Record Linkage Techniques Research Articles

Use of record linkage techniques to maintain the leicestershire diabetes register

The Leicestershire Diabetes Register is a computerised live register of approximately 5000 insulin users who are resident in Leicestershire. Data capture is from clinic notes and is input by two part-time clerical officers. The register is used for epidemiological and planning purposes. The computer software used for the register was developed in-house and is written in C, ensuring a fast response time even on large numbers of records, and will run on any IBM PC-compatible. Successful use is made of record linkage techniques, including phonetic name matching and a heuristic algorithm to improve reporting of results and to ensure the accuracy and completeness of the register, and much manual checking is thus avoided. Record linkage allows us to check for duplicates, deaths and migrations and allows cross-checks on the accuracy and completeness of patients' identifying details.

The incidence of child sexual abuse in Northern Ireland.

This is part of a major epidemiological survey of the reported incidence (new cases) of child sexual abuse in Northern Ireland. Based on multi-source methodology, the study used computer-based record linkage techniques to find the number of incident cases occurring in the Province in 1987. Reporters involved with the incident cases were interviewed by OEV who recorded details of the child, the abuse, and the abuser using a specially designed questionnaire. Age-sex specific incidence rates adjusted for potential misclassification were computed and allowance made for under-ascertainment. Overall, 408 Established cases were reported during 1987. The corresponding incidence rate for Northern Ireland was 0.9 cases per 1000 children. When Suspected and Alleged cases were compared with Established cases, a further 119 cases were allocated by discriminant function analysis to the Established group, and the rate rose to 1.16 per thousand. Selected reporters were interviewed about ascertainment which was estimated to lie between 62% and 74%. If the lowest level of ascertainment applied the corresponding incidence rate for Northern Ireland would increase to 1.87. These figures are higher than those currently reported elsewhere in the UK, but methodological differences in study design hamper interregional comparisons. A number of relevant methodological problems are discussed.

Getty's Synoname™ and its cousins: A survey of applications of personal name-matching algorithms

The study reported in this article was commissioned by the Getty Art History Information Program (AHIP) as a background investigation of personal name-matching programs in fields other than art history, for purposes of comparing them and their approaches with AHIP's Synoname™ project. We review techniques employed in a variety of applications, including art history, bibliography, genealogy, commerce, and government, providing a framework of personal name characteristics, factors in selecting matching techniques, and types of applications. Personal names, as data elements in information systems, vary for a wide range of legitimate reasons, including cultural and historical traditions, translation and transliteration, reporting and recording variations, as well as typographical and phonetic errors. Some matching applications seek to link variants, while others seek to correct errors. The choice of matching techniques will vary in the amount of domain knowledge about the names that is incorporated, the sources of data, and the human and computing resources required. Personal name-matching techniques may be included in name authority work, information retrieval, or duplicate detection, with some applications matching on name only, and others combining personal names with other data elements in record linkage techniques. We discuss both phonetic- and pattern-matching techniques, reviewing a range of implemented and proposed name-matching techniques in the context of these factors. © 1992 John Wiley & Sons, Inc.

Supravaginal uterine amputation in Denmark 1978–1988 and risk of cancer

In Denmark, total rather than subtotal hysterectomy has been recommended as treatment for benign conditions of the uterine corpus, primarily due to the risk of cervical cancer in the residual stump. Due to possible benefits during and after surgery, it has recently been argued that total hysterectomy for benign conditions could be substituted by subtotal hysterectomy, particularly since cancers of the cervix are declining as a consequence of screening. Cancer risk following supravaginal uterine amputation among 1104 Danish women undergoing operations during 1978–1988 was studied using record linkage techniques with the Danish Cancer Registry. No overall cancer risk, nor an overall increased risk of cancers coded as cervical, was observed. However, a 3.3 to 5.0-fold increased risk for cervical cancer was observed among women whose operations occurred at 50 years of age or older, possibly reflecting an increased risk of cancer in the endocervical area or in remaining uterine tissue. This observation, together with the difficulties in treating a cervical cancer of the stump when it does occur, should be taken into account before a change in surgical procedures from total to subtotal hysterectomies is recommended.

Retinopathy of Prematurity-Induced Blindness: Birth Weight-Specific Survival and the New Epidemic

A recent population-based study in the Canadian province of British Columbia showed that, since the mid-1960s, there has been a significant increase in the incidence of retinopathy of prematurity-induced blindness in infants weighing 750 to 999 g at birth. To determine the impact of changing birth weight-specific survival on this new epidemic, all infants born in the province in the period 1952 through 1986 and known to the British Columbia Health Surveillance Registry as having retinopathy of prematurity-induced blindness were identified. In addition, the birth registration records for the 1,299 740 infants born in British Columbia in the same period and the death records of the 22,940 British Columbia-born infants who died in the province before the end of their first year of life were linked using a combination of probabilistic and manual record linkage techniques. These linked records and the records from the Health Surveillance Registry were used to calculate birth weight-specific incidence rates of retinopathy of prematurity-induced blindness in liveborn infants and first-year-of-life survivors. The rates, in 5-year intervals, showed that, in both liveborn infants and first-year survivors, the highest birth weight-specific rates occurred during the first epidemic of retinopathy of prematurity, which ended in British Columbia in 1954. Since the mid-to late-1960s, the incidence of retinopathy of prematurity-induced blindness in liveborn infants weighing less than 1000 g increased steadily whereas in infants weighting 1000 to 1499 g, incidence decreased slightly since the original epidemic ended. However, the experience of first-year-of-life survivors is substantially different.(ABSTRACT TRUNCATED AT 250 WORDS)

Marital state, alcohol consumption, and liability to myocardial infarction, stroke, diabetes mellitus, or hypertension in men from Gothenburg.

