CoChat
Sign Up
Cited by 74Statistical Methods in Clinical TrialsAdvanced Causal Inference TechniquesStatistical Methods and Bayesian Inference

A comparison of methods to adjust for continuous covariates in the analysis of randomised trials

Open access
Kahan, Brennan C., Rushton, Helen, Morris, Tim P., Daniel, Rhian M.·2016·BMC Medical Research Methodology
DOI 10.1186/s12874-016-0141-3PMC PMC4827223
Open this paper in CoChat — free accountOpen PDF

Abstract

BACKGROUND: Although covariate adjustment in the analysis of randomised trials can be beneficial, adjustment for continuous covariates is complicated by the fact that the association between covariate and outcome must be specified. Misspecification of this association can lead to reduced power, and potentially incorrect conclusions regarding treatment efficacy. METHODS: We compared several methods of adjustment to determine which is best when the association between covariate and outcome is unknown. We assessed (a) dichotomisation or categorisation; (b) assuming a linear association with outcome; (c) using fractional polynomials with one (FP1) or two (FP2) polynomial terms; and (d) using restricted cubic splines with 3 or 5 knots. We evaluated each method using simulation and through a re-analysis of trial datasets. RESULTS: Methods which kept covariates as continuous typically had higher power than methods which used categorisation. Dichotomisation, categorisation, and assuming a linear association all led to large reductions in power when the true association was non-linear. FP2 models and restricted cubic splines with 3 or 5 knots performed best overall. CONCLUSIONS: For the analysis of randomised trials we recommend (1) adjusting for continuous covariates even if their association with outcome is unknown; (2) keeping covariates as continuous; and (3) using fractional polynomials with two polynomial terms or restricted cubic splines with 3 to 5 knots when a linear association is in doubt.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s12874-016-0141-3

Cites

4 of 4 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Methods that keep covariates as continuous typically have higher power than methods which use categorization.

    “In addition, methods that keep covariates as continuous typically have higher power than methods which use categorization (36).”

    From A Method to Combine Neurofilament Light Measurements From Blood Serum and Plasma in Clinical and Population-Based Studies (2022) · cites this paper as [36]

  • Regression coefficients based on categorized independent variables are not comparable across studies.

    “However, regression coefficients based on categorized independent variables are not comparable across studies.”

    From A Method to Combine Neurofilament Light Measurements From Blood Serum and Plasma in Clinical and Population-Based Studies (2022) · cites this paper as [36]

    The Methods section defines coefficients for BMI categories relative to an underweight reference group, and the Results discuss power and treatment-effect estimates. Neither addresses whether coefficients based on categorized independent variables are comparable across studies; searches for that claim and for differing category cut-points found no supporting passage. The supplied copy gives no page numbers. (LLM, conf=0.72)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 33

