Biomedical subjects

S G Baker

Publications and source records attributed to S G Baker.

58 records · Page 4Linked to original sources

PubMedDate not supplied

Composite linear models for incomplete multinomial data.

A composite linear model (CLM) is a matrix model for incomplete multinomial data. A CLM provides a unified approach for maximum likelihood inference which is applicable to a wide variety of problems involving incomplete multinomial data. By formulating a model as a CLM, one can simplify computation of maximum likelihood estimates and asymptotic standard errors. As an example, we use CLM to test marginal homogeneity for ordered categories, subject to both ignorable and non-ignorable missing-data mechanisms.

Algorithms↗

PubMedDate not supplied

Low density lipoprotein receptor mutations in South African homozygous familial hypercholesterolemic patients.

The binding and catabolism of low density lipoprotein (LDL) were studied in cultured skin fibroblasts and/or Epstein-Barr virus-transformed lymphoblastoid cells from 20 familial hypercholesterolemic (FH) homozygotes of 14 Afrikaner kindreds. Cells from available heterozygotes were also analyzed. Good resolution between the normal, heterozygote, and homozygote groups was obtained on the basis of these assays using either cell type. Results obtained with fibroblasts allowed the classification of 13 of these kindreds as being typically receptor-defective, and one kindred with two homozygotes as being receptor-negative. Fibroblasts from homozygotes of the receptor-defective class expressed LDL receptor activities which varied between 3% and 25% of normal. Where two homozygotes from the same family were available for assays (six families), both yielded similar activities. In contrast to fibroblasts, all the lymphoblastoid cells derived from FH homozygotes yielded LDL receptor activities equal to or less than 10% of normal; in a number of cases no significant 125I-LDL binding was detectable. The use of lymphoblastoid cells provides a convenient means for screening for FH at the cellular level. Our results indicate a predominance of a receptor-defective type of abnormality, which is consistent with a founder gene effect in the Afrikaner population.

Adolescent↗

PubMedDate not supplied

Abnormalities of high density lipoproteins in homozygous familial hypercholesterolemia.

The possibility that low high density lipoprotein (HDL) levels may add to the risk of occlusive atherosclerosis in familial hypercholesterolemia (FH) is supported by the frequency and severity of this finding in these patients, particularly homozygotes. To investigate this abnormality in greater detail, we measured HDL lipid and apoprotein values in nine homozygotes with FH. Compared to their unaffected relatives, they had markedly reduced HDL cholesterol and apo A-I and A-II levels. The values found in heterozygote relatives were between those of homozygotes and those of controls. Density gradient ultracentrifugation of serum or isolated HDL from homozygotes demonstrated little or no HDL2; there was also a shift of homozygotes' HDL to an above normal peak density (d = 1.15 g/ml). Heterozygotes had milder abnormalities. The reduced HDL in FH seems to be related to elevated low density lipoprotein levels although the pathophysiology is unknown. These findings add weight to the concept that HDL abnormalities contribute to the relentless course of atherosclerosis in familial hypercholesterolemia.

Adolescent↗

PubMedDate not supplied

Choosing sample sizes to maximize expected health benefits subject to a constraint on total trial costs.

The authors present a method for choosing sample sizes for randomized controlled trials that maximizes expected health benefits (measured in expected discounted life years gained) subject to the decision maker's budget constraint. In comparison with similar approaches, the method introduces richer and more realistic models for the following quantities: costs and benefits during and after the trial, rates of adopting interventions after a positive recommendation, based on the results of the trial. Although the methodology is applicable to any type of trial, the emphasis in the paper is on prevention trials. Calculations involve Monte Carlo methods. An example is provided.

Bayes Theorem↗