Disease Detectives B/C

[GrayEpi]

I checked out the Wiki. Those who put it together did a really good job. I did make a few additions and changes that will hopefully help but I did not see anything that was clearly wrong. For those of you trying to understand what this is all about, hang in there. I really suggest the Principles of Epidemiology 3rd edition (available free online). It is a self-study course - has examples and questions. With respect to the pure epi stuff, it has everything that Division B should know and more. It should at least enable Division C teams to play in the league although there it does not include bias, adjusting, stratified analysis or a few of the other newer concepts.

If you are just starting DD this year, don't feel bad. I did not understand it very much the first time I was exposed either (though I did "ace" the class). The second time it began to make sense, the third time the lights went on and they keep getting brighter. This stuff really works. I have investigated a lot of outbreaks over the years using the same methods in DD and have been amazed at how well it works. This is not a "learn the facts" type of event (certainly not at the national level) but more of a develop a way to look at things and solve problems. You need to know the facts and formulas - but that is what the cheat sheets are for. The thing that I find so neat is how you folks have risen to the challenge over the years. This will be my 16th national competition with DD and y'all keep on stepping up and knocking our socks off.

I look forward to seeing some of you in Orlando this May. Please do say "hi".

[GrayEpi]

I have been asked to explain the difference between the T-test, the paired T-test and Fisher's exact test. The first two deal with continuous data - things like weight, height, time (e.g. incubation periods). The latter is used for categorical data - things like sick, well, exposed, gender...

Lets say you are investigating an outbreak of salmonella and you have 50 cases from folks who attended the same wedding reception. You know when the reception took place and when each person became ill and are comfortable that you have a good measure of the incubation period (IP)for each. You plot the data and get an epi curve that appears to have two peaks. You hypothesis that those folks who took antacids or were on meds that lowered the pH of their stomachs were more susceptible, got sick sooner and had shorter incubation periods than those who did not take those meds. You contact each of your cases and ask if they took the meds. You calculate the mean IPs for each and there turns out to be a difference but is it statistically significant? Your null hypothesis is that there is no difference between the mean IPs of each of the two groups. You use a T-test to test that hypothesis.

Lets say you think that handedness affects the number of bacteria on one's right or left hand - right handed people tend to have more bugs on their right hand than on their left and the reverse for left-handed people. So you recruit a bunch of folks who are right handed and left handed and measure the number of bacteria on each of their hands. Here each person represents a distinct observation - some folks may be dirtier than others and what affects one hand probably affects the other - the count on the right hand and left hand are not independent. Here the null hypothesis is that the mean of the difference in number of bacteria on the dominant (d) and nondominant (nd) hands is 0 or (mean (d - nd)) = 0. You use the paired t-test to test that hypothesis. The lack of independence between your two observations is what makes the difference.

Fishers exact test is similar to the Chi square test in that you are interested in significant differences in rates. The Chi square test requires that each cell have a certain number of observations in each cell in order to be valid. If you have a small number of observations (e.g. a small party) or if they are clearly slanted (e.g., almost no one ate the liver) - you are likely to wind up with one or more cells in your 2x2 table that have at least 5 observations and the results of your Chi square test may be off. That is when you use the Fishers exact test.

Hope this is helpful.

[BipolarEconomist]

@GrayEpi:

Let me tell you just how delighted and grateful that you are here.
When I joined Science Olympiad at my school, I immediately had the desire to reach a Nationals level of competition.
After signing up for this event, I spent a week scouring the SciOly website PowerPoint, memorizing terms, memorizing the CD, memorizing the statistics information, etc.
For the past few months, I did a bunch of past tests from the Test Exchange. In December and January, I started checking out books from my library on Epidemiology.

However, I still need help. I took your recommendation of the Principles on Epidemiology 3rd edition and am currently going through that right now.
When I take practice tests, I notice that many times there are questions that I cannot answer.

