Missing not at random in end of life care studies: multiple imputation and sensitivity analysis on data from the ACTION study.
Advance care planning
MAR
MNAR
Missing data
Oncology
Quality of life
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:
09 01 2021
09 01 2021
Historique:
received:
19
06
2020
accepted:
26
11
2020
entrez:
10
1
2021
pubmed:
11
1
2021
medline:
25
6
2021
Statut:
epublish
Résumé
Missing data are common in end-of-life care studies, but there is still relatively little exploration of which is the best method to deal with them, and, in particular, if the missing at random (MAR) assumption is valid or missing not at random (MNAR) mechanisms should be assumed. In this paper we investigated this issue through a sensitivity analysis within the ACTION study, a multicenter cluster randomized controlled trial testing advance care planning in patients with advanced lung or colorectal cancer. Multiple imputation procedures under MAR and MNAR assumptions were implemented. Possible violation of the MAR assumption was addressed with reference to variables measuring quality of life and symptoms. The MNAR model assumed that patients with worse health were more likely to have missing questionnaires, making a distinction between single missing items, which were assumed to satisfy the MAR assumption, and missing values due to completely missing questionnaire for which a MNAR mechanism was hypothesized. We explored the sensitivity to possible departures from MAR on gender differences between key indicators and on simple correlations. Up to 39% of follow-up data were missing. Results under MAR reflected that missingness was related to poorer health status. Correlations between variables, although very small, changed according to the imputation method, as well as the differences in scores by gender, indicating a certain sensitivity of the results to the violation of the MAR assumption. The findings confirmed the importance of undertaking this kind of analysis in end-of-life care studies.
Sections du résumé
BACKGROUND
Missing data are common in end-of-life care studies, but there is still relatively little exploration of which is the best method to deal with them, and, in particular, if the missing at random (MAR) assumption is valid or missing not at random (MNAR) mechanisms should be assumed. In this paper we investigated this issue through a sensitivity analysis within the ACTION study, a multicenter cluster randomized controlled trial testing advance care planning in patients with advanced lung or colorectal cancer.
METHODS
Multiple imputation procedures under MAR and MNAR assumptions were implemented. Possible violation of the MAR assumption was addressed with reference to variables measuring quality of life and symptoms. The MNAR model assumed that patients with worse health were more likely to have missing questionnaires, making a distinction between single missing items, which were assumed to satisfy the MAR assumption, and missing values due to completely missing questionnaire for which a MNAR mechanism was hypothesized. We explored the sensitivity to possible departures from MAR on gender differences between key indicators and on simple correlations.
RESULTS
Up to 39% of follow-up data were missing. Results under MAR reflected that missingness was related to poorer health status. Correlations between variables, although very small, changed according to the imputation method, as well as the differences in scores by gender, indicating a certain sensitivity of the results to the violation of the MAR assumption.
CONCLUSIONS
The findings confirmed the importance of undertaking this kind of analysis in end-of-life care studies.
