Last year, a study came out showing that beef jerky and other cured meats, could trigger mania in bipolar disorder (paper, popular article). It was a pretty big deal, getting coverage in the national press and affecting the advice psychiatrists (including me) gave their patients.
The study was pretty simple: psychiatrists at a mental hospital in Baltimore asked new patients if they had ever eaten any of a variety of foods. After getting a few hundred responses, they compared answers to controls and across diagnostic categories. The only hit that came up was that people in the hospital for bipolar mania were more likely to have said they ate dry cured meat like beef jerky (odds ratio 3.49). This survived various statistical comparisons and made some biological sense.
The methodology was a little bit weird, because they only asked if they’d ever had the food, not if they’d eaten a lot of it just before becoming sick. If you had beef jerky once when you were fourteen, and ended up in the psych hospital when you were fifty-five, that counted. Either they were hoping that “ever had beef jerky at all” was a good proxy for “eats a lot of beef jerky right now”, or that past consumption produced lasting changes in gut bacteria. In any case, they found a strong effect even after adjusting for confounders and doing the necessary Bonferroni corrections, so it’s hard to argue with success.
Since the study was so simple, and already starting to guide psychiatric practice, I decided to replicate it with the 2019 Slate Star Codex survey.
In a longer section on psychiatric issues, I asked participants “Have you ever been hospitalized for bipolar mania?”. They could answer “Yes, many times”, “Yes, once”, or “No”. 3040 people answered the question, of whom 26 had been hospitalized once, 13 many times, and 3001 not at all.
I also asked participants “How often do you eat beef jerky, meat sticks, or other similar nitrate-cured meats?”. They could answer “Never”, “less than once a year”, “A few times a year”, “A few times a month”, A few times a week”, or “Daily or almost daily”. 5,334 participants had eaten these at least once, 2,363 participants had never eaten them.
(for the rest of this post, I’ll use “beef jerky” as shorthand for this longer and more complicated question)
Power calculation: the original study found odds ratio of 3.5x; because the percent of my sample who had been hospitalized for mania was so low, OR = RR; I decided to test for an odds ratio of 3. About 1.2% of non-jerky-eaters had been hospitalized for mania, so I used this site to calculate necessary sample size with Group 1 as 1.2%, Group 2 as 3.6% (=1.2×3), enrollment ratio of 0.46 (ratio of the 921 jerky-never-eaters to 2015 jerky eaters), alpha of 0.05, and power of 80%. It recommended a total sample of 1375, well below the 2974 people I had who answered both questions.
Of 932 jerky non-eaters, 11 were hospitalized for mania, or 1.2%. Of 2042 jerky-eaters, 27 were hospitalized for mania, or 1.3%. Odds ratio was 1.12, chi-square statistic was 0.102, p = 0.75. The 95% confidence interval was (.55, 2.23). So there was no significant difference in mania hospitalizations between jerky-eaters and non-eaters.
I also tried to do the opposite comparison, seeing if there was a difference in beef jerky consumption between people with a history of hospitalization for mania and people without such a history. I recoded the “beef jerky” variable to a very rough estimate to how many times per year people ate jerky (“never” = 0, “daily” = 400, etc). The rough estimate wasn’t very principled, but I came up with my unprincipled system before looking at any results. People who had never been hospitalized for mania ate beef jerky an average of 16 times per year; people who had been hospitalized ate it an average of 8 times per year. This is the opposite direction predicted by the original study, and was not significant.
I tried looking at people who had a bipolar diagnosis (which requires at least one episode of mania or hypomania) rather than just people who had been hospitalized for bipolar mania. This gave me four times the sample size of bipolar cases, but there was still no effect. 63% of cases (vs. 69% of controls) had ever eaten jerky, and cases on average ate jerky 15 times a year (compared to 20 times for controls). Neither of these findings was significant.
Why were my survey results so different from the original paper?
My data had some serious limitations. First, I was relying on self-report about mania hospitalization, which is less reliable than catching manic patients in the hospital. Second, I had a much smaller sample size of manic patients (though a larger sample size of controls). Third, I had a different population (SSC readers are probably more homogenous in terms of class, but less homogenous in terms of nationality) than the original study, and did not adjust for confounders.
There were also some strengths to this dataset. I had a finer-grained measure of beef jerky consumption than the original study. I had a larger control group. I was able to be more towards the confirmatory side of confirmatory/exploratory analysis.
Despite the limitations, there was a pretty striking lack of effect for jerky consumption. This is despite the dataset being sufficiently well-powered to confirm other effects that are classically known to exist (for example, people hospitalized by mania had higher self-rated childhood trauma than controls, p < 0.001). This is an important finding and should be easy to test by anyone with access to psychiatric patients or who is surveying a large population. I urge other people (hint to psychiatry residents reading this blog who have to do a research project) to look into this further. I welcome people trying to replicate or expand on these results. All of the data used in this post are freely available and can be downloaded here.