Improving Adult ADHD Diagnosis Across the Population:
An Interview With Stephen Faraone, 

 

Adult attention deficit/hyperactivity disorder (ADHD) is generally understood to be underdiagnosed and undertreated in the US population, with an estimated prevalence of 4.4% but a rate of diagnosis of 1.4%.[1,2] In a poster[3] presented at the US Psychiatric and Mental Health Congress, researchers used a model population to evaluate how various interventions affected the rate of correct diagnosis and how those interventions might work together. Co-author Stephen Faraone, PhD, discussed the research and its implications with Jane Lowers, editorial director for psychiatry at Medscape.

Medscape: In your poster, you used a model population to test out different educational initiatives to improve diagnosis of adult ADHD. Could you outline the premise for the research?

Stephen Faraone, PhD: This is a model of a population. We're not talking about analysis, real data, real patients. What we're talking about is modeling a process that we believe is going on in the population, and this is the process whereby people have a disorder and, if it's ADHD, a change from being unaware that one has ADHD to being diagnosed and treated for ADHD. What the mathematical model describes are the processes that underlie a person moving from 1 state to the next to get to the endpoint of diagnosis and treatment.

The reason we're interested in this was because, depending on different assumptions, we can then examine what would be the effect of making changes in specific areas. For example, patient outreach would help patients become more aware of their disorder, which would make them see physicians. Physician training would make them more knowledgeable on how to diagnose and prevent a misdiagnosis. So although it's a population model and not real data, I should point out that the assumptions that guide the model were derived from what we know about ADHD in adults. I also emphasize we're talking about adults with ADHD, not children with ADHD. These underlying assumptions derived mostly from epidemiological studies or health services studies that gave us some estimates; for example, we know that the prevalence of ADHD in adults is roughly in the 3% to 5% range of the population.

Medscape: That's the Kessler data and some of your previous work as well?[1,2,4]

Dr. Faraone: Exactly. The work that we did with Ron Kessler, the work that Joe Biederman and I published, a different study, as well as other studies of prevalence of ADHD and what we know about the specifics of the disorder, they all converge on a population rate around 4% of the population of adults that should have ADHD that would warrant treatment if seen by a clinician.

Medscape: For the purposes of our readers, could you narrate through what the modeling entailed and what you discovered along the way?

Dr. Faraone: Sure, I guess the best way to describe it would be to talk about the systems. When patients start outside the medical system, they're not involved in any treatment or diagnosis. The first step would be patients who are unaware they have ADHD but who do have it, and there are certainly people in the population out there like that. And there are people who are aware that they have an underlying problem, but they can't actually identify what that problem is. Because of some personal distress in their life or dysfunction that's occurring, they know there's a problem that needs to be dealt with. And then there are people who may actually suspect they have ADHD, because they've read something in a newspaper article or they've seen something on television or a friend told them about it. Most people are outside the medical system. Then inside the medical system there are people who are already seeking a diagnosis, meaning that they know they have a problem and they are going into a medical healthcare system, either seeing a primary care provider or a specialist to help them understand what their problem is and what can be done about it. When we think about people inside the system, you've got people who are misdiagnosed, people who are not yet diagnosed, or people who are correctly diagnosed; and then within the correctly diagnosed there are people who are treated appropriately and people who are not treated. So what the modeling process does is model that flow of patients through that system.

Medscape: Over time, what you look at as sources of change in that model are both patient outreach and physician education. Both had an effect on the number of people who would receive appropriate diagnosis and treatment.

Dr. Faraone: The model used 2 main variables. One was the patient activation, which meant that the patient became aware of their disorder and therefore sought treatment, and then there was the physician effectiveness -- how good were physicians at making the diagnosis. With those 2 variables, we looked at a few different outcomes. We started with 2007 data, and assumed that if there were no intervention at all, you have roughly 5.7 million patients with undiagnosed ADHD out of the whole US population who were outside the medical system, and if you made no changes at all in patient activation over the next 5 years, in 2012 you would have roughly the same number of people outside the medical system, so there would be no improvement.

If you made changes in physician effectiveness alone, that wouldn't change the number of people outside the system because you're not changing patient awareness. Physician effectiveness is only changing how the doctors are. To get patients inside the system, you have to change patient activation, making people more aware of ADHD with educational programs or outreach programs to patients and families to make them understand the nature of the disorder.

We also looked at how you could change the rate of patients actually seeking diagnoses, and over 2007 to 2012 as we're modeling it, you could increase the number of patients seeking a diagnosis by activating patients to think more about their disorder. One of the most important outcomes was how many patients are correctly diagnosed. And what we found is that as you increase both patient activation and physician effectiveness, both of those variables together would increase the number of correctly diagnosed patients. For example, starting with our population base in 2007, where there would roughly be about 2.3 million correctly diagnosed patients, that number would jump up to about 3.5 million if we implemented both physician effectiveness training and patient activation programs. It would increase with either of them, either patient activation or physician effectiveness alone, but not as much as it would if you used both of them.

