How can data be used in retrospective studies?

How can data be used in retrospective studies?

Primarily at academic and teaching hospitals, I do research most patients when they, you know, sign up for care, they sign the contract for care that has many provisions for it.

Some of it allows for research to be used in de-identified manner for retrospective, research studies down the line. What do all those big words mean?

So I think retrospective research we had alluded to earlier, that means that, you know, I'm not reaching out to the patient directly. I'm just looking at the medical record, at the computer charts for all these patients are looking back in time and saying, hey, over the last five years, what's happened when I look at this variable or this drug or this treatment paradigm, what happened with these patients?

And so, in terms of how our research use in those regards, you know, having done many, many, many of these studies since for over a decade, because it's often even our trainees, this is how trainees get involved with like meeting medical students and residents and fellows. How do they learn to do research? Through retrospective studies like these, where they think of questions that are important to them?

You know, for me, for example, again, going back to the ones that I'm very important to me, dexamethasone dosing and the side effects of dex in myeloma.

And a good way to look at that is to say, well, let me go back and look at all the patients treated at my cancer center, Fred Hutchison, Seattle, and look at the dexamethasone dosing and look for x, Y or Z side effects.

So I surgeries for cataracts, you know, blood sugar issues. you know, weight gain, things like that, issues that dexamethasone can cause. And that's a good example of the type of question that I would ask.

The patients whose chart that's being looked at for that reason, are they aware of the research being done? Not directly. To be totally frank.

Are their data protected? Absolutely. Because the team is doing the research, are themselves affiliated with the medical center. So these are patients we may see in real world practice. You know, I'm not sending the patient's name or contact information to anybody else.

What I'm doing is I'm using a unique code for that patient and looking at their, different outcomes.

And then, you know, in terms of what actually gets submitted or what gets used to generate real world evidence of submitted for a publication or an abstract is only the de-identified summary data saying at the end we looked at 300 patients and 200 of them, they had this, 100 of them had this. And here's how they did.

The other thing to keep in mind is that there is oversight or there's a very rigorous oversight for this.

many of you who are listening have heard of IRBs before Institutional Review Board. So they review all research. They are very involved with prospective trials. So anyone who wants to do a prospective trial meaning, an actual trial of a new drug and wants to, you know, work with the patient and the patient has to sign an informed consent form to do it.

That's all through the IRB. But even the retrospective project, even if it's for patients in my own practice. So even if I was a hey, I want to look at just all Doctor Banerjee’s patients who got dexamethasone, how did they do? I would still need IRB approval for that because I'm now not using this looking at patients charts for clinical care.

I'm looking at for a research perspective.

So I have to submit an application to the IRB with exactly what I'm planning to do, why I'm planning to do it, how am I going to maintain data security? So what are all the steps I'm taking to make sure that no one breaks into my office and sees the data, which is a very real, you know, thing that we actually answer in the IRB application and go from there.

So it's actually very regulated more than people realize. you know, if you're listening to this and you're like, well, I don't want my research to use for any purpose, I want to look at my charts unless they need to talk to your doctor about it. There are ways around that.

So many health care centers can, you know, have patients who are occluded from these research, or at least many medical records have a way to block off patient charts.

So you need to have a direct reason to access the chart to access their chart. But apart from that, I would say research retrospective research is often how we answer many of our most important questions.

So for all of you who have contributed to these projects, either prospectively or, you know, being at an academic center and your data being used retrospectively, thank you.