How to publish during M1 summer?

[subdermallight]

---

Members don't see this ad.

Coming from a background in bench research, I'm likely going to stay away from that. How do I start planning for clinical research, or even just case write-ups for the summer? I'm pretty unfamiliar with any internships available or if I could just contact like 3-4 physicians and just work with them throughout the summer. What would you all suggest?

(as far as pubs I'm looking for quantity over quality ie. high impact)

 

[kopftonmd]

Depends on your interests. For example, humanities stuff can be fairly easy to churn out (if that's your thing). Otherwise, see if your school has a formal program for getting med students into research; if not, just looking through the school/hospital's website to see who's publishing in stuff you're interested in and then emailing around works for lots of people.

 

[NickNaylor]

Projects in education, quality improvement, and other studies that are less experimental and more observational will be more likely to result in something tangible. A project that is pretty narrow and focused will be more likely to result in something tangible than something extremely broad and lofty. It also tends to be easier to work on things that don't require you to be in certain places at certain times, which can make it easier to get something written and published.

Those are the kinds of things that I would be on the lookout for, but the details would depend on your specific interests and what's available.

 

[aldol16]

You should look into the research publications that faculty have and see if there are any doing work that you are interested in. If you're interested in a competitive specialty, then you should ideally look at faculty in that field. The easiest projects to get involved in are typically projects for which the data is already collected, i.e. in a database or registry, that you can begin to analyze. A lot of clinical data is often collected for one purpose but then is also conducive to multiple analyses. Alternatively, large claims databases are also an option. These types of papers you can do at your own pace and typically end up having results quickly without a huge amount of overhead.

 

[subdermallight]

You should look into the research publications that faculty have and see if there are any doing work that you are interested in. If you're interested in a competitive specialty, then you should ideally look at faculty in that field. The easiest projects to get involved in are typically projects for which the data is already collected, i.e. in a database or registry, that you can begin to analyze. A lot of clinical data is often collected for one purpose but then is also conducive to multiple analyses. Alternatively, large claims databases are also an option. These types of papers you can do at your own pace and typically end up having results quickly without a huge amount of overhead.

How would you suggest I go about getting involved in projects using claims databases? Should I just be on the lookout for that in the Methods of recently published papers from faculty and clinicians I look up?

 

[aldol16]

How would you suggest I go about getting involved in projects using claims databases? Should I just be on the lookout for that in the Methods of recently published papers from faculty and clinicians I look up?

Yeah that's certainly an option. See what infrastructure your school has. Usually these databases are expensive to buy access to and so what happens is multiple departments end up pooling funds for research use. Or the school buys it. So there might actually be a point person who's responsible for access.

 

[deleted936470]

And to reiterate this thread's implied message, bench work typically takes "too long" to publish relative to clinical research, yes? This is because bench work is experimental in nature i.e. is results driven, not observational, merely gathers data rather than analyzes it, and requires the researcher to be somewhere at a certain time. Or am I misunderstanding the consensus? Is most bench work in med school "slow"?

 

[Dr G Oogle]

Some government databases like those from HCUP May be purchased by students at a deep discount which alone may justify including a student on a project using it. However, in order to do big data research you need to know how to do stats using statistical software. Generally the only real work of doing this sort of research is the data analysis the rest is just writing the paper. In general including students on big data research carries very little advantage for the PI unless student 1)provides the data somehow (collects it, able to use student discount to purchase data) 2)able to use stats software that PI is unable to use.

 

[WhereMyLiberalsAt]

And to reiterate this thread's implied message, bench work typically takes "too long" to publish relative to clinical research, yes? This is because bench work is experimental in nature i.e. is results driven, not observational, merely gathers data rather than analyzes it, and requires the researcher to be somewhere at a certain time. Or am I misunderstanding the consensus? Is most bench work in med school "slow"?

The bench is slow no matter where the column is being ran...

Also in our academic/research climate null results equal no results.

It’s hard to pivot your hypothesis on the bench.

 

[aldol16]

And to reiterate this thread's implied message, bench work typically takes "too long" to publish relative to clinical research, yes? This is because bench work is experimental in nature i.e. is results driven, not observational, merely gathers data rather than analyzes it, and requires the researcher to be somewhere at a certain time. Or am I misunderstanding the consensus? Is most bench work in med school "slow"?

All research is "experimental" in nature and results-driven. Bench work also includes analyzing data. Bench research is longer because of the infrastructure and logistics involved. You still have to design good experiments - this stage takes longer because if you don't design it well to begin with, then you may end up doing the project and figuring out halfway through that you can't use the results. And basic science is hard. You may design 20 good projects/ideas only to find that you don't get positive results. So there's a lot of failed experiments/projects. That's why it takes so long.

