ME: Counting Heads

Data Brief Number 488

Copyright 2024 Hillary Johnson. All Rights Reserved.

Last December 2023, the Centers for Disease Control issued a new prevalence estimate for ME in the United States. According to the CDC, 3.3 million people have been diagnosed with the disease by a doctor. This is a significantly revised number, adding nearly one million more cases of ME to the CDC’s long-standing official estimate of 836,000 to 2.5 million.

The Atlanta agency’s earlier estimate came from the Institute of Medicine’s (IOM) report on ME, published with fanfare in 2015.

Notably, in 2015 the IOM’s estimate was sixteen years old. That’s because IOM reviewers lifted the 836,000 to 2.5 million estimate from a 1999 paper by DePaul University psychologist Leonard Jason. The IOM authors failed to clarify that the prevalence numbers they cited were already quite old. Because of inadequate epidemiology on ME they likely were at a loss for more recent or more reliable numbers.

Nevertheless, the 836,000 to 2.5 million prevalence figure is a quarter-century old.

The CDC’s new, broader estimate works out to be 1.3 percent of the US population. In other words, more than one person out of every hundred is sick with ME in the United States.

For perspective on the government’s new prevalence figures, it’s worth noting that at the height of the U.S. polio epidemic in 1952, the attack rate of paralytic polio was .17 percent. The total number of paralytic polio cases in 1952 was 20,000. This attack rate data covers the incidence of polio during a specific year. Also, typical of most infectious diseases, in the case of polio there were orders of magnitude more people infected with the polio virus who had either no symptoms or extremely mild symptoms.

The CDC’s Head Counts of Yore

The CDC’s new estimate of ME prevalence is newsworthy, and not only for its significant upgrade of the head count. For decades, the CDC has obstinately refused to set up a clinician reporting system for the disease and has rarely made a serious effort to count heads even though Congress repeatedly asked it to do so when the disease became epidemic in the 1980s.

The CDC’s recent prevalence figure is 330 times it’s first estimate in 1992.

In May of that year, the agency released an estimate in response to a query from a member of the house appropriations subcommittee hearing. Illinois Congressman John Porter, who represented a tony suburban region outside of Chicago, asked a direct question of CDC Director William Roper.

“How many Americans suffer from CFS?”

In response, the CDC director told Porter, “A preliminary estimate…is that approximately ten thousand persons over age eighteen in the U.S. meet the CDC case definition of chronic fatigue syndrome.”  

Walter Gunn, then CDC’s principal investigator for the disease, privately scoffed at the number. “Totally incorrect, off by a factor of ten,” he told me, adding with mock gaiety, “Nobody’s perfect!”

“It’s the greatest underestimate of all time,” ME expert Paul Cheney, MD. commented to me when he heard the number. “It means that, certainly, I have all the cases in the state of North Carolina in my practice. I spoke to a support group in Monroe, Louisiana recently. Four hundred patients came. There were people being pushed into the auditorium in wheelchairs. Now, the disease is there, in a small town, by the hundreds if not the thousands. And there may be a thousand times that amount with subclinical illness.”

In 1996, as a result of revelations in Osler’s Web, both the U.S. Congress and the investigator general of the Department of Health and Human Services investigated the CDC for its “misappropriation” of all ME research funds sent to it by Congress over the previous eight years. William Roper was the CDC director when my reporting was corroborated in 1998. Roper gamely explained that agency staff spent the money on “more important diseases like measles and polio.” Roper must have experienced significant blowback because the next day, he retracted that comment and issued an apology.

How they Did It

The CDC counted doctor-diagnosed cases of both “chronic fatigue syndrome” and ME. The new estimate is in line with a 2018 prevalence study, which also looked at doctor-diagnosed cases of “chronic fatigue syndrome” and ME. In that study, investigators—who included familiar names like clinicians Lucinda Bateman, Charles Lapp and Peter Rowe—employed machine learning techniques to look for cases of disease in a cohort of fifty-million Americans. The study relied on Optum, a medical claims processing database.

The Optum database included people up to 89 years of age, but excluded Medicaid patients. The years covered were 2011 to 2017. The authors found “…a rough estimate” of between 1.7 million to 3.38 million for both diagnoses [combined],” and 2.8 million for ME exclusively.

The Medicaid exclusion may have led to an undercount. Demographic data have repeatedly indicated that ME is more common among low income populations. Which came first—the disease or poverty—remains an open question.

CDC used survey data from 2021 and 2022. Interviewers posed their questions to residents of a total of 57,113 households. In person interviews were conducted approximately half the time, the rest were undertaken by telephone after the Covid pandemic began. CDC pooled the data from both years.

The interviews were a tiny part of a much larger enterprise, the National Health Information Survey, or NHIS. Established in 1957 and headquartered in Hyattsville, Maryland, the NHIS has sought information about Americans’ health, mental health and behaviors every year for nearly six decades. Different centers at the CDC can choose to “sponsor” questions on this survey. It’s a relatively cheap way of acquiring information.