Identifiants
pubmed: 33422019
doi: 10.1186/s12874-020-01180-y
pii: 10.1186/s12874-020-01180-y
pmc: PMC7796568
doi:
Types de publication
Journal Article
Multicenter Study
Randomized Controlled Trial
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
13Subventions
Organisme : Seventh Framework Programme ()
ID : 602541-2
Investigateurs
Agnes van der Heide
(A)
Ida J Korfage
(IJ)
Judith A C Rietjens
(JAC)
Lea J Jabbarian
(LJ)
Suzanne Polinder
(S)
Hans van Delden
(H)
Marijke Kars
(M)
Marieke Zwakman
(M)
Luc Deliens
(L)
Mariëtte N Verkissen
(MN)
Kim Eecloo
(K)
Kristof Faes
(K)
Kristian Pollock
(K)
Jane Seymour
(J)
Glenys Caswell
(G)
Andrew Wilcock
(A)
Louise Bramley
(L)
Sheila Payne
(S)
Nancy Preston
(N)
Lesley Dunleavy
(L)
Eleanor Sowerby
(E)
Guido Miccinesi
(G)
Francesco Bulli
(F)
Francesca Ingravallo
(F)
Giulia Carreras
(G)
Alessandro Toccafondi
(A)
Giuseppe Gorini
(G)
Urška Lunder
(U)
Branka Červ
(B)
Anja Simonič
(A)
Alenka Mimić
(A)
Hana Kodba-Čeh
(H)
Polona Ozbič
(P)
Mogens Groenvold
(M)
Caroline Arnfeldt
(C)
Anna Thit Johnsen
(A)
Références
Clin Trials. 2017 Aug;14(4):387-395
pubmed: 28385071
Qual Life Res. 2010 Feb;19(1):137-48
pubmed: 20041307
Stat Med. 2018 Jul 10;37(15):2338-2353
pubmed: 29611205
Patient Educ Couns. 2009 Dec;77(3):404-12
pubmed: 19892508
Am J Epidemiol. 2014 Mar 15;179(6):764-74
pubmed: 24589914
Stat Med. 1999 Mar 30;18(6):681-94
pubmed: 10204197
Eur J Cancer. 2006 Jan;42(1):55-64
pubmed: 16162404
BMJ. 2009 Jun 29;338:b2393
pubmed: 19564179
Stat Methods Med Res. 2018 Sep;27(9):2610-2626
pubmed: 28034175
J Pers Soc Psychol. 2000 Jun;78(6):1150-69
pubmed: 10870915
J Pain Symptom Manage. 2004 Dec;28(6):612-8
pubmed: 15589087
Pharm Stat. 2013 Nov-Dec;12(6):337-47
pubmed: 23292975
J Clin Epidemiol. 2016 Jun;74:57-65
pubmed: 26718729
Palliat Med. 2013 Dec;27(10):899-907
pubmed: 23652842
Am J Epidemiol. 2017 Feb 15;185(4):304-315
pubmed: 28073767
Biostatistics. 2018 Oct 1;19(4):407-425
pubmed: 29028922
J Clin Epidemiol. 2015 Oct;68(10):1165-75
pubmed: 25920943
J Palliat Med. 2005;8 Suppl 1:S50-7
pubmed: 16499469
Stat Med. 2011 Feb 20;30(4):377-99
pubmed: 21225900
Stat Med. 2014 Sep 20;33(21):3601-28
pubmed: 24889022
Lung Cancer. 2015 Jun;88(3):304-9
pubmed: 25870155
Stat Med. 2016 Jul 30;35(17):2991-3006
pubmed: 26179943
Eur J Epidemiol. 2016 Apr;31(4):337-50
pubmed: 27209009
Am J Epidemiol. 2010 Nov 1;172(9):1070-6
pubmed: 20841346
Stat Anal Data Min. 2017 Dec;10(6):363-377
pubmed: 29403567
Pharm Stat. 2014 Jul-Aug;13(4):258-64
pubmed: 24931317
J Pers Soc Psychol. 1989 Feb;56(2):267-83
pubmed: 2926629
BMC Med Res Methodol. 2020 Jul 25;20(1):199
pubmed: 32711455
BMC Cancer. 2020 Mar 23;20(1):246
pubmed: 32293337
Stat Med. 2003 Mar 15;22(5):795-809
pubmed: 12587106
Biometrics. 2018 Dec;74(4):1232-1239
pubmed: 29665626
J Clin Epidemiol. 2017 Aug;88:81-91
pubmed: 28532739
Stat Methods Med Res. 2016 Aug;25(4):1471-89
pubmed: 23698867
BMC Cancer. 2016 Apr 08;16:264
pubmed: 27059593
Epidemiology. 2011 Mar;22(2):282
pubmed: 21293212
Eur J Cancer. 2005 Sep;41(14):2120-31
pubmed: 16182120
Stat Methods Med Res. 2017 Apr;26(2):674-690
pubmed: 25305196
Sci Rep. 2017 May 2;7(1):1339
pubmed: 28465554
Health Qual Life Outcomes. 2008 Aug 04;6:57
pubmed: 18680574
BMC Med Res Methodol. 2014 Nov 19;14:118
pubmed: 25407057