Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous

Sauzet O, Wright KC, Marston L, Brocklehurst P, Peacock JL (2013)
Statistics In Medicine 32(8): 1429-1438.

Journal Article | Published | English

No fulltext has been uploaded

Author
; ; ; ;
Abstract
In cluster-randomised trials, the problem of non-independence within clusters is well known, and appropriate statistical analysis documented. Clusters typically seen in cluster trials are large in size and few in number, whereas datasets of preterm infants incorporate clusters of size two (twins), size three (triplets) and so on, with the majority of infants being in clusters' of size one. In such situations, it is unclear whether adjustment for clustering is needed or even possible. In this paper, we compared analyses allowing for clustering (linear mixed model) with analyses ignoring clustering (linear regression). Through simulations based on two real datasets, we explored estimation bias in predictors of a continuous outcome in different size datasets typical of preterm samples, with varying percentages of twins. Overall, the biases for estimated coefficients were similar for linear regression and mixed models, but the standard errors were consistently much less well estimated when using a linear model. Non-convergence was rare but was observed in approximately 5% of mixed models for samples below 200 and percentage of twins 2% or less. We conclude that in datasets with small clusters, mixed models should be the method of choice irrespective of the percentage of twins. If the mixed model does not converge, a linear regression can be fitted, but standard error will be underestimated, and so type I error may be inflated. Copyright (c) 2012 John Wiley & Sons, Ltd.
Publishing Year
ISSN
PUB-ID

Cite this

Sauzet O, Wright KC, Marston L, Brocklehurst P, Peacock JL. Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous. Statistics In Medicine. 2013;32(8):1429-1438.
Sauzet, O., Wright, K. C., Marston, L., Brocklehurst, P., & Peacock, J. L. (2013). Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous. Statistics In Medicine, 32(8), 1429-1438.
Sauzet, O., Wright, K. C., Marston, L., Brocklehurst, P., and Peacock, J. L. (2013). Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous. Statistics In Medicine 32, 1429-1438.
Sauzet, O., et al., 2013. Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous. Statistics In Medicine, 32(8), p 1429-1438.
O. Sauzet, et al., “Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous”, Statistics In Medicine, vol. 32, 2013, pp. 1429-1438.
Sauzet, O., Wright, K.C., Marston, L., Brocklehurst, P., Peacock, J.L.: Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous. Statistics In Medicine. 32, 1429-1438 (2013).
Sauzet, Odile, Wright, K. C., Marston, L., Brocklehurst, P., and Peacock, J. L. “Modelling the hierarchical structure in datasets with very small clusters: a simulation study to explore the effect of the proportion of clusters when the outcome is continuous”. Statistics In Medicine 32.8 (2013): 1429-1438.
This data publication is cited in the following publications:
This publication cites the following data publications:

6 Citations in Europe PMC

Data provided by Europe PubMed Central.

The SafeBoosC II randomized trial: treatment guided by near-infrared spectroscopy reduces cerebral hypoxia without changing early biomarkers of brain injury.
Plomgaard AM, van Oeveren W, Petersen TH, Alderliesten T, Austin T, van Bel F, Benders M, Claris O, Dempsey E, Franz A, Fumagalli M, Gluud C, Hagmann C, Hyttel-Sorensen S, Lemmers P, Pellicer A, Pichler G, Winkel P, Greisen G., Pediatr. Res. 79(4), 2016
PMID: 26679155
Analysis of Randomised Trials Including Multiple Births When Birth Size Is Informative.
Yelland LN, Sullivan TR, Pavlou M, Seaman SR., Paediatr Perinat Epidemiol 29(6), 2015
PMID: 26332368
Accounting for multiple births in randomised trials: a systematic review.
Yelland LN, Sullivan TR, Makrides M., Arch. Dis. Child. Fetal Neonatal Ed. 100(2), 2015
PMID: 25389142
Late outcomes of a randomized trial of high-frequency oscillation in neonates.
Zivanovic S, Peacock J, Alcazar-Paris M, Lo JW, Lunt A, Marlow N, Calvert S, Greenough A; United Kingdom Oscillation Study Group, Zivanovic S, Peacock J, Alcazar-Paris M, Lo JW, Marlow N, Calvert S, Greenough A, Halliday H, Henderson J, Cunningham S, Vyas H, Kerry S, Dromgoole J, Coker B, Oedra R, Thomas F, D'eath T, Nguyen J, Lovestone J., N. Engl. J. Med. 370(12), 2014
PMID: 24645944
A phase II randomized clinical trial on cerebral near-infrared spectroscopy plus a treatment guideline versus treatment as usual for extremely preterm infants during the first three days of life (SafeBoosC): study protocol for a randomized controlled trial.
Hyttel-Sorensen S, Austin T, van Bel F, Benders M, Claris O, Dempsey E, Fumagalli M, Greisen G, Grevstad B, Hagmann C, Hellstrom-Westas L, Lemmers P, Lindschou J, Naulaers G, van Oeveren W, Pellicer A, Pichler G, Roll C, Skoog M, Winkel P, Wolf M, Gluud C., Trials 14(), 2013
PMID: 23782447

