Major Depressive Disorder Podcast

Socioeconomic Predictors of Treatment Outcomes in Patients With Major Depressive Disorder

Madhukar H. Trivedi, MD; Jeffrey R. Strawn, MD


May 09, 2023

This transcript has been edited for clarity. For more episodes, download the Medscape app or subscribe to the podcast on Apple Podcasts, Spotify, or your preferred podcast provider.

Madhukar H. Trivedi, MD: Hello, I'm Dr Madhukar Trivedi. Welcome to season two of Medscape's InDiscussion series on Major Depressive Disorder. Today, I am very excited to discuss the topic of socioeconomic predictors of treatment outcomes among adults with major depressive disorder, and we have a treat. Let me introduce our guest, Dr Strawn. He is a tenured professor in the Department of Psychiatry and Behavioral Neuroscience and holds secondary appointments in pediatrics and clinical pharmacology. He's also the co-director of the Center for Clinical and Translation Science and Training. Welcome, Dr Strawn.

Jeffrey R. Strawn, MD: Thank you so much. It's a pleasure to be here with you, Dr Trivedi.

Trivedi: Fantastic. I am particularly pleased because this topic of socioeconomic factors and predictors is discussed a lot, but I think that your work has helped us think through and make it a lot more concrete because in general, everybody understands that socioeconomic factors are associated with difficulties with health and healthcare. But I think you've really done a fantastic job of helping us understand the nuances. As a general strategy, I like to ask each of my guests a couple of initial questions. First and foremost, from your perspective, can you give us a sense of why should I and the audience be interested in this topic?

Strawn: Absolutely. My interest in the topic really comes from two directions. One is — as a clinician, I certainly know, like you and probably most of our listeners — that in the clinics, oftentimes, our treatments fail or they don't perform as they do in the randomized clinical trials. One of the issues that I'm always trying to understand is why is that? Is that something that I can predict? The other driver of my interest in this topic is really as a clinical trialist and as a researcher. Because one of the issues that we often see in our clinical trials is that we have all of these other factors that tend to drive response that, unfortunately, in many of the trials, particularly the industry-funded trials, we're not able to capture.

Trivedi: Wonderful. Maybe let's continue this and see if you can at least help summarize the findings of the recent study.

Strawn: Absolutely. Taking a step back, you really in some ways set this up already; we've long known that socioeconomic factors, as well as just other general factors, drive response in mental health, although that's really the case in the rest of medicine as well. Certainly in cardiovascular disease; diabetes; as well as depression, anxiety, bipolar disorder, etc. Historically, I think when we look at some of these socioeconomic variables, people think about poverty and they think about access. Unfortunately, the story has always been told as these are folks that just don't have the same access, they don't have the same resources, and therefore, they're not doing as well with our extant treatments. But really what these data show us is that this is much more complex than that. In the clinical trial, which is really the trial that you conceived and implemented, those factors were controlled for. There weren't differences in terms of access. There wasn't a specific cost associated with getting the medication or having the appropriate insurance to come to the clinician's office. Those things were all controlled for in the context of the clinical trial; yet, these factors still significantly drove response.

Trivedi: Thankfully, you raised the issue that this issue of socioeconomic factors actually affects more than just mental health, examples being diabetes, hypertension. I think it helps our listeners understand that this is about more than just mental health. Maybe any thoughts on how it affects, for example, diabetes or hypertension?

Strawn: Yeah, absolutely. You're hitting the nail on the head. This is pervasive and is something that has existed for decades, certainly likely longer. We just don't have the studies to necessarily support that. In terms of the causes, I think there are probably a number of them, and they're probably, unfortunately, very interrelated. I say "unfortunately" there from a research standpoint because it makes studying them and disentangling their specific influence quite difficult. But certainly, we can think about these factors as being driven by stress. We can think about the hypothalamic-pituitary-adrenal (HPA) axis and persistent hypercortisolemia and the effects there, probably interacting with the structural neurobiological aspects of depression as well as anxiety. But the other piece is really on a psychological side. When you have folks who are in a situation where they're dealing with chronic, often unpredictable, or what we might say is variable stress, it really complicates their ability to engage with treatment.

Trivedi: Wonderful. Thank you. Let's go back to your study, your new analysis. You found that several demographic and socioeconomic factors were associated with differences in improvement. Can you comment on how you disentangle these factors that are associated with one another, such as a person with low income perhaps being less likely to have a college education? Or are we looking at additive effects?

Strawn: Absolutely. That's a really important point. One of the things that I'm very fortunate to have here at the University of Cincinnati is some incredibly talented colleagues and two of those colleagues who are authors on this paper are actually economists, and they're specifically econometricians. They focus on really understanding, in a quantitative way, how economic or socioeconomic variables influence things over time. So not just a single remission or response, but that pattern or that trajectory of response, which is something that we really focused on here. Because the issue is that just understanding whether or not somebody gets better is certainly helpful to me as a clinician, but it's much more helpful if I can know whether that's a person who will get better quickly or whether that's a person who's going to take a bit more time to get better. But in terms of your actual question, Dr Trivedi, what we were able to do — or rather, what I should say, Dr Mills and Dr Chang and the graduate student, Mr Suresh, who worked on this study, were able to do is that they were able to use what are called Bayesian hierarchical models, wherein they're able to control the influence of these different factors that may be interrelated in order to parse out what is the unique contribution of being unemployed or the unique contribution of not having a college education. Also, in doing so, they're able to control for other factors. So sex, race, age, other things that we know are very important in terms of driving response to treatment.

