Improving analysis of cognitive outcomes in cardiovascular trials using different statistical approaches.
Beta-binomial
Bounded
Ceiling effect
Cognitive
Generalized linear regression
Tobit
Transformations
Journal
Trials
ISSN: 1745-6215
Titre abrégé: Trials
Pays: England
ID NLM: 101263253
Informations de publication
Date de publication:
02 Oct 2024
02 Oct 2024
Historique:
received:
24
01
2024
accepted:
17
09
2024
medline:
3
10
2024
pubmed:
3
10
2024
entrez:
2
10
2024
Statut:
epublish
Résumé
The Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA) questionnaires are commonly used to measure global cognition in clinical trials. Because these scales are discrete and bounded with ceiling and floor effects and highly skewed, their analysis as continuous outcomes presents challenges. Normality assumptions of linear regression models are usually violated, which may result in failure to detect associations with variables of interest. Alternative approaches to analyzing the results of these cognitive batteries include transformations (standardization, square root, or log transformation) of the scores in the multivariate linear regression (MLR) model, the use of nonlinear beta-binomial regression (which is not dependent on the assumption of normality), or Tobit regression, which adds a latent variable to account for bounded data. We aim to empirically compare the model performance of all proposed approaches using four large randomized controlled trials (ORIGIN, TRANSCEND, COMPASS, and NAVIGATE-ESUS), and using as metrics the Akaike information criterion (AIC). We also compared the treatment effects for the methods that have the same unit of measure (i.e., untransformed MLR, beta-binomial, and Tobit). The beta-binomial consistently demonstrated superior model performance, with the lowest AIC values among nearly all the approaches considered, followed by the MLR with square root and log transformations across all four studies. Notably, in ORIGIN, a substantial AIC reduction was observed when comparing the untransformed MLR to the beta-binomial, whereas other studies had relatively small AIC reductions. The beta-binomial model also resulted in a significant treatment effect in ORIGIN, while the untransformed MLR and Tobit regression showed no significance. The other three studies had similar and insignificant treatment effects among the three approaches. When analyzing discrete and bounded outcomes, such as cognitive scores, as continuous variables, a beta-binomial regression model improves model performance, avoids spurious significance, and allows for a direct interpretation of the actual cognitive measure. ORIGIN (NCT00069784). Registered on October 1, 2003; TRANSCEND (NCT00153101). Registered on September 9, 2005; COMPASS (NCT01776424). Registered on January 24, 2013; NAVIGATE-ESUS (NCT02313909). Registered on December 8, 2014.
Sections du résumé
BACKGROUND
BACKGROUND
The Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA) questionnaires are commonly used to measure global cognition in clinical trials. Because these scales are discrete and bounded with ceiling and floor effects and highly skewed, their analysis as continuous outcomes presents challenges. Normality assumptions of linear regression models are usually violated, which may result in failure to detect associations with variables of interest.
METHODS
METHODS
Alternative approaches to analyzing the results of these cognitive batteries include transformations (standardization, square root, or log transformation) of the scores in the multivariate linear regression (MLR) model, the use of nonlinear beta-binomial regression (which is not dependent on the assumption of normality), or Tobit regression, which adds a latent variable to account for bounded data. We aim to empirically compare the model performance of all proposed approaches using four large randomized controlled trials (ORIGIN, TRANSCEND, COMPASS, and NAVIGATE-ESUS), and using as metrics the Akaike information criterion (AIC). We also compared the treatment effects for the methods that have the same unit of measure (i.e., untransformed MLR, beta-binomial, and Tobit).
RESULTS
RESULTS
The beta-binomial consistently demonstrated superior model performance, with the lowest AIC values among nearly all the approaches considered, followed by the MLR with square root and log transformations across all four studies. Notably, in ORIGIN, a substantial AIC reduction was observed when comparing the untransformed MLR to the beta-binomial, whereas other studies had relatively small AIC reductions. The beta-binomial model also resulted in a significant treatment effect in ORIGIN, while the untransformed MLR and Tobit regression showed no significance. The other three studies had similar and insignificant treatment effects among the three approaches.