  1. [3]Hernandez AV Steyerberg EW Habbema JD Covariate adjustment in randomized controlled trials with dichotomous outcomes increases statistical power and reduces sample size requirements J Clin Epidemiol 2004 57 5 454 60 10.1016/j.jclinepi.2003.09.014 15196615 · DOI · Paper page
  2. [5]Turner EL Perel P Clayton T Edwards P Hernandez AV Roberts I Covariate adjustment increased power in randomized controlled trials: an example in traumatic brain injury J Clin Epidemiol 2012 65 5 474 81 10.1016/j.jclinepi.2011.08.012 22169080 PMC3589911 · DOI
  3. [6]Thompson DD Lingsma HF Whiteley WN Murray GD Steyerberg EW Covariate adjustment had similar benefits in small and large randomized controlled trials J Clin Epidemiol 2015 68 9 1068 75 10.1016/j.jclinepi.2014.11.001 25497979 PMC5708297 · DOI
  4. [7]Nicholas K Yeatts SD Zhao W Ciolino J Borg K Durkalski V The impact of covariate adjustment at randomization and analysis for binary outcomes: understanding differences between superiority and noninferiority trials Stat Med 2015 34 11 1834 40 10.1002/sim.6447 25641057 PMC4393769 · DOI
  5. [8]Kahan BC Morris TP Reporting and analysis of trials using stratified randomisation in leading medical journals: review and reanalysis BMJ 2012 345 e5840 10.1136/bmj.e5840 22983531 PMC3444136 · DOI
  6. [9]Kahan BC Morris TP Improper analysis of trials randomised using stratified blocks or minimisation Stat Med 2012 31 4 328 40 10.1002/sim.4431 22139891 · DOI
  7. [10]Kahan BC Morris TP Adjusting for multiple prognostic factors in the analysis of randomised trials BMC Med Res Methodol 2013 13 99 10.1186/1471-2288-13-99 23898993 PMC3733981 · DOI
  8. [11]Kahan BC Morris TP Assessing potential sources of clustering in individually randomised trials BMC Med Res Methodol 2013 13 58 10.1186/1471-2288-13-58 23590245 PMC3643875 · DOI
  9. [12]Parzen M Lipsitz S Dear K Does clustering affect the usual test statistics of no treatment effect in a randomized clinical trial? Biom J 1998 40 4 385 402 10.1002/(SICI)1521-4036(199808)40:4<385::AID-BIMJ385>3.0.CO;2-# · DOI
  10. [13]Royston P Saurbrei W Multivariable Model-Building 2008 Chichester Wiley
  11. [14]Durrleman S Simon R Flexible regression models with cubic splines Stat Med 1989 8 5 551 61 10.1002/sim.4780080504 2657958 · DOI
  12. [15]Harrell FE Jr Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis 2001 New York Springer
  13. [16]Brenner H Blettner M Controlling for continuous confounders in epidemiologic research Epidemiology (Cambridge, Mass) 1997 8 4 429 34 10.1097/00001648-199707000-00014 9209859 · DOI · Paper page
  14. [18]Royston P Altman DG Sauerbrei W Dichotomizing continuous predictors in multiple regression: a bad idea Stat Med 2006 25 1 127 41 10.1002/sim.2331 16217841 · DOI · Paper page
  15. [19]Sauerbrei W Royston P Binder H Selection of important variables and determination of functional form for continuous predictors in multivariable model building Stat Med 2007 26 30 5512 28 10.1002/sim.3148 18058845 · DOI
  16. [20]Committee for Proprietary Medicinal Products (CPMP) Points to consider on adjustment for baseline covariates Stat Med 2004 23 5 701 9 10.1002/sim.1647 14981670 · DOI
  17. [21]Schmoor C Schumacher M Effects of covariate omission and categorization when analysing randomized trials with the Cox model Stat Med 1997 16 1–3 225 37 10.1002/(SICI)1097-0258(19970215)16:3<225::AID-SIM482>3.0.CO;2-C 9004394 · DOI
  18. [22]Berger VW Valid adjustment of randomized comparisons for binary covariates Biom J 2004 46 5 589 94 10.1002/bimj.200410055 · DOI
  19. [23]Rosenbaum PR The Consequences of Adjustment for a Concomitant Variable That Has Been Affected by the Treatment J Roy Stat Soc a Sta 1984 147 656 66 10.2307/2981697 · DOI · Paper page
  20. [24]Rosenberger WF Lachin JM Randomization in Clinical Trials 2005 New York John Wiley & Sons, Inc.
  21. [25]Yang L Tsiatis A Efficiency Study of Estimators for a Treatment Effect in a Pretest-Posttest Trial The American Statistician 2001 55 4 314 21 10.1198/000313001753272466 · DOI
  22. [26]Gail M Wieand S Piantadosi S Biased estimates of treatment effect in randomized experiments with nonlinear regressions and omitted covariates Biometrika 1984 71 3 431 44 10.1093/biomet/71.3.431 · DOI
  23. [27]Hauck WW Anderson S Marcus SM Should we adjust for covariates in nonlinear regression analyses of randomized trials? Control Clin Trials 1998 19 3 249 56 10.1016/S0197-2456(97)00147-5 9620808 · DOI
  24. [28]Robinson LD Jewell NP Some surprising results about covariate adjustment in logistic regression models Int Stat Rev 1991 58 227 40 10.2307/1403444 · DOI
  25. [29]Altman DG Royston P The cost of dichotomising continuous variables BMJ 2006 332 7549 1080 10.1136/bmj.332.7549.1080 16675816 PMC1458573 · DOI · Paper page
  26. [30]Raab GM Day S Sales J How to select covariates to include in the analysis of a clinical trial Control Clin Trials 2000 21 4 330 42 10.1016/S0197-2456(00)00061-1 10913808 · DOI
  27. [31]Morris TP White IR Carpenter JR Stanworth SJ Royston P Combining fractional polynomial model building with multiple imputation Stat Med 2015 34 25 3298 317 10.1002/sim.6553 26095614 PMC4871237 · DOI
  28. [32]Rahman NM Maskell NA West A Teoh R Arnold A Mackinlay C Intrapleural use of tissue plasminogen activator and DNase in pleural infection N Engl J Med 2011 365 6 518 26 10.1056/NEJMoa1012740 21830966 · DOI
  29. [33]Kahan BC Jairath V Murphy MF Dore CJ Update on the transfusion in gastrointestinal bleeding (TRIGGER) trial: statistical analysis plan for a cluster-randomised feasibility trial Trials 2013 14 206 10.1186/1745-6215-14-206 23837630 PMC3716978 · DOI
  30. [34]Chan AW Tetzlaff JM Gotzsche PC Altman DG Mann H Berlin JA SPIRIT 2013 explanation and elaboration: guidance for protocols of clinical trials BMJ 2013 346 e7586 10.1136/bmj.e7586 23303884 PMC3541470 · DOI · Paper page
  31. [35]Freeman PR The performance of the two-stage analysis of two-treatment, two-period crossover trials Stat Med 1989 8 12 1421 32 10.1002/sim.4780081202 2616932 · DOI
  32. [36]Kahan BC Bias in randomised factorial trials Stat Med 2013 32 26 4540 9 10.1002/sim.5869 23733397 · DOI
  33. [37]Shuster JJ Diagnostics for assumptions in moderate to large simple clinical trials: do they really help? Stat Med 2005 24 16 2431 8 10.1002/sim.2175 15977289 · DOI