I've seen Ward Melville school take 1st place in so many events at the YUSO (Yale Invitational) this January, and I just don't know how they are able to do this.
Do they devote a whole class/course period to studying? If so, can anyone let us in on their curriculum/study methods?

GrayEpi, what are your suggestions for this event for someone who has already done the "basic training" and wish to take some more rigorous levels of studying to succeed?
Is there a list somewhere of basic epidemiology facts that I should memorize? I know that there were questions on the YUSO about number of people with a specific disease or epidemic, or other logistical information on outbreaks in history.
Furthermore, there was a fill in the blank part which gave lists of diseases and asked you to categorize them into pathogens, viruses, prions, bacteria, parasites, etc- where do you recommend I look for to study these diseases? I know some of them were more common, but I was incredibly depressed to see that a few were ones that I had never even heard of, and had no clue what to answer.

If anyone has successfully competed in Disease Detectives C, mind sharing with me your "advanced" studying methods? As of right now, I'm sort of treating this like SATs, just memorizing everything from sight on and cramming practice tests. I've had this huge desire to place a medal, and I'm so worried that I'm just not studying enough or correctly.
Could any seniors/experienced Science Olympians give me some advice? GrayEpi, I would love to hear some more from you!

Thanks, and again, incredibly humbled and grateful for this forum

[narwhals360]

I'm brand new to this event, could someone explain to me what we have to know regarding environmental quality? Is it just the theme the test revolves around or do we actually have to know about diseases associated with it? Also, how are my partner and i supposed to split up the work?

[navigator]

I'm brand new to this event, could someone explain to me what we have to know regarding environmental quality? Is it just the theme the test revolves around or do we actually have to know about diseases associated with it? Also, how are my partner and i supposed to split up the work?

Disease Detectives has been around for as long as I can remember, and it's kept the same underlying format, with case studies asking you to answer questions based on charts, graphs, or written information. In general, I would focus more on learning the new statistical methods this year than memorizing or finding diseases related to the environment, since at both the Invites and Regionals my school has competed in so far, the exam was heavily concentrated on statistics.

It's definitely important to know several key facts about the environment such as examples of environmental diseases or types of pollutants, but remember that (to answer your second question) environmental quality is the theme the test revolves around, not the core or focus. It's best to learn the newly introduced topic, because statistics will definitely be sticking around in this event for a while, if not forever.

To answer your last question, to make the most of a partnership, you and your partner should both know the basic facts and skillsets required (including the statistical methods) so that, in the case that you divide the physical test on competition day, both of you can hold your weight. It's never a good feeling when you or your partner have to do all the work because the test is concentrated on a subject which only one of you focused on.
It's really up to you and your partner on how you split up the rest of the information. A suggestion is to allocate diseases (albeit caused by the environment) to one person, while another person does types of bias, types of errors, etc.
Dunno if that helped... but yep, that's just my two cents.

[GrayEpi]

Bipolar economist - The WHO has a good epi manual that has more on sources of error than the CDC Principles of Epi. Not sure about the title but it should be included in the DD reference material. Another place to go is the Epi Supercourse. There are tons of lessons and materials there. Some are good - some are bad. Many are irrelevant to DD but a few are pretty good. There are a couple of errors that are pretty good. You may want to look into ActivEpi by David Kleinbaum. David is a biometrician and has a real passion for education and making things easy to understand. He has been at this a long time (I used one of his books when I was a student). I just got his CD and book - have not had the time to really dig into either but I expect they are excellent. There really are two levels of epi - applied epi - like that done at CDC and most health departments and research epi - like that done it universities. (CDC also does research but most outbreak investigations are pretty applied). Disease Detectives sort of represents the applied side although the sources of error and stratified analysis have a bit of research. Modeling, multivariable analysis where you look at the association with one variable while controlling for 3-4 others and more complex statistics where you address the lack of independence are all pretty heavy research types of things and not something we are likely to have in Disease Detectives anytime soon (Supercourse has a lot of this stuff).

How does the question bank work? I may try to put a few up there but don't want to horn in on any one else's turf. The organizers, managers and contributors for this site have done very well without my 2 cents.