CONCLUSION
CONCLUSIONS
When analyzing discrete and bounded outcomes, such as cognitive scores, as continuous variables, a beta-binomial regression model improves model performance, avoids spurious significance, and allows for a direct interpretation of the actual cognitive measure.
TRIALS REGISTRATION
BACKGROUND
ORIGIN (NCT00069784). Registered on October 1, 2003; TRANSCEND (NCT00153101). Registered on September 9, 2005; COMPASS (NCT01776424). Registered on January 24, 2013; NAVIGATE-ESUS (NCT02313909). Registered on December 8, 2014.
Identifiants
pubmed: 39358761
doi: 10.1186/s13063-024-08482-2
pii: 10.1186/s13063-024-08482-2
doi:
Banques de données
ClinicalTrials.gov
['NCT02313909']
Types de publication
Journal Article
Comparative Study
Langues
eng
Sous-ensembles de citation
IM
Pagination
644Informations de copyright
© 2024. The Author(s).
Références
Folstein MF, Folstein SE, McHugh PR. Mini-mental state. J Psychiatr Res. 1975;12(3):189–98.
doi: 10.1016/0022-3956(75)90026-6
pubmed: 1202204
Tombaugh TN, McIntyre NJ. The mini-mental state examination: a comprehensive review. J Am Geriatr Soc. 1992;40(9):922-35.
Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695-9.
Proust-Lima C, Amieva H, Dartigues JF, Jacqmin-Gadda H. Sensitivity of four psychometric tests to measure cognitive changes in brain aging-population-based studies. Am J Epidemiol. 2007;165(3):344–50.
doi: 10.1093/aje/kwk017
pubmed: 17105962
Gerstein HC, Colhoun HM, Dagenais GR, Diaz R, Lakshmanan M, Pais P, Probstfield J, Riesmeyer JS, Riddle MC, Rydén L, et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. The Lancet. 2019;394(10193):121–30.
doi: 10.1016/S0140-6736(19)31149-3
Tabachnick BG, Fidell LS, Ullman JB. Using multivariate statistics. In., Seventh edition edn. New York: Pearson; 2019.
Proust-Lima C, Dartigues JF, Jacqmin-Gadda H. Misuse of the linear mixed model when evaluating risk factors of cognitive decline. Am J Epidemiol. 2011;174(9):1077–88.
doi: 10.1093/aje/kwr243
pubmed: 21965187
Muniz-Terrera G, Hout AVD, Rigby RA, Stasinopoulos DM. Analysing cognitive test data: distributions and non-parametric random effects. Stat Methods Med Res. 2016;25(2):741–53.
doi: 10.1177/0962280212465500
pubmed: 23136147
Najera-Zuloaga J, Lee DJ, Arostegui I. Comparison of beta-binomial regression model approaches to analyze health-related quality of life data. Stat Methods Med Res. 2018;27(10):2989–3009.
doi: 10.1177/0962280217690413
pubmed: 29298606
Tobin J. Estimation of Relationships for Limited Dependent Variables. Econometrica. 1958;26(1):24–24.
doi: 10.2307/1907382
Twisk J, Rijmen F. Longitudinal tobit regression: A new approach to analyze outcome variables with floor or ceiling effects. J Clin Epidemiol. 2009;62(9):953–8.
doi: 10.1016/j.jclinepi.2008.10.003
pubmed: 19211221
Terrera GM, Minett T, Brayne C, Matthews FE. Education associated with a delayed onset of terminal decline. Age Ageing. 2014;43(1):26–31.
doi: 10.1093/ageing/aft150
pubmed: 24136340
Akaike H. On the Likelihood of a Time Series Model. J Roy Stat Soc. 1978;27(3/4):217–35.
