Advanced Epidemiologic Methods

L11: Emulating target trials

Thu, 07 Apr 2022 00:00:00 +0000

Readings

García-Albéniz X, Hernán MA, Logan RW, Price M, Armstrong K, Hsu J. Continuation of annual screening mammography and breast cancer mortality in women older than 70 years. Annals of internal medicine. 2020 Mar 17;172(6):381-9.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

There is no exercise this week. Instead, please work on your final project to design a target trial. Come to the lab session prepared to do a dry run of your final presentation and get feedback from your peers.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…

L10: Per protocol analyses

Thu, 31 Mar 2022 00:00:00 +0000

Readings

Lu H, Cole SR, Westreich D, Hudgens MG, Adimora AA, Althoff KN, Silverberg MJ, Buchacz K, Li J, Edwards JK, Rebeiro PF. Clinical Effectiveness of Integrase Strand Transfer Inhibitor–Based Antiretroviral Regimens Among Adults With Human Immunodeficiency Virus: A Collaboration of Cohort Studies in the United States and Canada. Clinical Infectious Diseases. 2021 Oct 1;73(7):e1408-14.

Notes

We are often interested in estimating the effects of treatments or interventions that are maintained over time, but exposures to these treatments and interventions often vary over time for participants in our datasets. Reasons may include

treatment changes or discontinuation
nonadherence
loss to follow-up
mobility and migration

among others.

This week, we will focus on a simple scenario in which we want to compare risk of death under continuous treatment and under no treatment, but some people in the treatment arm discontinue treatment. In the exercise, we will compare risk under 2 treatments, where people in both arms discontinue treatment.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

Exercise 10 may be found on the course Teams site.

Optional readings

In the reading and during lecture, we discussed using IPW to estimate “per protocol” effects in the presence of a simple time-varying exposure. The reading below illustrates use of g computation to achieve the same goals.

Naimi AI, Perkins NJ, Sjaarda LA, Mumford SL, Platt RW, Silver RM, Schisterman EF. The Effect of Preconception-Initiated Low-Dose Aspirin on Human Chorionic Gonadotropin–Detected Pregnancy, Pregnancy Loss, and Live Birth: Per Protocol Analysis of a Randomized Trial. Annals of Internal Medicine. 2021 Jan 26.

The reading also introduced the idea of emulating a randomized trial using observational data. We will discuss this more next week. In the meantime, you can read more about this idea here:

Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial is not available. American journal of epidemiology. 2016 Apr 15;183(8):758-64.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…

L8: Intro to part 2

Tue, 22 Mar 2022 00:00:00 +0000

Readings

No reading for this class session.

Notes

Today we discuss what is coming up in part 2 of the course. The goal of today’s lecture is to explain what we will discuss (and why) in the second part of the course, give you an overview of time varying exposures, and discuss why framing our questions becomes more challenging when factors vary over time.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exam info

Tue, 08 Mar 2022 00:00:00 +0000

The exam will be held in person on March 8th. Additional information will be provided here and in class prior to the exam.

Designing a time-to-event study

Thu, 24 Feb 2022 00:00:00 +0000

Readings

Hernán MA, Sauer BC, Hernández-Díaz S, Platt R, Shrier I. Specifying a target trial prevents immortal time bias and other self-inflicted injuries in observational analyses. Journal of clinical epidemiology. 2016 Nov 1;79:70-5.

Notes and slides

So far, we have discussed the nuts and bolts of analysis of time-to-event data. But in practice, using these tools effectively requires knowledge of study design. This week, we cover some highlights of study design, including overviews of the “target trial” framework and what to do about late entries.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

This week’s exercise is available on the EPID 722 Teams site. Please note that you will need to download the dataset from the data folder. The exercise questions can be found under /Class Materials/Exercises/Exercise 7/Exercise 7.docx. Sample SAS and R code can be found in the same folder.

Optional Reading

This reading presents a nice example of a target trial.

