Literature Review--how to conduct?

[appliepie789]

Hello Everyone! Fellow Pre-med here I am interested in conducting a literature review regarding some public health practices in Tijuana, Mexico. However, I am not really sure how to reach out to someone to do so. How did you get the opportunity to conduct a literature review? Did you just email professors? Thank you!

---

Members don't see this ad.

 

[Hzreio]

Definitely reach out to someone before doing so. The hardest part is finding the right idea. You also have to ask yourself what you’re going to add? Summarizing the data alone is not enough.

Once you have an idea, you then think of inclusion/exclusion criteria for a search of databases (pubmed, medline, etc.). Then you start figuring out which studies you want to include. After that you extract all the data from these studies.

Once you have all the data in front of you, you then want to think about what can you say about this topic? Does the data as a whole say something that one study couldn’t? Are there possibilities to explore different practices? You need to mention these things.

If you’re doing a systematic review, it usually needs to follow PRISMA guidelines. Same with a meta analysis.

 

[appliepie789]

Definitely reach out to someone before doing so. The hardest part is finding the right idea. You also have to ask yourself what you’re going to add? Summarizing the data alone is not enough.

Once you have an idea, you then think of inclusion/exclusion criteria for a search of databases (pubmed, medline, etc.). Then you start figuring out which studies you want to include. After that you extract all the data from these studies.

Once you have all the data in front of you, you then want to think about what can you say about this topic? Does the data as a whole say something that one study couldn’t? Are there possibilities to explore different practices? You need to mention these things.

If you’re doing a systematic review, it usually needs to follow PRISMA guidelines. Same with a meta analysis.

Thank you! Would you say I should contact a professor first before I begin? Or should I start executing the review myself? Thank you!

 

[Hzreio]

Thank you! Would you say I should contact a professor first before I begin? Or should I start executing the review myself? Thank you!

If you’re inexperienced, you probably almost need an advisor. The fear is doing all the work just for it to be unfeasible for publication. You can just look at the literature and see what’s been said. Then if you identify an advisor you can report your preliminary findings and see what they think.

 

[appliepie789]

If you’re inexperienced, you probably almost need an advisor. The fear is doing all the work just for it to be unfeasible for publication. You can just look at the literature and see what’s been said. Then if you identify an advisor you can report your preliminary findings and see what they think.

Thank you!! You have been a big help to me

 

[pazzer2]

Literature reviews (aka review papers), meta-analyses and case series are less appealing from an editor's standpoint for publication. Main reason is that it is difficult to drawn conclusions from studies without randomization, adequate controls and sample size.

Have you considered doing an analysis using publicly available databases? There are loads of clinical trial, imaging, genomic and other databases available to researchers. The most difficult part is finding a hypothesis that is worth analyzing. For example you could take data from pharma company sponsored studies and combine them together to analyze secondary questions that weren't answered in the primary publication. This takes a bit of clinical intuition along with computing/statistical knowledge. Have you thought about this?

 

[Hzreio]

Literature reviews (aka review papers), meta-analyses and case series are less appealing from an editor's standpoint for publication. Main reason is that it is difficult to drawn conclusions from studies without randomization, adequate controls and sample size.

Have you considered doing an analysis using publicly available databases? There are loads of clinical trial, imaging, genomic and other databases available to researchers. The most difficult part is finding a hypothesis that is worth analyzing. For example you could take data from pharma company sponsored studies and combine them together to analyze secondary questions that weren't answered in the primary publication. This takes a bit of clinical intuition along with computing/statistical knowledge. Have you thought about this?

You’re referring to a meta analysis. And not reviews necessarily need to have quantitative analysis. And yes, reviews are less commonly accepted but this doesn’t mean editors don’t like them. It’s harder to criticize reviews when you do it correctly. It’s literally just data being presented with limitations and strengths with our current knowledge. This is why systematic reviews literally have set forth guidelines to make sure authors do them right.

 

[pazzer2]

You’re referring to a meta analysis. And not reviews necessarily need to have quantitative analysis. And yes, reviews are less commonly accepted but this doesn’t mean editors don’t like them. It’s harder to criticize reviews when you do it correctly. It’s literally just data being presented with limitations and strengths with our current knowledge. This is why systematic reviews literally have set forth guidelines to make sure authors do them right.

Actually I'm not referring to meta-analyses. A meta-analysis relies on combination of measures of central tendency (mean, median, proportion) along with variance to create a standardized estimate of the effect. A meta-analysis does not require access to the underlying data itself, just the means and variance.

What I was referring to is combining the raw datasets together. This is often useful to characterize baseline characteristics, or for regression techniques that rely on limited number of events. This is completely different than a meta-analysis.

 

[pazzer2]

I think I see the source of confusion now. I refer to these analysis combining datasets together as a pooled analysis. I see that some authors refer to these as Individual Patient Data (IPD) meta-analyses. As compared to a meta-analysis of aggregate data. Sorry for the lack of clarity.