Robert Wack

You’ve heard me vent about disinformation and how science is easily misrepresented and misunderstood, not because it’s hard, but because of the channels we receive information through. We also have a challenge with our own human limits dealing with ambiguity, complexity, and some counterintuitive aspects of reality that science illuminates.

Examining and evaluating the “how” of consuming information is a never ending chore. Our biases are often hidden from us, and it’s only after we make a mistake of some kind that those blind spots are revealed.

An article came across my screen recently which illustrated these challenges in a variety of ways. The American Academy of Pediatrics (AAP) has a news feed where they highlight interesting articles relevant to pediatric practice. This past week I saw this headline:

“Urine Test Might Help Identify Children Likely To Have Autism Earlier Than The Best Assessment Tools Now Available, Study Suggests”

Before we dig in further, check out the qualifiers they loaded into this headline – 

• “might” help identify
• children “likely” to have autism
• “study suggests”

These are the signs of responsible science journalism. Lots of conditionality, lots of cautionary qualifiers. It basically says, “don’t get too worked up, but this might have interesting implications”. It caught my eye because of the mention of a urine test, which implies they think they have some kind of metabolic or biochemical marker for autism. Hmmm….what could that be about?

So I clicked through to the linked article, from a publication called Health Day, and found this headline:

“Urine Test Can Detect Autism, Study Says”

Whoa, hold up a sec…. that’s not what the AAP says the study is about. The Health Day article is by a science journalist named Dennis Thompson, and Health Day is a multimedia company which appears legit, staffed by reporters and editors with a clearly stated editorial standard and track record in journalism since 1998. The fact that the AAP feels comfortable linking to their stuff adds to their credibility. The article has direct quotes from the researchers and does a decent job explaining the actual work, an evaluation of a urine test that tracks metabolites of gut bacteria that MAY BE linked with autism. Left unstated in the article are several important questions:

• how are the researchers defining “autistic”?
• what is the mechanism behind the purported connection between gut microbe metabolites and the clinical features of autism?
• who are the people doing the research, and why?

To his credit, Dennis Thompson has a statement at the bottom of the article: “SOURCE: Arizona State University, news release, May 26, 2026”

This may indicate that he didn’t actually speak to the scientists quoted in his article, but rather was regurgitating what was in the ASU press release. So I went and tracked that down. Yup, he was just restating the original press release. He may not have actually spoken to anyone. Still, he appears to have exerted some care in what he included, but it’s an incomplete picture. Of note, the ASU press release concludes with some very cautious language from the researchers about the history of this line of inquiry, the preliminary nature of their findings, and the need for further research, replication, and confirmation of their findings. Good stuff.

So I went to the actual article, in Molecular Psychiatry, a Nature journal and this is the title of the study:

“Elevated microbially-derived metabolites in autism: a possible diagnostic screening test for a distinct ASD phenotype”

Ah, okay, this is more like it – cautious, qualified research-speak. Here are the relevant qualifiers in this title:

• “a possible diagnostic screening test”
• “for a distinct ASD phenotype”

Much better, but still, what are they talking about?

First, some background on the gut, autism, and how the urine test might be useful. Autism and gastrointestinal issues (constipation, diarrhea, abdominal pain, and eating disorders) have long been known to be associated. This is a great review article summarizing that association.

Unfortunately, it has also been the jumping off point for a lot of junk science and outright grifting by those preying on the families of people with autism. The King of Autism Scams, Andrew Wakefield, predicated his attacks on the MMR vaccines, the launch point for the anti-vax movement theory of autism, on a spurious link between the vaccine and “gut inflammation”.

This is why my spider sense was jangled by the initial assertion that this urine test was connected to the gut biome. Uh-oh, here we go, I thought. But as I dug in, maybe not. We need to go further.

Researchers have been looking at gut bacteria (the gut biome) for decades, gradually building a case that there are important connections between the millions of bacteria living inside our bodies and the function of multiple organ systems. It’s a fascinating area of research that will likely produce big discoveries, but it’s early on, and definitely not soup yet.

The gut biome and autism have an extensive literature. This review article has a mountain of references addressing that research. But there’s a major challenge: are the gut biome findings the CAUSE of autism, or the RESULT of autism? For example, the altered eating behaviors that manifest very early in the evolution of autism symptoms could just as easily be the reason the gut biome is altered in people with autism rather than being the cause of those behaviors. It’s a chicken and egg problem, common in scientific inquiry, which requires careful analysis to figure out.