Cited by

Sorted by citations

Authors
Kahan, B. C., Rushton, H., Morris, T. P., & Daniel, R. M.
Affiliations
  • Queen Mary University of London (GB)
  • University of Manchester (GB)
  • London School of Hygiene & Tropical Medicine (GB)
Corresponding author
Kahan, Brennan C.
ORCID
Venue
BMC Medical Research Methodology
Publisher
Springer Science and Business Media LLC
Published
2016-04-11
Volume
16
Issue
1
Pages
42-42
ISSN
1471-2288 (ISSN-L)
DOI
10.1186/s12874-016-0141-3
PMC
PMC4827223
OpenAlex
W2323405795
Cited by
74
References
4
Citations by year
2026: 9 · 2025: 10 · 2024: 7 · 2023: 9 · 2022: 12 · 2021: 6 · 2020: 7 · 2019: 6 · 2018: 5 · 2017: 3
FWCI
4.73
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • BMC Medical Research Methodology (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Research Explorer (The University of Manchester) (publishedVersion, cc-by)
  • ORCA Online Research @Cardiff (Cardiff University) (acceptedVersion)
  • LSHTM Research Online (London School of Hygiene and Tropical Medicine) (acceptedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • UCL Discovery (University College London) (submittedVersion)
  • Queen Mary Research Online (Queen Mary University of London) (submittedVersion, cc-by)
Funders
Medical Research Council (MC_EX_G0800814, MC_UU_12023/21)
Topics
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Clinical Trials
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
Keywords
covariate adjustment, continuous covariates, randomised trials, fractional polynomials, restricted cubic splines, nonlinear relationship, statistical power, treatment effect estimation
MeSH terms
Algorithms, Computer Simulation, Diethylstilbestrol, Humans, Male, Neoplasm Invasiveness, Neoplasm Staging, Prostatic Neoplasms, Models, Statistical, Linear Models, Proportional Hazards Models, Survival Analysis, Randomized Controlled Trials as Topic, Treatment Outcome, Disease-Free Survival

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review