Stanley Xu1, Sophia R Newcomer2, Martin Kulldorff3, Matthew F Daley4, Bruce Fireman5, Jason M Glanz6. 1. Institute for Health Research, Kaiser Permanente Colorado, United States; University of Colorado Denver, School of Public Health, United States. Electronic address: stan.xu@kp.org. 2. University of Montana, School of Public and Community Health Sciences, United States. 3. Brigham and Women's Hospital and Harvard Medical School, Division of Pharmacoepidemiology and Pharmacoeconomics, United States. 4. Institute for Health Research, Kaiser Permanente Colorado, United States; University of Colorado Denver, School of Medicine, Department of Pediatrics, United States. 5. Kaiser Permanente Northern California, Division of Research, Vaccine Study Center, United States. 6. Institute for Health Research, Kaiser Permanente Colorado, United States; University of Colorado Denver, School of Public Health, United States.
Abstract
BACKGROUND: Summary measures such as number of vaccine antigens, number of vaccines, and vaccine aluminum exposure by the 2nd birth day are directly related to parents' concerns that children receive too many vaccines over a brief period. High correlation among summary measures could cause problems in regression models that examine their associations with outcomes. OBJECTIVES: To evaluate the performance of multiple regression models using summary measures as risk factors to simulated binary outcomes. METHODS: We calculated summary measures for a cohort of 232,627 children born between 1/1/2003 and 9/31/2013. Correlation and variance inflation factors (VIFs) were calculated. We conducted simulations (1) to examine the extent to which an association can be detected using a summary measure other than the true risk factor; (2) to evaluate the performance of multiple regression models including true and redundant risk factors; (3) to evaluate the performance of multiple regression models when all three were risk factors; (4) to examine the performance of multiple regression models with incorrect relationship between risk factors and outcome. RESULTS: These summary measures were highly correlated. VIFs were 7.14, 6.25 and 2.17 for number of vaccine antigens, number of vaccines, and vaccine aluminum exposure, respectively. In simulations, an association would be detected if a summary measure other than the true risk factor was used. The power to detect the association between the true risk factor and outcome significantly decreased if redundant risk factors were included. When all three were risk factors, multiple regression model was appropriate to detect the stronger risk factor. Correctly specifying the relationship between risk factors and the outcome was crucial. CONCLUSIONS: Multiple regression models can be used to examine the association between summary measures and outcome despite of high correlation among summary measures. It is important to correctly specify the relationship between risk factors and outcome.
BACKGROUND: Summary measures such as number of vaccine antigens, number of vaccines, and vaccine aluminum exposure by the 2nd birth day are directly related to parents' concerns that children receive too many vaccines over a brief period. High correlation among summary measures could cause problems in regression models that examine their associations with outcomes. OBJECTIVES: To evaluate the performance of multiple regression models using summary measures as risk factors to simulated binary outcomes. METHODS: We calculated summary measures for a cohort of 232,627 children born between 1/1/2003 and 9/31/2013. Correlation and variance inflation factors (VIFs) were calculated. We conducted simulations (1) to examine the extent to which an association can be detected using a summary measure other than the true risk factor; (2) to evaluate the performance of multiple regression models including true and redundant risk factors; (3) to evaluate the performance of multiple regression models when all three were risk factors; (4) to examine the performance of multiple regression models with incorrect relationship between risk factors and outcome. RESULTS: These summary measures were highly correlated. VIFs were 7.14, 6.25 and 2.17 for number of vaccine antigens, number of vaccines, and vaccine aluminum exposure, respectively. In simulations, an association would be detected if a summary measure other than the true risk factor was used. The power to detect the association between the true risk factor and outcome significantly decreased if redundant risk factors were included. When all three were risk factors, multiple regression model was appropriate to detect the stronger risk factor. Correctly specifying the relationship between risk factors and the outcome was crucial. CONCLUSIONS: Multiple regression models can be used to examine the association between summary measures and outcome despite of high correlation among summary measures. It is important to correctly specify the relationship between risk factors and outcome.
Authors: Jason M Glanz; Sophia R Newcomer; Michael L Jackson; Saad B Omer; Robert A Bednarczyk; Jo Ann Shoup; Frank DeStefano; Matthew F Daley Journal: Vaccine Date: 2016-02-15 Impact factor: 3.641
Authors: Jason M Glanz; Sophia R Newcomer; Matthew F Daley; David L McClure; Roger P Baxter; Michael L Jackson; Allison L Naleway; Marlene M Lugg; Frank DeStefano Journal: Vaccine Date: 2015-10-27 Impact factor: 3.641
Authors: Jason M Glanz; Sophia R Newcomer; Matthew F Daley; Frank DeStefano; Holly C Groom; Michael L Jackson; Bruno J Lewin; Natalie L McCarthy; David L McClure; Komal J Narwaney; James D Nordin; Ousseny Zerbo Journal: JAMA Date: 2018-03-06 Impact factor: 56.272
Authors: James Baggs; Julianne Gee; Edwin Lewis; Gabrielle Fowler; Patti Benson; Tracy Lieu; Allison Naleway; Nicola P Klein; Roger Baxter; Edward Belongia; Jason Glanz; Simon J Hambidge; Steven J Jacobsen; Lisa Jackson; Jim Nordin; Eric Weintraub Journal: Pediatrics Date: 2011-04-18 Impact factor: 7.124
Authors: Tamara D Simon; Mary Lawrence Cawthon; Susan Stanford; Jean Popalisky; Dorothy Lyons; Peter Woodcox; Margaret Hood; Alex Y Chen; Rita Mangione-Smith Journal: Pediatrics Date: 2014-05-12 Impact factor: 7.124
Authors: Elizabeth T Luman; Lawrence E Barker; Kate M Shaw; Mary Mason McCauley; James W Buehler; Larry K Pickering Journal: JAMA Date: 2005-03-09 Impact factor: 56.272
Authors: Jason M Glanz; Sophia R Newcomer; Komal J Narwaney; Simon J Hambidge; Matthew F Daley; Nicole M Wagner; David L McClure; Stan Xu; Ali Rowhani-Rahbar; Grace M Lee; Jennifer C Nelson; James G Donahue; Allison L Naleway; James D Nordin; Marlene M Lugg; Eric S Weintraub Journal: JAMA Pediatr Date: 2013-03-01 Impact factor: 16.193
Authors: Michael M McNeil; Julianne Gee; Eric S Weintraub; Edward A Belongia; Grace M Lee; Jason M Glanz; James D Nordin; Nicola P Klein; Roger Baxter; Allison L Naleway; Lisa A Jackson; Saad B Omer; Steven J Jacobsen; Frank DeStefano Journal: Vaccine Date: 2014-08-06 Impact factor: 3.641
Authors: Jason M Glanz; Christina L Clarke; Matthew F Daley; Jo Ann Shoup; Simon J Hambidge; Joshua T B Williams; Holly C Groom; Elyse O Kharbanda; Nicola P Klein; Lisa A Jackson; Bruno J Lewin; David L McClure; Stanley Xu; Frank DeStefano Journal: Pediatrics Date: 2021-12-01 Impact factor: 9.703