights that incomplete “responders-only crossover” structures (e.g., White-type designs) can reintroduce carryover contamination and reduce efficiency due to response-dependent second-period sample sizes. Using LARGO-motivated parameters for change in daily off-time (rasagiline ≈ −1.2 hours; placebo ≈ −1.0 hours; SD ≈ 2; NI margin example −0.5), we outline a simulation framework comparing a parallel non-inferiority analysis to a Balaam-style crossover embedded in a randomized-withdrawal design to exploit within-subject contrasts while respecting ethical continuity of benefit. These design principles motivate a PD trial blueprint that maintains LARGO’s clinical endpoint while reducing sensitivity to carryover and improving inferential efficiency for rasagiline versus placebo. The anticipated contribution is a PD-focused trial strategy that strengthens the interpretability of treatment comparisons by design, separating direct effects from carryover sensitivity while transparently characterizing efficiency trade-offs for realistic PD heterogeneity.
