How to choose a research topic from scratch?

[socrates89]

---

Members don't see this ad.

I just started to work with a Cardiohoracic surgeon as a research intern. I thought he would give me a topic and data to work on, on which I could write a paper. But he told me to design my own study, collect fresh data from patients and then write up the study. I'm kinda lost since I don't know where to begin in choosing a topic, something related to Cardiology or Cardiothoracic Surgery.

I'm thinking of doing a cross-sectional (analayitcal) study since that would be the easiest to collect data on from newly admitted patients to the hospital and could be done in 2-3 months.

Can someone give me practical advice on how I could choose a topic ( where should I look ie journals), I could search pubmed for studies put that would be unrealistic since there are so many topics in cardiology. He is giving me total freedom, but I don't have a clue on how to start please help

 

[ShinyDome19]

Unlikely to get any fresh study up and running, let alone completed in 2-3 months taking into account that you'll have to get IRB approval before doing so, and that can be a pain in the *****.

However, the best way to get a study going is to read journals and see what has been done already and what questions remain from their studies. Review articles are a good place to start and often outline where future studies "need" to focus or where work in the future may be headed on those topics.

 

[tco]

I just started to work with a Cardiohoracic surgeon as a research intern. I thought he would give me a topic and data to work on, on which I could write a paper. But he told me to design my own study, collect fresh data from patients and then write up the study. I'm kinda lost since I don't know where to begin in choosing a topic, something related to Cardiology or Cardiothoracic Surgery.

I'm thinking of doing a cross-sectional (analayitcal) study since that would be the easiest to collect data on from newly admitted patients to the hospital and could be done in 2-3 months.

Can someone give me practical advice on how I could choose a topic ( where should I look ie journals), I could search pubmed for studies put that would be unrealistic since there are so many topics in cardiology. He is giving me total freedom, but I don't have a clue on how to start please help

My advice is to dump him and find someone who will be helpful to you and has a project already up and running.

 

[sinombre]

Unlikely to get any fresh study up and running, let alone completed in 2-3 months taking into account that you'll have to get IRB approval before doing so, and that can be a pain in the *****.

However, the best way to get a study going is to read journals and see what has been done already and what questions remain from their studies. Review articles are a good place to start and often outline where future studies "need" to focus or where work in the future may be headed on those topics.

Being that this has to do with a universally dysfunctional regulatory body and not medical school, I figure I can chime in. The amount of time it takes to get anything through the IRB I've worked with is mind-boggling. I'm just about to start a project that was initially submitted a year ago. This isn't only the fault of the IRB, but the IRB is more at fault than anyone else involved.

 

[IlDestriero]

The easiest thing for you to do is a retrospective analysis. The data is already there and you can be doing most of the project while you get approval. If it's a quality improvement project you may not even need approval. There is no way you could design and complete a new project in 3 months. At best, from scratch, you could get approval for a project to begin data collection in that time frame, assuming you already had an idea.

 

[socrates89]

your right about the retrospective analysis, but i'm looking for a topic in which I would only need like 20-30 patient files at most, some of the retrospective or corss-sectional studies require you to have collected data from 300-500 patients and then do analysis, and with my busy schedule that is going to be very difficult. Does anyone have sugggestions on how to pick a retrospective alaytical study that doesn't require more than having 30 cases?

 

[Lucidity]

your right about the retrospective analysis, but i'm looking for a topic in which I would only need like 20-30 patient files at most, some of the retrospective or corss-sectional studies require you to have collected data from 300-500 patients and then do analysis, and with my busy schedule that is going to be very difficult. Does anyone have sugggestions on how to pick a retrospective alaytical study that doesn't require more than having 30 cases?

Honestly I always shoot for >100 at least for any sort of retrospective study. Anything under that is only if it is a rare procedure/illness that hasn't been reported on yet. You would be doing a disservice to yourself especially if you hope to submit for an abstract or publish a paper.

I agree with the above posters that it might be best to just drop this surgeon and find another PI. Most PI's I work with have projects that are on the back burner or waiting for some bright eyed individual. Do this because this project will be a waste of time...

 

[michaelrack]

it's a quality improvement project you may not even need approval. There is no way you could design and complete a new project in 3 months.

Only if it is truely quality improvement (for internal use only). Any intent to publish/present (ie abstract) requires IRB approval, though you may be able to get expedited approval for a retrospective study.

 

[45408]

your right about the retrospective analysis, but i'm looking for a topic in which I would only need like 20-30 patient files at most, some of the retrospective or corss-sectional studies require you to have collected data from 300-500 patients and then do analysis, and with my busy schedule that is going to be very difficult. Does anyone have sugggestions on how to pick a retrospective alaytical study that doesn't require more than having 30 cases?

It's not likely to be of much/any use with 30 patients. I did a study with nearly a thousand patients, which attained adequate power, and it took me about a year and a half from concept to manuscript, and now I need to submit it.