A question arises: is self-report a more or less accurate method than using AI to mine enormous data bases of medical records? So far, I don’t have the answer. In the meantime, the results of both efforts were reasonably close. The machine learning technique estimate’s high number was 3.8 million ME cases, compared to CDC’s 3.3 million.

“Are you now—or have you ever been…”

Staff at CDC arrived at its new data by attaching two questions to the National Health Information Survey:

• “Have you ever been told by a doctor or other health professional that you had Chronic Fatigue Syndrome (CFS) or Myalgic Encephalomyelitis (ME)?”

• “Do you still have Chronic Fatigue Syndrome (CFS) or ME?”

The agency published its new prevalence figure on a CDC website, the National Center for Health Statistics, as Data Brief Number 488.

Compared to What? (Tryin’ to Make It Real)

Helpfully, the CDC’s new estimate clarifies the widespread nature of ME for people who were unaware. Compared to prevalence estimates for several other neurological diseases, ME is vastly more common. (“Prevalence” refers to the number of people living with a disease at any given time.)

• In the U.S., the number of people with Multiple Sclerosis is less than one million.

• The Parkinson’s Foundation estimates one-million people are living with Parkinson’s in the U.S. today.

• The Myasthenia Gravis Foundation estimates there are between 36,000 and 60,000 cases in the US, although apparently that disease, like ME, is underdiagnosed, too.

• Amyotrophic Lateral Sclerosis (ALS) afflicts 30,000 US residents at any given point in time.

• According to CDC, 2.5 million people sustain a traumatic brain injury (TBI) per year. That’s not a prevalence number, it’s an “incidence” number. No one knows how many people are coming down with ME in any given year, nor has CDC ever wanted to know. However, I mention TBI because some researchers compare ME brain damage to traumatic brain injury.

Presumably, the diseases just listed are not maladies in which a majority of patients are undiagnosed. Their diagnostic criteria do not rely on preposterously subjective symptoms like “unrefreshing sleep.” Even post exertional malaise, once ballyhooed as the hallmark of ME, is seen in Gulf War and Long Covid using identical technology. It’s possible that post exertional malaise—"malaise” landing on my ears as another non-descriptive euphemism like the overused “brain fog”—may even turn out to be a generic sequelae of other infections.

Another way to appreciate the 3.3 million number is to compare it to geographical populations. The new estimate is 670,000 more people than inhabit Chicago, the country’s third largest city after New York and Los Angeles.

(If one relied on the higher end of the 2018 machine learning model—or 3.8 million—then the number of people with doctor-diagnosed ME or “chronic fatigue syndrome” exceeds the population of Chicago by 200,000.)

The CDC’s revised estimate is also more than double the population of Phoenix.

The government’s number also amounts to one-third more people than comprise the entire US military establishment world wide, which is 2,079,142.

That’s an extraordinary prevalence for a disease that at its worst results in lifetime disability, an extreme risk of suicide, and, based on a sample of just 56 patients, appears to result in premature death. That 2016 study was undertaken by DePaul’s Jason, who cautioned in his conclusion: “The results suggest there is an increase in risk for earlier mortality in patients with ME and CFS. Due to the small sample size and over-representation of severely ill patients, the findings should be replicated to determine if the directional differences for suicide and cancer mortality are significantly different from the overall U.S. population.”

As far as I know, no further investigation of premature death has been undertaken by anyone.

The Great Undiagnosed

For several reasons, as large as the new prevalence number might be, it’s still an undercount.

CDC understates the problem in its data brief, “As the prevalence estimates in this report are based on a doctor’s diagnosis, adults who are undiagnosed are not included in this analysis. True prevalence estimates may be higher, as previous studies suggest that many people with ME/CFS are undiagnosed.”

Seeking clarity, I asked the first author on the data brief, Anjel Vahratian, whether the CDC had adjusted the prevalence estimates to account for undiagnosed patients, and her answer didn’t surprise me. How to factor in a negative?

She responded to my query via email, a correspondence coordinated by a CDC press intermediary.

“We only capture self-report[ed] doctor-diagnosed ME/CFS,” Vahratian confirmed.

The data brief avoids hinting at the potentially staggering number of undiagnosed patients, but the CDC publishes a figure on one of its ME website pages: “More than 9 in 10 people with ME/CFS have not been diagnosed by a doctor.”

CDC again appears to be relying on the 2015 IOM report for this reference. The IOM authors noted that “An estimated 84 to 91 percent of patients affected by the condition are not yet diagnosed.”

The IOM cited two sources for its claim. Once again, De Paul’s Jason, is one. His 2006 article, called “The Face of CFS in the United States,” offers the 84-91 percent estimate. The IOM’s second source was a 2004 paper by the late William Reeves of the CDC, based on his agency’s random digit dialing telephone survey of Kansas City, Kansas. Obviously, the literature on undiagnosed cases is limited.

I spent three days trying to continue writing this post at this juncture. I’ve since put aside the pages I generated. I’ve written it all before, for different venues and in different decades.

