Named Genetic Illnesses are More Harmful than Good

Great Debates
1

Let us assume that the human body falls on a bell curve, from the macro to the micro level. If nearly everyone has an averageish amount of serotonin, then a small fraction will have too much, another small fraction will have too little, and a smallish number of people will fall in between. And, in general, we would expect these values to be on a smooth continuum, not clusters. I.e., we would expect samples like this:

1,2,3,3,4,4,4,4,5,5,5,5,5,5,5,5,5,5,5,5,6,6,6,6,6,7,7,9

And not like this:

1,1,1,1,1,1,1,1,5,5,5,5,5,5,5,5,5,5,5,5,9,9,9,9,9,9,9,9

The way that we evaluate and treat disease however, you are either “diagnosed” or “not”. Even though reality creates us on a smooth continuum from “healthy” to “whacked out” we treat health as a boolean state of true or false.

Going further, we should expect to see our smooth bell curve manifest across every part of our body: Hormone levels, protein synthesis, t-cell production, and so on. If there are thousands of types of all of these things, added together, then some small percentage are fully out of whack, some larger percentage are somewhat out of whack, and so on. Some combination of these probably balance out - the body seems to have some amount of redundancy - but the total number of permutations when we’re talking about thousands of different chemicals and processes is sufficiently large as to call the potential space of disease infinite.

For example, let’s say that we have 1000 bioelements (steroids, hormones, killer cells, and so on) and assume each one has ten possible settings between too low and too high, with the highest (10) and lowest (1) settings being unhealthy. In this setup, there are 10[sup]1000[/sup] different permutations that your body could fall under and 2[sup]1000[/sup] of those will be unhealthy. That’s a small percentage of everything but it’s still so many bad setups that you could never name them all.

If you look at anything, from depression to lupus, they all present in different ways and the diagnostic will say “if more than 5 of these 10 symptoms present, then the patient has this disease”. A person with depression will have to try a half dozen medications to find the one that works for her, with the first 5 having done nothing or made things worse. If someone could have a zero intersection set of symptoms with another person, in their diagnoses for the same disease, and need treatment with an entirely different medicine - then do they really have the same disease? If one of them comes across as “depressed” because their serotonin level is off, another comes across as “depressed” because their dopamine level is off, and yet another comes across as “depressed” because they have a lithium deficiency - then these aren’t the same disease in any way beyond something very superficial.

Having a named disease leads the doctor towards a diagnosis on “is the person depressed”? Minus the name, and the doctor might simply test the patient’s bioelements and treat based on what is too high and what is too low.

Drugs tend to be targeted to a specific bioelement. They raise Il-6 or block citric acid production (whatever it might be) - just one target and one action - while we’ve already determined that we could have hundreds of different bioelements that are subtly or strongly out of order (whether other bioelements are compensating for them or not). A true treatment for any one person, logic would seem to say, should be more like a key - specifically modulating a wide variety of bioelements in a custom mix of increases and decreases. A good reason not to do that and not to treat the body in that way is because the scientific process tends us towards changing one variable and seeing what happens, rather than trying to go scattershot. But, in this case, that may well be the wrong approach and the reason that medicine has not had the same impact on longevity as basic sanitation and tends to create side effects that can be worse or more annoying than the original problem.

The not-so-good reason to try one thing and observe is that testing is still expensive - particularly if we are talking thousands of bioelements to check. But, should we have that ability, the need to test one variable goes away since we have a specific target to aim for across all variables. And, likely, we can narrow in on which subset to test based on symptoms even if not based on disease names.

Ultimately, naming illnesses - particularly when discussing those caused by genetic makeup - is unlikely to be as useful as simply testing the person’s body and correcting everything that looks wrong. Names sound nice but they mislead more than they help.

I am not a doctor and this is merely a topic for debate.

2

Why would we expect that?

3

Technically, most biomarkers seem to follow a skew curve, of greater or lesser skewness. E.g.:

https://images.app.goo.gl/swFbtcT6DpvFFPWQ6
https://images.app.goo.gl/kroXxf5Pt9k3fR8U9
https://images.app.goo.gl/HShobtKduPUnKnu68

I’m unaware of any biological test where everyone gets the same value - it’s always some sort of probability cloud. Possibly there might be some things out there that are bimodal but, if there is, I would tend to think that there’s some other bioelement that impacts the one we’re looking at, and it’s the other one that’s the cause of the issue.

