BACKGROUND: Observational data on composite scores often comes with missing component information. When a complete-case (CC) analysis of composite scores is unbiased, preferable approaches of dealing with missing component information should also be unbiased and provide a more precise estimate. We assessed the performance of several methods compared to CC analysis in estimating the means of common composite scores used in axial spondyloarthritis research. METHODS: Individual mean imputation (IMI), the modified formula method (MF), overall mean imputation (OMI), and multiple imputation of missing component values (MI) were assessed either analytically or by means of simulations from available data collected across Europe. Their performance in estimating the means of the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), the Bath Ankylosing Spondylitis Functional Index (BASFI), and the Ankylosing Spondylitis Disease Activity Score based on C-reactive protein (ASDAS-CRP) in cases where component information was set missing completely at random was compared to the CC approach based on bias, variance, and coverage. RESULTS: Like the MF method, IMI uses a modified formula for observations with missing components resulting in modified composite scores. In the case of an unbiased CC approach, these two methods yielded representative samples of the distribution arising from a mixture of the original and modified composite scores, which, however, could not be considered the same as the distribution of the original score. The IMI and MF method are, thus, intrinsically biased. OMI provided an unbiased mean but displayed a complex dependence structure among observations that, if not accounted for, resulted in severe coverage issues. MI improved precision compared to CC and gave unbiased means and proper coverage as long as the extent of missingness was not too large. CONCLUSIONS: MI of missing component values was the only method found successful in retaining CC's unbiasedness and in providing increased precision for estimating the means of BASDAI, BASFI, and ASDAS-CRP. However, since MI is susceptible to incorrect implementation and its performance may become questionable with increasing missingness, we consider the implementation of an error-free CC approach a valid and valuable option. TRIAL REGISTRATION: Not applicable as study uses data from patient registries.

Aberdeen Centre for Arthritis and Musculoskeletal Health University of Aberdeen Aberdeen UK

Amsterdam UMC Department of Rheumatology and Clinical Immunology and Department of Experimental Immunology Amsterdam Institute for Infection and Immunity University of Amsterdam Amsterdam The Netherlands

Center for Treatment of Rheumatic and Musculoskeletal Diseases Diakonhjemmet Hospital Oslo Norway

Centre for Rheumatology Research Landspitali University Hospital Reykjavik Iceland

Copenhagen Centre for Arthritis Research and the DANBIO Registry Centre for Rheumatology and Spine Diseases Centre for Head and Orthopaedics Rigshospitalet Glostrup Denmark

Copenhagen Centre for Arthritis Research Centre for Rheumatology and Spine Diseases Centre for Head and Orthopaedics Rigshospitalet Glostrup Denmark

Department of Clinical Sciences Lund Rheumatology Skåne University Hospital Lund University Lund Sweden

Department of Medicine Solna Clinical Epidemiology Division Karolinska Institutet Stockholm Sweden

Department of Rheumatology Aarhus University Hospital Aarhus Denmark

Department of Rheumatology Geneva University Hospital Geneva Switzerland

Department of Rheumatology Kuopio University Hospital Kuopio Finland

Department of Rheumatology University Hospital Zurich University of Zurich Zurich Switzerland

Department Rheumatology and Immunology Inselspital University Hospital Bern Bern Switzerland

Division of Rheumatology Department of Internal Medicine School of Medicine Kocaeli University Kocaeli Turkey

Faculty of Medicine University of Iceland Reykjavik Iceland

Faculty of Medicine University of Ljubljana Ljubljana Slovenia

Institute of Rheumatology Prague Czech Republic

Rheumatology Inflammation Center Helsinki University Hospital and University of Helsink Helsinki Finland

Rheumatology Radboud University Medical Centre Nijmegen The Netherlands

Section for Rheumatology Department for Neurology Rheumatology and Physical Medicine Helse Førde Førde Norway

Sociedade Portuguesa de Reumatologia Reuma Pt Lisbon Portugal

Statistics Group Swiss Clinical Quality Management Foundation Aargauerstrasse 250 Zurich 8048 Switzerland

Zobrazit více v PubMed

Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P, Calin A. A new approach to defining disease status in ankylosing spondylitis: the Bath Ankylosing Spondylitis Disease Activity Index. J Rheumatol. 1994:21(12):2286–91. PubMed

Calin A, Garrett S, Whitelock H, Kennedy LG, O’hea J, Mallorie P et al. A new approach to defining functional ability in ankylosing spondylitis: the development of the Bath Ankylosing Spondylitis Functional Index. J Rheumatol. 1994:21(12):2281–2285. PubMed

