The computerization of the Cardiff Births Survey.

The Cardiff Births Survey (CBS) is a major ongoing study of childbirth in a well defined population, and has formed the basis for a number of important investigations in its field. It started at a time when computers were about to become readily available in United Kingdom universities, and it has made use of them from its earliest days. The changing ways in which this has been done over the years are here described in parallel with the changes that have taken place to the survey itself. Conclusions are drawn concerning the characteristics of other studies to which similar methods might be applied, and the inherent advantages and limitations of these methods are considered. Finally brief discussion is made of future prospects for the use of computers in the Cardiff Births Survey itself.

The Programming Environment for Statistical Tabulation.

The Programming Environment for Statistical Tabulation (PEST) is a system which adopts an unusual approach to statistical computing. Most such systems present one or more complete programs which are controlled by parameters of varying degrees of complexity; PEST adopts a different method and provides an incomplete program to which extra instructions must be added to specify the analysis that is required. Designed originally as a replacement for a similar system written in Algol 60, it now uses instead the more powerful and flexible Algol 68. This much neglected language is ideal for the purpose, enabling an ordinary user to obtain results with a limited range of simple statements and at the same time allowing an experienced programmer free access to facilities that compare most favourably with those of comparable languages. The efficiency of modern Algol 68 compilers means that the overhead of compiling the user's instructions as often as required is small and incurs no disadvantage by comparison with conventional packages.

Pre-registration appointment in the 1980s--a replacement for a well-tried computerized system.

A new computerized system for the allocation of pre-registration house-officer appointments to medical graduates at the Welsh National School of Medicine is described. The need to withdraw an earlier system, used through much of the 1970s, is explained, and the advantages possessed by the new system are discussed.

Implementation of a free-format request profile in a system for retrieving references to scientific publications.

A process is described whereby a set of instructions in free format can be used to control a search of a data file in an efficient manner. The implications of the effectiveness of this process are considered in relation to a system for the storage and retrieval of references to published works in a typical scientific research establishment. Details of this system have been described earlier.

An effective and reliable computerised system for storage and retrieval of references to scientific publications.

A suite of programs has been implemented to facilitate the storage and retrieval of references to published works in a typical scientific research establishment. The retrieval process is governed by searches for specified combinations of keywords, author names, individual references, and lengths of time that references have been held on file. The system currently caters for approximately 1290 journals together with abstracts, reports, supplements, symposia, proceedings and books. Five years experience has proved the system to be reliable and effective, and it requires only 1.5 technician-days per week to operate it and to input about 25 000 articles annually. This system fulfills a requirement felt by many specialised institutions in providing localised control over quantities of information which are too large to be processed manually.

Obstetric practice and outcome of pregnancy in Cardiff residents 1965-73.

Trends in management and outcome of pregnancy in Cardiff residents from 1965 to 1973 were reviewed. The mean age and parity of parturients fell. Hospital delivery became almost universal, monitoring the fetus during pregnancy was introduced, and induction and acceleration of labour became commonplace. These developments were not associated with any striking change in either the total perinatal death rate or the timing or cause of perinatal death. Possibly a real change in perinatal mortality between 1965 and 1973 was masked by random fluctuation of small numbers, or possibly factors peculiar to the Cardiff population prevented a decrease in perinatal mortality that would otherwise have resulted from improved medical care. Only by large-scale randomised trials can the true value of induction and other medical developments be assessed.

Factors associated with preterm birth in Cardiff, Wales. I. Univariable and multivariable analysis.

OBJECTIVE: Our purpose was to examine the associations of demographic, social, and medical factors with risk of preterm birth. STUDY DESIGN: By use of the Cardiff Births Survey, a large database of largely homogeneous (white) births in Wales, multivariable analysis by logistic regression examined the relative importance of risk variables associated with preterm birth. RESULTS: Significant independent associations with preterm birth were found (in decreasing order of magnitude) for late pregnancy bleeding, preeclampsia-proteinuria, low maternal weight, low maternal age, early pregnancy bleeding, history of previous stillbirth, smoking, high parity, low or high hemoglobin concentration, history of previous abortion, low social class, bacteriuria, and nulliparity. CONCLUSION: In this population demographic, social, and medical characteristics of the pregnancies showed significant associations with preterm birth.

Factors associated with preterm birth in Cardiff, Wales. II. Indicated and spontaneous preterm birth.

OBJECTIVE: Our purpose was to examine and contrast associations of risk factors with spontaneous preterm birth and indicated preterm birth. STUDY DESIGN: Separate multiple logistic regression analyses were performed of indicated and spontaneous preterm births in a large database of births in Cardiff, Wales. RESULTS: Spontaneous preterm births were associated with young maternal age, low maternal weight, low or high parity, previous abortion, smoking, and early pregnancy bleeding. Indicated preterm births were associated with older age, low weight, previous stillbirth, bacteriuria, and early pregnancy bleeding. CONCLUSION: Spontaneous and indicated preterm births have different overall profiles of association with pregnancy risk factors.

Factors associated with term low birthweight in Cardiff, Wales.

This project was undertaken to examine the associations of a number of risk factors with impaired fetal growth, exemplified by delivery of a low-birthweight infant at term (TLBW). Utilising a large database of largely homogeneous (white) births in Wales, the Cardiff Births Survey, multivariable analysis by logistic regression examined the relative importance of these risk variables. Significant independent associations with TLBW were found (in decreasing order of magnitude) for low maternal weight, pre-eclampsia, smoking, short maternal height, late pregnancy bleeding, early pregnancy bleeding, history of a previous stillbirth, primiparity and older maternal age. In this population, risk of TBLW was associated with a number of demographic, medical and social factors. Of these, only smoking is likely to be modified by intervention.