There is a reason most ME patients go undiagnosed, and it has nothing to do with the disease or the patients. It has everything to do with the government’s failure to commit an appropriate research budget to solve it.

The flaws in multiple diagnostic criteria are many, mostly having to do with their use of vague symptom requirements and a lack of biomarkers. But the biggest problem is that the U.S. government has insisted on controlling diagnostic criteria, taking the matter out of the hands of scientists and doctors and putting it in the hands of bureaucrats. I have written a great deal about the ease with which the government historically has dispensed with important studies performed by academic researchers by complaining about which criteria were used and patient selection. Ironically, the NIH and CDC are the unconscious—or conscious—masterminds of the confusion.

In fact, on that particular issue, I recently noted the ease with which the NIH authors of the recent “deep phenotyping” study of seventeen ME patients dismissed virtually all research that’s been done prior to their own in their opening paragraphs:

“A major obstacle to rigorous ME/CFS research is case assessment due to the absence of a diagnostic biomarker. Over 20 different diagnostic criteria underscore the difficulty in defining the clinical symptom content of ME/CFS⁷. The usual symptoms are non-specific and overlap with other diseases, hence misattribution is frequent…” they wrote.

This subject matter has festered for so many years that, for the moment, I’m putting aside the age-old controversies about diagnostic criteria because the topic is an entire chapter of its own, one that includes interwoven narratives of bio-politics, extreme bias, woman-hating, incompetence, Orwellian bureaucracies, and a wholesale absence of human empathy—with a few influential nut jobs thrown in at critical junctures.

I’ll just point out, probably unnecessarily, that the vast majority of doctors are either clueless about the existence of ME or have no idea how to diagnose it. In addition, doctors frequently dismiss ME wholesale or chalk symptoms up to personality disorders of women. The number of undiagnosed cases may be uncertain, but it is most likely huge. Far too many ME sufferers are being left out of the calculus.

Iceberg Epidemiology

Icebergs, too, reveal just ten percent of their mass to the untrained eye. One might ask, wouldn’t a tax-payer funded agency with a mandate to prevent disease want to know more about this vast uncounted population and undertake efforts to make it visible? That would involve an investment of money, however, a rescaling of the government’s commitment to ME and a massive physician education effort. Above all, it would demand a definition with actual biomarkers, not to mention a more serious, retooled effort at epidemiology.

The CDC clearly has zero intentions to fund any of that, nor does NIH, whose scientists are insisting “difficult” ME patients owe them thank you notes for NIH’s first ME study in three decades. The CDC, in particular, has settled comfortably into a role in which it maintains cleanly-designed websites written in robotic AI-style prose, notable for promoting the status quo. The agency is unashamed to point out that as many as 80-90 percent of patients are undiagnosed, but its policy makers propose nothing by way of remediating this decades-long tragedy.

Nonetheless, it’s interesting that CDC seems less afraid to admit that a lot of people have ME than it’s been willing to admit in the past. But why now? What is CDC up to this time?

The timing of this effort speaks volumes.

The ME questions on the annual health survey were sponsored by the National Center for Zoonotic and Emerging Infectious Diseases (NCZEID). ME has held an awkward place within this center for decades considering the agency has never given an inch on matters of infection or emergence, staunchly denying both for decades. The disease itself has been reduced to fine print on a web page within this center titled, “Chronic Symptoms Following Infections.”

When I asked the NHIS’s Vahratian why the agency decided to add these questions to the survey now when the CDC has had several decades to do so, she directed me to page 70 of an online PDF about the survey, where I found the following sentence:

“There are few national population-based prevalence estimates of CFS/ME. Additionally, the COVID-19 pandemic has increased the need for national prevalence estimates because post-infection CFS/ME symptoms have been reported to be associated with COVID-19 (Bansal, Gubbi and Koch, 2022). Estimates of lifetime and current CFS/ME can provide data to assist in planning for the care of individuals living with this condition.”

Given this imprecise language, it’s unclear which of the two “conditions” CDC thinks might be assisted by prevalence data, but because of recent history I’m guessing it’s Long Covid.

Judge for yourself whether ME or Long Covid is the point of interest:

The timing of the questions, coming not in 1990 but during the first full year of the Covid pandemic, suggests the concern is Long Covid.

In addition, the agency added the same two ME questions to the 2023 and 2024 survey, the results of which have not yet been reported. When was the last time the CDC pursued ME prevalence four years in a row?

The conflation of ME and Long Covid continues apace. Where it stops, nobody knows. I may already be writing about a chimera when I reference “ME.” Will it continue to exist as a discrete entity in the future as the machinery churns, shredding decades of scientific literature as it goes?

That said, there is a final upside to the CDC’s prevalence estimate: for every ME case identified, the agency had access to a wealth of demographic data. The statistics go beyond gender and ethnicity. They are telling us something new and important about ME, something that predates Covid by decades.

Stay tuned for Part Deux.

Copyright 2024 Hillary Johnson. All Rights Reserved,