Thank you @ng0rge. The following points come to mind as I search not so much for a predictively accurate aging clock ā although that would be valuable ā but for a way to protect, preserve, and if possible, extend my healthspan over what it would otherwise be were I not to intervene. I see these two goals more different than they might appear to some. I also think the second goal might be attained more easily than the first and, even there, only in the individual case (i.e., not for all of us). Following are some points offered less as certainties than to share my thinking and invite other points of view.
As do most of us, I have a few relatives who lived active lives well into their 90s. Some of these relatives lived to that age with one or typically several health metrics suggestive of a much shorter life span. My paternal grandmother who died just short of 100, had high blood pressure for her last 50 years of life (recall that the threshold for that diagnosis was much higher when she first received it, in the 1950s I think). Yet, she outlived several family physicians even though she generally chose not to medicate herself for her hypertension. One mentor and friend (not a relative) was a high functioning intellectual until a matter of weeks before his death at 95. He was overweight for much of his life, was not particularly fit, and had T2D at least for his last 40 years. Writing this reminds me that I had a great uncle having the same general profile as my mentor friend who died of a coronary at age 50. I have a good friend with poor eyesight due to a lifetime of working in front of a computer screen and poor hearing due to unprotected machine gun and mortar fire in Vietnam. How do we assess his sensory metrics in relation to longevity? I have high VO2Max for a male in his 70s but the structural CV remodeling caused by many decades of competitive distance running beginning in college may (or may not) account for my VO2Max more than my genetic predisposition. I also have substantial CAC and a very low resting heartrate coupled with a rapid rate of recovery from exercise. How should we account for those diverging metrics, or my mentorās unfavorable metrics, in an aging algorithm? Would improving any of these metrics improve my health- or lifespan? Do we know?
Because we lack the robust forms of validity that would come from assessing the current crop of aging and longevity metrics ā whether individually or factor analyzed for clusters ā against the hard dependent variables of mortality or a loss of functional health, we are trapped necessarily in a kind of circularity where we look for intercorrelations among competing metrics to assess against a proxy measure for a real dependent variable. The more sophisticated analysts might look for statistical factors, alpha coefficients, or correspondence indices, but their hunting will be hampered by not having a true dependent variable.
We sometimes assume that any slowing of a so-called aging clock metric will result in increased health- or lifespan. We take that assumption as a matter of logic. I think the question is more empirical and only logical insofar as we construct operational definitions. Even if we improve our condition by improving certain metrics, will the same apply to everyone else in the same way, to most of us, or to a select few of us. As I said, the question may have a much larger empirical component than we think and we have few empirical answers to it.
Thanks to those who bore the expense of testing themselves on many aging metrics, including so-called biological aging clocks, the large and sometimes contradictory lack of correspondence among these metrics suggests that we are a long way from possessing a type and level of understanding that would generalize to actionable principles for us, much less be applicable to all or most of us.
If we knew much more than we know today, we might construct and algorithmically derived index of individual metrics which, if improved through behavioral intervention, would correlate highly with health- or lifespan. However, I think such an algorithm would of necessity be dynamic taking the form, āIf metric X is greater than n, then metric Y becomes relevant and loads in at a coefficient of 0.62ā and so on for many other metrics defining complex interactions. I also think the final product might apply mostly to us and not necessarily to anyone else. In other words, I think the search for modifiable metrics that will or are likely to increase our health- and/or lifespan is a personal one, at least until we arrive at a future where such a dynamic algorithm will demonstrate predictive, discriminant, and convergent validity.
How would we begin if we wanted to focus on developing a set of personal aging or lifespan metrics that we believe will have decent predictive validity (barring being run over by a truck)? I am proceeding by ranking each of my personal health metrics along a few dimensions including how far the metricās value departs from optimal (if such is known), the degree to which it can be modified, the effort required to modify it, where if at all it shows up in my family history, ditto for genetic profile, and current thinking about the probable return from modifying the specific metric. (One general area I am ignoring for now, is the so-called aging clock metrics such those based on methylation. I am not denying the potential value of these metrics. My decision is based on a few considerations including but not limited to the lack of consistency and concordance in this general pursuit, including its commercial products.) By placing these metrics and ratings in a spreadsheet (all in standardized scores, the raw data is meaningless in this context and would result in unintended weighting) it seems possible to visualize the data in various ways that support a data-driven plan of action. Moreover, this plan of action can be tailored to be consistent with how much I am willing to take on at the moment and it can be reassessed if I have more or less time available at some future time. This plan will change frequently as we learn more as a community and through the scientific research.
I have more thoughts on this topic but this seems like a good place to stop for now.
