We can’t wait for our #EBM chat w/ @MedStudentChat on Thurs, 3/19, at 9 pm EST. Here are our updated objectives, tips to participate in #PHMFellowJC & a link to a fun article on a RCT looking at the association b/w eating ice cream quickly and headache! https://t.co/DzJEJgz9SH

#PHMFellowJC #MedStudentChat #MedStudentTwitter T7: D) @DrJenChen4kids already talked about generalizability. Middle schoolers do not represent whole population No blinding: not sure how you would have done that for this study That said, amazing thiswas done by an 8th grader!!

#PHMFellowJC #MedStudentChat #MedStudentTwitter T7: C) For NNH, it’s the number of people who need to be exposed for there to be a person to have the negative outcome, who would not have had the outcome if not exposed. Related term: number needed to treat

#PHMFellowJC #MedStudentChat #MedStudentTwitter T7: B) For RR, if I have a patient who was exposed to something (eg eating 🍨 quickly), it is the increased (or decreased) chance of having an outcome (headache) compared to the unexposed person (eating ice cream slowly)

#PHMFellowJC #MedStudentChat #MedStudentTwitter T7: A) continued: The other way people look at p < 0.05 is that there’s a less than 5% chance we mistakenly rejected the null hypothesis. The null hypothesis usually says no difference

#PHMFellowJC #MedStudentChat #MedStudentTwitter T7: A)p<0.05 is threshold researchers have agreed upon for statistical significance. It means there’s less than 5% chance that the differences aren’t true (Type I error). But statist. significant doesn’t mean clinical significant

Not only does this article have great teaching points but is short and fun to read, the first author was in 8th grade when she wrote it! And she still claims to love ice cream! https://t.co/o3fKRNS9lK #PHMFellowJC #MedStudentChat

#PHMFellowJC #MedStudentChat

#PHMfellowJC #MedstudentChat T6 a) Helps show the two groups are similar prior to intervention (mostly were, so probably valid) b) ⬆️ headache in rapid eating group

Anyone want to take a stab at answering these questions? You don't even need to read the article to answer 7a-c! #PHMFellowJC #MedStudentChat

For example, if a kid in the slow ice cream group had wolfed their ice cream down (It me 🙋🏻‍♀️), they would still be analyzed as part of the group they were assigned to. #rulebreaker #PHMFellowJC #MedStudentChat

@PHMFellowJC If they’d analyzed per protocol, only those participants who followed the 🍦 eating rules could be included in the final analysis, decreasing the sample size and introducing bias. Here’s an article about why ITT is important! #PHMFellowJC #MedStudentChat https://t.co/KZSfSElMJ3

@PHMFellowJC #PHMFellowJC #MedStudentChat Generalizability is the idea of whether the research is applicable (or not) to your own patient/population, eg a study on asthma in ex-preemies (who likely have other issues such as CLD) might not be helpful for treating normal full-term kids w asthma

@PHMFellowJC They generally look similar, except for gender. Somehow, the slow eating group ended up w more girls. So that might make you question the results; what if girls are less likely to have headaches in general? Or less likely to complain? 🤪 #jk #PHMFellowJC #MedStudentChat

Great way of explaining it, @maha_kaissi: Concealment: Blinding the randomization process so you the system can't be played Blinding: Blinding the participants and researchers from knowing who got the intervention. #PHMFellowJC #MedStudentChat

Agree! “Table 1” is often overlooked during critical appraisal of a manuscript! #PHMFellowJC

#PHMfellowJC #MedStudentChat T6c: ITT means students kept in orig group even if they didn’t follow directions eg someone in fast-eating group ate slowly instead. Not sure how much it mattered in this study but here’s a quick ref of pros/cons for RCTs https://t.co/QRawxdeFz3

Whew! We made it! We hope you learned something and had fun during our chat! We covered a lot so go back and search #PHMFellowJC and #MedStudentChat at your own pace to review the material. We will post a moment and transcript when we can. DM @PHMFellowJC w/ ?s! Thanks all!

T6c: Intention to treat means that analysis is completed based on the initial group assignment, regardless of what happens to individuals along the way. #PHMFellowJC #MedStudentChat

#PHMFellowJC #MedStudentChat The other null hypothesis is that the pace of eating ice cream doesn't affect frequency of headaches.

a. Ice Cream headache only occurs in hot weather. b. Statistical power deals with Type II errors. The main purpose is to help the researcher to determine the smallest sample size that is suitable to detect the effect of a given test at the desired level of significance.