Working Paper: NBER ID: w23900
Authors: Kristopher J. Hult
Abstract: Individuals respond to pharmaceutical treatments differently due to the heterogeneity of patient populations. This heterogeneity can make it difficult to determine how efficacious or burdensome a treatment is for an individual patient. Personalized medicine involves using patient characteristics, therapeutics, or diagnostic testing to understand how individual patients respond to a given treatment. Personalized medicine increases the health impact of existing treatments by improving the matching process between patients and treatments and by improving a patient's understanding of the risk of serious side effects. In this paper, I compare the health impact of new treatment innovations with the potential health impact of personalized medicine. I find that the impact of personalized medicine depends on the number of treatments, the correlation between treatment effects, and the amount of noise in a patient's individual treatment effect signal. For multiple sclerosis treatments, I find that personalized medicine has the potential to increase the health impact of existing treatments by roughly 50 percent by informing patients of their individual treatment effect and risk of serious side effects.
Keywords: personalized medicine; health impact; multiple sclerosis; treatment efficacy; patient heterogeneity
JEL Codes: I10; I11; O30; O31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| personalized medicine (I11) | health impact (I12) |
| personalized medicine (I11) | treatment efficacy (C22) |
| informed patients (I11) | health impact (I12) |
| accurately identifying risks (D81) | health impact (I12) |
| number of available treatments (I12) | treatment efficacy (C22) |
| variance in treatment effects (C21) | treatment efficacy (C22) |
| correlation of treatment effects (C32) | treatment efficacy (C22) |
| reducing noise in treatment effect signals (C22) | health outcomes (I14) |