García-Albéniz X, Hsu J, Hernán MA. The value of explicitly emulating a target trial when using real world evidence: an application to colorectal cancer screening. European journal of epidemiology. 2017 Jun;32(6):495-500.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…

Competing events

Thu, 17 Feb 2022 00:00:00 +0000

Readings

Andersen PK, Geskus RB, De Witte T, Putter H. Competing risks in epidemiology: possibilities and pitfalls. International journal of epidemiology. 2012 Jun 1;41(3):861-70.

Slides

The slides for today’s class are available in the class Teams/Sharepoint folder.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

This week’s exercise is available on the EPID 722 Teams site. Please note that you will need to download the dataset from the data folder. The exercise questions can be found under /Class Materials/Exercises/Exercise 6/Exercise 6.docx. Sample SAS and R code can be found in the same folder.

Optional readings

Below are some additional readings about competing events

Cole SR, Lau B, Eron JJ, Brookhart MA, Kitahata MM, Martin JN, Mathews WC, Mugavero MJ. Estimation of the standardized risk difference and ratio in a competing risks framework: application to injection drug use and progression to AIDS after initiation of antiretroviral therapy. American journal of epidemiology. 2015 Feb 15;181(4):238-45.
Edwards JK, Hester LL, Gokhale M, Lesko CR. Methodologic issues when estimating risks in pharmacoepidemiology. Current epidemiology reports. 2016 Dec;3(4):285-96.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…

Generalizability

Thu, 10 Feb 2022 00:00:00 +0000

Readings

Stuart EA, Rhodes A. Generalizing treatment effect estimates from sample to population: A case study in the difficulties of finding sufficient data. Evaluation review. 2017 Aug;41(4):357-88.

Slides

The slides for today’s class are available in the class Teams/Sharepoint folder.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

This week’s exercise is available on the EPID 722 Teams site. Please note that you will need to download the dataset from the data folder. The exercise questions can be found under /Class Materials/Exercises/Exercise 5/Exercise 5.docx. Sample SAS and R code can be found in the same folder.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…

Comparisons of risk

Thu, 03 Feb 2022 00:00:00 +0000

Readings

Sato T, Matsuyama Y. Marginal structural models as a tool for standardization. Epidemiology 2003;14(6):680-6.

Notes

The notes for today’s class are available in the class Teams/Sharepoint folder.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

This week’s exercise is available on the EPID 722 Teams site. Please note that you will need to download the dataset from the data folder. The exercise questions can be found under /Class Materials/Exercises/Exercise 4/Exercise 4.docx. Sample SAS and R code can be found in the same folder.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…

Cox Models

Thu, 27 Jan 2022 00:00:00 +0000

Readings

Bradburn MJ, Clark TG, Love SB, Altman DG. Survival analysis part II: multivariate data analysis–an introduction to concepts and methods. British journal of cancer. 2003 Aug;89(3):431-6

Slides

Download PDF of all slides

Download PDF of Notes

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

This week’s exercise is available on the EPID 722 Teams site. Please note that you will need to download the dataset data.dat. The exercise questions can be found under /Class Materials/Exercises/Exercise 3/Exercise 3.docx. Sample SAS and R code can be found in the same folder.

Optional Readings

Hernán MA. The hazards of hazard ratios. Epidemiology (Cambridge, Mass.). 2010;21(1):13–5.

Questions

Please use the form below to submit your questions about this week’s reading.

NOTE: Please submit your question by Wednesday at 5pm.

Loading…

Nonparametric estimators of risk

Thu, 20 Jan 2022 00:00:00 +0000

Readings

Clark TG, Bradburn MJ, Love SB, Altman DG. Survival analysis part I: basic concepts and first analyses. British journal of cancer. 2003 Jul;89(2):232-8.

