The following is from Boomsday (page 119) by Christopher Buckley (2007) Comment by Claude
BADMAP is an acronym for Bio-Actuarial Dyna-Metric Age Predictor. It works like this:
” A person’s DNA profile, family history, mental history, lifestyle profile, every variable –how many trips to the grocery per week, how many airplane flights, hobbies, food, booze, number of times per month you had sex and with whom, everything down to what color socks you put on in the morning– were all fed into the software. RIP-ware would then calculate and predict how and when you’d die. In the testing, they had programmed it retroactively with the DNA and lifestyle profile of thousand of people who had already died. RIP-ware predicted their deaths with an accuracy of 99.07 percent. In a simulation, it predicted the death of Elvis Presley — just four months from he actual date of his demise. The ultimate “killer app.”
Insurance companies had been working on similar programs. What a windfall it would be for them if they could sell life insurance to someone they knew was going to live another forty years–and conversely decline life insurance to someone the computer predicted would be pushing up daisies within two years.
Another field of vast potential were the old folks’ homes. typically, these demanded that a prospective resident turn over his and her entire net worth in return for perpetual care. You could live two years or twenty years; that was their gamble. But if a nursing home knew,in advance, that John Q. smith was going to have a fatal heart attack in 2.3 years while watching an ad for toenail fungus ointment on the evening news, they would much rather have his nest egg as advance payment than that of, say, Jane Q. Jones, who RIP-ware predicted would live another twenty-five years and die at the ripe old age of 105.
Claude’s analysis: It’s a good satirical target because the pieces are all real now, just not yet assembled into one product.
Polygenic risk scores already give reasonably calibrated probabilities for things like cardiovascular disease and some cancers. Wearables and phones passively log the “lifestyle profile” — steps, sleep, heart rate variability, even irregular gait. Insurers have been experimenting with wearable-linked discounts (John Hancock’s Vitality program is the obvious example) and with using consumer data brokers for underwriting, which regulators have pushed back on precisely because it edges toward what Buckley is joking about. The “sex, socks, and grocery trips” granularity is still satire, but the *category* of input isn’t — data brokers already hold shopping and location histories that granular.
Where the novel’s premise still breaks down is the 99.07% accuracy figure. Mortality prediction at the individual level, even with genomics plus continuous biometric data, is nowhere near that — you can sharpen actuarial tables at the population level, but pinning an individual’s death to within a few months (the Elvis bit) requires modeling contingent events — accidents, acute illness, homicide — that aren’t encoded in DNA or habits. That’s the part that’s aged as pure fiction rather than “not so far-fetched.”
The part that’s genuinely closer to reality than 2007 is the institutional incentive structure Buckley describes — insurers and nursing homes wanting exactly this tool for adverse selection. That’s not speculative at all; it’s just underwriting with better data, and it’s a live regulatory fight right now (genetic non-discrimination law covers health insurance and employment but has gaps around life and long-term care insurance).”