1
|
Roberts EK, Elliott MR, Taylor JMG. Surrogacy validation for time-to-event outcomes with illness-death frailty models. Biom J 2024; 66:e2200324. [PMID: 37776057 PMCID: PMC10873101 DOI: 10.1002/bimj.202200324] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/29/2022] [Revised: 04/20/2023] [Accepted: 06/15/2023] [Indexed: 10/01/2023]
Abstract
A common practice in clinical trials is to evaluate a treatment effect on an intermediate outcome when the true outcome of interest would be difficult or costly to measure. We consider how to validate intermediate outcomes in a causally-valid way when the trial outcomes are time-to-event. Using counterfactual outcomes, those that would be observed if the counterfactual treatment had been given, the causal association paradigm assesses the relationship of the treatment effect on the surrogate outcome with the treatment effect on the true, primary outcome. In particular, we propose illness-death models to accommodate the censored and semicompeting risk structure of survival data. The proposed causal version of these models involves estimable and counterfactual frailty terms. Via these multistate models, we characterize what a valid surrogate would look like using a causal effect predictiveness plot. We evaluate the estimation properties of a Bayesian method using Markov chain Monte Carlo and assess the sensitivity of our model assumptions. Our motivating data source is a localized prostate cancer clinical trial where the two survival outcomes are time to distant metastasis and time to death.
Collapse
Affiliation(s)
| | - Michael R. Elliott
- Department of Biostatistics, University of Michigan, Ann Arbor, MI
- Survey Methodology Program, Institute for Social Research Ann Arbor, MI
| | | |
Collapse
|
2
|
Roberts EK, Elliott MR, Taylor JMG. Solutions for surrogacy validation with longitudinal outcomes for a gene therapy. Biometrics 2023; 79:1840-1852. [PMID: 35833874 DOI: 10.1111/biom.13720] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/29/2021] [Accepted: 07/01/2022] [Indexed: 11/30/2022]
Abstract
Valid surrogate endpoints S can be used as a substitute for a true outcome of interest T to measure treatment efficacy in a clinical trial. We propose a causal inference approach to validate a surrogate by incorporating longitudinal measurements of the true outcomes using a mixed modeling approach, and we define models and quantities for validation that may vary across the study period using principal surrogacy criteria. We consider a surrogate-dependent treatment efficacy curve that allows us to validate the surrogate at different time points. We extend these methods to accommodate a delayed-start treatment design where all patients eventually receive the treatment. Not all parameters are identified in the general setting. We apply a Bayesian approach for estimation and inference, utilizing more informative prior distributions for selected parameters. We consider the sensitivity of these prior assumptions as well as assumptions of independence among certain counterfactual quantities conditional on pretreatment covariates to improve identifiability. We examine the frequentist properties (bias of point and variance estimates, credible interval coverage) of a Bayesian imputation method. Our work is motivated by a clinical trial of a gene therapy where the functional outcomes are measured repeatedly throughout the trial.
Collapse
Affiliation(s)
- Emily K Roberts
- Department of Biostatistics, University Michigan, Ann Arbor, Michigan, USA
| | - Michael R Elliott
- Department of Biostatistics, University Michigan, Ann Arbor, Michigan, USA
- Survey Methodology Program, Institute for Social Research, Ann Arbor, Michigan, USA
| | - Jeremy M G Taylor
- Department of Biostatistics, University Michigan, Ann Arbor, Michigan, USA
| |
Collapse
|
3
|
Roberts EK, Gu T, Wagner AL, Mukherjee B, Fritsche LG. Estimating COVID-19 Vaccination and Booster Effectiveness Using Electronic Health Records From an Academic Medical Center in Michigan. AJPM Focus 2022; 1:100015. [PMID: 36942016 PMCID: PMC9323299 DOI: 10.1016/j.focus.2022.100015] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Indexed: 11/19/2022]
Abstract
Introduction Observational studies of COVID-19 vaccines' effectiveness can provide crucial information regarding the strength and durability of protection against SARS-CoV-2 infection and whether the protective response varies across different patient subpopulations and in the context of different SARS-CoV-2 variants. Methods We used a test-negative study design to assess vaccine effectiveness against SARS-CoV-2 infection and severe COVID-19 resulting in hospitalization, intensive care unit admission, or death using electronic health records data of 170,741 adults who had been tested for COVID-19 at the University of Michigan Medical Center between January 1 and December 31, 2021. We estimated vaccine effectiveness by comparing the odds of vaccination between cases and controls during each 2021 calendar quarter and stratified all outcomes by vaccine type, patient demographic and clinical characteristics, and booster status. Results Unvaccinated individuals had more than double the rate of infections (12.1% vs 4.7%) and >3 times the rate of severe COVID-19 outcomes (1.4% vs 0.4%) than vaccinated individuals. COVID-19 vaccines were 62.1% (95% CI=60.3, 63.8) effective against a new infection, with protection waning in the last 2 quarters of 2021. The vaccine effectiveness against severe disease overall was 73.7% (95% CI=69.6, 77.3) and remained high throughout 2021. Data from the last quarter of 2021 indicated that adding a booster dose augmented effectiveness against infection up to 87.3% (95% CI=85.0, 89.2) and against severe outcomes up to 94.0% (95% CI=89.5, 96.6). Pfizer-BioNTech and Moderna vaccines showed comparable performance when controlling for vaccination timing. Vaccine effectiveness was greater in more socioeconomically affluent areas and among healthcare workers; otherwise, we did not detect any significant modification of vaccine effectiveness by covariates, including gender, race, and SES. Conclusions COVID-19 vaccines were highly protective against infection and severe COVID-19 resulting in hospitalization, intensive care unit admission, or death. Administration of a booster dose significantly increased vaccine effectiveness against both outcomes. Ongoing surveillance is required to assess the durability of these findings.
