Retrospective Study Guidance

[Chrism611]

Hello,

I hope that someone may be able to give me a little bit of advice on this topic, and I hope I am able to give enough information so that you can.

First off about me. I am a Junior nursing student and want to go into medicine. I have worked in the ED for six years and have a passion for emergency electrocardiography. I have, unfortunately, never had a research course and do not know much about research, other than what I have read.

I am looking to do a retrospective study on an EKG change that a few case studies have noted to be significant for a specific coronary artery occlusion. I have also seen patient's with the EKG change personally that have had such findings.

I plan to look at EKGs of patients who presented to the ED over a time period. Exclude those with EKG changes that can interfere with the significance of this finding. Then I will look up those with the positive finding and evaluate what was found diagnostically through heart caths. The study will be conducted in a single facility.

So my first question concerns exclusion factors. Am I correct by planning to exclude patients who had the EKG change but were not admitted or had no diagnostic testing to test for coronary artery occlusion?

My next question concerns the evaluation of data. I plan to use chi-square test or Pearson's, depending on what size of population to evaluate the null hypothesis of: The EKG change is not associated with coronary artery occlusion.
This is where I get confused, because I'm not sure if this is the right method to use, or is there something else that would be better?

I have spoken with an EM physician that I work with and a cardiologist at our facility who think that this is a good idea. They are also willing to help me with the study. The EM doc has been walking me through some of it statistically but I still have some gray areas.

In the study, do I need to just focus on the general population? Or should I divide the patients into groups, for example three age groups. Also should I evaluate other characteristics (sex, smoking status, history, symptoms, race, etc.).

Or would it be more reasonable to complete the first study, and if a group of patients with certain characteristics are noted to have the coronary artery occlusion with the EKG change, a subsequent study would be done to evaluate the significance of the change in that specific group?

I may have not been specific as I should have been and my main question that I really need guidance on is the last paragraph. I apologize if I was not clear enough about what I plan to do or if I am completely off because I don't have experience with research.

Thanks in advance.

---

Members don't see this ad.

 

[SpartanWolverine]

Hi there,

I would assume there are similar studies out there. You can probably find a good list of exclusionary criteria and statistical analyses from some of those studies -- you'll want to use the same general criteria and analyses as those studies to have the most generalizability (e.g., here's one study I found with a quick search that looks like the same general idea http://content.onlinejacc.org/article.aspx?articleid=1125899)

Lots of questions for you...
Where are you currently in the process of conducting the study (e.g., have you received IRB approval?)
Is your facility associated with a medical school?
Is there a lab you could join?
Has one of the physicians committed to seeing this project through with you? Depending on your goals with this project, you will want a clinician/professor to help you work through the process and be a co-author on anything you may publish.

Again, see relevant articles to see their methodology and factors they used to stratify the groups.

 

[Chrism611]

Hi there,

I would assume there are similar studies out there. You can probably find a good list of exclusionary criteria and statistical analyses from some of those studies -- you'll want to use the same general criteria and analyses as those studies to have the most generalizability (e.g., here's one study I found with a quick search that looks like the same general idea http://content.onlinejacc.org/article.aspx?articleid=1125899)

Lots of questions for you...
Where are you currently in the process of conducting the study (e.g., have you received IRB approval?)
Is your facility associated with a medical school?
Is there a lab you could join?
Has one of the physicians committed to seeing this project through with you? Depending on your goals with this project, you will want a clinician/professor to help you work through the process and be a co-author on anything you may publish.

Again, see relevant articles to see their methodology and factors they used to stratify the groups.

Thank you, I will look through their way of statistical analysis for ideas. I was unsure if I needed to be broad about the concept or go into sub group analysis. I know what I'm looking for and how to find the data for it, just have some confusion about how exactly I need to statistically analyze the data and how in depth I should go.

I have not submitted the proposal to the IRB yet, just because of the fact I lack the statistical knowledge to carry it out.
We are not affiliated with a medical school, however we do have some medical students rotate in our facility.
As I am in nursing school, I don't really have access to any labs. Our school has a science dept but I do not think they are strong on research.
Both of the physicians are willing to go through the process with me and want to be involved in the study.

 

[deleted343839]

You're asking a lot of good questions. You really need an experienced mentor to make this project worthwhile and a good learning experience for you. Unless your collaborators are methods savvy you might run into some pitfalls. Perhaps you, or one of your collaborators, has access to a methodologist or a biostatistician who could meet with you. Could you do this project as part of a research methods course? That would probably be ideal.

 

[Chrism611]

You're asking a lot of good questions. You really need an experienced mentor to make this project worthwhile and a good learning experience for you. Unless your collaborators are methods savvy you might run into some pitfalls. Perhaps you, or one of your collaborators, has access to a methodologist or a biostatistician who could meet with you. Could you do this project as part of a research methods course? That would probably be ideal.

A major problem is a lack of resources in my area. We are a large community hospital, but we don't have residents and aren't a huge teaching hospital. Therefore from a medical aspect, most of the guys are old internists that have practiced in the area for years and aren't really involved with research. I feel that a methodologist or biostatician would be very beneficial to have to help me, however, our university isn't huge on research, especially this type.

Another problem I have is my undergrad. I'm a nursing major, so we have a nursing undergrad research class that is more of read this paper, write what you think and cite with APA. Not the actual methods of research. I did contact graduate research instructor who was just interested in what my results would be. Another research course is not really ideal outside of my major because we meet every day and I don't really have capabilities of taking another course due to time overlaps. And contacting someone in the biology dept probably would't be too interested in this type of research. Most pre-meds here are either biology or chemistry, there is no BMD sciences or anything like that.

Thank you for your time and I am continuing to search for someone who can assist me.

 

[SpartanWolverine]

Unfortunately, I think it would be supremely difficult for an undergraduate student to independently write a proposal, collect, analyze, write up, and publish research in a credible journal without the help of a mentor with significant experience. Even in the best of circumstances, you would likely need well over a year to accomplish such a thing (and it sounds like you're going to be done with your degree around then).

A research course (depending on the resources available at your school) is also not something that I expect would help you much with pushing forward a specific project. As you note, it's likely more about methodology and how to interpret primary sources.

So, it sounds like you know the limitations and drawbacks to the environment you're currently in. Perhaps you can put your idea for a study on the back burner until you have a bit more experience? Is it possible there are ongoing research projects that you could join to have a hand in? You seem to be very motivated, so I'm sure you could get some good work done. ... ...Heck, I could use some help with a couple of projects...

 

[deleted343839]

Many psychology departments have upper-level undergraduate research methods courses that are project-based and require original data collection, hence the suggestion. It sounds as though the offerings in your program are different, though. I do think this sounds extremely ambitious and risky without a research mentor. This sounds equivalent to an honors thesis-type project in my field.