DDMODEL00000194: Likert Pain Score Modeling: A Markov Integer Model and an Autoregressive Continuous Model

  public model
Short description:
Likert Pain Score Modeling: A Markov Integer Model and an Autoregressive Continuous Model
Original code
  • Likert pain score modeling: a Markov integer model and an autoregressive continuous model.
  • Plan EL, Elshoff JP, Stockis A, Sargentini-Maier ML, Karlsson MO
  • Clinical pharmacology and therapeutics, 5/2012, Volume 91, Issue 5, pages: 820-828
  • Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden. elodie.plan@farmbio.uu.se
  • Pain intensity is principally assessed using rating scales such as the 11-point Likert scale. In general, frequent pain assessments are serially correlated and underdispersed. The aim of this investigation was to develop population models adapted to fit the 11-point pain scale. Daily Likert scores were recorded over 18 weeks by 231 patients with neuropathic pain from a clinical trial placebo group. An integer model consisting of a truncated generalized Poisson (GP) distribution with Markovian transition probability inflation was implemented in NONMEM 7.1.0. It was compared to a logit-transformed autoregressive continuous model with correlated residual errors. In both models, the score baseline was estimated to be 6.2 and the placebo effect to be 19%. Developed models similarly retrieved consistent underlying features of the data and therefore correspond to platform models for drug effect detection. The integer model was complex but flexible, whereas the continuous model can more easily be developed, although requires longer runtimes.
Anders Thorsted
Context of model development: Clinical end-point;
Discrepancy between implemented model and original publication: The publication describes development of two models: A Markov Integer Model (fast but complex) and an Autoregressive Continuous Model (slow but simpler). This upload contains the Markov Integer Model;
Long technical model description: The model was based on 231 subjects randomized to the placebo arm of three Phase III trials (daily measurements of 11-point Likert pain scale over 18 weeks, a total of 22 492 measurements). The structural and statistical components were first developed assuming a constant baseline and independence of observations. Following this, the time course of the scores was investigated and fixed for the exploration of elements accounting for the nonindependence between scores. Lastly, the underdispersion of the scores and inclusion of covariates were tested. The integer model was developed based on the Poisson distribution, with right-truncation implemented to keep scores in the finite range [0, 10] while ensuring that the sum of all possibile probabilities remained 1. For the time-course, the mean count (lambda) can drift over time, and in the model an exponential decay over time was implemented to describe this. As count models assume that observations are statistically mutually independent (i.e. that they are not conditioned on each other), discrete Markov properties were included using First-order Markov components (only conditioned on the preceeding observation). Finally, underdispersion (where nonequality is observed between mean and variance) were implemented and found to better account for observed within-individual variability.;
Model compliance with original publication: Yes;
Model implementation requiring submitter’s additional knowledge: No;
Modelling context description: Development of platform models dealing with inferences about stochastic relationships between random variables, exemplified for description of pain scores.;
Modelling task in scope: simulation; estimation;
Nature of research: Approval phase/Registration trial (Phase III);
Therapeutic/disease area: CNS;
Annotations are correct.
This model is not certified.
  • Model owner: Anders Thorsted
  • Submitted: May 26, 2016 10:36:32 AM
  • Last Modified: May 27, 2016 10:03:45 AM
Revisions
  • Version: 14 public model Download this version
    • Submitted on: May 27, 2016 10:03:45 AM
    • Submitted by: Anders Thorsted
    • With comment: Edited model metadata online.
  • Version: 11 public model Download this version
    • Submitted on: May 26, 2016 10:36:32 AM
    • Submitted by: Anders Thorsted
    • With comment: Edited model metadata online.
 
Help