Journal Club

[PharmD62294]

---

Members do not see this ad.

Does anyone here have any experience doing Journal Clubs using academic-focused articles? I'm doing one on the use of case-based learning in a clin pharm course.

I'm really struggling because it seems as though the things I'm "supposed" to be evaluating as I would in a traditional scientific article... don't apply. I'm just not sure what direction to go in.

😕😕

 

[rxlea]

What specifically did they tell you that you are "supposed" to be evaluating? What was the method of data collection? Survey? Performance? What statistical analysis did they use? Likert scale if survey?

How did they control for covariates? What type of data? Internal/external validity? Sample size?

 

[PharmD62294]

They haven't told us anything. They gave us an article and told us we're doing a Journal Club presentation.

The only data collection was examination scores, and a likert scale assessing confidence in performing different tasks.

They used Mann-Whitney (p < 0.05) to compare current to historical exam scores and linear regression for the likert scale data.

 

[rxlea]

They haven't told us anything. They gave us an article and told us we're doing a Journal Club presentation.

The only data collection was examination scores, and a likert scale assessing confidence in performing different tasks.

They used Mann-Whitney (p < 0.05) to compare current to historical exam scores and linear regression for the likert scale data.

Well what type of data?

 

[rxlea]

Ok I quoted you right before you edited your post.

 

[rxlea]

So is mann whit u the appropriate test? Look at the assumptions for using that test... What are your groups?

 

[PharmD62294]

The likert scale data was of course 1-5 and thus is ordinal data. Repeated measures regression is appropriate depending on what one may consider confounding variables.

The exam data is continuous data (ratio) because it has an absolute zero. I don't really see Mann Whitney U being appropriate here? I would think ANCOVA or repeated measures regression would be more appropriate... but I'm not sure.

 

[PharmD62294]

I have three groups:

case-based learning in small groups
traditional lectures
case-based learning in large groups

They attempted to eliminate prior experience as a confounding variable by a pre-course survey that would help them distribute students by exposure to certain medications in IPPE (such as vanc, aminoglycosides, etc.).

 

[rxlea]

I'm never a fan of a 5 pointlikert. Middle variables like "don't know/don't care/etc" tell us nothing about the variable/outcome..

See, you're working it out though.



Can you generalize the findings?

How were the surveys administered? What was response rate? Met power?

 

[PharmD62294]

I would say the results are fairly generalizable.

The majority of students were female (distribution same between historical and current) - this is what is seen across most schools to my knowledge.

The study included all enrolled students so they didn't cherry-pick the best students. They also didn't exclude students based on performance (or other factors).

The intervention is not completely novel to the curriculum, just to the science classes. They may need to look at IPPE/APPE performance to get a full grasp on how they improved student learning and application skills...

There's not really a "NNT" for this, but since overall the exam scores were better and attitudinal surveys improved - it may suggest that the intervention was beneficial.

 

[PharmD62294]

The surveys were administered periodically through the course after examinations and also a final survey. There was a 96.1% response rate overall so the data -should- be applicable to the class as a whole. But, I don't think they set power..

 

[hereticmnk]

Great link

 

[PharmD62294]

Great link

Thanks!

 

[All4MyDaughter]

I also have trouble evaluating these types of articles.

Any thread where rxlea talks nerdy to us gets 👍 from me.

 

[rxlea]

The surveys were administered periodically through the course after examinations and also a final survey. There was a 96.1% response rate overall so the data -should- be applicable to the class as a whole. But, I don't think they set power..

Ok in journal club you'll need to point out the strengths/weaknesses of the study...how they set a baseline (which you pointed out)...strength...mann whitney appropriate or not? Was the primary outcome the actual performance since it's interval data(why wasn't ANOVA done?) Or was it perceived confidence on performance? Unless they are using A,B,C,D,E or some other ordinal categorization, ANOVA would seem appropriate. Sample size is very important with surveys. It affects how you can generalize your data and whether your study means something significant. Also, a representative sample is important. Talk about whether or not your study has those strengths. Then make suggestions for a future study that would be stronger.

Depending on your journal club preceptor, it might feel like you're being interrogated. Just know it inside and out. Know if it's significant (p value), speak to it's validity (also was the survey validated already?) External and internal, and the usefulness of the statistical test.

Sorry formessed up "typing". Am on my phone with swype 🙄

Edit: I'll say this again regarding likert...a middle variable is ******ed. Use 4 or 6. Never 5. Especially of your middle is any equivalent of "don't care"

 

[rxlea]

I also have trouble evaluating these types of articles.

Any thread where rxlea talks nerdy to us gets 👍 from me.

yeah they are convoluted with a lot of blahblah. It takes forever to really get through it sometimes.

 

[PharmD62294]

If this was only a scientific journal article with clear methods, endpoints, and results. 😀

I'm in my last year of didactic work and I'm in a teaching concentration (we do a teaching APPE as well) and we have to do a journal club at the end of the class. I'm hoping it goes well - I've only done one and it was a "group" journal club.

Thank you for all of your help. I'm going to get something together over the next week or so... may post it here if you don't mind.

 

[rxlea]

If this was only a scientific journal article with clear methods, endpoints, and results. 😀

I'm in my last year of didactic work and I'm in a teaching concentration (we do a teaching APPE as well) and we have to do a journal club at the end of the class. I'm hoping it goes well - I've only done one and it was a "group" journal club.

Thank you for all of your help. I'm going to get something together over the next week or so... may post it here if you don't mind.

You can PM me and I'd be happy to help.

 

[rxlea]

So dude, I was running a wilcoxon signed rank test and thought of you.

How skewed is the data?