%0 Journal Article %J Stat Med %D 2009 %T Systematically missing confounders in individual participant data meta-analysis of observational cohort studies. %A Jackson, Dan %A White, Ian %A Kostis, J B %A Wilson, A C %A Folsom, A R %A Wu, K %A Chambless, L %A Benderly, M %A Goldbourt, U %A Willeit, J %A Kiechl, S %A Yarnell, J W G %A Sweetnam, P M %A Elwood, P C %A Cushman, M %A Psaty, B M %A Tracy, R P %A Tybjaerg-Hansen, A %A Haverkate, F %A de Maat, M P M %A Thompson, S G %A Fowkes, F G R %A Lee, A J %A Smith, F B %A Salomaa, V %A Harald, K %A Rasi, V %A Vahtera, E %A Jousilahti, P %A D'Agostino, R %A Kannel, W B %A Wilson, P W F %A Tofler, G %A Levy, D %A Marchioli, R %A Valagussa, F %A Rosengren, A %A Wilhelmsen, L %A Lappas, G %A Eriksson, H %A Cremer, P %A Nagel, D %A Curb, J D %A Rodriguez, B %A Yano, K %A Salonen, J T %A Nyyssönen, K %A Tuomainen, T-P %A Hedblad, B %A Engstrom, G %A Berglund, G %A Loewel, H %A Koenig, W %A Hense, H W %A Meade, T W %A Cooper, J A %A De Stavola, B %A Knottenbelt, C %A Miller, G J %A Cooper, J A %A Bauer, K A %A Rosenberg, R D %A Sato, S %A Kitamura, A %A Naito, Y %A Iso, H %A Salomaa, V %A Harald, K %A Rasi, V %A Vahtera, E %A Jousilahti, P %A Palosuo, T %A Ducimetiere, P %A Amouyel, P %A Arveiler, D %A Evans, A E %A Ferrieres, J %A Juhan-Vague, I %A Bingham, A %A Schulte, H %A Assmann, G %A Cantin, B %A Lamarche, B %A Després, J-P %A Dagenais, G R %A Tunstall-Pedoe, H %A Lowe, G D O %A Woodward, M %A Ben-Shlomo, Y %A Davey Smith, G %A Palmieri, V %A Yeh, J L %A Meade, T W %A Rudnicka, A %A Brennan, P %A Knottenbelt, C %A Cooper, J A %A Ridker, P %A Rodeghiero, F %A Tosetto, A %A Shepherd, J %A Lowe, G D O %A Ford, I %A Robertson, M %A Brunner, E %A Shipley, M %A Feskens, E J M %A Di Angelantonio, E %A Kaptoge, S %A Lewington, S %A Lowe, G D O %A Sarwar, N %A Thompson, S G %A Walker, M %A Watson, S %A White, I R %A Wood, A M %A Danesh, J %K Cohort Studies %K Computer Simulation %K Coronary Disease %K Data Interpretation, Statistical %K Female %K Fibrinogen %K Humans %K Male %K Meta-Analysis as Topic %K Models, Statistical %X

One difficulty in performing meta-analyses of observational cohort studies is that the availability of confounders may vary between cohorts, so that some cohorts provide fully adjusted analyses while others only provide partially adjusted analyses. Commonly, analyses of the association between an exposure and disease either are restricted to cohorts with full confounder information, or use all cohorts but do not fully adjust for confounding. We propose using a bivariate random-effects meta-analysis model to use information from all available cohorts while still adjusting for all the potential confounders. Our method uses both the fully adjusted and the partially adjusted estimated effects in the cohorts with full confounder information, together with an estimate of their within-cohort correlation. The method is applied to estimate the association between fibrinogen level and coronary heart disease incidence using data from 154,012 participants in 31 cohorts

%B Stat Med %V 28 %P 1218-37 %8 2009 Apr 15 %G eng %N 8 %1 https://www.ncbi.nlm.nih.gov/pubmed/19222087?dopt=Abstract %R 10.1002/sim.3540