Research Guidance

[MossPoh]

Hey all,

I'm a premed that is currently doing research with the psych department in conjunction with the medical school at my current uni. (Doing a post-bacc type thing) We are doing a study on the difference between "experts", who rarely make mistakes, versus "so-called experts" and those with less experience. Our goal is find the point at which the thought pattern kind of diverges leading one to the correct answer and the other to an incorrect answer. (Kind of the paraphrased version of it) This research is involving chest x-rays to begin with. I was wondering if there were any solid websites where one could get a high quality image along with a case-study type analysis of it. In addition, if any of you who are at least residents want to offer input, it would be more than welcome. Frankly, me being a pre-med and the people I'm working with being psychologists, we don't really have the appropriate base knowledge to write authentic sounding case studies.

I left significant details like the university and names out of it on purpose, but if you want more details just message me. Also, anyone that gives input should leave some sort of contact information. The contact information is so that we can have a method to reach you and trace any trail of information we can get. Thanks a bunch, and if you have any questions just message me. I'll give you a real email for communication there.

---

Members don't see this ad.

 

[116408]

Hey all,

was wondering if there were any solid websites where one could get a high quality image along with a case-study type analysis of it

http://www.learningradiology.com/

I you do some googling you'll find alot of other sources.

For some hands on work i would just contact the local hospital radiology department.

 

[hans19]

Not to discourage you, but there are so many variables and so much subjectivity in your proposed project. It will be an extremely difficult task.

Here are some issues that immediately come to mind.

How do you objectively quantify or characterize a thought process?

As you have alluded to there are at least 2 types of errors that are the subject of malpractice suits. Errors of perception (I didn't see the lesion.) Versus errors of interpretation (I saw that lesion, but I thought it was non-aggressive/benign). It sounds like you are focusing on the latter.

Often even with a chest x ray, if you are testing whether or not an observer picks up on a nodule. Unless you know that the patient developed path-proven lung cancer 5 years later, I don't know if you can definitively say whether a particular interpretation is right or wrong.

Further more if you have 50 different patients with <1 cm lung nodules on CXR, the majority will be benign. Because of this, lung nodules and other indeterminate lesions fall into a followup to assess for stability algorithm. Its hard to fault a radiologist for 'missing' a cancer when >99% of the time, that appearance on imaging turns out to be benign.

Also what kind of things are you assessing for? Findings can range from blatantly obvious to very subtle (only in retrospect). If you test for something obvious like a huge pneumothorax, every one should pick up on it.

If you are testing people for subtle findings most may see it, but some may disregard it as not being significant and others may call it significant. Even the same observer, if you showed him the same image 1 year later, may either not pick up on the finding, or interpret it a different way.

So you see, unless you narrow the parameters of your study, I don't know what kind of conclusions you will be able to make with any confidence.

 

[Docxter]

Agree with Hans.

The design seems to be very vague (even considering the standards for psych experiments). How do you define "experts"? Just because someone has been doing something for a while, it doesn't make them expert readers. Nor does having written many papers on the subject. I know of attendings who have been very prolific in writing about certain diseases, findings, etc., but when it comes to routine clinical work, they "miss" things all the time.

I'm sure you've done your lit search, but just in case, refer to the works of Hal Kundel, Calvin Nodine, and K. Andriole, and more recently, Elizabeth Krupinski.

 

[MossPoh]

Not to discourage you, but there are so many variables and so much subjectivity in your proposed project. It will be an extremely difficult task.

Here are some issues that immediately come to mind.

How do you objectively quantify or characterize a thought process?

As you have alluded to there are at least 2 types of errors that are the subject of malpractice suits. Errors of perception (I didn't see the lesion.) Versus errors of interpretation (I saw that lesion, but I thought it was non-aggressive/benign). It sounds like you are focusing on the latter.

Often even with a chest x ray, if you are testing whether or not an observer picks up on a nodule. Unless you know that the patient developed path-proven lung cancer 5 years later, I don't know if you can definitively say whether a particular interpretation is right or wrong.

Further more if you have 50 different patients with <1 cm lung nodules on CXR, the majority will be benign. Because of this, lung nodules and other indeterminate lesions fall into a followup to assess for stability algorithm. Its hard to fault a radiologist for 'missing' a cancer when >99% of the time, that appearance on imaging turns out to be benign.

Also what kind of things are you assessing for? Findings can range from blatantly obvious to very subtle (only in retrospect). If you test for something obvious like a huge pneumothorax, every one should pick up on it.

If you are testing people for subtle findings most may see it, but some may disregard it as not being significant and others may call it significant. Even the same observer, if you showed him the same image 1 year later, may either not pick up on the finding, or interpret it a different way.

So you see, unless you narrow the parameters of your study, I don't know what kind of conclusions you will be able to make with any confidence.

Yea...I know all the difficulties. I'm more of a get stuff together. I've kind of stated to them what you said and they give me the "just make it happen" approach. I guess tons of funding and cool toys makes you feel invinceable.

I've mentioned the whole concept of a differential, and not being too many definitive "oh it's THAT" , excluding the said pneumothorax and whatnot. I am getting the impression that it isn't so much the direct answer as it is the thought process involved. They are trying to find whether there is a definitive thought process that separates the good from the bad from the inexperienced. When you look at eye-tracking studies with radiologists. You see that the first year residents have a pretty eratic method of viewing a chest radiograph versus an experienced radiologist who has been reading films for 20+ years. (It makes a pretty nice hourglass shape)

Right now I'm just trying to gather images with confirmed diagnoses that can be cited and tracked down. Also, images would preferably be near full size and of course...no hippa issues. I really feel they don't grasp the difficulty of the task they have given to me, but Ill just smile and nod and see how it evolves.

*Also, note that I gave a very high level view, since that is all I have really been supplied with so far. My explanation wasn't the greatest. If you talk to the cognitive psychologist in charge he'll give you very specific details to what he means. He literally talked non-stop for an hour about it. As a result, I can't really sum it up too well since I'm not a psych person. I'm just a tech guy pre-med who has done basic usability studies involving PACS.