Emulating a Sequence of Target Trials to Avoid Immortal Time Bias in Pregnancy Studies — an Application to Antibiotic Initiation and Preterm Delivery

Saturday, August 27, 2022

8:45 AM – 9:00 AM CEST

Location: C1M5

Publication Number: 117

Background: Immortal time bias is introduced when treatment initiation occurs after time zero of follow-up and is common in observational studies of medication use during pregnancy.

Objectives: To describe how emulating a series of hypothetical randomized trials (target trials) avoids immortal time bias and apply the approach to antibiotic initiation and preterm delivery.

Methods: The Tsepamo Study captured birth outcomes at up to 18 delivery sites in Botswana from 2014-2021. To investigate the relationship between antibiotic initiation after 24 weeks gestation and preterm delivery (24-37 weeks gestation) among women presenting for antenatal care < 24 weeks gestation we: 1) started follow-up at 24 weeks and defined exposure as antibiotic initiation between 24-37 weeks gestation (susceptible to immortal time bias) and 2) conducted 13 sequential target trials of antibiotic initiation versus no initiation during each week from 24-37 weeks gestation. For each trial, eligible women had not previously initiated antibiotics and remained pregnant at the start of that week. We used log-binomial models to calculate risk ratios (RR) and 95% confidence intervals (CI).

Results: Of 111,403 eligible women, 17,009 (15%) initiated antibiotics between 24-37 weeks gestation. The RR (95% CI) for preterm delivery comparing antibiotic initiation to no initiation between 24-37 weeks was 0.92 (0.89, 0.96). In the series of target trials which address immortal time bias, the RRs (95% CIs) ranged from 1.06 (0.93, 1.22) to 1.28 (1.14, 1.43).

Conclusions: Because the time period between time zero and treatment initiation is "immortal" to the outcome, defining exposure based on treatment initiation after time zero biased the RR for the association between antibiotic initiation and preterm delivery downwards. Conducting a series of target trials can avoid this immortal time bias in pregnancy studies by aligning exposure initiation and start of follow-up.

Presenting Author(s)

Ellen Caniglia

University of Pennsylvania Perelman School of Medicine
Philadelphia, Pennsylvania, United States

Slides

Co-Author(s)

RZ

Rebecca Zash

Beth Israel Deaconess Medical Center, United States
CF

Christina Fennell

Harvard T.H. Chan School of Public Health, United States
MD

Modiegi Diseko

Botswana Harvard AIDS Institute Partnership
Gaborone, South-East, Botswana
GM

Gloria Mayondi

Botswana Harvard AIDS Institute Partnership, Botswana
JH

Jonathan Heintz

University of Pennsylvania Perelman School of Medicine, United States
MM

Mompati Mmalane

Botswana Harvard AIDS Institute Partnership, Botswana
JM

Joseph Makhema, F.R.C.P

Chief Executive Officer
Botswana Harvard AIDS Institute Partnership
Gaborone, Botswana
SL

Shahin Lockman

Harvard T.H. Chan School of Public Health, United States
SM

Sunni L. Mumford

University of Pennsylvania Perelman School of Medicine, United States
EM

Eleanor Murray

Boston University School of Public Health, United States
Sonia Hernandez-Diaz, MD, MPH, FISPE

Professor of Epidemiology
Harvard School of Public Health
Boston, Massachusetts, United States
RS

Roger Shapiro

Harvard T.H. Chan School of Public Health, United States