By David Tuller, DrPH
A Brief Update: Berkeley’s crowdfunding period closed on April 30th–Monday night. I ended the campaign with $87,580. After Berkeley’s 7.5% in fees, the funds will cover my salary/benefit from July 1, 2018 to June 30, 2019, and some travel costs. I really, really appreciate the fantastic support. Thanks to everyone! I’ve taken a few days to regroup from my Australia trip and catch up on my time zones.
Leonard Jason is a professor of psychology at DePaul University in Chicago. He has served as vice president of the International American Association of CFS/ME and as chairperson of the Research Subcommittee of the U.S. Chronic Fatigue Syndrome Advisory Committee. Professor Jason began investigating chronic fatigue syndrome almost 30 years ago. Much of his work has focused on the epidemiology and prevalence of the illness and on the impact of using various case definitions. He has long been concerned that the lack of a uniform set of criteria for identifying study participants has hindered progress in the science.
Dr. Jason recently shared his thoughts about these issues. (This Q-and-A has been edited for clarity and length.)
How common is fatigue?
If you were to ask people right now if they are “fatigued,” which means feeling weak, tired, or lacking energy, about 25% of the population would say yes, so this symptom is very common. In contrast, “chronic fatigue” means that a person has had fatigue for 6 or more months. Only about 4-5% of the population has chronic fatigue.
There are multiple reasons for people to be fatigued–for example depression, anxiety, over-exertion, people working three jobs, medications, sleep deprivation, weight problems, poor diet, inactivity, and deconditioning. These are just a few of the many causes of fatigue and chronic fatigue.
Physicians see lots of people coming into their practices, where the patients are seeking help for their fatigue, and in fact it is one of the most common reasons for seeing a doctor. But it’s very hard for many physicians to differentiate complaints of general or chronic fatigue versus the illness known as ME [myalgic encephalomyelitis]. Yet it is of critical importance to make a differential diagnosis between those with purely chronic fatigue versus those who have ME. In fact, it is this failure to differentiate these two conditions that has caused so many problems, and the culprit is a flawed and imprecise case definitions as well as failures to gain an international consensus for one research case definition.
So what is a case definition, and why are there different research and clinical definitions?
A case definition is a set of rules that helps a researcher or a clinician make a decision about whether someone has a particular illness or does not have the illness. It’s that simple. A good case definition is critical for the assessment process, to identify those people who actually have an illness or disease. It is the cornerstone of medicine.
A research case definition tries to identify a homogeneous group of people who have the illness and can be recruited for research purposes. In contrast, a clinical case definition is used to identify or diagnose a broader group of patients for treatment purposes, and many of these wouldn’t qualify for research studies. For example, if someone is very obese, a research case definition might exclude that person because the weight issue could be causing the person’s problems. In other words, for research purposes, we want to select only patients who do not have other psychological or medical conditions that could be causing the illness we are studying.
For science to progress, the research case definition is critical, as it can standardize the selection of patient samples so that research groups around the world are all studying the patients with the same disease. So gaining consensus among international scientists for a research case definition is a most critical task, and one that unfortunately has still not been accomplished for our field.
One of the parameters that’s important for a research case definition for this illness, in your view, is that psychiatric co-morbidities should be excluded. Can you explain the reason for that?
Yes, and let me give an example that illustrates this issue. A patient with a major depressive disorder with melancholic features would probably have fatigue, aches and pains, as well as sleep and cognitive problems. Yet these are also symptoms of ME, so some clinicians and researchers could easily confuse these two conditions. But they are very different illnesses, as people with a major depressive disorder feel self-reproach, whereas those with ME do not. If you ask people with a major depressive disorder what they would do tomorrow if they were well, they would not be sure. In contrast, if you asked people with ME what they would do if they were well, they’d give you a long list of all the things they have wanted to do but been unable due to their illness.
If you are studying ME, you need to exclude people who have a primary psychiatric disorder from your study. If researchers misclassify people with a major depressive disorder as having ME, this will have serious negative consequences for identifying biomarkers, estimating prevalence rates, and determining outcomes of treatment trials. The issue of selecting patients who really have ME is the most important issue facing our field. In a sense, the lack of a consensus on a ME research case definition is like building a pyramid of playing cards with a very shaky bottom, and then everything built on top of this foundation is vulnerable to collapsing.
