Chapter 11.4
1. Ambiguous dropouts
a. 2 phenomena:
i. Non adherence: study participants who no longer comply with the
intervention but who are willing to continue with follow-up for the
outcome events of interest
ii. Loss to follow up: study participants
Chapter 11.3
1. RCT stopped early for benefit play a prominent role in the medical literature
1. RCT can be stopped early because of the perceived harm from the intervention, or
they lose hope in finding positive results, or the sponsor wishes to save mon
Chapter 11.2
1. Major basic science and preclinical promises for effective interventions disappoint in clinical
trials
a. Randomized trials are usually unavailable for interventions that need to be applied
for specialized decisions after some major first
Chapter 13.4 surrogate end points vs target end points
1. What is a surrogate outcome?
a. Surrogate end points= outcomes that substitute for direct measures of how a patient
feels, functions, or survives
b. Ex: HbA1c as a surrogate end point for outcome r
Chapter 4
1. Three ways to use the medical literature
a. Staying alert to important new evidence
i. Browsing medical literature by skimming the table of contents and reading
relevant articles is inefficient and fails to meet the critical appraisal criteri
Chapter 6
1. Research studies attempt to estimate the underlying truth by sampling
a. However, due to flaws in study design systematic error (bias) as well as random error,
we will never know the exact truth
2. Random error
a. Ex: RCT finds that 10 of 100
Chapter 3
1. What is evidence?
a. Evidence= grounds for belief, it provides support for a belief
b. It is not tied to any specific theory of knowledge
c. Whether systematically or unsystematically collected
d.
2. empirical evidence vs theory
a. the role o
Chapter 2
1. Clinical epidemiology is the critical appraisal and application of the evidence to decisions
about patients (individuals level) and populations (policy-level)
2. EBM involves thoroughly working with patients to help them resolve or cope with
Chapter 11.1
1. Large sample size does not make a study more valid or reduce its risk of bias. Small sample
size does not produce bias, but it can increase the likelihood of a misleading result thru
random error.
2. an error can be random (random error wi
Chapter 12.1
1. The hypothesis-testing approach to statistical exploration is to begin with a null hypothesis
and try to disprove that hypothesis.
a. Null hypothesis: there is no difference between the interventions being compared
2. The role of chance
a.
Assessment
of Interaction
Zou, GY
From
Statistical
Models to
Biologic
Interactions
Statistical
Inference
for
Biologic
Interaction
MOVER,
Method
Of
Variance
Estimates
Recovery
Examples
Evaluation
Discussion
Statistical Assessment of
Biologic Interaction
G
E9660A
GY Zou
Baron-Kenny
framework
Limitation of
Baron-Kenny
Causal mediation
analysis
Estimation by
prediction
(Standardization)
Estimation by inverse
prob weighting
(MSM)
Estimation by model
tting (SAS macro)
E9660A: Advanced Statistical Methods for
Ep
E9660A
GY Zou
AN EXAMPLE
INTRODUCTION
PRELIMINARIES FOR
CAUSAL GRAPHS
GRAPHIC MODELS
GRAPHICAL
REPRESENTATION OF
BIAS AND ITS
CONTROL
E9660A: Advanced Statistical Methods
for Epidemiology
Causal Diagrams
SOME APPLICATIONS
Conventional confounding
Adjustin
In 1976, it was recognized that [a]s more risk factors
become established as probable causes in the elaboration of
disease etiology, scientists will turn their attention increasingly
to the question of interaction (synergy or antagonism)
of the causes (1,
Example 3: exaggerated interaction using ORs in
a cohort study
This data set arose from a cohort study in which it was of
interest to investigate the effect of age and body mass index
(weight (kg)/height (cm)2) on diastolic blood pressure (17).
To form a
Example 2: falsely claimed interaction resulting from
the SA method
Consider a data set arising from a case-cohort design in
the Atherosclerosis Risk in Communities (ARIC) Study (21,
22), where it is of interest to determine the interaction between
a susc
SIMULATION STUDY
Despite the justification provided in the Appendix, the
proposed procedure for measures of interaction is based
on asymptotic theory. Simulation studies were therefore undertaken
to evaluate its performance.
For AP, a method based on ln(1
Interaction of the causes
`As more risk factors become established as probable causes in the
elaboration of disease etiology, scientists will turn their attention
increasingly to the question of interaction of the causes '
(Rothman 1976 Am J Epidemiol 103
Epistasis: interaction in genetics
Bateson in 1909 used `epistasis' to describe cases where a variant or
allele at one locus prevents the variant at another locus from manifesting
its e_ect.
Biologic Interaction-Notation
_ Y -outcome, A-factor 1, B-factor
Validity: precision and bias
Different fields in epidemiology have different levels of validity. One way to
assess the validity of findings is the ratio of false-positives (claimed effects that are
not correct) to false-negatives (studies which fail to su
Case series
Case-series may refer to the qualititative study of the experience of a single patient,
or small group of patients with a similar diagnosis, or to a statistical technique
comparing periods during which patients are exposed to some factor with
Modern era
Dr. John Snow is famous for his investigations into the causes of the 19th century
cholera epidemics, and is also known as the father of (modern) epidemiology. He
began with noticing the significantly higher death rates in two areas supplied by
SAS codes the modi_ed Poisson from
UCLA website
`Zou ([2]) suggests using a \modi_ed Poisson" approach to estimate the
relative risk and con_dence intervals by using robust error variances. Using
a
Poisson model without robust error variances will result