How many of the fifteen men in the First Presidency and the Quorum of the Twelve are currently suffering from Alzheimer’s disease? Of course I don’t know the answer to this question. I can give you an estimate, though. Since I’ve been crunching numbers recently to predict which Q15 members might become Church President, I have all these data on their ages and life expectancies lying around, and given that age is a strong predictor of Alzheimer’s, I thought it would be an interesting exercise to match up the age data with an Alzheimer’s prevalence table to see what proportion of the quorum might suffer from it in the past, present, and future.
The major data sources I used are (1) ldsfacts.net for birth, calling, and death dates for historical Q15 member ages, (2) the simulations I did for my post last month on predicting who will become Church President, for future Q15 member ages, and (3) this paper from the journal Alzheimer’s and Dementia for the Alzheimer’s and dementia prevalence rates. If you’re interested, I’ve described the process I followed in more detail at the end of this post.
Here’s a graph showing the average age of the Q15 and the age of the Church President from 1835 to 2014 (taken at the end of each year), and predicted ages for the Q15 and for the Church President for the next 15 years. I calculated predicted ages in two ways, one using the SOA mortality table that I used for my post last month about predicting which Q15 members would become Church President (labeled “not adjusted” in the graph, with darker colored lines), and the other (labeled “adjusted,” with lighter colored lines) with the mortality rates in the SOA table multiplied by 0.89 because I found in analyses for another post that this provided better fit to actual historical mortality rates of Q15 members.
I’ve blogged about the age of the Church President and the Q15 before, so if you read ZD a lot, you may have already seen a graph that looks like the white part of this graph. The gray part of the graph is new, though. It shows predicted ages for the Church President and the average for the entire Q15. These are averages across 1000 simulations. The Church President’s age is predicted to continue to increase for a few more years because President Monson is likely to live for a few more years, but then it declines because when he dies, he is likely to be replaced by a younger man. The predictions for the entire Q15 go downward pretty quickly because several current members are old enough that their expected life spans are quite short, and when they die, it seems very likely that they will be replaced by much younger men. (In the simulation, when a Q15 member died, I replaced him with a new member who was between 52 and 67 years old. See more detail at the end of the post.)
Here’s a graph showing estimated Alzheimer’s prevalence. Like for the age graph, I’ve shown the Church President’s and the Q15 lines separately, and there are two predicted lines for each, one using the unadjusted SOA table, and the other using the 0.89 adjustment. Note that the values for the Church President are probabilities–the expected probability that he has Alzheimer’s–and the values for the Q15 are rates–the expected percentage of the quorum that has Alzheimer’s.
In the early history of the Church, the Q15 generally had very young men, so probabilities and rates are all zero. It wasn’t until near the end of the nineteenth century that any members got old enough for Alzheimer’s to become an issue. (Of course I’m assuming that Alzheimer’s prevalence was the same in the past as they are today, and I really have no way of checking this.) The President line has a big spike whenever the Church President reaches 80, and an even bigger one when he enters his 90s (e.g., David O. McKay in the 1960s). It’s not surprising that the Church President’s probability is almost always higher than the rate for the entire Q15, as the President is almost always one of the older men in the quorum. The prediction lines show the President’s probability increasing dramatically in the next few years, as President Monson is near 90, when rates go up even more steeply.
For the Q15 line, the expected rate has gone up dramatically in the past few years as by far the majority (9 of 15) are now 80 or older. The current estimate is about 12%, which means about 2 current members of the Q15 have Alzheimer’s (12% x 15 = 1.8). The prediction lines show a sharp dropoff, though, that corresponds with the age prediction lines. This is because when the oldest current Q15 members die, the newly-called members who replace them are likely to come from an age range in which Alzheimer’s is rare.
Finally, here’s a graph showing estimated dementia prevalence. It’s laid out the same way as the Alzheimer’s graph is.
The shapes of the lines look very similar to the lines in the Alzheimer’s graph, but the scale has shifted. Here the spikes for the Church President aren’t around 35%, they’re approaching 50%. The rate for the whole quorum has been above 10% more often than not in the past twenty years, and it’s currently above 15%.