4

Addendum: It’s plausible that, rather than a full bell or skew curve going off into infinity in both directions that we would have some hard caps where the curve ends in a cliff.

1/3rd of pregnancies are autoterminated and it’s possibly reasonable to think that it’s because the mother’s body determined that the baby was not going to be viable in some way. Similarly, we would expect the extreme ends of any bioelelement - where we’re talking about something directly correlated to health - to be self-removing as time continues. When we get most of our values by testing college students and adults, rather than fetuses and newborns, we’re liable to see fewer extreme values.

5

It would be lovely if we could measure human health by testing a bunch of bioelements and measuring how far they are from a defined optimum level. Maybe that will be possible one day, but given current medical science, you might as well say “Cars are dangerous, slow and polluting, we should just use teleporters”.

To take one example, serotonin and depression:
We can’t measure serotonin in the brain, and we don’t know how well blood serotonin correlates to brain serotonin.
We don’t know whether low serotonin causes depression, or depression causes low serotonin, or a third factor causes both, or some of each.
The seratonin level that is “too low” isn’t consistent from person to person, and we don’t know how to predict it in an individual.
We don’t have a good understanding of what serotonin does, and monkeying with it can have other effects on systems we don’t understand.

Until we can completely and accurately model the human organism and all its variables, basing diagnoses on only what can be quantitatively measured will just lead to results like classifying bodybuilders as obese because their BMI is too high.

6

False.

7

That is all reasonable and fair, but:

  1. That is a problem specific to a subset of all types of issues and so not a valid basis for denying the general concept just an encouragement to work harder to overcome that obstacle in that particular space.
  2. Serotonin is not the only thing which can cause depression. If we could test for everything except serotonin, for example, then the test would be to test everything except serotonin and - by confirming that they are all in the clear - we know that serotonin is the issue. Obviously, that’s not how it breaks down but eliminating the things which we can test for does narrow down on what the possible causes and remedies are and may lead to finding that the cause was something on this side of the blood-brain barrier or something which is correlated across the blood-brain barrier.
8

Then my understanding is incorrect. I would encourage you to give a better description of the current state of affairs beyond the one word.

9

I’m not a doctor but as I understand it you diagnose a disease based on a number of things (symptoms, personal history, family history, environmental factors) and ideally treatment is even more personalized. Some diseases are classified by pathogenesis, especially when you get in to specific cancers. Most are classified by symptoms or organ systems because that’s how the patient presents themselves for care. When making medical decisions, the doctor is supposed to go from the nominal diagnosis (eg: R05 cough) to a differential diagnosis (eg: J20 acute bronchitis) and then to specific treatment.

With genetic screening it’s all backwards. You can detect risk factors and in rare cases actual disorders before any symptoms present. It is usually irresponsible to diagnose a patient with a disease based on genetic tests alone. But it is a much worse idea to jump into general treatment for a disease like cancer just because the risk factors show up on a genetic test. At the very least, treatment for genetic diseases must incorporate family history. We’ve come a long way with medicine, but treatment for systemic diseases can be life changing and not always in a good way, even when the mechanism is known there are always risks and side effects.

A more reasonable approach is to tell the patient, straightforward, your genetic tests indicate that you are at higher risk than most for this kind of disease so we would like you to come in for screening more often. Here are the symptoms of the disease, here is information about it, you do not have the disease but you are at higher risk to develop it. If you notice X, Y, Z don’t panic just give me a call. Here are references for second opinions, remember that it is always your decision. etc.

~Max

10

I think he means something like this: Take a disease such as diabetes. You can have “normal” blood sugar levels but still have risk factors that warrant treatment, or you can have abnormal blood sugar levels but not require medication. You can be pre-diabetic, or have a mild case, or a severe case. There are two main types of diabetes, or maybe five, or seven, plus for example the eleven different genetic changes that cause Maturity Onset Diabetes of the Young (MODY), which can be confused with each other and with other types. It’s not as simple as “he has diabetes” versus “he does not have diabetes.”

