Latent class instrumental variables: A clinical and biostatistical perspective

Stuart G. Baker, Barnett S. Kramer, Karen Sue Lindeman

Research output: Contribution to journalArticle

Abstract

In some two-arm randomized trials, some participants receive the treatment assigned to the other arm as a result of technical problems, refusal of a treatment invitation, or a choice of treatment in an encouragement design. In some before-and-after studies, the availability of a new treatment changes from one time period to this next. Under assumptions that are often reasonable, the latent class instrumental variable (IV) method estimates the effect of treatment received in the aforementioned scenarios involving all-or-none compliance and all-or-none availability. Key aspects are four initial latent classes (sometimes called principal strata) based on treatment received if in each randomization group or time period, the exclusion restriction assumption (in which randomization group or time period is an instrumental variable), the monotonicity assumption (which drops an implausible latent class from the analysis), and the estimated effect of receiving treatment in one latent class (sometimes called efficacy, the local average treatment effect, or the complier average causal effect). Since its independent formulations in the biostatistics and econometrics literatures, the latent class IV method (which has no well-established name) has gained increasing popularity. We review the latent class IV method from a clinical and biostatistical perspective, focusing on underlying assumptions, methodological extensions, and applications in our fields of obstetrics and cancer research.

Original languageEnglish (US)
Pages (from-to)147-160
Number of pages14
JournalStatistics in Medicine
Volume35
Issue number1
DOIs
StatePublished - Jan 15 2016

Fingerprint

Latent Class
Instrumental Variables
Random Allocation
Treatment Refusal
Biostatistics
Compliance
Obstetrics
Names
Randomisation
Availability
Average Treatment Effect
Causal Effect
Randomized Trial
Research
Econometrics
Neoplasms
Monotonicity
Efficacy
Cancer
Restriction

Keywords

  • All-or-none compliance
  • Causal inference
  • Encouragement design, observational
  • Paired availability design
  • Principal stratification, randomized trial

ASJC Scopus subject areas

  • Epidemiology
  • Statistics and Probability

Cite this

Latent class instrumental variables : A clinical and biostatistical perspective. / Baker, Stuart G.; Kramer, Barnett S.; Lindeman, Karen Sue.

In: Statistics in Medicine, Vol. 35, No. 1, 15.01.2016, p. 147-160.

Research output: Contribution to journalArticle

Baker, Stuart G. ; Kramer, Barnett S. ; Lindeman, Karen Sue. / Latent class instrumental variables : A clinical and biostatistical perspective. In: Statistics in Medicine. 2016 ; Vol. 35, No. 1. pp. 147-160.
@article{e89049e475e1449bbba901ed89fdfb3c,
title = "Latent class instrumental variables: A clinical and biostatistical perspective",
abstract = "In some two-arm randomized trials, some participants receive the treatment assigned to the other arm as a result of technical problems, refusal of a treatment invitation, or a choice of treatment in an encouragement design. In some before-and-after studies, the availability of a new treatment changes from one time period to this next. Under assumptions that are often reasonable, the latent class instrumental variable (IV) method estimates the effect of treatment received in the aforementioned scenarios involving all-or-none compliance and all-or-none availability. Key aspects are four initial latent classes (sometimes called principal strata) based on treatment received if in each randomization group or time period, the exclusion restriction assumption (in which randomization group or time period is an instrumental variable), the monotonicity assumption (which drops an implausible latent class from the analysis), and the estimated effect of receiving treatment in one latent class (sometimes called efficacy, the local average treatment effect, or the complier average causal effect). Since its independent formulations in the biostatistics and econometrics literatures, the latent class IV method (which has no well-established name) has gained increasing popularity. We review the latent class IV method from a clinical and biostatistical perspective, focusing on underlying assumptions, methodological extensions, and applications in our fields of obstetrics and cancer research.",
keywords = "All-or-none compliance, Causal inference, Encouragement design, observational, Paired availability design, Principal stratification, randomized trial",
author = "Baker, {Stuart G.} and Kramer, {Barnett S.} and Lindeman, {Karen Sue}",
year = "2016",
month = "1",
day = "15",
doi = "10.1002/sim.6612",
language = "English (US)",
volume = "35",
pages = "147--160",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "1",

}

TY - JOUR

T1 - Latent class instrumental variables

T2 - A clinical and biostatistical perspective

AU - Baker, Stuart G.

AU - Kramer, Barnett S.

AU - Lindeman, Karen Sue

PY - 2016/1/15

Y1 - 2016/1/15

N2 - In some two-arm randomized trials, some participants receive the treatment assigned to the other arm as a result of technical problems, refusal of a treatment invitation, or a choice of treatment in an encouragement design. In some before-and-after studies, the availability of a new treatment changes from one time period to this next. Under assumptions that are often reasonable, the latent class instrumental variable (IV) method estimates the effect of treatment received in the aforementioned scenarios involving all-or-none compliance and all-or-none availability. Key aspects are four initial latent classes (sometimes called principal strata) based on treatment received if in each randomization group or time period, the exclusion restriction assumption (in which randomization group or time period is an instrumental variable), the monotonicity assumption (which drops an implausible latent class from the analysis), and the estimated effect of receiving treatment in one latent class (sometimes called efficacy, the local average treatment effect, or the complier average causal effect). Since its independent formulations in the biostatistics and econometrics literatures, the latent class IV method (which has no well-established name) has gained increasing popularity. We review the latent class IV method from a clinical and biostatistical perspective, focusing on underlying assumptions, methodological extensions, and applications in our fields of obstetrics and cancer research.

AB - In some two-arm randomized trials, some participants receive the treatment assigned to the other arm as a result of technical problems, refusal of a treatment invitation, or a choice of treatment in an encouragement design. In some before-and-after studies, the availability of a new treatment changes from one time period to this next. Under assumptions that are often reasonable, the latent class instrumental variable (IV) method estimates the effect of treatment received in the aforementioned scenarios involving all-or-none compliance and all-or-none availability. Key aspects are four initial latent classes (sometimes called principal strata) based on treatment received if in each randomization group or time period, the exclusion restriction assumption (in which randomization group or time period is an instrumental variable), the monotonicity assumption (which drops an implausible latent class from the analysis), and the estimated effect of receiving treatment in one latent class (sometimes called efficacy, the local average treatment effect, or the complier average causal effect). Since its independent formulations in the biostatistics and econometrics literatures, the latent class IV method (which has no well-established name) has gained increasing popularity. We review the latent class IV method from a clinical and biostatistical perspective, focusing on underlying assumptions, methodological extensions, and applications in our fields of obstetrics and cancer research.

KW - All-or-none compliance

KW - Causal inference

KW - Encouragement design, observational

KW - Paired availability design

KW - Principal stratification, randomized trial

UR - http://www.scopus.com/inward/record.url?scp=84954384808&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84954384808&partnerID=8YFLogxK

U2 - 10.1002/sim.6612

DO - 10.1002/sim.6612

M3 - Article

C2 - 26239275

AN - SCOPUS:84954384808

VL - 35

SP - 147

EP - 160

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 1

ER -