Literature DB >> 12040738

Using correspondence analysis in pharmacy practice.

John F Inciardi1, Theo Stijnen, Kay McMahon.   

Abstract

Correspondence analysis (CA) and some of its uses in pharmacy practice are described. CA is a multivariate and graphic form of exploratory data analysis that reduces a multidimensional contingency table to a two-dimensional plot with minimal loss of information. An exploratory association between medication use and the number and severity of falls among the elderly living at home is used to illustrate the process of CA. Row profiles are constructed by dividing the count data in each cell by the total of the corresponding row, converting frequency data into relative frequencies. A graph of the data can be made by recalculating the frequencies to illustrate how the data might appear in a three-dimensional space with the axes defined by the three categories of falling (minimum, moderate, and major). For example, the number of patients using central nervous system (CNS) agents is plotted using the relative frequencies as vectors to locate a position in the three-dimensional space. All drug profiles and the average profile will lie exactly in a triangular plane defined by the terminal points of each axis. In this manner, data with three dimensions can be projected onto a two-dimensional space. Transferring the vectors named by pharmacologic class onto the plot requires creating two axes that cross at a common point or the origin. This point locates the average profile and defines an important reference for making comparisons. CA revealed that CNS agents may be associated with moderate to major falls, psychotherapeutic agents with moderate falls, and anticoagulants with moderate to minimum falls. CA has widespread uses in pharmacy practice. It can identify patients at risk for serious but preventable drug-related complications, enabling pharmacists to allocate pharmacy resources to the areas in most need and suggest hypotheses for future research. Clinicians can also use CA to analyze vast amounts of patient-related data to uncover hard-to-detect associations. The use of CA in pharmacy practice will allow new strategies for improved patient care to be more readily appreciated and implemented.

Entities:  

Mesh:

Substances:

Year:  2002        PMID: 12040738     DOI: 10.1093/ajhp/59.10.968

Source DB:  PubMed          Journal:  Am J Health Syst Pharm        ISSN: 1079-2082            Impact factor:   2.637


  2 in total

1.  Host selection and niche differentiation in sucking lice (Insecta: Anoplura) among small mammals in southwestern China.

Authors:  Xiao-Hua Zuo; Xian-Guo Guo; Yin-Zhu Zhan; Dian Wu; Zhi-Hua Yang; Wen-Ge Dong; Li-Qin Huang; Tian-Guang Ren; Yong-Guang Jing; Qiao-Hua Wang; Xiao-Mei Sun; Shang-Jin Lin
Journal:  Parasitol Res       Date:  2010-12-08       Impact factor: 2.289

2.  Using latent class analysis to model prescription medications in the measurement of falling among a community elderly population.

Authors:  Patrick C Hardigan; David C Schwartz; William D Hardigan
Journal:  BMC Med Inform Decis Mak       Date:  2013-05-25       Impact factor: 2.796

  2 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.