Collapse
Affiliation(s)
- Emily K. Roberts
- Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, Michigan
- Center for Precision Health Data Science, School of Public Health, University of Michigan, Ann Arbor, Michigan
| | - Tian Gu
- Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, Massachusetts
| | - Abram L. Wagner
- Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan
| | - Bhramar Mukherjee
- Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, Michigan
- Center for Precision Health Data Science, School of Public Health, University of Michigan, Ann Arbor, Michigan
- Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan
- Michigan Institute for Data Science, University of Michigan, Ann Arbor, Michigan
| | - Lars G. Fritsche
- Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, Michigan
- Center for Precision Health Data Science, School of Public Health, University of Michigan, Ann Arbor, Michigan
| |
Collapse
|
4
|
Vornhagen J, Roberts EK, Unverdorben L, Mason S, Patel A, Crawford R, Holmes CL, Sun Y, Teodorescu A, Snitkin ES, Zhao L, Simner PJ, Tamma PD, Rao K, Kaye KS, Bachman MA. Combined comparative genomics and clinical modeling reveals plasmid-encoded genes are independently associated with Klebsiella infection. Nat Commun 2022; 13:4459. [PMID: 35915063 PMCID: PMC9343666 DOI: 10.1038/s41467-022-31990-1] [Citation(s) in RCA: 8] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/12/2022] [Accepted: 07/12/2022] [Indexed: 11/28/2022] Open
Abstract
Members of the Klebsiella pneumoniae species complex frequently colonize the gut and colonization is associated with subsequent infection. To identify genes associated with progression from colonization to infection, we undertook a case-control comparative genomics study. Concordant cases (N = 85), where colonizing and invasive isolates were identical strain types, were matched to asymptomatically colonizing controls (N = 160). Thirty-seven genes are associated with infection, 27 of which remain significant following adjustment for patient variables and bacterial phylogeny. Infection-associated genes are not previously characterized virulence factors, but instead a diverse group of stress resistance, regulatory and antibiotic resistance genes, despite careful adjustment for antibiotic exposure. Many genes are plasmid borne, and for some, the relationship with infection is mediated by gut dominance. Five genes were validated in a geographically-independent cohort of colonized patients. This study identifies several genes reproducibly associated with progression to infection in patients colonized by diverse Klebsiella.
Collapse
Affiliation(s)
- Jay Vornhagen
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
- Department of Microbiology & Immunology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Emily K Roberts
- Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, MI, USA
| | - Lavinia Unverdorben
- Department of Microbiology & Immunology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Sophia Mason
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Alieysa Patel
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Ryan Crawford
- Department of Computational Medicine and Bioinformatics, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Caitlyn L Holmes
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
- Department of Microbiology & Immunology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Yuang Sun
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Alexandra Teodorescu
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Evan S Snitkin
- Department of Microbiology & Immunology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
- Department of Internal Medicine/Infectious Diseases Division, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Lili Zhao
- Department of Biostatistics, School of Public Health, University of Michigan, Ann Arbor, MI, USA
| | - Patricia J Simner
- Division of Medical Microbiology, Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, MI, USA
| | - Pranita D Tamma
- Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, MI, USA
| | - Krishna Rao
- Department of Internal Medicine/Infectious Diseases Division, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Keith S Kaye
- Department of Internal Medicine/Infectious Diseases Division, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA
| | - Michael A Bachman
- Department of Pathology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA.
