Welcome to (probably) the best medical school in the world. Now, let’s get on with what really matters.
“There is absolutely no difference between statistics and gambling” says Prof. John Orav. Just maybe a little bit or money…. and gambling is more addictive I believe.
Statistics class has moved from the dryness of a fine white wine to that of a biscuit barrel (assuming there’s a log regression for that).
However, one cannot criticise the teaching dedication, availability or detail. EVERYTHING is worked through step by step, no assumptions, no brushing over. I think that’s where much statistic teaching leaves people only half way. Not here though with two hours of statistics everyday. For such a clever guy he is really happy to spell it out.
“…and so why is it important to understand this derivation of the F-distribution in testing if (sigma)x = (sigma)y ? ….. it’s not….. just make the right decision regarding the degrees of freedom and let SAS do the rest.” Hey, still nice to know.
For those who couldn’t give a monkeys but would like to sound they do – here’s today’s statistic tips for continuous data:
a. Validity before Power. That applies to politicians too.
b. Normality is a relative term. Don’t assume anyone or anything is normal. When you run Shapiro-Wilk you will find nothing is normal.
c. Central Limit Theorem deals with everything if n is large anyway. Wahey. (If n>200 the T distribution will be essentially Z. Yeah, 200 is a great number. I love 200 man).
d. N is small? Wilcoxon gets you home and dry for a two way analysis (okay, you may have dropped a bit of power on the way but you are home and dry)
They take research seriously here and the publications prove it. No less on the teaching.
Many of the lectures are observed by the other course directors and the major modules have ‘office hours’ every day – open access to discuss with the lecturers any thing you want plus individual 1:1 slots bookable online with practically everyone on faculty so you can get someone to pull your project apart before actually standing up. Granted, some of the associate professors look like they are 18 but a review of their publications soon dispels this. I kind of like it how they often use their own publications to prove the concepts – New England Journal of Medicine…..of course.
The Program for Clinical Effectiveness is essentially the fundamentals of f 1/3 of the Masters in Public Health. You want to do research out here? You need an MPH or PhD or you won’t get a look in. MPH is not what I thought back home. That probably explains why this building is 10 floors at the heart of the medical school and not some poxy department no one has ever seen before.
This is where you learn the tools for clinical research. Lots of the faculty are profs at Brigham and Womens, Mass General and Beth Israel. No surprise then that 75% of the guys on this course are from the surrounding Harvard hospitals about to start or complete their specialty fellowship. A lot have funding from their department to complete the MPH. Anyone got a bit of spare wonga and I’ll be more than happy to finish the remaining modules.
A common theme. “So what do you do then?” “Well I’m a Paediatric Neurological Oncological Pharmacologist. What do you do?” “Oh, I’m a GP.” The only GP in this entire building as far as I can see.
Specialists rule here but people are beginning to realise the shortage of GPs means the ship is starting to keel. Hey, one more GP and the risk ratio will be (just about) halved.
What is helpful is the assignments are really applied and you get to find out what everyone else is doing. In fact the idea is after you have got your grant proposal assignment sorted, you get the grant, do the research and publish it. A lot of it’s pretty niche specialised stuff but the methodology and analysis applies across the board.
Next class, measuring patient centred health outcomes. If you don’t know the difference between an SF36 and EQ5D then go and grab a cup of tea.