Now that is different from saying, “he has a transcription variant on the KLF11 gene” or “she has karotype 47,XX,+21”; those are yes/no statements. However, the effects of those genetic variations can vary quite widely.

11

For both of you, I believe that you’re missing the distinction between an actual disease (e.g. cancer, diabetes, coronavirus, etc.) and “construction tolerances”.

Any healthy person can get diabetes, if they work for it hard enough. It’s not an issue with how they were constructed any more than slamming my car into a wall at 200 miles per hour somehow proves that it was made deficiently. Or, at least, that the Almighty One didn’t make us invincible isn’t my point.

Having gotten diabetes, it has a specific cause, a specific course, and a specific effect that is knowable and relatively uniform. Likewise, if I crash my car, we can say exactly what broke it, we can use physics to determine what effect that will have on the car, and we can say exactly what all needs to be done to make the car better again. Heck, I don’t even need to slam it into the wall for us to talk about what the result would be.

Whereas with construction tolerances, if we know that every single wire, bolt, hole, pipe, panel, and everything else could be arbitrarily different from its design by up to 5% in any imaginable way…can you look at it from 100 yards away and predict which tire is the most likely to fall off if the car started moving? Can you predict whether it will even start? Whether it will tend towards veering left or right? If it’s all just random, without any sort of rhyme nor reason for one car to behave in a particular way versus the next one off the line, and there’s no reason for any particular thing to be more nor less wrong than the next since we’re just talking about construction tolerances under chaotic rules, then does it make sense to apply names to some arbitrary set of results and then complain that not everything fits into your buckets?

12

I suppose the point I was trying to make is that treatment is supposed to be based on the actual disease and other patient-specific factors, not solely whatever this thing is that you call “construction tolerances”.

Your example of a psychiatrist making a broad diagnosis of depression based on say, self-answers to a depression questionnaire and then prescribing every antidepressant under the sun until something sticks - with the given rationale of ‘depression is a broad disease so let’s just try all the antidepressants’ - is particularly striking to me as bad medicine. If the decision-making is so shallow, we wouldn’t need to send psychiatrists to medical school.

~Max

13

I tend to think a lot of what the OP is talking about is just a consequence of how primitive our knowledge still is in a lot of ways. I mean, we didn’t get a good handle on what DNA was and what it did until the 1950s, and we’ve been constantly learning ever since. Not too long before that, we figured out what vitamins were. Vitamin C of all things, was discovered and explored around 1930. The Krebs cycle that details how our cells respire was discovered in 1937. Dopamine wasn’t discovered until 1957.

So in a lot of ways, we’re still in the early phases of learning how to identify and manipulate the body’s chemistry. What we do in a lot of cases is more along the lines of “That dumpster’s on fire; we don’t know if it’s a grease dumpster fire, or a paper dumpster fire, or something else we don’t even know about, so we’ll just use foam on it because we know it’ll put it out.” And it may not be the best thing for the parking lot, or for the grass nearby, but at least the fire’s out.

14

This is certainly not true of diabetes. It’s not even true that diabetes is a single disease, as opposed to a whole cluster of diseases with various causes, varying courses and outcomes, and symptoms that can also vary but have as a common denominator “blood sugar levels are chronically out of whack.”

No matter how hard your healthy person tries, for example, they’re not going to get Maturity Onset Diabetes of the Young type 2 unless they have a mutation on a specific gene. Even if they have that mutation, the course of the disease depends on a variety of factors, including their weight and diet, whether their mother also had the mutation (which can affect fetal development), whether they are heterozygous or homozygous for that mutation, what other mutations they might have, etc.

MODY 2 is a particularly good example here because it’s a situation where the body produces and uses insulin in the “normal” fashion, but the sensor that tells the body when to produce insulin doesn’t respond at the “normal” threshold, instead allowing blood sugar levels to rise to higher levels before insulin kicks in. Since insulin resistance is not typical of this form of diabetes, the medications commonly used for type 2 diabetes aren’t usually helpful, and depending on how much higher their threshold is, medication may not be needed at all.

