The Study of Factors Affecting the Levels of Plasma Beta-Carotene and Plasma Retinol Using the Logistic Model

Abstract

The study of plasma beta-carotene and plasma retinol levels data obtained from website
http://lib.stat.cmu.edu/ on ???Determinants of Plasma Retinol and Beta-Carotene Levels??, department
of Statistics, Carnegie Mellon University (2001) Pittsburgh, PA 15213. The experimental units or
patients, who had consulted at the beginning, might take the risk of being cancer. The research
objective was to study the explanatory variables affecting the response variables (Y), each of which
was the plasma beta-carotene and plasma retinol levels. There were medical evidences that betacarotene
will be modified to be vitamin A or plasma retinol in blood. They were antioxidants which
may cause cancer. The response variable (Y) was divided into two groups. One was patients who
had the plasma beta-carotene level less than 243 ng/ml, other of which the plasma beta-carotene
level more than and equal to 243 ng/ml. As well as the plasma retinol level less than 555 ng/ml
represented the first group and the other group was the plasma retinol level more than and equal to
555 ng/ml. The explanatory variables were Age, Sex, Smoking status (Smoke), Quetelet, vitamin use (Vituse), Number of calories consumed/day(Calories), Grams of fat consumed/day(Fat ), Fiber
consumed/day (Fiber), Number of alcoholic drinks consumed/week (Alcohol), Cholesterol
consumed/day(Cholesterol), Dietary beta-carotene consumed (Betadiet) and Dietary retinol
consumed (Retinoidiet). The data analysis was based on logistic model. The selection of models
under stepwise selection method was performed using SAS version 9.0
The results revealed that the statistically significant factors affecting plasma beta-carotene
levels were Age, Dietary beta-carotene (Betadiet), Quetelet, Dietary retinol consumed (Retinoidiet),
Smoking status (Smoke), and Vitamin use (Vituse). The estimated logit model was
The affecting factors to plasma retinol levels were Age, Number of alcoholic drinks
consumed/week (Alcohol), Fiber consumed/day (Fiber) and Quetelet. The estimated logit model was
Keywords: Conditional Probability Plasma beta-carotene Plasma Retinol Logistic Model

Keywords: cadmium, Zephyranthes rosea (Spreng) Lind., Eucrosia bicolor Ker-Gawl, mitotic index, gennotoxicity

Keywords: Conditional Probability Plasma beta-carotene Plasma Retinol Logistic Model