I've been wanting to spruce up my biostats knowledge. Really understanding things like sensitivity, specificity, PPV, confidence intervals, etc. are really important for clinical medicine. And I've never actually taken a stats class.
Unfortunately, I gave myself just this one day to study these topics because (1) I don't think they're really heavily represented on the exam and (2) there are other things to pay attention to that are more important.
Resources: First Aid for Step 1 Behavioral Sciences Chapter, High-Yield Biostatistics and Google search. USMLERx qbank as well.
Method: Mostly just reading and questions. The qbank is my way of solidifying the knowledge I acquire through reading. I made a couple of Anki cards for key things that I want to remember about epidemiology in the future, but that's more for me than the test. As I said in my intro post a couple of days back, this study time is as much about filling the gaps in my knowledge base as it is about prepping for this exam.
Here are some nuggets from notes I took today.
Sensitivity and Specificity
-There is a trade off between specificity and sensitivity. As the one increases, the other usually decreases. In order to have high specificity and high sensitivity, there must be no overlap between the populations of diseased and disease-free people
- High sensitivity = low false negative rate
- High specificity = low false positive rate
- High sensitivity tests are used to rule out - if a test is very sensitive, it will pick up disease more easily. But that also means it have a greater chance of picking up disease free people mistakenly. But if a high sensitivity test is negative, one can be highly certain that the patient does not have the disease
- High specificity tests are used to rule in because if a person is negative with a high Sp test, then it’s very unlikely that they will have that disease;
- Generally, a high sensitivity test is used first for screening, and then a high specificity test is used for weeding out false positives from the initial screening population of positives
- PPV and NPV are dependent on the context, they depend on the prevalence of disease in a population; SN and SP are intrinsic properties of a test.
Also, I added some of the Anki cards I made to my Step 1 Wiki
So far so good. These topics are light. I goofed a lot today though. Being in my house, as much as I like it, can be distracting. I play on the net. I go make food. I do house projects. It's easy to stray. I need to get that under control. But all in all, spirits are high. And I just remind myself every day that I'm not standing in the OR for 8 hours or on the floor rounding forever. It's so good to control my time and do what I want, when I want to. And to be learning lots of information rather than just showing up to school to make an appearance in the clinic.
I think the schedule I've set up will be plenty sufficient, and I think the way I've broken things down, with the USMLERx/content review up front is going to be good too.
Anyway, until tomorrow.