Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

The need for mixed-effects modeling with population dichotomous data.

I Yano1, S L Beal, L B Sheiner

  • 1Department of Pharmacy, Kyoto University Hospital, Kyoto, Japan.

Journal of Pharmacokinetics and Pharmacodynamics
|October 27, 2001
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Association between dexamethasone treatment and alterations in serum concentrations of trace metals.

Die Pharmazie·2020
Same author

Population pharmacokinetics and pharmacodynamics of mycophenolic acid using the prospective data in patients undergoing hematopoietic stem cell transplantation.

Bone marrow transplantation·2017
Same author

Efficacy of protocol-based pharmacotherapy management on anticoagulation with warfarin for patients with cardiovascular surgery.

Journal of clinical pharmacy and therapeutics·2017
Same author

Potentially inappropriate medications in elderly Japanese patients: effects of pharmacists' assessment and intervention based on Screening Tool of Older Persons' Potentially Inappropriate Prescriptions criteria ver.2.

Journal of clinical pharmacy and therapeutics·2017
Same author

Monitoring mycophenolate mofetil is necessary for the effective prophylaxis of acute GVHD after cord blood transplantation.

Bone marrow transplantation·2014
Same author

Assessment of four methodologies (microparticle enzyme immunoassay, chemiluminescent enzyme immunoassay, affinity column-mediated immunoassay, and flow injection assay-tandem mass spectrometry) for measuring tacrolimus blood concentration in Japanese liver transplant recipients.

Transplantation proceedings·2014

Sophisticated mixed-effect (ME) models do not always improve pharmacokinetic and pharmacodynamic data analysis over naive (NA) models. Simulation results show ME models can sometimes yield larger errors for fixed-effect parameters.

Area of Science:

  • Pharmacometrics
  • Statistical Modeling
  • Data Analysis

Background:

  • Advanced data analytic techniques, including mixed-effect (ME) models with interindividual random effects, have been developed over 25 years.
  • While ME models can improve population pharmacokinetic (PK) data analysis, they may not benefit all data types, such as some pharmacodynamic (PD) data.
  • These advanced methods incur additional computational costs.

Purpose of the Study:

  • To investigate whether mixed-effect (ME) models with interindividual random effects consistently outperform naive (NA) models without such effects.
  • To analyze simulated dichotomous and continuous population data using both ME and NA models.
  • To evaluate the impact of model choice on the accuracy of fixed-effect parameter estimates.

Main Methods:

Related Experiment Videos

  • Simulation of simple population dichotomous and related continuous data.
  • Analysis of simulated data using mixed-effect (ME) models incorporating interindividual random effects.
  • Analysis of simulated data using naive (NA) models excluding interindividual random effects.
  • Comparison of model performance using maximum likelihood estimation (MLE) and restricted maximum likelihood estimation (REML).

Main Results:

  • Mixed-effect (ME) models did not invariably provide superior results compared to naive (NA) models.
  • Under maximum likelihood estimation (MLE), the root mean square estimation errors for fixed-effect parameters were sometimes larger when using ME models than NA models.
  • Using restricted maximum likelihood estimation (REML) with ME models resulted in comparable root mean square errors to NA models, but still did not demonstrate a consistent, marked advantage.

Conclusions:

  • The assumption that mixed-effect models (ME) always yield improved parameter estimates over naive models (NA) is not universally true.
  • The choice between ME and NA models should be carefully considered based on the specific data characteristics (e.g., PK vs. PD data).
  • Further investigation into the conditions under which ME models offer significant advantages is warranted, considering both computational cost and analytical benefits.