There are a lot of things that could be wrong with the analyses I’ve done. The predicted lifespans from my simulation could be wrong. The ages at which I’ve assumed new Q15 members will be called could be wrong. The Alzheimer’s and dementia rates I used might not be the best estimates. But even if all the predicted parts are off and the real Alzheimer’s and dementia rates are (to be optimistic) lower, I think the bottom line would be the same. Because of the ages they’re living to, the First Presidency and Quorum of the Twelve are at substantial risk of suffering from Alzheimer’s and dementia. As medical science and technology get better, unless the very next thing to be solved is Alzheimer’s, they’ll be at even greater risk. For me, this just underscores the need to give Q15 members emeritus status at some point, and not ask them anymore to serve for life.
Here’s some more detail on where the numbers came from. Let’s start with the age data. The historical data I used are ages of all Q15 members at the end of each calendar year from 1835 through 2014. I got these from Q15 member birth, calling, and death dates at ldsfacts.net. The future age data come from the simulation I did for my post where I was using mortality tables to predict who in the Q15 would become Church President. For that post, I looked at 1000 simulations of how long each current Q15 member will live, and for each simulation, checked which members would get to become President because they had outlived all quorum members senior to them. For this post, instead of just looking at who would get to become President, I also used the yearly future ages for each of the simulations.
Of course it’s easy to predict a quorum member’s age at any particular time in the future assuming he’s still living. When the simulation said a quorum member would die at a particular age, though, I had to replace his age with the age of a replacement member. To choose ages for these replacement members, I looked back at the age at calling for Q15 members called in the last 25 years. They are distributed pretty evenly between 52 (Bednar) and 67 (Cook). Therefore, I assigned each replacement member to have 1/16 probability of being at each age from 52 to 67 in his first year. The replacement member’s age then progressed from year to year in the simulation. After one year, for example, a replacement member had a nearly 1/16 probability of being each age from 53 to 68. (The reason the probabilities weren’t exactly 1/16 is that the mortality rate at each age was taken into account, so there was a small probability that the replacement member would die and be replaced by another replacement member, who again would be started out with a 1/16 probability of being each age from 52 to 67.)
I think the way I handled adding replacement members to the Q15 is reasonable, but it could certainly be done other ways. It was for this reason that I only ran the simulations 15 years into the future for this analysis, rather than the 30 years that I did in the post on predicting who will become Church President. In that post, the analysis showed that there was a 50% probability that all current Q15 members would have died in 30 more years. I wanted this analysis to be mostly driven by the current Q15 members, and not much affected by simulated replacement members, so I chose to have it go only half as far into the future.
I took the Alzheimer’s and dementia prevalence rates from this 2011 paper from the journal Alzheimer’s and Dementia. I used this paper because a couple of the resources I found when searching online recommended the Aging, Demographics, and Memory Study (ADAMS), on which this paper is based, as a good source because it uses a nationally (US) representative sample of older adults. The Alzheimer’s and dementia rates appear in Table 3 of the paper. They are broken down by age range and sex. I used the rates for men, and applied them to Q15 members whose ages matched the ranges in the table. Because the table begins at age 71, I assigned Q15 members who were 70 and younger at a particular time a 0% probability of having Alzheimer’s or dementia at that time. I found the estimated rate for the entire Q15 by taking the average of their individual estimated rates. For example, the Q15 right now has three members aged up to 70 (0% probability of Alzheimer’s), three members aged 71-79 (2.30%), six members aged 80-89 (12.33%), and three members aged 90+ (33.89%). The average of the 15 rates (or the weighted average of the four rates, with the number of members being the weight) is 12.17%, so this is the estimated Alzheimer’s prevalence rate for the current Q15.
I made one adjustment to both the Alzheimer’s and dementia prevalence rates in the table when making the lines for the Church President only. The age categories are wide and have dramatically different rates, so the jumps between them are dramatic. For example, when turning 80, a Church President moves suddenly from having a 2.30% rate (or probability of having Alzheimer’s) to a 12.33% probability. To make these changes over time look more realistic for the graphs, I converted the single estimate for each age category into a linear increase through the decade the age category covered. I set the single estimate to be at the midpoint of the decade, and made the endpoints of the linear increase meet each other at the boundaries between the age ranges. Again, I only did this for Church President lines. I didn’t need to do it for Q15 rate lines because with 15 people’s data going into the rate calculations, there’s much less problem with the sudden jumps at the boundaries between age categories.