Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You and Your Team.

Learn More →

There Is Nothing Personal

There Is Nothing Personal Ioannidis1 nicely addressed key challenges of “personal” genetic prediction for common diseases. Expectations are huge in this domain. I argue that some of these expectations may be favored by the term personal and that it would be better to use the term stratified.2-4 Patient's characteristics such as age, sex, lifestyle, socioeconomic status, biomarkers, past environmental exposure, or genetic variants can help identify groups or strata of patients who are more (or less) likely to develop a disease or respond to a treatment.3 Such characteristics can improve our ability to estimate the probability of getting a common disease. Nevertheless, probability is a group property and should not be confounded with individual determinism. At the individual level, either you get or do not get the disease; there is no probability. Suppose there are 2 patients with exactly the same characteristics, including genetic makeup, and these characteristics are predictive of getting a disease. These 2 patients are in the same risk stratum, which is associated with a given—and sometimes quantifiable—likelihood of getting the disease. Still, 1 of these 2 patients could get the disease and not the other, and it is not possible to know a priori which one will be afflicted eventually. Inference of the risk associated with the characteristics of these patients is to the corresponding group or stratum level, not to the personal or individual level where near-random events occur.2 Prediction can be strong at stratum level and poor at individual level, especially for remote outcomes (eg, common chronic diseases requiring years to occur) compared with imminent outcomes (eg, death in critically ill patients).5 There is nothing personal—bad luck, chance, or randomness is at work,2,6 and randomness gets cancelled out at a stratum level, not at an individual level. Back to top Article Information Correspondence: Dr Chiolero, Institute of Social and Preventive Medicine, University Hospital of Lausanne, Biopôle 2, Route de la Corniche 10, 1010 Lausanne, Switzerland (arnaud.chiolero@chuv.ch). Conflict of Interest Disclosures: None reported. References 1. Ioannidis JPA. Genetic prediction for common diseases: will personal genomics ever works? Arch Intern Med. 2012;172(9):744-74622782208PubMedGoogle ScholarCrossref 2. Smith GD. Epidemiology, epigenetics and the “Gloomy Prospect”: embracing randomness in population health research and practice. Int J Epidemiol. 2011;40(3):537-56221807641PubMedGoogle ScholarCrossref 3. Trusheim MR, Berndt ER, Douglas FL. Stratified medicine: strategic and economic implications of combining drugs and clinical biomarkers. Nat Rev Drug Discov. 2007;6(4):287-29317380152PubMedGoogle ScholarCrossref 4. Chiolero A. Stratified vs personalised medicine. http://www.bmj.com/rapid-response/2011/11/03/stratified-vs-personalised-medicine. Accessed June 5, 2012 5. Ioannidis JPA. Limits to forecasting in personalized medicine: an overview. Int J Forecast. 2009;25(4):773-783Google ScholarCrossref 6. Coggon DI, Martyn CN. Time and chance: the stochastic nature of disease causation. Lancet. 2005;365(9468):1434-143715836893PubMedGoogle ScholarCrossref http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Archives of Internal Medicine American Medical Association

There Is Nothing Personal

Archives of Internal Medicine , Volume 172 (21) – Nov 26, 2012

Loading next page...
 
/lp/american-medical-association/there-is-nothing-personal-YU4ljQM1nL
Publisher
American Medical Association
Copyright
Copyright © 2012 American Medical Association. All Rights Reserved.
ISSN
0003-9926
eISSN
1538-3679
DOI
10.1001/archinternmed.2012.4430
Publisher site
See Article on Publisher Site

Abstract

Ioannidis1 nicely addressed key challenges of “personal” genetic prediction for common diseases. Expectations are huge in this domain. I argue that some of these expectations may be favored by the term personal and that it would be better to use the term stratified.2-4 Patient's characteristics such as age, sex, lifestyle, socioeconomic status, biomarkers, past environmental exposure, or genetic variants can help identify groups or strata of patients who are more (or less) likely to develop a disease or respond to a treatment.3 Such characteristics can improve our ability to estimate the probability of getting a common disease. Nevertheless, probability is a group property and should not be confounded with individual determinism. At the individual level, either you get or do not get the disease; there is no probability. Suppose there are 2 patients with exactly the same characteristics, including genetic makeup, and these characteristics are predictive of getting a disease. These 2 patients are in the same risk stratum, which is associated with a given—and sometimes quantifiable—likelihood of getting the disease. Still, 1 of these 2 patients could get the disease and not the other, and it is not possible to know a priori which one will be afflicted eventually. Inference of the risk associated with the characteristics of these patients is to the corresponding group or stratum level, not to the personal or individual level where near-random events occur.2 Prediction can be strong at stratum level and poor at individual level, especially for remote outcomes (eg, common chronic diseases requiring years to occur) compared with imminent outcomes (eg, death in critically ill patients).5 There is nothing personal—bad luck, chance, or randomness is at work,2,6 and randomness gets cancelled out at a stratum level, not at an individual level. Back to top Article Information Correspondence: Dr Chiolero, Institute of Social and Preventive Medicine, University Hospital of Lausanne, Biopôle 2, Route de la Corniche 10, 1010 Lausanne, Switzerland (arnaud.chiolero@chuv.ch). Conflict of Interest Disclosures: None reported. References 1. Ioannidis JPA. Genetic prediction for common diseases: will personal genomics ever works? Arch Intern Med. 2012;172(9):744-74622782208PubMedGoogle ScholarCrossref 2. Smith GD. Epidemiology, epigenetics and the “Gloomy Prospect”: embracing randomness in population health research and practice. Int J Epidemiol. 2011;40(3):537-56221807641PubMedGoogle ScholarCrossref 3. Trusheim MR, Berndt ER, Douglas FL. Stratified medicine: strategic and economic implications of combining drugs and clinical biomarkers. Nat Rev Drug Discov. 2007;6(4):287-29317380152PubMedGoogle ScholarCrossref 4. Chiolero A. Stratified vs personalised medicine. http://www.bmj.com/rapid-response/2011/11/03/stratified-vs-personalised-medicine. Accessed June 5, 2012 5. Ioannidis JPA. Limits to forecasting in personalized medicine: an overview. Int J Forecast. 2009;25(4):773-783Google ScholarCrossref 6. Coggon DI, Martyn CN. Time and chance: the stochastic nature of disease causation. Lancet. 2005;365(9468):1434-143715836893PubMedGoogle ScholarCrossref

Journal

Archives of Internal MedicineAmerican Medical Association

Published: Nov 26, 2012

Keywords: biological markers,chronic disease,critical illness,life style,genetics,inference,socioeconomic factors,environmental exposure

References