14 References

Data provided by Europe PubMed Central.

The statistical analysis of data from small groups.
Kenny DA, Mannetti L, Pierro A, Livi S, Kashy DA., J Pers Soc Psychol 83(1), 2002
PMID: 12088122
Analysis of a trial randomised in clusters.
Kerry SM, Bland JM., BMJ 316(7124), 1998
PMID: 9451271
Comparison of methods for analysing cluster randomized trials: an example involving a factorial design.
Peters TJ, Richards SH, Bankhead CR, Ades AE, Sterne JA., Int J Epidemiol 32(5), 2003
PMID: 14559762
High-frequency oscillatory ventilation for the prevention of chronic lung disease of prematurity.
Johnson AH, Peacock JL, Greenough A, Marlow N, Limb ES, Marston L, Calvert SA; United Kingdom Oscillation Study Group., N. Engl. J. Med. 347(9), 2002
PMID: 12200550
A survey of methods for analyzing clustered binary response data
Pendergast, International Statistical Review 64(), 1996
How should randomised trials including multiple pregnancies be analysed?
Gates, BJOG-an International Journal of Obstetrics and Gynaecology 111(), 2004
Regression models for twin studies: a critical review.
Carlin JB, Gurrin LC, Sterne JA, Morley R, Dwyer T., Int J Epidemiol 34(5), 2005
PMID: 16087687
Analysis of repeated pregnancy outcomes.
Louis GB, Dukic V, Heagerty PJ, Louis TA, Lynch CD, Ryan LM, Schisterman EF, Trumble A; Pregnancy Modeling Working Group., Stat Methods Med Res 15(2), 2006
PMID: 16615652
Analysis of neonatal clinical trials with twin births.
Shaffer ML, Kunselman AR, Watterberg KL., BMC Med Res Methodol 9(), 2009
PMID: 19245713
Comparing methods of analysing datasets with small clusters: case studies using four paediatric datasets.
Marston L, Peacock JL, Yu K, Brocklehurst P, Calvert SA, Greenough A, Marlow N., Paediatr Perinat Epidemiol 23(4), 2009
PMID: 19523085
Factors affecting vocabulary acquisition at age 2 in children born between 23 and 28 weeks' gestation
Marston, Development Medicine and Child Neurology 49(), 2007
Randomised trial of high frequency oscillatory ventilation or conventional ventilation in babies of gestational age 28 weeks or less: respiratory and neurological outcomes at 2 years.
Marlow N, Greenough A, Peacock JL, Marston L, Limb ES, Johnson AH, Calvert SA., Arch. Dis. Child. Fetal Neonatal Ed. 91(5), 2006
PMID: 16690640
The INIS Study. International Neonatal Immunotherapy Study: non-specific intravenous immunoglobulin therapy for suspected or proven neonatal sepsis: an international, placebo controlled, multicentre randomised trial.
INIS Study Collaborative Group, Brocklehurst P, Brearley S, Haque K, Leslie A, Salt A, Stenson B, Stephenson J, Tarnow-Mordi W., BMC Pregnancy Childbirth 8(), 2008
PMID: 19063731

Export

0 Marked Publications

Open Data PUB

Web of Science

View record in Web of Science®

Sources

PMID: 23027676
PubMed | Europe PMC

Search this title in

Google Scholar