Notes

The notes for today’s class are available via Teams.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Notes

As a self check, reflect on how what you learned helps answer the following questions:

What is immortal person time?
What is risk? What is survival?
What is the hazard?
What is a nonparametric estimator and how does it differ from a parametric estimator?
How does the Kaplan-Meier account for late entry and censoring?
What are the pros and cons of allowing late entry?
How can we assess the possible effects of informative drop out?
What is a line diagram and how is it drawn?
Fill in a table like Table 2 in the Cole and Hudgens paper in AIDS given sample data
When is censoring informative?
What do we assume about censoring when using the Kaplan-Meier?
What is the difference between a parameter and an estimator?

Optional Readings:

Cole SR, Hudgens MG. Survival analysis in infectious disease research: describing events in time. AIDS (London, England). 2010 Oct 23;24(16):2423.

Exercise

This week’s exercise is available on the EPID 722 Teams site. The exercise questions can be found under /Class Materials/Exercises/Exercise 2/Exercise 2.docx. Sample SAS and R code can be found in the same folder.

Questions

Please use the form below to submit your questions about this week’s reading.

NOTE: Please submit your question by Wednesday at 5pm.

Loading…

L1: Time in epidemiology: Questions and parameters

Thu, 13 Jan 2022 00:00:00 +0000

Readings

Samet JM. Concepts of time in clinical research. Annals of internal medicine. 2000 Jan 4;132(1):37-44.

Notes

Download PDF of Notes

Below is the code to produce line diagrams for each scenario described at the end of the notes. Please check out the course notes and attempt the scenarios outlined there by hand (using pencil and paper) before looking at the answers posted here.

Say have a (small) cohort study of 5 soldiers selected at random from all people enlisting in the armed services between 2004 and 2014 and followed for mortality up to 10 years.

id	age	enlist	last_info	vital_status
1	24	2008-07-01	2022-01-13	alive
2	18	2008-07-01	2011-01-01	dead
3	21	2011-01-01	2016-06-15	dead
4	35	2011-01-01	2022-01-13	alive
5	19	2013-07-01	2022-01-13	alive

Practice drawing 3 line diagrams:

one with age (starting at age 18) on the \(x\) axis
one with calendar time (starting 1 Jan 2008) on the \(x\) axis
one with time since enlistment on the \(x\) axis.

id <- c(1:5)
age <- c(24, 18, 21, 35, 19)
enlist <- c(as.Date("2008-07-01"),
            as.Date("2008-07-01"),
            as.Date("2011-01-01"),
            as.Date("2011-01-01"),
            as.Date("2013-07-01"))
last_info <- c(Sys.Date(),
               as.Date("2011-01-01"),
               as.Date("2016-06-15"),
               Sys.Date(),
               Sys.Date())
vital_status <- c("alive", "dead", "dead", "alive", "alive")
dat <- tibble(id, age, enlist, last_info, vital_status)

#line diagram 1
dat$t1 <- (last_info - enlist)/365.25 + age
dat$w1 <- age
dat <- dat %>%
  mutate(delta1 = ifelse(vital_status == "dead" & t1<=(age+10), 1, 0))
#library(grid)
library(ggthemr)
ggthemr('flat')
line <- ggplot() +
 # geom_segment(aes(x = 0, y = newid, xend = yearw, yend = newid), lty = "dotted") +
  geom_segment(data = dat %>% filter(delta1 == 0),
               aes(x = w1, y = id, xend = t1, yend = id),
               arrow = arrow(length=unit(0.2, "cm"))) +
  geom_segment(data = dat %>% filter(delta1 == 1),
               aes(x = w1, y = id, xend = t1, yend = id)) +
  ylab("ID") +
  scale_x_continuous(name = "Age" ) +
   geom_point(data = dat %>% filter(delta1 == 1),
              aes(x = t1, y = id),
              color = "red", size = 1.5)

