Analysis of self-report and biochemically verified tobacco abstinence outcomes with missing data: A sensitivity analysis using two-stage imputation

Yiwen Zhang, Xianghua Luo, Chap T Le, Jasjit S. Ahluwalia, Janet L Thomas

Research output: Contribution to journalArticle

Abstract

Background: Missing data are common in tobacco studies. It is well known that from the observed data alone, it is impossible to distinguish between missing mechanisms such as missing at random (MAR) and missing not at random (MNAR). In this paper, we propose a sensitivity analysis method to accommodate different missing mechanisms in cessation outcomes determined by self-report and urine validation results. Methods: We propose a two-stage imputation procedure, allowing survey and urine data to be missing under different mechanisms. The motivating data were from a tobacco cessation trial examining the effects of the extended vs. standard Quit and Win contests and counseling vs. no counseling under a 2-by-2 factorial design. The primary outcome was 6-month biochemically verified tobacco abstinence. Results: Our proposed method covers a wide spectrum of missing scenarios, including the widely adopted "missing = smoking" imputation by assuming a perfect smoking-missing correlation (an extreme case of MNAR), the MAR case by assuming a zero smoking-missing correlation, and many more in between. The analysis of the data example shows that the estimated effects of the studied interventions are sensitive to the different missing assumptions on the survey and urine data. Conclusions: Sensitivity analysis has played a crucial role in assessing the robustness of the findings in clinical trials with missing data. The proposed method provides an effective tool for analyzing missing data introduced at two different stages of outcome assessment, the self-report and validation time. Our methods are applicable to trials studying biochemically verified abstinence from alcohol and other substances.

Original languageEnglish (US)
Article number170
JournalBMC Medical Research Methodology
Volume18
Issue number1
DOIs
StatePublished - Dec 18 2018

Fingerprint

Self Report
Tobacco
Smoking
Urine
Counseling
Alcohol Abstinence
Tobacco Use Cessation
Outcome Assessment (Health Care)
Clinical Trials
Surveys and Questionnaires

Keywords

  • Abstinence outcome
  • Imputation
  • Missing data
  • Sensitivity analysis

PubMed: MeSH publication types

  • Journal Article
  • Research Support, N.I.H., Extramural

Cite this

Analysis of self-report and biochemically verified tobacco abstinence outcomes with missing data : A sensitivity analysis using two-stage imputation. / Zhang, Yiwen; Luo, Xianghua; Le, Chap T; Ahluwalia, Jasjit S.; Thomas, Janet L.

In: BMC Medical Research Methodology, Vol. 18, No. 1, 170, 18.12.2018.

Research output: Contribution to journalArticle

@article{72216b7f5c5d4997ac17d0c4d49bb76b,
title = "Analysis of self-report and biochemically verified tobacco abstinence outcomes with missing data: A sensitivity analysis using two-stage imputation",
abstract = "Background: Missing data are common in tobacco studies. It is well known that from the observed data alone, it is impossible to distinguish between missing mechanisms such as missing at random (MAR) and missing not at random (MNAR). In this paper, we propose a sensitivity analysis method to accommodate different missing mechanisms in cessation outcomes determined by self-report and urine validation results. Methods: We propose a two-stage imputation procedure, allowing survey and urine data to be missing under different mechanisms. The motivating data were from a tobacco cessation trial examining the effects of the extended vs. standard Quit and Win contests and counseling vs. no counseling under a 2-by-2 factorial design. The primary outcome was 6-month biochemically verified tobacco abstinence. Results: Our proposed method covers a wide spectrum of missing scenarios, including the widely adopted {"}missing = smoking{"} imputation by assuming a perfect smoking-missing correlation (an extreme case of MNAR), the MAR case by assuming a zero smoking-missing correlation, and many more in between. The analysis of the data example shows that the estimated effects of the studied interventions are sensitive to the different missing assumptions on the survey and urine data. Conclusions: Sensitivity analysis has played a crucial role in assessing the robustness of the findings in clinical trials with missing data. The proposed method provides an effective tool for analyzing missing data introduced at two different stages of outcome assessment, the self-report and validation time. Our methods are applicable to trials studying biochemically verified abstinence from alcohol and other substances.",
keywords = "Abstinence outcome, Imputation, Missing data, Sensitivity analysis",
author = "Yiwen Zhang and Xianghua Luo and Le, {Chap T} and Ahluwalia, {Jasjit S.} and Thomas, {Janet L}",
year = "2018",
month = "12",
day = "18",
doi = "10.1186/s12874-018-0635-2",
language = "English (US)",
volume = "18",
journal = "BMC Medical Research Methodology",
issn = "1471-2288",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - Analysis of self-report and biochemically verified tobacco abstinence outcomes with missing data

