Premium
A ‘missing not at random’ ( MNAR ) and ‘missing at random’ ( MAR ) growth model comparison with a buprenorphine/naloxone clinical trial
Author(s) -
McPherson Sterling,
BarbosaLeiker Celestina,
Mamey Mary Rose,
McDonell Michael,
Enders Craig K.,
Roll John
Publication year - 2015
Publication title -
addiction
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.424
H-Index - 193
eISSN - 1360-0443
pISSN - 0965-2140
DOI - 10.1111/add.12714
Subject(s) - missing data , buprenorphine , dropout (neural networks) , (+) naloxone , psychology , medicine , opioid , statistics , mathematics , receptor , machine learning , computer science
Aims To compare three missing data strategies: (i) the latent growth model that assumes the data are missing at random ( MAR ) model; (ii) the D iggle– K enward missing not at random ( MNAR ) model, where dropout is a function of previous/concurrent urinalysis ( UA ) submissions; and (iii) the W u– C arroll MNAR model where dropout is a function of the growth factors. Design Secondary data analysis of a N ational D rug A buse T reatment C linical T rials N etwork trial that examined a 7‐day versus 28‐day taper (i.e. stepwise decrease in buprenorphine/naloxone) on the likelihood of submitting an opioid‐positive UA during treatment. Setting 11 out‐patient treatment settings in 10 US cities. Participants A total of 516 opioid‐dependent participants. Measurements Opioid UAs provided across the 4‐week treatment period. Findings The MAR model showed a significant effect ( B = −0.45, P < 0.05) of trial arm on the opioid‐positive UA slope (i.e. 28‐day taper participants were less likely to submit a positive UA over time) with a small effect size ( d = 0.20). The MNAR D iggle– K enward model demonstrated a significant ( B = −0.64, P < 0.01) effect of trial arm on the slope with a large effect size ( d = 0.82). The MNAR W u– C arroll model showed a significant ( B = −0.41, P < 0.05) effect of trial arm on the UA slope that was relatively small ( d = 0.31). Conclusions This performance comparison of three missing data strategies (latent growth model, D iggle– K enward selection model, W u– C arrol selection model) on sample data indicates a need for increased use of sensitivity analyses in clinical trial research. Given the potential sensitivity of the trial arm effect to missing data assumptions, it is critical for researchers to consider whether the assumptions associated with each model are defensible.