Let’s start with what is the broadest case definition that has been used, the so-called Oxford criteria for CFS. Can you describe that and explain why it presents a problem?
If you have six or more months of fatigue, then you meet this case definition, so it’s a very broad category. Clearly, as I mentioned earlier, a lot of people who meet this criteria have medical or lifestyle reasons causing their fatigue. One of my students, Madison Sunnquist, just published her master’s thesis that indicated how the CBT theoretical model only works if you identify people with a very wide case definition, but when you have a better and more restricted case definition that requires core symptoms of ME, then the CBT model no longer works. In contrast to the CBT approach, my research group for the past 20 years has been doing research on what we call the energy envelope. But this pacing approach is not a cure, just a strategy to help better cope with ME. Our approach involves helping patients to better monitor their energy levels, learn how to stay within their energy envelope, and sustain lifestyle changes that involve reprioritizing activities.
So how did the CDC come up with the Holmes and then the Fukuda case definitions?
The Holmes case definition came out in 1988. The CDC investigators had gone to Incline Village and ultimately named this illness CFS. Their first case definition included too many symptoms. In fact, to meet their case definition, a patient would have needed to have eight or more symptoms out of a list of 11. But here is the problem that soon emerged–if you develop a case definition that requires so many unexplained somatic symptoms, you have a very high probability of unwittingly selecting people who have a somatoform disorder. And you don’t want to select people who have a purely psychiatric condition.
So in 1994, the Fukuda case definition was developed to replace the Holmes definition. For the 1994 case definition, the authors selected eight of the symptoms that had been listed in the Holmes criteria, and a patient needed to have any four of those eight symptoms to meet the new Fukuda case definition.
But here is the problem with the Fukuda CFS case definition–patients are not required to have post-exertional malaise, cognitive problems and unrefreshing sleep, and as we know, these are core symptoms of ME. So, a person could have four of the eight Fukuda symptoms and be diagnosed with CFS, and not have any of the three critical symptoms. In that case, you would be including in your sample a person who does not have the core elements of the illness.
From 1994 and on, I have been doing research that shows some of the diagnostic problems with the Fukuda case definition. And remember, the Fukuda case definition is the research case definition that has been used throughout the world for the past 25 years. But this Fukuda case definition identifies a heterogenous group of patients, because core symptoms are not required of all patients. So, as a consequence, samples of patients with CFS based on Fukuda case definition vary widely in different research groups and labs.
What is the impact of the case definitions on prevalence rates?
In the late 1980s and early 1990s, the CDC conducted a prevalence study where they started by asking physicians in four cities to identify patients they thought had CFS. At that time, a lot of physicians didn’t believe CFS existed, so putting physicians as gatekeepers in the selection of patients for this study resulted in a prevalence rate that was very low. Also, many people in the US do not have the financial resources to have a physician, so relying on primary care doctors to identify patients was another reason for low prevalence rates. The study suggested that CFS was a rare disease that affected fewer than 20,000 people in the US.
At that point, a group of researchers in Chicago began working on a study that involved finding patients from a random community sample, rather than a sample referred from physicians. In 1995, with NIH funding, our Chicago research team conducted a community-based prevalence study, which found that about a million people in the US had CFS. We also found that CFS affected all ethnic and socioeconomic groups, and thus we helped shatter the myth that CFS was a “Yuppie Flu” disease.
What did William Reeves [then-head of the CDC division in charge of the illness] do with the so-called “empiric” criteria? And why did this increase the CDC’s estimate of disease prevalence by a factor of 10?
In the early 2000s, Bill Reeves felt there was a need to operationalize the Fukuda case definition. For example, he tried to standardize the way we measure a patient’s disability or a substantial reduction in functioning. He used one instrument that has been referred to as the SF-36. According to Reeves, if a patient met criteria for one of several sub-scales within the SF-36, the patient would meet the disability criteria for having CFS.