So at the end of the day, to get more correctly diagnosed people into the system it is very important to have both physician effectiveness and patient activation programs. It's very important because what it tells us is that if you only try to activate patients somehow, for example a pharmaceutical company tries to do direct marketing to patients, and activate underlying base of the patients, it's not going to help -- the number of correctly diagnosed people will go up but that's only because more people are seeking help and you have to actually change the physician effectiveness to really dramatically improve the correct diagnosis rate.

Medscape: So the raw number would increase but the percentage wouldn't necessarily go up at the same time unless you implement some physician changes as well?

Dr. Faraone: Right. If you simply activate patients what you're going to have is more patients in physician offices, but you're going to increase both the number of correctly diagnosed and the number of misdiagnosed patients because all you're getting is bigger numbers of people face-to-face with the doctor. You actually need to work on physician effectiveness to improve the correct diagnosis.

Medscape: One of the conclusions of the research is that the physician education needs to precede the patient outreach so that the physicians are prepared to deal appropriately with the influx.

Dr. Faraone: Absolutely. You have to have doctors prepared for these patients coming into their office. What you don't want to do is have physicians flooded with patients that they don't know how to diagnose properly.

Medscape: Let's talk a little bit about some of the assumptions that you've made, some of the numbers that you worked with in terms of modeling an increase in number of patients seeking diagnosis, increasing percentage of correct diagnosis, etc. Some of those rates seem ambitious.

I'm thinking of the threefold increase in the number of patients seeking diagnosis.

Dr. Faraone: The model that we've run obviously is making assumptions as to how well we can activate patients and how well we can train physicians. We haven't actually studied that process. We do know that physicians can be trained, but whether that can be done on a population basis is unknown because there's been no research on that and it's a massive undertaking. So the bottom line from this report is not so much that we should take the absolute numbers to be meaningful in a sense that in 2012 we know this many people are going to be diagnosed with ADHD correctly. The research should be taken as evidence of how this population system works. If you turn up the temperature here, you have this effect downstream. If you also turn the temperature here, you have another effect downstream. As you might expect, the more you change activation and the more you change physician effectiveness, the better the outcomes are, meaning the more people are correctly diagnosed and the more people in doctors' office are getting the treatment that they need. One of the important results is that the accumulative effects are changing. One thing we're interested in, if you change the patient activation and physician effectiveness at the same time, would those effects be simply additive or would they be more synergistic, and would they multiply the effects. We found the latter: It wasn't simply an additive effect.

Medscape: Is the premise then that you can make clinicians more aware that they should be looking for ADHD in their adult patients, but that effect is enhanced if patients are coming into the office primed to say, "I think I may have signs of ADHD?"

Dr. Faraone: Absolutely. As patients become more aware, that will help because you're getting more people into the system. It's important that physicians are doing a good job making the diagnosis because some patients will come in thinking they have ADHD because they've heard about it, but you still have to do a complete full workup because some of those patients won't actually have ADHD. They'll think they do just because it's something they've read about and they really have an anxiety disorder. So it's extremely important for the physicians to do a good differential diagnosis because the psychiatric disorders tend to co-occur and you want to be sure you diagnose the correct disorder.

Medscape: Is your model based on an assumption of a single point of intervention with the patient population and physician population, or is it based on the assumption of repeated exposure and different kinds of exposure?

Dr. Faraone: It's repeated in the sense that the patient can move through the system in different ways. For example, you start with a patient who is unaware of having an issue at all. Some of them become aware of it and go see a doctor. The others who don't become aware of it, they go to a different pool of people who are unaware of the diagnosis. They then get sent to the awareness pool because over time they get exposed to patient education again, they have a second opportunity. They continue to have opportunities to become aware. Patients might be misdiagnosed for example, or patients who are diagnosed might see a primary care physician and be sent to a specialist, or some might just simply get diagnosed and treated. People can go through different pipelines of potential, different clinical trajectories where they have continued to have the opportunity to become diagnosed or become self-aware of a diagnosis or become misdiagnosed. For example, a patient who initially is misdiagnosed might eventually become correctly diagnosed. There is feedback between the different systems.

Medscape: Were any of the premises for intervention based on smaller studies of actual effects of physician education or patient outreach over time? Have there been studies that would help us gauge how much real success might be possible here?