line

#line diagram 2
dat$t2 <- (last_info)
dat$w2 <- enlist
dat <- dat %>%
  mutate(delta2 = ifelse(vital_status == "dead" & t2<=(enlist+10*365.25),
                         1, 0))
line <- ggplot() +
  geom_segment(data = dat %>% filter(delta2 == 0),
               aes(x = w2, y = id, xend = t2, yend = id),
               arrow = arrow(length=unit(0.2, "cm"))) +
  geom_segment(data = dat %>% filter(delta2 == 1),
               aes(x = w2, y = id, xend = t2, yend = id)) +
  ylab("ID") +
  scale_x_date(name = "Calendar Date" ) +
   geom_point(data = dat %>% filter(delta2 == 1),
              aes(x = t2, y = id),
              color = "red", size = 1.5)

line

#line diagram 3
dat$t3 <- (last_info - enlist)/365.25
dat$w3 <- 0
dat <- dat %>%
  mutate(delta3 = ifelse(vital_status == "dead" & t3<=(10), 1, 0))

library(ggthemr)
ggthemr('flat')
line <- ggplot() +
  geom_segment(data = dat %>% filter(delta3 == 0),
               aes(x = w3, y = id, xend = t3, yend = id),
               arrow = arrow(length=unit(0.2, "cm"))) +
  geom_segment(data = dat %>% filter(delta3 == 1),
               aes(x = w3, y = id, xend = t3, yend = id)) +
  ylab("ID") +
  scale_x_continuous(name = "Years since enrollment") +
   geom_point(data = dat %>% filter(delta3 == 1),
              aes(x = t3, y = id),
              color = "red", size = 1.5)

line

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Optional readings

Further reading about time in epidemiologic research:

Cole SR, Hudgens MG, Brookhart MA, Westreich D. Risk Am J Epidemiol. 2015 Feb 15;181(4):246-50

Other resources

Parts of this week’s lecture are based on a talk I gave at SER in 2020.

Exercise

This week’s exercise, along with SAS and R code to get you started, is posted on the Microsoft Teams site for this course.

Questions

Please use the form below to submit your questions about this week’s reading.

Questions are due by Wednesdays at 5pm.

Loading…

L0: Introduction

Tue, 11 Jan 2022 00:00:00 +0000

Readings

The syllabus page for this class

Slides

Below are links to the two sets of slides used this week: an introduction to the course and Lecture 1. You can download all slides by clinking the download link below.

Download PDF of all slides

Notes

Notation

Everyone in this course has the preparation required to master the material presented. There is no advanced math in this class, but you may see notation you have not seen before. Get in the habit of translating all notation you come across in this course into English language sentences and ask questions about notation that is unfamiliar.

Some examples:

\(\sum_{i=1}^nX_i\): Sum \(X_i\) over all units (up to unit \(n\)) indexed by \(i\)

\(\prod_{k=1}^m d_k\): multiply \(d\) over all units (up to unit \(m\)) indexed by \(k\)

What you should know after this week

Each week, we will post a list of skills you should have developed (e.g., drawing line diagrams, defining risk, coding a Cox model). Check the Notes section each week to be sure you are keeping up.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Optional readings

Further reading about time in epidemiologic research

Cole SR, Hudgens MG, Brookhart MA, Westreich D. Risk Am J Epidemiol. 2015 Feb 15;181(4):246-50

Optional reading (for fun) about learning science and math

Hamming RW. Mathematics on a distant planet. The American Mathematical Monthly. 1998 Aug 1;105(7):640-50.

Exercise

There is no exercise today.

Questions

Each week, you will use a form below to submit your questions about the week’s reading. By tomorrow at 12pm, please navigate to the page for Lecture 1 to find the assigned reading and submit your question.

L9: Informative censoring

Wed, 24 Mar 2021 00:00:00 +0000

Readings

Howe CJ, Cole SR, Lau B, Napravnik S, Eron Jr JJ. Selection bias due to loss to follow up in cohort studies. Epidemiology (Cambridge, Mass.). 2016 Jan;27(1):91.

Slides

Coming soon!

Notes

Coming soon.

Videos

Recordings of lectures are available on the EPID 722 Sakai site.

Exercise

Exercise 9 may be found on the course Teams site.

Questions

Please use the form below to submit your questions about this week’s reading.

Loading…