Model selection for component network meta-analysis in connected and disconnected networks: a simulation study.
Component network meta-analysis
Disconnected networks
Model selection
Multicomponent interventions
Simulation
Journal
BMC medical research methodology
ISSN: 1471-2288
Titre abrégé: BMC Med Res Methodol
Pays: England
ID NLM: 100968545
Informations de publication
Date de publication:
14 06 2023
14 06 2023
Historique:
received:
11
10
2022
accepted:
29
05
2023
medline:
16
6
2023
pubmed:
15
6
2023
entrez:
14
6
2023
Statut:
epublish
Résumé
Network meta-analysis (NMA) allows estimating and ranking the effects of several interventions for a clinical condition. Component network meta-analysis (CNMA) is an extension of NMA which considers the individual components of multicomponent interventions. CNMA allows to "reconnect" a disconnected network with common components in subnetworks. An additive CNMA assumes that component effects are additive. This assumption can be relaxed by including interaction terms in the CNMA. We evaluate a forward model selection strategy for component network meta-analysis to relax the additivity assumption that can be used in connected or disconnected networks. In addition, we describe a procedure to create disconnected networks in order to evaluate the properties of the model selection in connected and disconnected networks. We apply the methods to simulated data and a Cochrane review on interventions for postoperative nausea and vomiting in adults after general anaesthesia. Model performance is compared using average mean squared errors and coverage probabilities. CNMA models provide good performance for connected networks and can be an alternative to standard NMA if additivity holds. For disconnected networks, we recommend to use additive CNMA only if strong clinical arguments for additivity exist. CNMA methods are feasible for connected networks but questionable for disconnected networks.
Sections du résumé
BACKGROUND
Network meta-analysis (NMA) allows estimating and ranking the effects of several interventions for a clinical condition. Component network meta-analysis (CNMA) is an extension of NMA which considers the individual components of multicomponent interventions. CNMA allows to "reconnect" a disconnected network with common components in subnetworks. An additive CNMA assumes that component effects are additive. This assumption can be relaxed by including interaction terms in the CNMA.
METHODS
We evaluate a forward model selection strategy for component network meta-analysis to relax the additivity assumption that can be used in connected or disconnected networks. In addition, we describe a procedure to create disconnected networks in order to evaluate the properties of the model selection in connected and disconnected networks. We apply the methods to simulated data and a Cochrane review on interventions for postoperative nausea and vomiting in adults after general anaesthesia. Model performance is compared using average mean squared errors and coverage probabilities.
RESULTS
CNMA models provide good performance for connected networks and can be an alternative to standard NMA if additivity holds. For disconnected networks, we recommend to use additive CNMA only if strong clinical arguments for additivity exist.
CONCLUSIONS
CNMA methods are feasible for connected networks but questionable for disconnected networks.
Identifiants
pubmed: 37316775
doi: 10.1186/s12874-023-01959-9
pii: 10.1186/s12874-023-01959-9
pmc: PMC10268445
doi:
Types de publication
Meta-Analysis
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
140Informations de copyright
© 2023. The Author(s).
Références
Stat Med. 2022 Sep 30;41(22):4444-4466
pubmed: 35844085
Stat Med. 2014 Nov 10;33(25):4353-69
pubmed: 24942211
Cochrane Database Syst Rev. 2020 Oct 19;10:CD012859
pubmed: 33075160
BMC Med Res Methodol. 2016 Apr 27;16:47
pubmed: 27116943
Value Health. 2019 Jan;22(1):85-91
pubmed: 30661638
PLoS One. 2021 Feb 8;16(2):e0246631
pubmed: 33556155
Anaesthesia. 2021 Jul;76(7):962-973
pubmed: 33170514
Am J Epidemiol. 2009 May 1;169(9):1158-65
pubmed: 19258485
Res Synth Methods. 2012 Dec;3(4):312-24
pubmed: 26053424
BMC Med Res Methodol. 2018 Jun 28;18(1):66
pubmed: 29954322
Pharmacoeconomics. 2010;28(10):935-45
pubmed: 20831302
Biom J. 2021 Feb;63(2):447-461
pubmed: 32596834
Res Synth Methods. 2016 Sep;7(3):236-63
pubmed: 26754852
Value Health. 2012 Sep-Oct;15(6):940-7
pubmed: 22999145
Clin Epidemiol. 2012;4:75-85
pubmed: 22570567
Res Synth Methods. 2016 Dec;7(4):420-432
pubmed: 27061025
Res Synth Methods. 2012 Jun;3(2):80-97
pubmed: 26062083
Res Synth Methods. 2016 Mar;7(1):6-22
pubmed: 26536149
Pharm Stat. 2018 May;17(3):264-277
pubmed: 29676023
Res Synth Methods. 2018 Jun;9(2):148-162
pubmed: 29073336
Med Decis Making. 2018 Feb;38(2):200-211
pubmed: 28823204
Stat Methods Med Res. 2008 Jun;17(3):279-301
pubmed: 17925316
Med Decis Making. 2021 Feb;41(2):194-208
pubmed: 33448252
Stat Med. 2022 Jun 30;41(14):2586-2601
pubmed: 35261053
BMC Med Res Methodol. 2015 Apr 12;15:34
pubmed: 25887646
Biom J. 2020 May;62(3):808-821
pubmed: 31021449
Res Synth Methods. 2016 Sep;7(3):306-13
pubmed: 26610409
Biom J. 2018 May;60(3):431-449
pubmed: 29292533
Res Synth Methods. 2017 Dec;8(4):465-474
pubmed: 28737842
BMC Med Res Methodol. 2020 Feb 24;20(1):36
pubmed: 32093605