Developing a good event takes a lot of time as does scoring. I know we supervise a regional middle school event as well as the nationals. Regional events (including ours) are often scored by undergrad volunteers. We have to make sure that most questions can be scored by someone who may not be a seasoned epidemiologist. That means multiple choice, fill-in-the blank, short answer or simple calculations. It is tough to fit those types of questions to an event that emphasizes critical thinking and the ability to analyze and interpret data. The nationals are a bit different. We've had some really great local volunteers work with us - the team last year had 4 MDs, a PhD and 2-3 MPHs. Even there we have to be careful about the types of questions we ask. We usually start working on the events in early January - sometimes earlier.

[GrayEpi]

Navigator is right about statistics though we are hoping to start out slow at first. I believe national posted some guidance on this a few weeks ago. We don't want to give a problem that will take 30 of the 50 minutes in this event to solve. Statistics is an important part of epidemiology and epidemiology is a great example of how statistics is used to solve real world problems. We understand this is not the math olympics and have no desire to make it such.

[penelope]

I'm in tenth grade and competing in C for the first time, although I've been doing DD since I joined up in fifth grade. I feel out of my depth in DD for the first time in my life. The C Div tests require a lot of knowledge that I don't have and don't know where to look for. Can somebody recommend some good resources for updating my knowledge to C Div level? I don't know anything about some of the more sophisticated statistical calculations such as Fisher's Exact Test, I don't know anything about lab criteria like whether Staphylococcus is Gram-negative or -positive, and I have no idea whether the outbreak in Springfield ten years ago was E. coli or listeria. Help!

[colorado mtn science]

I'm in tenth grade and competing in C for the first time, although I've been doing DD since I joined up in fifth grade. I feel out of my depth in DD for the first time in my life. The C Div tests require a lot of knowledge that I don't have and don't know where to look for. Can somebody recommend some good resources for updating my knowledge to C Div level? I don't know anything about some of the more sophisticated statistical calculations such as Fisher's Exact Test, I don't know anything about lab criteria like whether Staphylococcus is Gram-negative or -positive, and I have no idea whether the outbreak in Springfield ten years ago was E. coli or listeria. Help!

Even the nation's best in this event will inevitably run into something they don't know. The problem with this subject is that there is new information every day. Your success in this event is ultimately determined by the test writers' mood while writing the test. You could win one gold in DD and lose the very next test. My point? It's fine not to know everything in epidemiology (but still try your best to) because no one can. Just study hard and cross your fingers

[Infinity Flat]

I'm in tenth grade and competing in C for the first time, although I've been doing DD since I joined up in fifth grade. I feel out of my depth in DD for the first time in my life. The C Div tests require a lot of knowledge that I don't have and don't know where to look for. Can somebody recommend some good resources for updating my knowledge to C Div level? I don't know anything about some of the more sophisticated statistical calculations such as Fisher's Exact Test, I don't know anything about lab criteria like whether Staphylococcus is Gram-negative or -positive, and I have no idea whether the outbreak in Springfield ten years ago was E. coli or listeria. Help!

For the statistic part go through the guide uploaded on the soinc national site.

Minor historical events and very disease-specific things I feel shouldn't really be a large part of this event, but I've decided to put them in my notes anyway because test writers can be lazy / don't understand the event. I'd recommend making a table including basic information on a lot of the common diseases, namely symptoms, mode of transmission, and type (bacteria/virus etc.). When the topic was foodborne I also included some more in depth info on each of those such as gram+/-, method of treatment etc. but since that's not the focus now you (hopefully) shouldn't have to worry about it. However, I would recommend being familiar with lab practice in general, e.g. what the gram test is etc. Checking the validity of the diagnoses is an important part of any outbreak investigation, and often times the cases are sorted based on the results of laboratory tests into categories such as "suspected", "probable", and "confirmed." Understanding some of the basic principles going on here can only help you. I can't think of anything offhand that does a good job going into this, but if I find something I'll post it here.