Another challenge is figuring out who to study. The diagnostic category of Autistic Spectrum Disorder has come under recent criticism that it is now so broad as to be almost meaningless. No less than Dr. Uta Frith, an early pioneer in autism research, is now raising concerns (paywall) about the overly broad use of “autism” as a diagnosis for things that may explained by anxiety, personality differences, social awkwardness, and sensory sensitivities. The Psychology Today article linked above quotes Frith from the interview, saying: “… the autism spectrum has become so broad that it is “close to collapse” and has “lost all meaning” as a medical category (Amass, 2026).”

This is an example of the age old “lumping versus splitting” problem in science that goes all the way back to Darwin and Linnaeus’s early classifications in the animal kingdom and fossil record. Most recently, we see it in data science with the challenge of data elements, dividing continuous variables into subsets, and loosing specificity vs. creating unnecessary complexity. If you want to try to find some key marker for a group, how do you define that group, and who do you compare them to for ensuring the marker is real and reliable?

Back to our study with the gut biome, a key question is how did they define patients with “ASD” in this study? What were the inclusion and exclusion criteria? What other variables did they look at in their analysis? They could have done everything right and still have missed something important. More importantly, if they were very conservative in their definition, even if their results hold up and are replicated, the utility of that test is almost certainly going to be minimal for the individuals who are part of the overly broad current definitions of ASD. This is why they likely chose the wording “a distinct ASD phenotype” (a phenotype is a fancy way of saying the list of characteristics of a group or diagnosis) in the study title. 

Last, who are these people, the researchers, and why are they working on this? Here’s the part that gets murky. Almost all of the researchers on this paper have some commercial interest in developing autism testing technologies. You find this at the verrrrrry end of the paper, not quite in fine print, but not the most upfront disclosure. You be the judge:

“JBA, CKF, RKB, JH, PW, and KC have patents and/or pending patents related to the diagnosis of ASD. JBA and RKB are co-founders of Autism Diagnostics LLC. JBA is the founder of Autism Diagnostics Lab Inc. PW and KC are co-owners of Analutos. JBA and RKB are co-founders of Gut-Brain-Axis Therapeutics Inc.”

Analutos is the company that performed the secondary confirmation of the metabolites in the urine specimens. 

In the Methods section, they mention this proprietary system for doing the data analysis:

A novel multivariate analysis called the Microbially-Derived Metabolite System^TM, or “The MDM System™, was developed as a diagnostic screening tool to identify the subset of children with ASD who have intestinal dysbiosis.

Any time you see the trademark symbol ™ in a science paper, look very, very closely as to why. Which of the many commercial entities the authors are involved with owns “The MDM System”? Can’t tell.

This is a good example of the intersection of capitalism and science, where the incentives to commercialize research lurk in the background as a possible corrupting influence. I’ve got nothing against researchers making a buck off their work, but don’t loose sight of the possible problems resulting from the temptations of money, as well as the pressures those researchers are exposed to from their sponsors, the University hosting their labs and the investors paying for their stuff. Everyone is human, and everyone has their breaking point when enough pressure is applied.

Let’s bring this home: this long post isn’t an attack on this particular paper. This looks to be legitimate science by legitimate scientists. The takeaway points:

• beware simplistic headlines about cutting edge science research;
• if you really want to understand how legit some piece of research is, you have to poke at the strengths and weaknesses of the study, which means really digging into the assumptions behind how they put the study together;
• if the media source you read something in doesn’t do this, get out your salt shaker;
• what’s the impact? This study appears to add to the already large body of data building a case for a possible association between gut dysbiosis (the gut bacteria misbehaving and causing problems) and autism, but association does NOT mean causality. It could just as easily be the other way around. This study doesn’t do anything to answer that question;
• except for the possible utility of “The MDM System” ™ for confirming the diagnosis in people ALREADY diagnosed, the paper is almost a big “so what”, which means it could be construed as merely an advertisement for the technology.
• back to where we started, the headlines we initially looked at do NOTHING to communicate any of these complexities. If you are only looking at headlines, or even the top line summaries, you are almost certainly going to be mislead, either inadvertently or intentionally.

This post first appeared on Your Angry Pediatrician on Substack.