But one of these domains was “role emotional” functioning. It turns out that every person with a major depressive disorder meets the criteria for “role emotional” functioning. So you can’t just specify instruments such as the SF-36; you have to specify which sub-scales of the instruments you are going to use, and what are the cut-off points. And if any of these choices are wrong, you will identify people who have another illness. My team gathered data on this point, and we conducted a study that assessed people with major depressive disorder, and found that over one-third of them could be inappropriately classified as having CFS under the so-called Reeves empiric criteria.
So, I think in the attempt to operationalize the Fukuda criteria, Reeves made mistakes, and I believe that is one of the reasons the estimated CDC prevalence estimates increased ten-fold, from .24% in a 2003 sample to 2.54% in 2007. They operationalized the Fukuda criteria in a way that classified many people as having CFS when they really had other illnesses.
At that time, many thought this increase in prevalence figures that Reeves proposed was constructive as it suggested that far more people had the illness, and thus these findings could be used to argue for more attention and funding due to this illness being so widespread. But if you use a very broad criteria, and bring into the illness case definition people who don’t have the disease, then the entire research effort is seriously compromised. Fortunately, over the past decade, few researchers have used the Reeves way of operationalizing CFS.
What about the CCC and ICC criteria?
The CCC case definition for ME/CFS in 2003 was better because it specified key symptoms such as PEM. It was developed as a clinical case definition, and now it’s being used by several teams as a research case definition. With the 2011 ME-ICC, I have noticed problems, and in part this is due to them once again requiring too many symptoms that could, as with the Holmes criteria of 1988, bring into the ME category some individuals who have a primary psychiatric disorder. In addition, the ME-ICC criteria is complicated to use, and many clinicians and scientists will have a difficult time reliably using it with patients.
What is the problem you see with the IOM case definition, apart from the name?
Well, it is true that Systemic Exertion Intolerance Disease (SEID) is a name most patients dislike. However, the IOM report was correct in requiring several core symptoms, such as PEM. But I believe these authors made a mistake in indicating that a patient could have either cognitive impairment or orthostatic intolerance—one or the other. Cognitive impairment should have been required for all patients to have. But a more serious problem is that they inadvertently expanded the case definition by having just about no exclusionary illnesses, such as primary psychiatric disorders. My team recently conducted a study where about half the people with a variety of medical and psychiatric illnesses met the IOM criteria.
Now the IOM criteria was developed as a clinical case definition, but there was no federal effort to develop a research criteria that selects a more homogenous group of patients. The failure to develop an international consensus on a research case definition means that many researchers will continue to use the problematic Fukuda case definition, or they might use the IOM clinical criteria to select patients for research purposes, and this process has already begun.
To summarize, for research purposes, if a person has the core symptoms of the IOM definition, it would be important to exclude those with a primary medical or psychiatric condition, but this is not what the IOM authors recommended. So, the clinical IOM case definition once again over-identifies people as having the illness. That means what occurred with the Reeves criteria of a decade ago has once again occurred with the IOM, as these criteria broaden the types of patients identified as having the illness.
What is at stake in this debate?
The stakes are high, for if you have an inappropriately wide case definition for research purposes, you will bring into your studies many fatigued people with a variety of conditions. In other words, if you identify the wrong patients, then your study will make conclusions about people who do not have ME, and you will have significant barriers to engaging in critical scientific activities such as estimating accurate prevalence rates or identifying biological markers. Also, if you bring in lots of people who don’t have this illness but lifestyle issues and/or a solely depressive disorder, a good percentage of them will respond favorably to psychogenically oriented treatments. As I have been writing about for many years, this will ultimately lead to some researchers making conclusions about CBT and GET that are not true for patients with ME.
My case is simple. You need to have one research case definition that is used by scientists throughout the world. The clinical case definition can be broader, but the research case definition has to be tightly focused on those with the illness so that results can be replicated in different laboratories. This scientific achievement has been accomplished with every illness or disease except for ME.
We can do better. After working in this area for almost three decades, I am confident that we have the tools and methods to use psychometrically sound procedures to develop a consensus on one research case definition. I am optimistic that one day this will occur, and for me, there is literally nothing as important for our scientific field.