Data from the Swedish personal number system were used to examine rates of admissions to hospital, particularly with myocardial infarction, stroke, diabetes mellitus, or hypertension, in relation to marital state and histories of admission with alcoholism in 77843 men from Gothenburg born in 1911-40. The number of admissions to hospital with the four diseases varied with marital state and history of admission with alcoholism, but by considering each subset from the total group it was possible to show that among alcoholic men no material variation in the number of admissions was evident. By contrast, among the non-alcoholic men those who were married were particularly prone to myocardial infarction but less likely to be admitted with stroke or diabetes mellitus than the unmarried or divorced men. The data show the value of common personal and health statistical numbering systems in generating epidemiological information beyond that obtainable from aggregated data without a record linkage technique.

‘Community’ and the social geography of Victorian cities

Quantitative studies of nineteenth-century cities have made rapid advances in the last decade in terms of both the complexity of techniques and the range of sources to which techniques have been applied. Simple representations of material from the census enumerators' books like graphs and pie diagrams have been superseded by location quotients, indices of segregation and factor analyses. Synchronic descriptions of urban structure have been supplemented by assessments of change (for example, of residential persistence and mobility) and analyses of relationships (for example, journeys to work and marriage patterns). Researchers have ventured beyond the census to examine other nominal listings—in directories, electoral registers, ratebooks and church records—often using record linkage techniques to check or supplement information from one source with another. Our enthusiasm to apply new techniques to new sources has often left us exposed to Anderson's criticism that ‘we have been paying too little attention to the question of why we are measuring what we are measuring’. Many quantitative historical geographers justify their enterprise as the testing of modern theory on such subjects as migration and modernization. Some such theories, particularly those concerned with Sjoberg's and Burgess' stereotypes of pre-industrial and industrial cities and the ‘transitional’ nature of Victorian cities, are little more than rough generalizations and some urban historians have been more than sceptical about them. Certainly the retrospective application of the methods, concepts and theories of contemporary social research is a critical issue. In this paper we consider the significance of ‘what we are measuring’ for a particularly contentious concept, that of ‘community’. We assess the usefulness or relevance of more sophisticated methodological techniques to the identification and investigation of communities in Victorian cities.

Record linkage and the identification of long-term environmental hazards.

The long-term effects of toxic substances in man that have been discovered so far have involved gross relative risks or bizarre effects, or have stumbled upon by chance or because of special circumstances. These facts and some recent epidemiological evidence together suggest that a systematic approach with better methods would reveal the effects of many more toxic substances, particularly in manufacturing industry. Record linkage is a powerful tool because it makes possible the correlation of indicators of exposure with indicators of the biological effect of such exposure in the same persons or in their progeny even after considerable periods of time have elapsed. A system of linked records exists in England and Wales which is at present used by research workers to follow up samples of persons defined in various ways, e.g. in respect of exposure to a suspected toxic factor. In this way hypotheses about substances causing cancer or other lethal effects can be tested. It is suggested that there are two additional ways in which record linkage techniques could be used to identify substances with long-term toxic effects: the first would be by setting up a register of women employed in industry during pregnancy and linking this register to records of the occurrence of congenital malformations and of stillbirth or death in their children; the second would be to follow samples of workers in manufacturing industry, notably those engaged in the manufacture of products from raw materials including the chemical industry, to death and to the development of cancer. Regular analyses of material from these two systems of linked records would provide the basis for a monitoring system for certain gross effects of long-term toxic substances in man. There are two principal obstacles to further progress in this field. The first is the lack of a clear statement of public policy concerning the issues of confidentiality and informed consent in the use of identifiable personal records for medical research. A settlement is needed which defines the proper limits of their use in the interests of health with safeguards to privacy. The second obstacle is a lack of resources to improve the quality, accessibility and organization of the appropriate data.

Evaluation of 'infant at risk' registers.

The operational effectiveness of Infant At Risk Registers was examined using record linkage techniques. Two operations were carried out: (<i>a</i>) the linkage of the 1965 Birmingham Observation Register with the 1968 file of `Defective Children9 as notified to the Education Department; and (<i>b</i>) the file of maternity data for the whole of Birmingham in 1956, linked with the list of children receiving special education in 1968. Positive associations were found between assessment of risk in the Observation Register and subsequent defects as notified to the Education Department at age 3; and between assessment of risk made on the basis of the maternity data and the likelihood of requiring special education at age 13. The main predictors were low birthweight, traumatic delivery, and the indices of social underprivilege. However, none of these determinants were sufficient to justify the continuation of selective At Risk Registers whose purpose is defined in terms of prediction. They should be withdrawn or should be redesigned to the different purposes of registering the known handicapped, and extending developmental surveillance to all children, rather than concentrating it upon a few.