Lukas C, Landewé R, Sieper J, Dougados M, Davis J, Braun J, et al. Development of an ASAS-endorsed disease activity score (ASDAS) in patients with ankylosing spondylitis. Ann Rheum Dis. 2009:68:18–24. PubMed

van der Heijde D, Lie E, Kvien T, Sieper J, Van den Bosch F, Listing J, et al. ASDAS, a highly discriminatory ASAS-endorsed disease activity score in patients with ankylosing spondylitis. Ann Rheum Dis. 2009:68:1811–8. PubMed

Jones GT, Dean LE, Pathan E, Hollick RJ, Macfarlane GJ. Real-word evidence of TNF inhibition in axial spondyloarthritis: can we generalise the results from clinical trials? Ann Rheum Dis. 2020:79:914–9. PubMed

Marzo-Ortega H, Sieper J, Kivitz AJ, Blanco R, Cohen M, Pavelka K, et al. 5-year efficacy and safety of secukinumab in patients with ankylosing spondylitis: end-of-study results from the phase 3 MEASURE 2 trial. Lancet Rheumatol. 2020;2(6):e339–46. PubMed

Schafer JL, Graham JW. Missing data: our view of the state of the art. Psychol Methods. 2002;7(2):147. PubMed

Braun J, McHugh N, Singh A, Wajdula JS, Sato R. Improvement in patient-reported outcomes for patients with ankylosing spondylitis treated with etanercept 50 mg once-weekly and 25 mg twice-weekly. Rheumatology. 2007;46(6):999–1004. PubMed

Horton NJ, Kleinman KP. Much ado about nothing: A comparison of missing data methods and software to fit incomplete data regression models. Am Stat. 2007;61(1):79–90. PubMed PMC

Carpenter JR, Kenward MG. Multiple imputation and its application. 1st ed. Hoboken: Wiley & Sons; 2013.

Rombach I, Gray AM, Jenkins C, Murray DW, Rivero-Arias O. Multiple imputation for patient reported outcome measures in randomized controlled trials: advantages and disadvantages of imputing at the item, subscale or composite score level. BMC Med Res Methodol. 2018;18:87. PubMed PMC

O’Keeffe AG, Farewell DM, Tom BDM, Farewell VT. Multiple imputation of missing composite outcomes in longitudinal data. Stat Biosci. 2016;8:310–32. PubMed PMC

Simons CL, Rivero-Arias O, Yu LM, Simon J. Multiple imputation to deal with missing EQ-5D-3L data: should we impute individual domains or the actual index? Qual Life Res. 2015;24(4):805–15. PubMed

White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30(4):377–99. PubMed

Hughes RA, Heron J, Sterne JAC, Tilling K. Accounting for missing data in statistical analyses: multiple imputation is not always the answer. Int J Epidemiol. 2019;48(4):1294–304. PubMed PMC

Morris TP, White IR, Crowther MJ. Using simulation studies to evaluate statistical methods. Stat Med. 2019;38:2074–102. PubMed PMC

The EuroSpA Research Collaboration Network. URL https://eurospa.eu (2016). Accessed 2 Feb 2023.

Ornbjerg LM, Linde L, Georgiadis S, Rasmussen SH, Lindström U, Askling J, et al. Predictors of ASDAS-CRP inactive disease in axial spondyloarthritis during treatment with TNF-inhibitors: Data from the EuroSpA collaboration. Semin Arthritis Rheum. 2022;56: 152081. PubMed

Van Buuren S. Flexible imputation of missing data. 2nd ed. London: Chapman and Hall/CRC; 2012.

van Buuren S, Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45(3):1–67.

Kleinke K. Multiple imputation by predictive mean matching when sample size is small. Methodology. 2018;14:3–15.

R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2022. URL https://www.r-project.org/.

RStudio Team. RStudio: Integrated development for R. RStudio, PBC, Boston, MA. 2020. URL http://www.rstudio.com.

Microsoft Corporation, Weston S. doParallel: foreach parallel adaptor for the ‘parallel’ package. 2022. URL https://cran.r-project.org/package=doParallel.

Microsoft Corporation, Weston S. foreach: provides foreach looping construct. 2022. URL https://cran.r-project.org/package=foreach.

Wickham H. ggplot2: elegant graphics for data analysis. 2nd ed. New York: Springer; 2016.

Ramiro S, van Tubergen A, van der Heijde D, van den Bosch F, Dougados M, Landewé R. How to deal with missing items in BASDAI and BASFI. Rheumatology. 2014;53(2):374–6. PubMed

Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338: b2393. PubMed PMC