Dr. Faraone: The different parameters of the model were based on the literature, broadly speaking. For example, we made the assumption that the roughly 15% of people would go from diagnosed and untreated to diagnosed and treated, and that was based on a Gallup study on ADHD and we estimated the rates from that.[5] Each of the model parameters were estimated from different datasets. It wasn't like a single study looking at this. It was us going into the literature and saying, "What's the best guess based upon what we know in the literature?" For example, we look at the time to diagnosis by a primary care physician, which we assumed was 3 years vs the time to diagnosis by a specialist, which we assume was 1 year. We took that data from my papers that were published in The Archives of Internal Medicine where we had done a survey of primary care practitioners and psychiatrists and asked them lots of questions about their treatment practices and the experiences of their patients in their practices.[6] Essentially all the parameters came from different studies but in academia what happens is that sometimes, it may be a very big problem and too big for 1 study. People look at different pieces of the pie. We took these different pieces of the pie, put them together, and created a theoretical model of how the population ought to behave.

Medscape: What are some of the interventions that have been done to influence improved prescribing and diagnosis for physicians?

Dr. Faraone: I'm not sure where to start with that question.

Medscape: Let's start with the patients. What sort of things have been done to study what will increase the likelihood that patients will seek diagnosis?

Dr. Faraone: This is part of the problem, there's not a lot of work done in the ADHD world. From Ron Kessler's study on treatment failures and treatment delays, we pulled out that of patients who are aware that they have a problem, roughly 5% of those people will go seek diagnosis.[7] We say seek diagnosis, which obviously means they will see a physician about it. It's actually pretty low if you think about it. There are lots of people who are having trouble basically in their lives in some ways, and they don't seek help for it in any way at all.

Another of our parameters is people who already suspect they have ADHD, and they go and seek diagnosis, and that's only 6%. I'm thinking that was based on Ron Kessler's data.[7] It's still a very low rate of people, I would have thought those numbers would be much higher than 5% and 6%. I would have thought that people, particularly people who expect they have a treatable disorder, maybe at least half of them go and seek help for it, but the answer is actually very few of them go and seek help.

Medscape: On the physician side, what sort of interventions have been studied for improving recognition and diagnosis and then treatment?

Dr. Faraone: I don't think anybody has ever actually studied specific interventions in the sense of trying to do an intervention and to say "How has this actually changed physician behavior?"

I can see how someone would do it. Typically there are many education programs for physicians around the topic of ADHD in adults and also in of course children, but I don't think there is anybody who has published an academic paper saying these physicians went through this training and at the end of the training, their skills were improved in this specific way. I don't know of any study that's actually done that. There are hundreds of education programs I'm sure that help physicians learn about ADHD, but some of those will have little test at the end of them I suppose. I've never actually seen any published results.

I did a paper with Len Adler on training in the context of a large clinical trial. The doctors in this trial were being trained to specifically use an ADHD rating scale to diagnose ADHD and we did show the effectiveness of the training, they did improve their ability to diagnose patients.[8]

Medscape: So some of the changes in behavior that you use in the model on the physician side, the decrease in misdiagnosis, reduction in time to diagnosis, etc, are maybe a little more theoretical than the patient population.

Dr. Faraone: It's theoretical in the sense that we believe that physicians need to learn a few things about ADHD. They clearly need to learn how to diagnose properly, and that's relatively straightforward. But they also need to understand the importance of making a diagnosis, and knowledge of that among physicians varies a lot, particularly between specialists, psychiatrists, vs primary care physicians. If you go back a decade, the majority view among primary care people was that ADHD basically was a childhood disorder, and they didn't need to be concerned about it. You didn't see very many primary care docs working with adult ADHD at all; it wasn't done. After lots of research and education and so forth, my guess is that's changed. I couldn't cite a paper that says that that's changed, but certainly we do see more, probably in part because of more continuing education programs for primary care doctors and more research published in primary care journals so the word is out now. Even though we didn't model this in this work we're talking about today, another important aspect of this prospect is that physicians themselves had to become aware of the disorder because some of the misdiagnoses were simply related to lack of awareness. If you're not aware of a disorder you don't diagnose it or you make the wrong diagnosis.

The other aspect of helping physicians with awareness is not just understanding that the disorder occurs in their practice, but that it's a disorder that's associated with a fair amount of disability, dysfunction, and distress -- that it's a disorder that is worthy of treatment, so to speak. They have to learn that these people are at risk for traffic accidents and more likely to engage in risky behaviors, and because of that there is a heightened risk, for example, for sexually transmitted diseases, they're more likely to get involved with drug abuse, and so forth.[9-11]

It's not to say that every person with ADHD has this , but it's important to remember that ADHD is a risk factor for some very substantial impairment that needs to be addressed with treatment. Part of training doctors involves making them aware of the real-life impairment associated with disorder instead of it being just an abstract of disorder of inattention, hyperactivity, and impulsivity, with meaningful effects in a person's life.

Medscape: It seems like based on some of these you said you could also extrapolate outward about changes in diagnosis of substance abuse or anxiety disorders or some of the other comorbidities that show up.

Dr. Faraone: That's right. We didn't for example address comorbidity -- you could try to model that in here as well.

Medscape: Thank you for your time.