Gerstein HC, Bosch J, Dagenais GR, Díaz R, Jung H, Maggioni AP, Pogue J, Probstfield J, Ramachandran A, Riddle MC, et al. Basal Insulin and Cardiovascular and Other Outcomes in Dysglycemia. N Engl J Med. 2012;367(4):319–28.
doi: 10.1056/NEJMoa1203858
pubmed: 22686416
Cukierman-Yaffe T, Bosch J, Diaz R, Dyal L, Hancu N, Hildebrandt P, Lanas F, Lewis BS, Marre M, Yale J-F, et al. Effects of basal insulin glargine and omega-3 fatty acid on cognitive decline and probable cognitive impairment in people with dysglycaemia: a substudy of the ORIGIN trial. Lancet Diabetes Endocrinol. 2014;2(7):562–72.
doi: 10.1016/S2213-8587(14)70062-2
pubmed: 24898834
Yusuf S, Teo K, Anderson C, Pogue J, Dyal L, Copland I, Schumacher H, Dagenais G, Sleight P. Effects of the angiotensin-receptor blocker telmisartan on cardiovascular events in high-risk patients intolerant to angiotensin-converting enzyme inhibitors: a randomised controlled trial. Lancet. 2008;372:1174–83.
doi: 10.1016/S0140-6736(08)61242-8
pubmed: 18757085
Anderson C, Teo K, Gao P, Arima H, Dans A, Unger T, Commerford P, Dyal L, Schumacher H, Pogue J, et al. Renin-angiotensin system blockade and cognitive function in patients at high risk of cardiovascular disease: analysis of data from the ONTARGET and TRANSCEND studies. Lancet Neurol. 2011;10(1):43–53.
doi: 10.1016/S1474-4422(10)70250-7
pubmed: 20980201
Eikelboom JW, Connolly SJ, Bosch J, Dagenais GR, Hart RG, Shestakovska O, Diaz R, Alings M, Lonn EM, Anand SS, et al. Rivaroxaban with or without Aspirin in Stable Cardiovascular Disease. N Engl J Med. 2017;377(14):1319–30.
doi: 10.1056/NEJMoa1709118
pubmed: 28844192
Hart RG, Sharma M, Mundl H, Kasner SE, Bangdiwala SI, Berkowitz SD, Swaminathan B, Lavados P, Wang Y, Wang Y, et al. Rivaroxaban for Stroke Prevention after Embolic Stroke of Undetermined Source. N Engl J Med. 2018;378(23):2191–201.
doi: 10.1056/NEJMoa1802686
pubmed: 29766772
Bosch J, Pearce LA, Sharma M, Canavan M, Whiteley WN, Mikulik R, Mundl H, Yusuf S, Hart RG, O’Donnell MJ. Rivaroxaban versus aspirin on functional and cognitive outcomes after embolic stroke of undetermined source: NAVIGATE ESUS trial. J Stroke Cerebrovasc Dis. 2022;31(5):106404.
doi: 10.1016/j.jstrokecerebrovasdis.2022.106404
pubmed: 35292423
Laird NM, Ware JH. Random-Effects Models for Longitudinal Data In. 1982;38:963–74.
Huppert FA, Cabelli ST, Matthews FE. Brief cognitive assessment in a UK population sample - Distributional properties and the relationship between the MMSE and an extended mental state examination. BMC Geriatrics .2005;5.
Simonoff J. Analyzing Categorical Data. New York: Springer; 2003.
Guimarães P. A Simple Approach to Fit the Beta-binomial Model. Stand Genomic Sci. 2005;5(3):385–94.
Liu CF, Burgess JF Jr, Manning WG, Maciejewski ML. Beta-binomial regression and bimodal utilization. Health Serv Res. 2013;48(5):1769–78.
doi: 10.1111/1475-6773.12055
pubmed: 23521600
pmcid: 3796113
Kim J, Lee JH. The Validation of a Beta-Binomial Model for Overdispersed Binomial Data. Commun Stat Simul Comput. 2017;46(2):807–14.
doi: 10.1080/03610918.2014.960091
pubmed: 29276335