For clinical research, a lot of time the data is already collected. The rate limiting step then is just how fast you can come up with a question and run stats on the data. It's still an experiment because you're still testing a question. But if the question amounts to nothing, you didn't invest a whole lot of time in it (i.e. just running the stats). Whereas in basic science, you would have had to invest a lot of time actually doing the experiments (not to mention resources) and analyzing the data to get to that point.

 

[aldol16]

Some government databases like those from HCUP May be purchased by students at a deep discount which alone may justify including a student on a project using it. However, in order to do big data research you need to know how to do stats using statistical software. Generally the only real work of doing this sort of research is the data analysis the rest is just writing the paper. In general including students on big data research carries very little advantage for the PI unless student 1)provides the data somehow (collects it, able to use student discount to purchase data) 2)able to use stats software that PI is unable to use.

The last sentence isn't true. A PI actually gets a lot of benefit from including students on database projects. You just tell the student to learn how to use statistical software (commonly R, SAS, Stata) and then you think of questions you would like that student to answer in the data. As a PI, your priority isn't being in the trenches and doing the data analysis yourself. Your bigger picture is clinical work and publishing and getting grant money. So the more students you have running your projects, the better off you'll be because the operating cost is so cheap.

So there is no need to be able to use software that no one else in the lab is able to use (although that would be a benefit and selling point).

 

[deleted936470]

Ok so the answer is “yes”, thanks.

 

[Dr G Oogle]

The last sentence isn't true. A PI actually gets a lot of benefit from including students on database projects. You just tell the student to learn how to use statistical software (commonly R, SAS, Stata) and then you think of questions you would like that student to answer in the data. As a PI, your priority isn't being in the trenches and doing the data analysis yourself. Your bigger picture is clinical work and publishing and getting grant money. So the more students you have running your projects, the better off you'll be because the operating cost is so cheap.

So there is no need to be able to use software that no one else in the lab is able to use (although that would be a benefit and selling point).

So I actually am the PI in these types studies and have a fair amount of experience working with students, particularly M1s and M2s (at a “highly ranked” Med school) and it has been my experience that most students have little motivation to spend time learning stats software or bio stats or how to clean data; I’ve made that offer many a time and often learn at the end of the research period they’ve not made any headway and I’ve lost time. It is often easier for me to do the stats myself or have one of my fellows do it then to teach a student how to use the software as the majority are coming from chemistry/bio/humanities background and have no experience with programming. Believe me I enjoy mentoring students and if I have someone very motivated who has a math or engineering background and understands not only biostatistics but the nuances of the clinical question we are asking early in their medical training then I willl spend a great deal of time mentoring that student. I’ve yet to meet them. Most students (because of the way the system is set up) have little interest in research for research’s sake and want to do it to get a first author publications so that they can match into a competitive specialty thus when presented with a task that does not promise immediate results of a publication often balk.

 

[aldol16]

So I actually am the PI in these types studies and have a fair amount of experience working with students, particularly M1s and M2s (at a “highly ranked” Med school) and it has been my experience that most students have little motivation to spend time learning stats software or bio stats or how to clean data; I’ve made that offer many a time and often learn at the end of the research period they’ve not made any headway and I’ve lost time. It is often easier for me to do the stats myself or have one of my fellows do it then to teach a student how to use the software as the majority are coming from chemistry/bio/humanities background and have no experience with programming. Believe me I enjoy mentoring students and if I have someone very motivated who has a math or engineering background and understands not only biostatistics but the nuances of the clinical question we are asking early in their medical training then I willl spend a great deal of time mentoring that student. I’ve yet to meet them. Most students (because of the way the system is set up) have little interest in research for research’s sake and want to do it to get a first author publications so that they can match into a competitive specialty thus when presented with a task that does not promise immediate results of a publication often balk.

Your experience at such a "highly ranked" med school seems non-ideal. I'm surprised that such students would balk at learning something that isn't incredibly difficult to learn. I think R even has a point and click interface now that facilitates learning hard-coding in a graduated fashion (R Commander). I don't necessarily think that all bioscience/humanities students have no experience with R or have no motivation to learn R. I actually think that many come into med school with at least some statistical software experience either from basic science research (e.g. MatLab) or a statistical course in college (which is being increasingly emphasized in the admissions process). If you get those students, then the job becomes a whole lot easier for you.

