Premium
A statistical comparison of different family history scores
Author(s) -
Murad H.,
KalterLeibovici O.,
Chetrit A.,
Freedman L. S.
Publication year - 2006
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.2750
Subject(s) - family history , confounding , demography , disease , medicine , relative risk , statistics , confidence interval , mathematics , sociology
Family history (FH) scores are used for estimating the familial risk (FR), i.e. the level of risk for a particular disease among members of that family. An FH score is created from reports about the disease status of the relatives in each family. The most commonly used score is the dichotomous score (positive when at least one relative is affected), which does not consider the family size, number of affected relatives nor each relative's risk factor profile. Authors have proposed many other FH scores that overcome these deficiencies by using external expected risks adjusted for important risk factors. We consider the use of FH scores in studies, which investigate risk factors for a disease and where family risk is considered as a confounder, and examine through simulations the performance of a variety of FH scores in controlling the FR status. We also examine performance in predicting true FR status. For both criteria, only small differences were found between most of the FH scores, although the dichotomous score performed the poorest. Since the proportion score (the proportion of first‐degree relatives of the index who have the disease) is the simplest to calculate, use of this score seems to be justified. Copyright © 2006 John Wiley & Sons, Ltd.