T2 - A sensitivity analysis using two-stage imputation

AU - Zhang, Yiwen

AU - Luo, Xianghua

AU - Le, Chap T

AU - Ahluwalia, Jasjit S.

AU - Thomas, Janet L

PY - 2018/12/18

Y1 - 2018/12/18

N2 - Background: Missing data are common in tobacco studies. It is well known that from the observed data alone, it is impossible to distinguish between missing mechanisms such as missing at random (MAR) and missing not at random (MNAR). In this paper, we propose a sensitivity analysis method to accommodate different missing mechanisms in cessation outcomes determined by self-report and urine validation results. Methods: We propose a two-stage imputation procedure, allowing survey and urine data to be missing under different mechanisms. The motivating data were from a tobacco cessation trial examining the effects of the extended vs. standard Quit and Win contests and counseling vs. no counseling under a 2-by-2 factorial design. The primary outcome was 6-month biochemically verified tobacco abstinence. Results: Our proposed method covers a wide spectrum of missing scenarios, including the widely adopted "missing = smoking" imputation by assuming a perfect smoking-missing correlation (an extreme case of MNAR), the MAR case by assuming a zero smoking-missing correlation, and many more in between. The analysis of the data example shows that the estimated effects of the studied interventions are sensitive to the different missing assumptions on the survey and urine data. Conclusions: Sensitivity analysis has played a crucial role in assessing the robustness of the findings in clinical trials with missing data. The proposed method provides an effective tool for analyzing missing data introduced at two different stages of outcome assessment, the self-report and validation time. Our methods are applicable to trials studying biochemically verified abstinence from alcohol and other substances.

AB - Background: Missing data are common in tobacco studies. It is well known that from the observed data alone, it is impossible to distinguish between missing mechanisms such as missing at random (MAR) and missing not at random (MNAR). In this paper, we propose a sensitivity analysis method to accommodate different missing mechanisms in cessation outcomes determined by self-report and urine validation results. Methods: We propose a two-stage imputation procedure, allowing survey and urine data to be missing under different mechanisms. The motivating data were from a tobacco cessation trial examining the effects of the extended vs. standard Quit and Win contests and counseling vs. no counseling under a 2-by-2 factorial design. The primary outcome was 6-month biochemically verified tobacco abstinence. Results: Our proposed method covers a wide spectrum of missing scenarios, including the widely adopted "missing = smoking" imputation by assuming a perfect smoking-missing correlation (an extreme case of MNAR), the MAR case by assuming a zero smoking-missing correlation, and many more in between. The analysis of the data example shows that the estimated effects of the studied interventions are sensitive to the different missing assumptions on the survey and urine data. Conclusions: Sensitivity analysis has played a crucial role in assessing the robustness of the findings in clinical trials with missing data. The proposed method provides an effective tool for analyzing missing data introduced at two different stages of outcome assessment, the self-report and validation time. Our methods are applicable to trials studying biochemically verified abstinence from alcohol and other substances.

KW - Abstinence outcome

KW - Imputation

KW - Missing data

KW - Sensitivity analysis

UR - http://www.scopus.com/inward/record.url?scp=85058817766&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058817766&partnerID=8YFLogxK

U2 - 10.1186/s12874-018-0635-2

DO - 10.1186/s12874-018-0635-2

M3 - Article

C2 - 30563473

AN - SCOPUS:85058817766

VL - 18

JO - BMC Medical Research Methodology

JF - BMC Medical Research Methodology

SN - 1471-2288

IS - 1

M1 - 170

ER -