Trivedi: I think one of the other things I would like your thoughts on is that I know you've been interested in understanding differences in improvements in patients with depression, and you recently examined the impact of age by combining patients from this and other large studies. Tell us a little more about these findings.

Strawn: Certainly, and maybe again, taking a step back. I'm a child and adolescent psychiatrist, although I treat a large number of adults as well. One of the things that fascinates me when I look at the randomized controlled trials that have been done is that we see differences in response across those studies. In the pediatric trials, particularly with regard to depression, patients tend not to do as well. They don't have as robust a response, perhaps they have a slower response. And we know, certainly, at the other end of the age range in terms of the older adult populations that we see something similar. Now, we know that neurobiologically there are significant differences between those two extremes. We know that in terms of lifestyle, in terms of other factors, there are also significant differences. But the question that I had as a clinician is why is that? The related question is why don't we see that age effect when we look at these studies? I think the issue is, and I'm going to try not to get into the weeds here from a statistical standpoint, although I may probably have to use a couple of statistical words here, but when we think about the relationship between age and response, people think about it in terms of a linear relationship. The older I get, the worse my outcomes, or the older I get, the better my outcomes. So for each, say, increase in age by a year or a decade, there's a proportional increase or decrease in whatever that outcome is. But certainly, we know that that doesn't capture it. In reality, what we tried to do in this paper was to understand that more complex relationship other than just being a simple line, which certainly doesn't make sense to us as physicians and clinicians or as scientists. What we actually found was that it's a very interesting — if you can almost imagine ­— a U-shaped curve and then maybe perhaps flip that U so that it's a hump. So we have the lower response in the younger patients, and then we increase up to a certain point. And that point seems to be middle age where we have the most robust and the fastest response to antidepressants across all of these studies. But then, as we age, we are now on the downhill slide of that inverted U. We see those responses, that trajectory of response, the magnitude of response, tend to degrade as we approach those older ages.

Trivedi: Your thoughts really remind me of another issue that we as a field have not really addressed. And maybe your thoughts would be good for our audience. We end up studying the same treatments that we find in one population across all populations and not paying much attention to trying to figure out if we need to be really looking at different types of treatments. Especially either early in the course of the illness, like with kids, and then later, in geriatric populations. Any thoughts on how one should think about this as we start planning future studies?

Strawn: Certainly. Maybe not to be too prescriptive, but I'll think about my own practice and my own research program. I think the results that we have, certainly from these two studies, suggest that whatever we're doing is not enough at those extremes. The argument there is for more intensity in terms of what we're doing, and the question could be: Is that pharmacologic intensity or is that psychotherapeutic intensity, or is that some other intervention there? So one of the issues is that we certainly know that adding psychotherapy, whether that's an interpersonal therapy or cognitive-behavioral therapy or even psychodynamic psychotherapy across the age span, tends to enhance outcomes. Now, that being said, the elephant in the room is really that there's a lot of heterogeneity. We know that when we're looking at these psychotherapeutic interventions, there are some patients that have fairly minimal response to one but great response to another. But unfortunately, as a field, we really don't have great tools in terms of identifying perhaps which patient would benefit from which adjunctive psychotherapy and why. That's something that we definitely have to understand. I can tell you in the office, as I'm thinking about individual patients, one of the things that I tend to look at is the context of their depression. And if we're noticing more of an interpersonal context and there's more of a discussion of relationships as opposed to thinking patterns, that's a patient that I tend to try to involve in an interpersonal psychotherapy, particularly on the adolescent side. Whereas if there are very clear distortions, I tend to go the route of cognitive-behavioral therapy.

Trivedi: This idea of thinking about combination treatments, especially when and what type of therapy, for example, to add, is, again, not as very commonly studied, especially in the two extremes: the younger and the older. I think maybe something that it sounds like you're pointing out is that we need to be thinking about being a little more precise in the timing of when you add a second treatment.

Strawn: Certainly. I think that's definitely the case. Maybe, just to jump from there to the pharmacologic side, one of the things that really is frustrating to me as a clinician is that when we look at many of our guidelines, they recommend things like a selective serotonin reuptake inhibitor (SSRI). And there's this idea that there's a one-size-fits-all approach to using an SSRI, you know, as if when I go into the department store and order a suit, they say, "Well, you're going to be getting a 42 suit coat size, and a size 35 waist, and a size 11 shoe, because that's the average American male and that's what we're going to give you." We know that that's not the case. We know that there are differences in terms of tolerability that are probably driven by genetic factors. The other piece is that we know that there are significant differences in terms of the metabolism of many of the SSRIs in particular. Yet, when we're recommending these medications, it's as if they're interchangeable and there's a single dose that should be prescribed for all patients. And this age factor really plays in here as well. We know that metabolic activity, particularly with regard to cytochrome P4502C19, decreases significantly after about age 60 years. So when we're using that same dose of escitalopram or citalopram or sertraline, we may be running into tolerability concerns in the older adult that then impairs our ability to see improvement. Similarly, on the younger side, we have other factors that affect tolerability. I think that there does, as you're alluding to, need to be a lot more of a precision medicine approach to how we're using these medications, kind of moving beyond the one-size-fits-all.