 

[Dr G Oogle]

I agree it is quite non-idea and frankly I was expecting a completely different experience when I joined the faculty. I’ve certainly had some excellent students, which is why I continue time to mentor them, but they’ve come from EECS or other math based backgrounds. The main problem I’ve encountered is this need for research to be able to match competitively and thus to have something done as quick as possible but not really understanding that research is work on top of class load and board studying and it’s often a lot of work if one has never done significant research and doesn’t really understand the clinical components of the field that are important to study. I’ve had more than my fair share of students who get sent to me by the deans who want research and when I lay out my expectations for first authorship (come up with your own research question, design study, learn stats all with mentorship and write the paper themselves with me as the editor/senior author) they balk or some have committed and then don’t finish projects.

 

[mvenus929]

And to reiterate this thread's implied message, bench work typically takes "too long" to publish relative to clinical research, yes? This is because bench work is experimental in nature i.e. is results driven, not observational, merely gathers data rather than analyzes it, and requires the researcher to be somewhere at a certain time. Or am I misunderstanding the consensus? Is most bench work in med school "slow"?

I mean, clinical research isn't super fast either. I've been working on a retrospective chart review for the past year and we now have a manuscript ready for submission... and now we play the submission game where we get rejected and/or have to do revisions, etc. I did a prospective clinical research project in residency, and from the time we came up with the idea til the time I got my manuscript accepted was 3.5 years, and my project was fairly straight forward.

Sure, case reports that don't require follow-up data can be written up and submitted in a couple months, but you still have to find a journal to submit to, and hope you get accepted. If not, it's more revisions, that have to get the approval of all the authors. And hopefully your authors are actually responsive to requests for feedback... I worked on one case report in med school and two in residency that died before the manuscript was fully written because my co-authors sucked. It can also be difficult for M1s and M2s to write because you need to have a fairly extensive background knowledge to know how to frame the case and what it is an important thing to publish.

 

[OnePunchBiopsy]

If you have residents affiliated with your institution try and reach out to them. I love having med students help with my projects and many earn authorships on posters, podium presentations, and manuscripts.

Residents appreciate some extra help on their projects, especially literature and chart review. Just make it clear you are interested in a manuscript authorship, this will help them decide how much you need to contribute.


Lastly, avoid case reports. No good journals accept them anymore and they will actually lower the quality of your CV.

 

[aldol16]

I agree it is quite non-idea and frankly I was expecting a completely different experience when I joined the faculty. I’ve certainly had some excellent students, which is why I continue time to mentor them, but they’ve come from EECS or other math based backgrounds. The main problem I’ve encountered is this need for research to be able to match competitively and thus to have something done as quick as possible but not really understanding that research is work on top of class load and board studying and it’s often a lot of work if one has never done significant research and doesn’t really understand the clinical components of the field that are important to study. I’ve had more than my fair share of students who get sent to me by the deans who want research and when I lay out my expectations for first authorship (come up with your own research question, design study, learn stats all with mentorship and write the paper themselves with me as the editor/senior author) they balk or some have committed and then don’t finish projects.

I think it would be incredibly frustrating to have someone on your project who is there solely because they are interested in matching a particular specialty and have no interest in the research itself or no motivation to learn how the research process and methodology actually works. At many of the top programs, a lot of med students come in having had at least some research experience. Those who had experiences in the biological sciences should at least understand research even if they have to learn more stats/coding! I think that anybody who doesn't want to at least help come up with the idea for a study, design it, then write the paper up doesn't necessarily deserve first authorship at all. But at the same time - and this is how things work in the hard sciences as well - some students can get first authorship if they are given an idea by the PI and then develop that idea further and actually carry out the research. I think that should be the minimum for first authorship.

 

[Dr G Oogle]

I think it would be incredibly frustrating to have someone on your project who is there solely because they are interested in matching a particular specialty and have no interest in the research itself or no motivation to learn how the research process and methodology actually works. At many of the top programs, a lot of med students come in having had at least some research experience. Those who had experiences in the biological sciences should at least understand research even if they have to learn more stats/coding! I think that anybody who doesn't want to at least help come up with the idea for a study, design it, then write the paper up doesn't necessarily deserve first authorship at all. But at the same time - and this is how things work in the hard sciences as well - some students can get first authorship if they are given an idea by the PI and then develop that idea further and actually carry out the research. I think that should be the minimum for first authorship.

agree with most of the above. My minimum for first authorship is doing the lit search, writing the entire study protocol, submitting to IRB, collecting data (if a database then either obtaining database, cleaning it up or doing the stats) and writing majority of paper.