AI Drinks from the Pool of Public Medical Information 





By Gary M. Pepper, M.D.

40% of the U.S. population uses the internet to obtain medically related information equating  to over 1 billion Google searches daily in this category. At the same time, artificial intelligence is rapidly assuming a dominant role in the search industry. Traditionally, in response to a query for medical information,  a search engine would produce a prioritized list of sources leaving the user with a choice of who to believe.  In distinction, the AI agent generally produces a pre-digested summary of what it perceives as the relevant facts scraped from the pool of public medical data. This underscores the importance of keeping the public data resource  pristine. One practical way an AI agent can differentiate good versus bad sources of medical information is by sorting each contributor by their assigned  “authority” quotient determined by an unbiased referee (think Google).  But what happens when a high authority participant pollutes the information pool? Who will clean up that mess?

During my 40 years as a physician it became increasingly necessary to attend to individuals who had ingested poor quality information.  Be assured, this is a draining interaction. The following is a current precautionary example of how a single top ranked  medical source  can jeopardize the integrity of the information pool. 

The Source of the Problem

The source in question is the venerable New England Journal of Medicine (NEJM). A Google search lists the NEJM as possessing the highest impact score of all medical journals. In their January 18th, 2024 edition, the lead article is entitled, “Testosterone Treatment and Fractures in Men with Hypogonadism”. The publication of the study (the TRAVERSE study)  was not an earth shattering development. What signals the importance of the report,  by definition, is the article’s top position in this prominent medical journal compounded by the surprisingly controversial and counter-intuitive nature of the findings.

Hypogonadism refers to the deficiency of sex hormones, testosterone in men and estrogen in women. For men with low testosterone it is common practice to prescribe testosterone supplementation, the treatment being studied in the NEJM report. In many prior studies, men similar to that of the TRAVERSE study receiving testosterone, showed robust bone improvement which implies improved bone strength. The findings in the TRAVERSE study were startlingly the reverse of this expectation, as the men receiving testosterone had a higher incidence of bone fracture than the placebo group. Crucially, standard radiologic measurements  of bone density for the groups were absent from this report.

The primary focus of the TRAVERSE study was to determine if giving testosterone to a group of older men with low testosterone levels (hypogonadism) increases the risk of cardiovascular events such as  heart attack or  stroke. The results published in the NEJM about bone fracture risk were secondary or a spin off of the primary purpose of the study. Secondary findings are generally regarded as less reliable and harder to draw valid conclusions from than primary results.

The Yellowish Ring

There are additional deficiencies in the study as published in the NEJM, some of which were pointed out in an editorial in the same edition of the same journal. Given the extraordinarily high profile granted to this information by the journal it is not at all surprising that a confusing headline soon followed in a prominent on-line medical source, “Testosterone Increases Fracture Risk in Hypogonadal Men? “ So expands the yellowish ring in the public information pool.

Underlying Factors

The testosterone product used in the NEJM report, Androgel, is marketed by the study’s corporate sponsor Abbvie.  The study’s backing by a pharmaceutical giant with vested interests suggests it may be more of a marketing endeavor than pure science. Additionally, the financial challenges faced by print journals may lead them to prioritize revenue from pharmaceutical industry advertising over scientific rigor. The linking of these interests is a growing and powerful threat to the health of the medical information ecosphere.

A Possible Solution

If a top medical journal conducts itself in this manner, how can we guard the integrity of the public pool of medical information? One thought is to separate sponsored research into a different pool from independent and less conflicted forms of research. This would help ensure that AIs draw from a cleaner source, Until then, we depend even more on the pipeline gatekeepers, to  remain vigilant to new sources of pollution within the medical information pool. Could a retooling of the algorithms for assigning domain authority, be in order? (Disclosure:  Since the 1980’s , Dr. Pepper has been a subscriber and infrequent contributor to the New England Journal of Medicine). 

Boyfriend Has Low Testosterone. What Can Lu Do?

Lu posts these concerns to

My boyfriend is in his early 40′s and has been taking testosterone therapy. Instead of his levels increasing, they have decreased…his total is now in the single digits. He takes very good care of himself as he is a fitness trainer and body builder (takes vitamins, etc.). Obviously, with his total level being in the single digits, he has all the “symptoms” of low-T and is frustrated that the therapy is having a reverse reaction. He also suffers from Migraines and has recently been in a car accident that he suffered brain trauma in. I’m wondering if there could be a connection between the trauma and low-T or lower T. Any advice or direction you can head us in would be much appreciated.

In reply Dr. Pepper writes:

Hi Lu

You can’t pour water into a cup and wind up with less water in the cup then what you put in. Likewise, if someone takes testosterone supplement they will have more testosterone in their body then they started with. However, some things can influence the blood levels so one person will have higher or lower levels then someone else taking an identical dose. I have seen a wide variation in how testosterone gels are absorbed through the skin. These products include Androgel, Androderm, Testim, Axilron and Fortesa. One person may not see much of an increase in blood levels of testosterone on one of these gels while another will see levels zoom up to a 1000. Absorption of testosterone that is injected with a needle is less variable. Levels go very high in the first few days after the the injection but after 2 or 3 weeks levels will be low again. Here’s an important point. Since testosterone replacement turns off the body’s production of testosterone, if you stop taking replacement your body will not be making testosterone for weeks to months after resulting in very low levels on blood tests. People who abuse testosterone know this and will have the doctor check their testosterone level a month or two after their last dose, so the doctor will see the low levels and give them a prescription for more medication.

Can head trauma effect the testosterone level? For that to occur the pituitary gland would have to be damaged and that will often be associated with other obvious brain damage. In children less severe trauma can hurt the pituitary.

Hope some of this information is helpful in trying to figure out what is going on with your boyfriend. Good luck.

Gary Pepper, Editor-in-Chief,

Verified by MonsterInsights