Trivedi: And a lot of us, including your group and others, are really spending a fair amount of our research effort on trying to identify and develop blood and brain tests that would help. But maybe your thoughts [on how] in the meantime, our measurement tools that we have are actually pretty accurate in terms of assessing these things. So, for example, the part you were talking about with the changes in cytochrome P450 systems, if you measure their symptoms and side effects on the rating instrument, then you would catch it and be able to modify your treatment. Does that make sense?

Strawn: I 100% agree, and one of the clinical variables that we don't often think about is the speed at which someone is improving. That's something that's really important, particularly when we look at some of these studies in adults, but also when we look at the studies in the pediatric population, we often think about the variables that we would look at as, perhaps age or race or family environment or many of these socioeconomic variables that we've now talked about in the first publication. But we unfortunately don't think about that trajectory of early response. That's one of the things that's been consistently shown to be a predictor of eventual response. One of the things for me in my clinic is if I have that patient that's really failing to demonstrate any significant improvement over those first 4 weeks, say, of an antidepressant treatment, it's probably time to start thinking about switching horses and moving to another treatment or doing something else differently. The issue is that, as you alluded to, our trials don't really allow us to address that.

Trivedi: How do you suggest we move forward? Do we do clinical trials differently? Collect different variables? What is your recommendation on that?

Strawn: Two initial recommendations: First, in terms of the paper that we're discussing, looking at socioeconomic factors, we have to measure these factors. We have to understand the impact of race as well as related factors. We have to understand the impact of education and access because in many cases, being able to account for the influence of those factors may mean the difference between being able to see a difference between two treatments in the study or between the treatment and placebo in the study. That's really the first piece of this. I think the second is moving beyond the way in which we traditionally do clinical trials, which is to compare two interventions or three interventions or look at superiority or noninferiority or something along those lines. I think if we were to re-envision the clinical trial, it might be how do we sequence treatments or how do we sequence the timing of intervention? For example, as I was talking about, that early improvement. Do we look at patients who fail to have early improvement and randomize them to a specific intervention or two interventions? So we're randomizing based on the timing of improvement as opposed to based on treatment or treatment and placebo.

Trivedi: Thank you very much. I think before we end I would like some synthesis of your thoughts on if I'm listening and I'm treating young kids like teenagers or on the other extreme geriatric patients, can you give some immediate pointers of what I should be thinking about as I plan treatment with them?

Strawn: For me, it's thinking about home. What I mean when I say that is where does the patient go after they leave our office in terms of their job, in terms of their relationships, in terms of their family environment, in terms of the neighborhood factors that are driving stress? How do we not discount those factors which we've seen here are incredibly important in terms of driving response? And again, is that something that's addressed through a family-based intervention? Is that something that's addressed through more frequent support in an older adult who maybe is struggling with loneliness that perpetuates their depressive symptoms?

Trivedi: Wonderful. Thank you so very much for taking the time, Dr Strawn, I appreciate your thoughts on this very important topic. This stumps a lot of us because we know that these factors affect patients' outcomes, but we don't exactly know how to address it. Thank you very much. Today we've talked to Dr Jeffrey Strawn about the impact of socioeconomic factors on outcomes of medication treatments for major depressive disorder. He shared that not having a college education, being unemployed, or being a member of a minority population were each associated with slower and lower improvement. Also, Dr Strawn shared fascinating data describing the complex relationship between age and antidepressant response in patients with major depressive disorder, finding that older and younger patients have less robust responses. Thank you for tuning in. If you have not done so already, take a moment to download the Medscape mobile app to listen and subscribe to this podcast series on Major Depressive Disorder. This is Dr Madhukar Trivedi for InDiscussion.

Listen to additional seasons of this podcast.



Center for Clinical and Translational Science and Training (CCTST)

Socioeconomic Predictors of Treatment Outcomes Among Adults With Major Depressive Disorder

Cortisol and Major Depressive Disorder-Translating Findings From Humans to Animal Models and Back

Bayesian Hierarchical Models

The Impact of Age on Antidepressant Response: A Mega-Analysis of Individuals With Major Depressive Disorder

The Effectiveness of Individual Interpersonal Psychotherapy as a Treatment for Major Depressive Disorder in Adult Outpatients: A Systematic Review

Cognitive Behavioral Therapy for Depression

Depression and Psychodynamic Psychotherapy

Selective Serotonin Reuptake Inhibitors

Cytochrome p450 Structure, Function and Clinical Significance: A Review

Optimizing Drug Selection in Psychopharmacology Based on 40 Significant CYP2C19- and CYP2D6-Biased Adverse Drug Reactions of Selective Serotonin Reuptake Inhibitors

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.