Abstract
This paper studies optimal treatment allocations for two treatment comparisons when the outcome is ordinal and analyzed by a proportional odds cumulative logits model. The variance of the treatment effect estimator is used as optimality criterion. The optimal design is sought so that this variance is minimal for a given total sample size or a given budget, meaning that the power for the test on treatment effect is maximal, or it is sought so that a required power level is achieved at a minimal total sample size or budget. Results are presented for three, five and seven ordered response categories, three treatment effect sizes and a skewed, bell-shaped or polarized distribution of the response probabilities. The optimal proportion subjects in the intervention condition decreases with the number of response categories and the costs for the intervention relative to those for the control. The relation between the optimal proportion and effect size depends on the distribution of the response probabilities. The widely used balanced design is not always the most efficient; its efficiency as compared to the optimal design decreases with increasing cost ratio. The optimal design is highly robust to misspecification of the response probabilities and treatment effect size. The optimal design methodology is illustrated using two pharmaceutical examples. A Shiny app is available to find the optimal treatment allocation, to evaluate the efficiency of the balanced design and to study the relation between budget or sample size and power.
Original language | English |
---|---|
Article number | e0250119 |
Number of pages | 17 |
Journal | PLoS One |
Volume | 16 |
Issue number | 4 |
DOIs | |
Publication status | Published - Apr 2021 |
Keywords
- Data Interpretation, Statistical
- Humans
- Logistic Models
- Models, Statistical
- Probability
- Sample Size
Fingerprint
Dive into the research topics of 'Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model'. Together they form a unique fingerprint.
View full fingerprint
Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver
Moerbeek, M. (2021). Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model. PLoS One, 16(4), Article e0250119. https://doi.org/10.1371/journal.pone.0250119
Moerbeek, Mirjam. / Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model. In: PLoS One. 2021 ; Vol. 16, No. 4.
@article{bb09d7ddf9234055b9ccc48f094b31f6,
title = "Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model",
abstract = "This paper studies optimal treatment allocations for two treatment comparisons when the outcome is ordinal and analyzed by a proportional odds cumulative logits model. The variance of the treatment effect estimator is used as optimality criterion. The optimal design is sought so that this variance is minimal for a given total sample size or a given budget, meaning that the power for the test on treatment effect is maximal, or it is sought so that a required power level is achieved at a minimal total sample size or budget. Results are presented for three, five and seven ordered response categories, three treatment effect sizes and a skewed, bell-shaped or polarized distribution of the response probabilities. The optimal proportion subjects in the intervention condition decreases with the number of response categories and the costs for the intervention relative to those for the control. The relation between the optimal proportion and effect size depends on the distribution of the response probabilities. The widely used balanced design is not always the most efficient; its efficiency as compared to the optimal design decreases with increasing cost ratio. The optimal design is highly robust to misspecification of the response probabilities and treatment effect size. The optimal design methodology is illustrated using two pharmaceutical examples. A Shiny app is available to find the optimal treatment allocation, to evaluate the efficiency of the balanced design and to study the relation between budget or sample size and power.",
keywords = "Data Interpretation, Statistical, Humans, Logistic Models, Models, Statistical, Probability, Sample Size",
author = "Mirjam Moerbeek",
note = "Publisher Copyright: Copyright: {\textcopyright} 2021 Mirjam Moerbeek. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.",
year = "2021",
month = apr,
doi = "10.1371/journal.pone.0250119",
language = "English",
volume = "16",
journal = "PLoS One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "4",
}
Moerbeek, M 2021, 'Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model', PLoS One, vol. 16, no. 4, e0250119. https://doi.org/10.1371/journal.pone.0250119
Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model. / Moerbeek, Mirjam.
In: PLoS One, Vol. 16, No. 4, e0250119, 04.2021.
Research output: Contribution to journal › Article › Academic › peer-review
TY - JOUR
T1 - Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model
AU - Moerbeek, Mirjam
N1 - Publisher Copyright:Copyright: © 2021 Mirjam Moerbeek. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
PY - 2021/4
Y1 - 2021/4
N2 - This paper studies optimal treatment allocations for two treatment comparisons when the outcome is ordinal and analyzed by a proportional odds cumulative logits model. The variance of the treatment effect estimator is used as optimality criterion. The optimal design is sought so that this variance is minimal for a given total sample size or a given budget, meaning that the power for the test on treatment effect is maximal, or it is sought so that a required power level is achieved at a minimal total sample size or budget. Results are presented for three, five and seven ordered response categories, three treatment effect sizes and a skewed, bell-shaped or polarized distribution of the response probabilities. The optimal proportion subjects in the intervention condition decreases with the number of response categories and the costs for the intervention relative to those for the control. The relation between the optimal proportion and effect size depends on the distribution of the response probabilities. The widely used balanced design is not always the most efficient; its efficiency as compared to the optimal design decreases with increasing cost ratio. The optimal design is highly robust to misspecification of the response probabilities and treatment effect size. The optimal design methodology is illustrated using two pharmaceutical examples. A Shiny app is available to find the optimal treatment allocation, to evaluate the efficiency of the balanced design and to study the relation between budget or sample size and power.
AB - This paper studies optimal treatment allocations for two treatment comparisons when the outcome is ordinal and analyzed by a proportional odds cumulative logits model. The variance of the treatment effect estimator is used as optimality criterion. The optimal design is sought so that this variance is minimal for a given total sample size or a given budget, meaning that the power for the test on treatment effect is maximal, or it is sought so that a required power level is achieved at a minimal total sample size or budget. Results are presented for three, five and seven ordered response categories, three treatment effect sizes and a skewed, bell-shaped or polarized distribution of the response probabilities. The optimal proportion subjects in the intervention condition decreases with the number of response categories and the costs for the intervention relative to those for the control. The relation between the optimal proportion and effect size depends on the distribution of the response probabilities. The widely used balanced design is not always the most efficient; its efficiency as compared to the optimal design decreases with increasing cost ratio. The optimal design is highly robust to misspecification of the response probabilities and treatment effect size. The optimal design methodology is illustrated using two pharmaceutical examples. A Shiny app is available to find the optimal treatment allocation, to evaluate the efficiency of the balanced design and to study the relation between budget or sample size and power.
KW - Data Interpretation, Statistical
KW - Humans
KW - Logistic Models
KW - Models, Statistical
KW - Probability
KW - Sample Size
UR - http://www.scopus.com/inward/record.url?scp=85104510504&partnerID=8YFLogxK
U2 - 10.1371/journal.pone.0250119
DO - 10.1371/journal.pone.0250119
M3 - Article
C2 - 33882086
SN - 1932-6203
VL - 16
JO - PLoS One
JF - PLoS One
IS - 4
M1 - e0250119
ER -
Moerbeek M. Optimal allocations for two treatment comparisons within the proportional odds cumulative logits model. PLoS One. 2021 Apr;16(4):e0250119. doi: 10.1371/journal.pone.0250119