- Department of Microbiology & Immunology, Michigan Medicine, University of Michigan, Ann Arbor, MI, USA.
| |
Collapse
|
5
|
Roberts EK, Elliott MR, Taylor JMG. Incorporating baseline covariates to validate surrogate endpoints with a constant biomarker under control arm. Stat Med 2021; 40:6605-6618. [PMID: 34528260 DOI: 10.1002/sim.9201] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/24/2021] [Revised: 07/20/2021] [Accepted: 08/31/2021] [Indexed: 11/09/2022]
Abstract
A surrogate endpoint S in a clinical trial is an outcome that may be measured earlier or more easily than the true outcome of interest T. In this work, we extend causal inference approaches to validate such a surrogate using potential outcomes. The causal association paradigm assesses the relationship of the treatment effect on the surrogate with the treatment effect on the true endpoint. Using the principal surrogacy criteria, we utilize the joint conditional distribution of the potential outcomes T, given the potential outcomes S. In particular, our setting of interest allows us to assume the surrogate under the placebo, S ( 0 ) , is zero-valued, and we incorporate baseline covariates in the setting of normally distributed endpoints. We develop Bayesian methods to incorporate conditional independence and other modeling assumptions and explore their impact on the assessment of surrogacy. We demonstrate our approach via simulation and data that mimics an ongoing study of a muscular dystrophy gene therapy.
Collapse
Affiliation(s)
- Emily K Roberts
- Department of Biostatistics, University of Michigan School of Public Health, Ann Arbor, Michigan, USA
| | - Michael R Elliott
- Department of Biostatistics, University of Michigan School of Public Health, Ann Arbor, Michigan, USA.,Survey Methodology Program, Institute for Social Research, Ann Arbor, Michigan, USA
| | - Jeremy M G Taylor
- Department of Biostatistics, University of Michigan School of Public Health, Ann Arbor, Michigan, USA
| |
Collapse
|
6
|
Steckler TL, Roberts EK, Doop DD, Lee TM, Padmanabhan V. Developmental programming in sheep: administration of testosterone during 60-90 days of pregnancy reduces breeding success and pregnancy outcome. Theriogenology 2006; 67:459-67. [PMID: 17010414 DOI: 10.1016/j.theriogenology.2006.08.010] [Citation(s) in RCA: 32] [Impact Index Per Article: 1.8] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/17/2006] [Accepted: 08/16/2006] [Indexed: 10/24/2022]
Abstract
Evidence suggests that exposure to excess steroids during critical periods of fetal development leads to reproductive disorders. Exposure of female lambs to excess testosterone (T) from Days 60 to 90 of gestation (T60-90; term, 147 days) delayed onset of the LH surge and resulted in to male-typical reproductive behavior. The objectives of this study were to test the ability of T60-90 ewes to mate, conceive and lamb during the first three breeding seasons (Years 1, 2 and 3). Pregnant Suffolk ewes were injected with T propionate in cottonseed oil (100mg, im twice weekly) or vehicle (control; C) from Days 60 to 90 of gestation. In Year 1, ewes (C=12, T60-90=12) were kept with a vasectomized ram for 3 months and markings/visual observation of copulations were recorded. Rams had paint applied to their chest to facilitate detection of estrus and mating. All C but only three T60-90 ewes were marked (P<0.001). All ewes were then estrus-synchronized with two injections of prostaglandin F2alpha (20mg, im) given 11 days apart and allowed to mate with a painted, fertile ram. Nine of 12 C and 4 of 12 T60-90 ewes (P=0.1) were mated. Based on estrus and long-term monitoring of progesterone, more C than T60-90 became pregnant (82 and 18%, respectively; P<0.01). In Year 2, to maximize ram exposure, two C and two T60-90 estrus-synchronized ewes were placed with a painted, fertile ram at a time and mated ewes were removed to a nearby pen to force mating with others. Twenty-four hour video monitoring revealed the rams mated more C than T60-90 ewes (83 and 25%, respectively; P=0.01). In both Years 1 and 2, the rams preferred C over T60-90 ewes; therefore in Year 3 rams were given access only to T60-90 ewes. Only four T60-90 estrus-synchronized ewes were placed with a painted ram at a time. Not given an option, 91% of the T60-90 ewes were marked resulting in 4 of 11 (36%; first-service pregnancy rate in the breeding herd was 91%) ewes becoming pregnant to the synchronized estrus. Collectively these studies showed that fertility in T60-90 females was severely compromised, even after overcoming ram preference for controls.
Collapse
Affiliation(s)
- T L Steckler
- Department of Pediatrics, University of Michigan, Ann Arbor, MI, United States
| | | | | | | | | |
Collapse
|