Getting the dose right: The dose selected for a phase 3 program was selected by the client's chief scientist, despite FDA's suggestion that the dose was wrong. Two expensive phase 3 trials failed. My analyses revealed that the dose was 2-fold too small. A new trial with the larger dose produced the exact balance between efficacy and safety.

Getting the dose right: An entire clinical program used weight-based dosing. A pharmacokinetic analysis, commissioned shortly before NDA submission, revealed that exposure did not vary as a function of weight. A high incidence of non-trivial adverse events in heavier subjects (with no increase in efficacy) supported non-weight-based dosing. Pharmacodynamic analyses led to improved dosing regimens.

Getting the dose right: Pharmacokinetic results from two phase 1 PK studies in healthy subjects were reported with weight-normalized units. A weight-based dosing regimen was then used in a phase 2 study in patients: plasma concentrations were 2-fold larger than predicted. The unexpected results led to a 2-year suspension in development. My analyses revealed that total clearance did NOT vary with weight; as a result, the higher-than-expected concentrations could be explained by patients weighing more than healthy subjects.

Antibody development: Simulations of a clinical trial for an antibody revealed that the proposed trial was likely to fail. The company ignored these recommendations; the trial failed as predicted. The company asked that I propose a new dosing regimen, simulate this regimen, and use this information to convince regulatory authorities to permit trials to resume.

Real-time analyses: An early-stage company needed rapid analyses from a first-in-man trial to permit selection of dosing regimens for subsequent trials (a decision was needed in < 5 days). In 3 days, I provided preliminary NONMEM analyses, simulated dosing regimens for the second trial, and recommended improved sampling regimens.

Diligence: A pharma company in-licensed a novel delivery system: mean data available during diligence suggested that the desired delivery profile would be easy to achieve. My analyses revealed that no subject had an acceptable release profile and that mean data were misleading. The company decided that additional phase I data was needed before further development.