Biological Aging & The Octo Tool!

We have talked before about how biological age often gets talked about as a simplified number, but it’s more complex than that, as we can age in different ways at different rates, for example:

• Visual markers of aging (e.g. wrinkles, graying hair)
• Performative markers of aging (e.g. mobility tests)
• Internal functional markers of aging (e.g. tests for cognitive decline, eyesight, hearing, etc)
• Cellular markers of aging (e.g. telomere length)
• …and more, but we only have so much room here

For more on that (including what we can do about each of them to slow or in some cases reverse biological aging), see:

Age & Aging: What Can (And Can’t) We Do About It?

So with that in mind, let’s examine…

The Octo Tool

It sounds like either something a Spider-Man villain would have, or possibly a middle-aged barbecue dad’s birthday present. However, it is neither of those things!

A team of researchers (Dr. Shabnam Salimi et al.) looked at data from several longitudinal studies combined (total n=42,683) and created a statistical model that can assess biological age and predict disability, geriatric syndrome, short physical performance battery, and mortality with ≥90% accuracy.

How? The tool looks at eight health measures using simple info from physical exams and lab tests. Instead of just focusing on single diseases, it takes a big-picture view of the whole body and how different organs are aging.

The idea is based on what the researchers call “health entropy”, in other words: how much wear and tear your body has experienced over time. By looking at how different organ systems are affected, the researchers created a set of aging clocks to reflect how quickly or slowly someone is aging on the inside.

You may be wondering: what are the eight health measures?

And the answer is: it’s a secret 🤭

Ok, it’s not really, and we are going to tell you the eight health measures. But, the pop-science article that brought this to our attention mysteriously did not mention what they are:

New tool uses eight health metrics to track biological aging

…which made us curious too; why would you use that headline and then not say what the 8 health metrics are?

Of course, we at 10almonds are the sort to read actual studies, not just press releases, so naturally our next stop was the paper itself:

Health octo tool matches personalized health with rate of aging

…which also does not make it very clear; look, here’s the abstract, which makes no mention of the 8 health metrics either:

❝Medical practice mainly addresses single diseases, neglecting multimorbidity as a heterogeneous health decline across organ systems. Aging is a multidimensional process and cannot be captured by a single metric. Therefore, we assessed global health in longitudinal studies, BLSA (n = 907), InCHIANTI (n = 986), and NHANES (n = 40,790), by examining disease severities in 13 bodily systems, generating the Body Organ Disease Number (BODN), reflecting progressive system morbidities. We used Bayesian ordinal models, regressing BODN over organ specific and all organs disease severities to obtain Body System-Specific Clocks and the Body Clock, respectively. The Body Clock is BODN weighted by the posterior coefficient of diseases for each individual. It supersedes the frailty index, predicting disability, geriatric syndrome, SPPB, and mortality with ≥90% accuracy. The Health Octo Tool, derived from Bodily System-Specific Clocks, the Body Clock and Clocks that incorporate walking speed and disability and their aging rates, captures multidimensional aging heterogeneity across organs and individuals.❞

In fact, not only does that not mention the 8 health metrics, it speaks of examining disease severities in 13 bodily systems.

If you scroll down a bit in the paper, you’ll even see their visual abstract including a body clock with 13 “hours” on it, representing these systems.

Later in the paper, it mentions finding 11 of these systems to be critically relevant for the purpose of the calculations, but, counting carefully on our fingers here, we find that’s still 3 more than 8.

So what gives, did they shave another 3 off?

No!

In fact, the answer is found by reading in much more detail, where we find the formula for the statistical analysis:

Fit_BLSA = (bodn ∼ mo(Hypertension) + mo(congestiveHeartFailure)

+ mo(IschemicHeartDisease) + mo(Arrhythmia) + mo(Kidney)

+ mo(Diabetes) + mo(Hyperlipidemia) + PrepheralArteryDisease

+ mo(Stroke) + mo(Anemia) + Thrombocytopnia

+ mo(GastrointestinalDisease) + mo(Liver) + mo(COPD) + Asthma

+ mo(OralHealth) + mo(Hypothyroisism) + Hyperthyroisism

+ mo(OsteoArtheristis) + mo(Osteoporosis) + mo(Hearing) + mo(Eye)

+ mo(Depression) + mo(sParkinsons) + mo(Cognition) + Cancer

+ yrs + (1 + 1|id), data = BLSA)

Body Clock_BLSA = posterior_predict(fit_BLSA, data = BLSA)

Note: mo = the monotonic function for the ordinal predictor

And there we have it, those are the 8 health metrics! Each of the 8 metrics is actually a composite of several others.

Maybe, dear reader, you do not love advanced mathematics, and/or do not wish to take the time to format all your personal data into the required numerical representations, apply the formula from the study, and then do various kinds of Bayesian analysis of the results, to get the promised predictions.

There are, then, two things you can do:

1. pay close attention to each of those items you see in (e.g. hypertension, oral health, depression, etc), and simply use the information we provide at 10almonds to keep them all in as good condition as you can
2. watch this space, because the research team is also working on creating an app that everyone can use that will do the math for you (in all likelihood that app will be free, and/but in return, will invite you to opt in to become part of the future study participant pool, effectively doing crowdsourced data science)

Want to get started already?

For pointers, check out:

6 Lifestyle Factors To Measurably Reduce Biological Age

Take care!