Even for type 2 diabetes (the kind commonly associated with obesity), a complex interplay of genetics, lifestyle choices, and pure luck–there is for example research on how the kinds of bacteria in your gut influence the development of diabetes–determine whether and when a particular person develops diabetes. Having high blood sugar levels is associated with increased risks of heart disease, kidney damage, blindness, Alzheimer’s and other dementias, and so forth, but an increased risk is not an inevitability, and the actual course and effect can vary quite substantially from one person to another, even with similar blood sugar levels.

15

Why use a heuristic process instead of a more precise analysis of cause and effect? Is that your fundamental question here?

The answer is that in the field of medicine and especially psychological disorders, “precise analysis of cause and effect” is often beyond our current level of technology. And what it is not beyond our understanding is often beyond our budget.

Your criticisms are valid where there are feasible alternatives to more accurately diagnose and treat more specific diseases. Some physicians (possibly all doctors who do not work for insurance companies) think it is irresponsible to use antibiotic tablets as a go-to first line of defense when patients present with chronic rhinitic symptoms. Not all rhinitis is caused by bacterial infection, and there may be diagnostic procedures like endoscopy or CT scans that can help make a more accurate diagnosis.

Your criticisms are also valid where the heuristic process generates false positives and treatment involves significant risk - for example, it would be horrifyingly irresponsible to order chemotherapy or mastectomy solely because a genetic test exposed BRCA mutations.

But in the general case, our understanding is not such that we know every mechanism of the disease. Even if the mechanism is known for a particular disease, neither the diagnostic procedure nor the treatment is ever 100% accurate. We’re talking about people’s health, so it is important to understand that medicine is not a theoretical science but an applied science. There will always be margins of error and tolerances, and huge swaths of the field we have no idea what’s going on or why. Where the solution to a problem cannot be solved deterministically, it is reasonable to resort to heuristic techniques, so long as everyone understands the risks involved.

~Max

16

I don’t believe that there have been any studies that found a correlation between blood serotonin and depression in people that were SSRI naive. I think they found a mild to moderate correlation between serotonin levels and depression in women that had been previously treated for depression.

17

While a lot of the OP is valid, I don’t think the conclusion and title of the thread is anywhere near warranted.

Not all physical traits are continuous; the reason height, skin color etc follow a bell curve is because they are encoded by multiple genes. But on the other hand, many genetic illnesses are caused by a single gene, and you simply have it or you don’t (with sometimes a discrete case for heterozygotes).
So yes, we are often justified in treating diseases as binary.

For other cases, I just don’t see the harm being done by naming the illness. Knowing that a patient has a specific genetic illness means knowing exactly what to look for and what treatments should be put on the table.
If there are doctors that prescribe exactly the same dose for everyone diagnosed with Sage-rat’s Disease, regardless of pathology, then that’s a problem. But that’s a problem with that particular doctor or methodology, not the fact that the disease was given a name.

Or…do you mean this the other way? That, since the most obvious diseases to identify and name are the binary ones, that encourages us to think of all disease as binary? There’s a degree of truth to that but still the issue is not with naming the binary ones.

18

I think you meant to say that naming genetic illnesses are more harmful than good.

We kind of already knew (or considered it known) that the illnesses themselves are more harmful than good, that’s why we call them illnesses.

19

This is a consequence of the Central Limit Theorem. If a feature (like height) is the result of many independent genes and non-genetic factors, the distribution of the feature is going to be bell shaped. The details are important, but it takes fewer independent factors then you may think to get a bell shape. For example, the sum of three dice is clearly bell-shaped.

20

I understand that, but the OP seems to imply a stronger connection. For example (I’m not a doctor, so forgive me if I get the details wrong) there’s probably some type of bell shaped curve on insulin production levels across the population. But there will be a strong outlying group of people who have Type 1 Diabetes and produce no insulin naturally. There are 30 million people with diabetes in the US (not sure how many are Type 1) but it sounds like enough to impact the shape of the curve.

In general, I understand that the bell-shaped curve applies. But it seems that for many types of illness, the outliers are important and we all benefit from naming them and treating them as such.