Search PubMed⌕ Search

Biomedical subjects

Tim Churches

Publications and source records attributed to Tim Churches.

6 recordsLinked to original sources
PubMed2003

A proposed architecture and method of operation for improving the protection of privacy and confidentiality in disease registers.

BACKGROUND: Disease registers aim to collect information about all instances of a disease or condition in a defined population of individuals. Traditionally methods of operating disease registers have required that notifications of cases be identified by unique identifiers such as social security number or national identification number, or by ensembles of non-unique identifying data items, such as name, sex and date of birth. However, growing concern over the privacy and confidentiality aspects of disease registers may hinder their future operation. Technical solutions to these legitimate concerns are needed. DISCUSSION: An alternative method of operation is proposed which involves splitting the personal identifiers from the medical details at the source of notification, and separately encrypting each part using asymmetrical (public key) cryptographic methods. The identifying information is sent to a single Population Register, and the medical details to the relevant disease register. The Population Register uses probabilistic record linkage to assign a unique personal identification (UPI) number to each person notified to it, although not necessarily everyone in the entire population. This UPI is shared only with a single trusted third party whose sole function is to translate between this UPI and separate series of personal identification numbers which are specific to each disease register. SUMMARY: The system proposed would significantly improve the protection of privacy and confidentiality, while still allowing the efficient linkage of records between disease registers, under the control and supervision of the trusted third party and independent ethics committees. The proposed architecture could accommodate genetic databases and tissue banks as well as a wide range of other health and social data collections. It is important that proposals such as this are subject to widespread scrutiny by information security experts, researchers and interested members of the general public, alike.

Computer Security↗
PubMed2003

Exploratory data analysis using set operations and ordinal mapping.

Exploratory data analysis requires the ability to issue ad hoc queries to filter and summarise data sets. As the sizes of health data sets grow, traditional methods of processing data have difficulty in providing acceptable response times for such queries. An alternative method is described which combines complete vertical partitioning of data with set operations on ordinal mappings (SOOM). An initial implementation of the technique provides significantly better performance than a conventional SQL database on typical exploratory data analysis queries. The use of parallel, distributed computation to further increase the performance of the technique appears to be feasible.

Algorithms↗
PubMed2002

Preparation of name and address data for record linkage using hidden Markov models.

BACKGROUND: Record linkage refers to the process of joining records that relate to the same entity or event in one or more data collections. In the absence of a shared, unique key, record linkage involves the comparison of ensembles of partially-identifying, non-unique data items between pairs of records. Data items with variable formats, such as names and addresses, need to be transformed and normalised in order to validly carry out these comparisons. Traditionally, deterministic rule-based data processing systems have been used to carry out this pre-processing, which is commonly referred to as "standardisation". This paper describes an alternative approach to standardisation, using a combination of lexicon-based tokenisation and probabilistic hidden Markov models (HMMs). METHODS: HMMs were trained to standardise typical Australian name and address data drawn from a range of health data collections. The accuracy of the results was compared to that produced by rule-based systems. RESULTS: Training of HMMs was found to be quick and did not require any specialised skills. For addresses, HMMs produced equal or better standardisation accuracy than a widely-used rule-based system. However, accuracy was worse when used with simpler name data. Possible reasons for this poorer performance are discussed. CONCLUSION: Lexicon-based tokenisation and HMMs provide a viable and effort-effective alternative to rule-based systems for pre-processing more complex variably formatted data such as addresses. Further work is required to improve the performance of this approach with simpler data such as names. Software which implements the methods described in this paper is freely available under an open source license for other researchers to use and improve.

Data Collection↗