Where to start with BioStatistics?

[Patassa]

I finished up my post-bacc last May and will be matriculating in August. I'm currently a Mechanical Engineer so I can speak some math, although it's been a while since I was fluent. I see some utility for knowing Biostatistic in my future but I'm not sure where I should start.

Stats and Prob. through Khan academy to get a basic understanding?
Jump right into a Biostats course through Hopkins School of Public Health, open course ware?
Some novel book off of Amazon geared specifically for health professionals?

Any one else tackle this subject solo in their free time and have any helpful approaches?

I'm trying to avoid wasting time as August is quickly approaching and I'd like to get this under my belt before I start.

Thanks

---

Members don't see this ad.

 

[GUH]

Basic biostatistics is pretty simple stuff and accounts for about 0.05% of medical school material. If you are interested in biostats just because, then take a class if you can. Learning from a textbook without any support can be boring and confusing, and some biostats concepts are important to understand correctly if you are going to use them.

 

[Kahreek]

Definetely get the basics of stats and probs on khan academy if you dont have those.
Getting to know how to work with a programs like PASW/SPSS or other statistical program is the bulk of the actual work.

 

[cs24]

I finished up my post-bacc last May and will be matriculating in August. I'm currently a Mechanical Engineer so I can speak some math, although it's been a while since I was fluent. I see some utility for knowing Biostatistic in my future but I'm not sure where I should start.

Stats and Prob. through Khan academy to get a basic understanding?
Jump right into a Biostats course through Hopkins School of Public Health, open course ware?
Some novel book off of Amazon geared specifically for health professionals?

Any one else tackle this subject solo in their free time and have any helpful approaches?

I'm trying to avoid wasting time as August is quickly approaching and I'd like to get this under my belt before I start.

Thanks

I took a great biostats class at the end of undergrad and our textbook was:

intuitive biostatistics by harvey motulsky, 2nd ed (make sure you look for this as the 3rd ed. is much more expensive)

It was incredibly well written and is advertised as a nonmathematical guide to statistical thinking geared towards health professionals. Because computers do all the math now, it focuses on how to interpret statistical tests and how to pick the proper test based on the data you have. You could read through it by yourself, although the class certainly helped. If you want to learn it on your own, my best advice is to get an idea of what basic types of data you can have (continuous vs categorical variables), and then learn how to select tests to apply to this data. I'm kind of rambling here and need to get back to studying, but I really can't recommend this book enough. I was able to jump right into performing my own clinical research statistics as early as October of my first year thanks to this book.

 

[Patassa]

I took a great biostats class at the end of undergrad and our textbook was:

intuitive biostatistics by harvey motulsky, 2nd ed (make sure you look for this as the 3rd ed. is much more expensive)

It was incredibly well written and is advertised as a nonmathematical guide to statistical thinking geared towards health professionals. Because computers do all the math now, it focuses on how to interpret statistical tests and how to pick the proper test based on the data you have. You could read through it by yourself, although the class certainly helped. If you want to learn it on your own, my best advice is to get an idea of what basic types of data you can have (continuous vs categorical variables), and then learn how to select tests to apply to this data. I'm kind of rambling here and need to get back to studying, but I really can't recommend this book enough. I was able to jump right into performing my own clinical research statistics as early as October of my first year thanks to this book.

Now we're takin, thank you.

 

[yesdee]

Take this one from coursera: https://www.coursera.org/course/casebasedbiostat

 

[Amygdarya]

A couple of online courses that I liked:
Stanford Statistics in Medicine (very well taught and really relevant to medicine/clinical research): http://online.stanford.edu/course/statistics-medicine (the course is over but I believe you can access the videos if you register)
Quantitative Methods for Clinical and Public Health Research (biostats+epidemiology for medicine and clinical research - an awesome course!): https://www.edx.org/course/harvardx/harvardx-ph207x-health-numbers-354 (the same, the class is over but you should be able to access the videos)
There are also some biostats courses on Coursera, but I don't think they're as rigorous and as well taught as the 2 above.

 

[cbrons]

Basic biostatistics is pretty simple stuff and accounts for about 0.05% of medical school material. If you are interested in biostats just because, then take a class if you can. Learning from a textbook without any support can be boring and confusing, and some biostats concepts are important to understand correctly if you are going to use them.

I wouldn't say .05%

Maybe like 0.25% So yeah it is pretty low yield. Our stats course was literally 8 classes

I still think its pretty important though in general to have a good background in this so you can adequately and efficiently judge the literature

 

[sazerac]

(1) Read the couple pages of biostats in First Aid.
(2) Do the 50 sample biostat questions in a QBank like RX.
(3) move on to something more important

If you can correctly draw a 2x2 square you will be in the 90th percentile in biostats.

 

[Patassa]

(1) Read the couple pages of biostats in First Aid.
(2) Do the 50 sample biostat questions in a QBank like RX.
(3) move on to something more important

If you can correctly draw a 2x2 square you will be in the 90th percentile in biostats.

Do you think this is adequate to scrutinize the analysis you'll find in the literature? Is that not a real issue for practicing MDs?

 

[evilbooyaa]

Do you think this is adequate to scrutinize the analysis you'll find in the literature? Is that not a real issue for practicing MDs?

The issue with literature is generally not the statistical analysis, but rather the lack of controls, unbalanced treatment groups, or general confounding (things the study doesn't even consider). Most times the statistical analysis, if shown at all, is in supplemental tables (while the real figures are graphs that just have p-values on them) which most people don't really look at.


For medical school, for the love of god, do not pre-study biostats, especially if you have any math background at all. It's a joke how much stuff you need to know. If you can learn PPV/NPV and the difference between sensitivity and specificity (and can draw a 2x2 square as written above) you will do fine in biostats.

 

[Patassa]

The issue with literature is generally not the statistical analysis, but rather the lack of controls, unbalanced treatment groups, or general confounding (things the study doesn't even consider). Most times the statistical analysis, if shown at all, is in supplemental tables (while the real figures are graphs that just have p-values on them) which most people don't really look at.


For medical school, for the love of god, do not pre-study biostats, especially if you have any math background at all. It's a joke how much stuff you need to know. If you can learn PPV/NPV and the difference between sensitivity and specificity (and can draw a 2x2 square as written above) you will do fine in biostats.

Thanks for that info.

I wasn't thinking of it in terms of prestudy for Med School, more in terms of learn before I start seeing trial studies for various reasons.

 

[mcloaf]

I usually start biostatistics with a drink.

 

[notbobtrustme]

Thanks for that info.

I wasn't thinking of it in terms of prestudy for Med School, more in terms of learn before I start seeing trial studies for various reasons.

protip: no one reads those, they just read the abstract and if they are feeling ambitious, the conclusions.

 

[Psai]

protip: no one reads those, they just read the abstract and if they are feeling ambitious, the conclusions.

I know what you're saying but there's value in being able to read between the lines. Seeing who funded the trial, looking at where everyone's from, etc. because you can massage data to say almost anything depending on your interpretation.

 

[Cytarabine]

The issue with literature is generally not the statistical analysis, but rather the lack of controls, unbalanced treatment groups, or general confounding (things the study doesn't even consider). Most times the statistical analysis, if shown at all, is in supplemental tables (while the real figures are graphs that just have p-values on them) which most people don't really look at.

Methodology and data analysis can and are certainly both used to influence a study's results, but the depth of knowledge needed to rigorously analyze the methodology is MUCH easier to obtain than that needed to analyze the data analysis choices. For that reason alone, I'd say focus your efforts on basic biostats + more in depth knowledge of research methodology, and leave it to the people at Cochrane review to analyze the stats