Little to No Relationship Observed between Physical Health Outcomes and Mental Health Outcomes

Suicide is a good proxy measure for mental health [1] and life expectancy is a good proxy measure for physical health [2]. Intuitively, I’d expect suicide per capita to anticorrelate with life expectancy as both suicide and short life expectancy could be expected symptoms of poor health outcomes in general. I was drawn to this subject in particular because suicide rates have been going up in America for two decades [3], but only very recently has life expectancy begun trending downward as well [4]. Could mental health be a leading indicator of physical health, or could physical health be a lagging indicator of mental health?

Figure 1 dashed my hopes of a positive finding.


Figure 1: Suicide rate per 100,000 people vs. life expectancy over time in the UK, Poland, Hungary, Lithuania, and Japan. Coloration from blue to yellow shows rising socioeconomic status over time in all five countries. Based on free material from GAPMINDER.ORG, CC-BY LICENSE. [5]

There are a number of observations we can draw from this graphic, but the one we can state most confidently is that there is not a simple, linear relationship between the two variables. If suicide and life expectancy were linearly correlated then we would see linear paths with a non-zero slope over time. If one were a lagging or a leading indicator for the other, we would see ‘J’-shaped curves. The United Kingdom’s linear path from left to right suggests there is little to no relationship between mental health and physical health, and the erratic up-and-down motion of Hungary, Lithuania, and Japan suggest this observation is not confounded by hidden dependencies. Poland, a near neighbor to Hungary and Lithuania with a comparable suicide rate in the modern day, has a straight-line path from left to right like the UK’s.

I’d like to add a disclaimer here to emphasize that absence of evidence is not evidence of absence, which is why I’ve gone out of my way to use hesitant language. It’s even possible, though unlikely, that a direct causal relationship exists in the absence of correlation due to unknown confounding variables, as in the co-occurrence model of the Flynn effect in psychometric theory [6]. This is only a preliminary observation.

I initially restricted the analysis to European countries to keep as many variables constant as possible. I chose the particular countries to observe the effects of socioeconomic status. The UK is a prosperous nation, Lithuania is and has been a very poor country, and Poland has been a very poor country that has experienced significant recovery after Communist rule. I included Hungary to observe whether cultural differences between Eastern and Western Europe (most notably, Communism) would confound simple socioeconomic differences, but was surprised to observe that Poland was much more UK-like than Hungary-like. Japan was added at the end to dispel concern that variance around a country’s mean may be a statistical artifact of small sample sizes, and to illustrate an intriguing sine wave pattern in suicide rates in some countries.

Including all countries in the analysis (Figure 2) reinforces this negative finding.


Figure 2: Suicide rate per 100,000 people vs. life expectancy over time for all country data available through Gapminder’s bubble chart tool. Low average income to high average income is displayed as a gradient from blue to yellow. Based on free material from GAPMINDER.ORG, CC-BY LICENSE. [5]

Life expectancies rose linearly with average income over time while suicide rates hovered around a fixed average for each country, with some countries being very steady (e.g. Poland, the UK) and others significantly varying around their averages (e.g. Hungary, Lithuania, Japan), and in several cases even appearing to be a sine wave (e.g. Japan). The latter note may interest those with a taste for cyclical theories of history.


[1] L. S. Too, M. J. Spittal, L. Bugeja, L. Reifels, P. Butterworth, and J. Pirkis, “The association between Mental Disorders and Suicide: A systematic review and meta-analysis of record linkage studies,” Journal of Affective Disorders, vol. 259, pp. 302–313, 2019.

[2] M. C. Stiefel, R. J. Perla, and B. L. Zell, “A healthy bottom line: Healthy life expectancy as an outcome measure for health improvement efforts,” Milbank Quarterly, vol. 88, no. 1, pp. 30–53, 2010.

[3] J. Elflein, “Suicide death rate U.S. by gender 1950-2018,” Statista, 01-Jul-2021. [Online]. Available: [Accessed: 12-Nov-2021].

[4] E. Arias, T.-V. Betzaida, F. Ahmad, and K. Kochanek, “Provisional life expectancy estimates for 2020,” NVSS Vital Statistics Rapid Release, vol. 10, Feb. 2021.

[5] “Gapminder Tools,” Gapminder. [Online]. Available: // [Accessed: 12-Nov-2021].

[6] M. A. Woodley of Menie and H. B. F. Fernandes, “Do opposing secular trends on backwards and forwards digit span evidence the co-occurrence model? A comment on gignac (2015),” Intelligence, vol. 50, pp. 125–130, 2015.

About Aeoli Pera

Maybe do this later?
This entry was posted in Uncategorized. Bookmark the permalink.

3 Responses to Little to No Relationship Observed between Physical Health Outcomes and Mental Health Outcomes

  1. Aeoli Pera says:

    NB: United States longevity is 2.9 years below what we’d predict, statistically

    Ref: T. Freeman, H. A. Gesesew, C. Bambra, E. R. Giugliani, J. Popay, D. Sanders, J. Macinko, C. Musolino, and F. Baum, “Why do some countries do better or worse in life expectancy relative to income? an analysis of Brazil, Ethiopia, and the United States of America,” International Journal for Equity in Health, vol. 19, no. 1, Nov. 2020.

  2. Aside, whatever happened to Woodley? What’s he doing nowadays?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s