Teaching Medical Students, Interns, and Residents to read Medical Literature: The STATS Method

1
Teaching Medical Students, Interns, and Residents
to read Medical Literature The STATS Method
2
Session Objectives
1) Examine the current trends in physician use of
medical literature.
2) Explain why teaching medical students,
interns, and residents how to read medical
literature is important.
3) Explore the STATS Method to research
evaluation.
3
Current Trends in Physician use of Medical
Literature

• The Good, the Bad, and the Lack of Probability

4
The Good
5
Growth in Clinical Trials
According to the Centre for Evidence Based
Medicine, there are now around 1 million clinical
trials that have been completed.
6
Trial Classification(NIH Standards)
Phase I Trials

• 20-80 Participants
• Examine drug/therapy safety, dosage
  recommendations, and possible side effects

7
Phase II Trials

• 100-300 Participants
• Purely a Safety Evaluation

Phase III Trials

• 1,000-3,000 Participants
• Confirms effectiveness, monitors side effects,
  compares to other treatments, collects info to
  allow for safe usage of drugs/treatments

8
Phase IV Trials

• Post Marketing Examine
• Collects additional information on risks,
  benefits and optimal use

9
Growth in Medical Research
In 2002 alone, there were 522,943 articles
published in some 4,600 medical journals
catalogued by PubMed.
There is an abundance of medical research being
conducted!
10
The Bad
11
Last 5 Years

• 1999 473,044
• 2000 512,226
• 2001 518,741
• 2002 522,943
• 2003 100,001
• As of March 24, 2003

12
Thats a total of 2,126,955 Articles in just the
last 5 years
13
According to the PubMed Website1, they have over
12 million citations dating back to the
1960s. So the last five years accounts for 18
of all of those citations.
14
To stay up-to-date with the current medical
literature, you would have to read about 6,000
research articles a day.
15
(No Transcript)
16
Medical Knowledge with Years Since Graduation3
17
The Probability
18
Statistics is simply a matter of mathematical
probability.
Most modern medical research is based in
statistics.
19
It is estimated that nearly 50 of all medical
research articles contain serious methodological
and/or statistical errors4.
20
Why teaching medical students, interns, and
residents about research is so important
21
Paradigm Shift5
E
Evidence
B
Based
M
Medicine
22
EBM Defined
23
Traditional vs. EBM Training
Sackett et al. (1977) Study on EBM and
Hypertension Treatment
24
EBM in the USA
In the early 1990s, JAMA established the
Evidence-Based Medicine Working Group5.
This group, released a series of articles about
reading medical literature throughout the 1990s
and early 2000s7.
25
The STATS Method
26
S
Sample
T
Treatment
A
Argument
T
Testing
S
Significance
27
SAMPLE
28
Questions to ask about the sample.
1. How were participants selected for inclusion
in the study?
2. How were they approached for inclusion?
3. Do study participants have special
characteristics?
29
Questions to ask about the sample.
4. Do the demographics of the study bias the
study in any direction?
5. Does the location of the participants bias
the study in any direction?
6. What could you do as a researcher to prevent
sampling bias?
30
TREATMENT
31
Treatment Defined
Treatment refers to those variables actively
manipulated or measured in a scientific study.
32
Variables
Any entity that can take on different values.
33
Dependent Variables
The variable that is effected or not effected by
another variable in a research study.

34
Independent Variables
The variable that is being manipulated or
examined in a study to see if it effects the
dependent variable.
35
Ultimately We want Causation
IV
DV
We want the IV to cause a difference in the DV.
36
FOR EXAMPLE
Bioloical Sex
Physical Aggression
Is there a difference between females and males
and their use of physical aggression?
The treatment is how we go about measuring and
explaining this process.
37
Questions to ask about the treatment.
1. What is the primary dependent variable studied
in the article?
2. How did the researchers go about measuring
this variable?
3. Did the researchers attempt to manipulate the
target variable? If so, how?
38
ARGUMENT
39
Logical argumentation should be the basis for all
scientific research.
40
A logical argument in science is called a
hypothesis.
41
Questions to ask about the arguments.
1. Was the argument for the study made clear in
the article?
2. Is the articles argument supported by current
medical research and practice?
3. Is the argument in the article based on
science or conjecture?
42
TESTNG
43
Testing is the process a researcher goes through
to either prove or not prove a hypothesis.
44
Two Types of Tests
1. Difference Tests
2. Relationship Tests
45
Difference Tests
A difference test attempts to see if there is a
difference between the means of two or more
groups in relation to one or more dependent
variable(s).
46
FOR EXAMPLE
Bioloical Sex
Physical Aggression
In this case, we wanted to see if there was a
difference in the mean scores of female and male
physical aggression.
47
Two Types of Difference Tests

• Non-Parametric Tests Involve Nominal and
  Ordinal Variables
• (x2 or rank order test)

2. Parametric Tests Involve interval or ratio
variables with each other or with nominal and
ordinal variables (t-tests or ANOVAs)
48
Relationship Tests
A relationship test attempts to see if two (or
more) variables can be plotted in a linear
fashion.
49
Relationships
50
Two Types of Relationship Tests

• Symmetric Unable to designate cause and effect
• (Correlation)

2. Asymmetric Existence of explanatory
variables or predictors (Regression)
51
BIG NOTE
Correlation does not mean causation
However, Causation does mean Correlation
52
Significance
53
What is the actual significance of this study?
If research doesnt say anything new, why should
we care?
54
Activity
55
Concluding Thoughts
1. EBM is a new direction for medical
practitioners.
2. With the abundance of research available we
have to be more cynical.
3. As people involved in post-graduate medical
education, we need to teach critical analysis of
research.
56
Concluding Thoughts cont
4. Ask questions about research based on the
STATS method.
5. Learn more about how research is actually
conducted. Dont be a sucker!!!
6. One study does not equal science.
57
Prepared by Jason S. Wrench, Ed. D. Medical
Educational Specialist West Virginia School of
Osteopathic Medicine