ABSTRACT
Disease risk score (DRS) can be used to adjust the confounding effects on data with high dimensions and can reduce related bias through balancing the risk or probability,regarding the development of some specific diseases,between the two compared groups.The DRS approach thus can be applied to studies of pharmacoepidemiology when administrative medical database is used for data analysis.Although DRS functions are similarly to the propensity scores (PS) under many situations,even with some advantages over PS or conventional analytical methods in some special exposure settings,the usage of DRS is far limited than the PS method.Considering the important application value of DRS in pharmacoepidemiologic studies,we are introducing the theory,model,estimation and application of DRS,to present reference for the development of DRS method in the pharmacoepidemiologic studies.
ABSTRACT
Disease risk score (DRS) can be used to adjust the confounding effects on data with high dimensions and can reduce related bias through balancing the risk or probability,regarding the development of some specific diseases,between the two compared groups.The DRS approach thus can be applied to studies of pharmacoepidemiology when administrative medical database is used for data analysis.Although DRS functions are similarly to the propensity scores (PS) under many situations,even with some advantages over PS or conventional analytical methods in some special exposure settings,the usage of DRS is far limited than the PS method.Considering the important application value of DRS in pharmacoepidemiologic studies,we are introducing the theory,model,estimation and application of DRS,to present reference for the development of DRS method in the pharmacoepidemiologic studies.
ABSTRACT
Objective To recommend the appropriate retrieval tools for scientific researchers and improve their re-trieval efficiency by comparatively analyzing the retrieval results from PubMed and Embase .com.Methods Eight retrieval terms from different subjects were used to retrieve PubMed and Embase .com by subject headings and basic retrieval methods than can automatically transfer their function .Results Six different results were retrieved from PubMed and Embase.com with 8 retrieval terms.Conclusion The recall ratio is higher in Embase.com than in PubMed when they are retrieved by subject headings while the accuracy ratio is higher in PubMed than in Embase . com when they are retrieved by basic retrieval methods than can automatically transfer their function although error transformation occurs in both PubMed and Embase.com.
ABSTRACT
In medical database systems of security environments, multiple versions are used so as to prevent covert channel between transactions with different security level. Like the existing database systems, multilevel secure medical database systems can also occur various failures. Although the system was fault, the consistency and security for medical information must be guaranteed to establish reliable system from failures. It is difficult issues to prevent covert channel and maintain the serializability of transactions from various failures. This paper proposes multilevel secure recovery algorithms to solve failures from dynamic multiversion control and proves the correctness as to them. The proposed algorithms are very good for the multilevel secure reliability of long read-only transactions.
Subject(s)
Hospital Information SystemsABSTRACT
The integrated management of medical data raised the efficiency of them and improved doctor' s diagnostic ability by processing correct data. Various medical information processing led out the development of new medical technology and is incurring huge changes for the medical examination and diagnosis. In medical database security, both serializability and security must be considered to maintain data consistency and prevent covert channel. In the security environments of heterogeneous medical database systems, the existing local autonomy and the security autonomy as a new constraint are required. This paper defines multilevel secure one-copy quasi-serializability (MLS/1QSR) necessary for concurrency control in heterogeneous medical database systems with replicated data to solve security problem among many objects of medical database systems. The proposed MLS/1QSR neither violates security autonomy nor establishes covert channel for transactions in heterogeneous medical database systems.