Statistics is something that we learn and forget fairly quickly. In post-graduate exams, we are expected not only to know it, but also to be able to apply it and discuss it with our peers. From this year onwards, in Part 2 FRACGP, a new station called Peer Level Discussion (PLC) is introduced. This may be the reason why so many of us failed this year. At this station, a metaanalysis result is summarised with all the diagrams and figures for you to read and interpret within 3 minutes. After 3 minutes the bell will ring and you walk in to face the 'expert' peer!!!
What do you understand by Relative Risk (RR)?
Well, I know many of us are not so good in statistics, can each of us put in something that you understood from your reading or research. Give examples if possible. By doing so, you are helping each other to understand the topics better. Mind you, the person who teach wholeheartedly is the person that will gain the most. Questions on statistics are also asked in the Single Best Answer of Part 1. Hopefully, among us, there is someone who is an expert in statistics and who is keen to teach us.
What is Odds Ratio (OR)?
What is Hazard Ratio (HR)?
What is Likelihood Ratio?
What is Surrogate Endpoint?
What is Confidence Interval?
What is P value?
What is Type 1 error?
What is Type 2 error?
What is Positive Predictive Value?
What is Negative Predictive Value?
What is Numbers Needed to Treat (NNT)?
6 comments:
Seems that everybody is not familiar with statistics. I've read up some books. Here's some concepts :
Odds ratio: The ratio of two odds, for example the odds of an event for males and the
odds of the same event for females.
Hazard ratio: The reduction of risk of death on treatment compared to control over the
period of follow-up. A type of relative risk.
Likelihood ratio test: A test for comparing two competing models for a data set. The
test statistics is -2 times the difference in the log-likelihoods of the two
models. Under the hypothesis that the two models fit the data equally well, the test
statistic has a chi-squared distribution with degrees of freedom equal to the
difference in the two models.
Surrogate endpoints: A term often encountered in discussions of clinical trials
to refer to an outcome measure that an investigator considers is correlated highly
with an endpoint of interest, but that can be measured at lower expense or at an
earlier time. In some cases, ethical issues may suggest the use of a surrogate.
Examples include measurement of blood pressure as a surrogate for cardiovascular
mortality, measurement of lipid levels as a surrogate for arteriosclerosis, and, in
cancer studies, measurement of time to relapse as a surrogate for total survival
time. Considerable controversy in interpretation can be generated when doubts
arise about the correlation of the surrogate endpoint with the endpoint of interest,
or over whether the surrogate endpoint should be considered as an endpoint of
primary interest in its own right.
Confidence interval: A range of values calculated from the sample observations that is
believed, with a particular probability, to contain the true parameter value. A 95%
confidence interval, for example, implies that if the estimation process was repeated
again and again, then 95% of the calculated intervals would be expected to contain
the true parameter value. Note that the stated probability level refers to properties
of the interval and not to the parameter itself, which is not considered a random
variable
P-value: The probability of the observed data (or data showing a more extreme departure
from the null hypothesis) when the null hypothesis is true.
Type I error: The error that results when the null hypothesis is rejected falsely.
Type II error: The error that results when the null hypothesis is accepted falsely.
Positive predictive value: The probability that a person having a positive result on a
diagnostic test for a particular condition actually has the condition. For example, in
a study of a screening tool for alcoholism, the positive predictive value was
estimated to be 0.85. Consequently, 15% of patients diagnosed by the test as
suffering from alcoholism will be misclassified.
Negative predictive value: The probability that a person having a negative result on a
diagnostic test for a particular condition does not have the condition. For example,
in a study of a screening tool for alcoholism, the negative predicted value was
estimated to be 0.93. Consequently, 7% of patients with a negative result are
actually likely to be alcoholic and so misclassified by the test.
Number needed to treat (NNT): A measure of the impact of a treatment or
intervention that is often used to communicate results to patients, clinicians, the
public and policymakers. It states how many patients need to be treated in order to
prevent an event that would occur otherwise (e.g. a death). Calculated as the inverse
of the absolute risk reduction. An NNT can help in making a decision
between treatment groups and in making choices for an individual patient. For
example, in a study into the effectiveness of intensive diabetes therapy on the
development and progression of neuropathy, 9.6% of patients randomized to usual
care and 2.8% patients randomized to intensive therapy suffered from neuropathy,
leading to an estimate of NNT of 1/6.4%, which rounded up gives the value 15. This
means that 15 diabetic patients need to be treated with intensive therapy to prevent
one patient from developing neuropathy. The concept of NNT can equally well be
applied to harmful outcomes as well as beneficial ones, when instead it becomes
number needed to harm (NNH).
It needs time to digest all these terms. I suggest we learn it systematically. I'm looking for a book "medical statistics made easy". Anyone can contribute?
Thank you sc for your contribution. The ability to say a few words about each of the statistical terms is very important especially in Part 2. I suggest all readers print out this summary and file it up for last minute reference.
There are a lot of information about statistic at Wikipedia, just login and read. But what sc provided is very concise and easier to remember than the lengthy description at Wikipedia.
Post a Comment