Search PubMed⌕ Search

Biomedical subjects

D G Altman

Publications and source records attributed to D G Altman.

At least 163 records · Page 9Linked to original sources
PubMed1994

Practical problems in fitting a proportional hazards model to data with updated measurements of the covariates.

We review and discuss the practical problems encountered when analysing the effect on survival of covariates which are measured repeatedly over time. Specific issues arise over and above those met with the standard proportional hazards model and concern all stages of data preparation, data analysis and interpretation of the results. Data from a randomized clinical trial of patients with primary biliary cirrhosis, on whom several measurements were taken at regular intervals after entry, are presented as an illustration.

Aged↗
PubMed1994

Development of Design-a-Trial, a knowledge-based critiquing system for authors of clinical trial protocols.

Many published clinical trials are poorly designed, suggesting that the protocol was incomplete, disorganised or contained errors. This fact, doctors' limited statistical skills and the shortage of medical statisticians, prompted us to develop a knowledge-based aid, Design-a-Trial, for authors of clinical trial protocols. This interviews a physician, prompts them with suitable design options, comments on the statistical rigour and feasibility of their proposed design and generates a 6-page draft protocol document. This paper outlines the process used to develop Design-a-Trial, presents preliminary evaluation results, and discusses lessons we learned which may apply to the developed of other medical decision-aids.

Clinical Trials as Topic↗
PubMed1994

Charts of fetal size: 1. Methodology.

OBJECTIVES: To discuss the features of study design and analysis which are necessary to derive valid reference centiles for fetal size. To describe a study which meets the stated criteria. DESIGN: Prospective study of 663 fetuses. RESULTS: The selection of the sample and adequate sample size are of great importance. Prospective collection of data specifically for the purpose of deriving centiles is recommended. It is essential to use statistical methods that take proper account of the increasing variation among fetuses as pregnancy proceeds; such methods are described and illustrated. A study is described which meets the stated criteria for design and analysis, and from which new fetal size centile charts have been derived and are presented in subsequent papers. CONCLUSIONS: Many published studies containing charts (standards) of fetal size are methodologically flawed. Research design and statistical analysis must adhere to sound principles for fitted centiles of size to be valid and so clinically relevant.

Anthropometry↗
PubMed1994

Charts of fetal size: 2. Head measurements.

OBJECTIVE: To construct new size charts for fetal head circumference, biparietal diameter and other head dimensions. DESIGN: A prospective, cross sectional study. SETTING: The routine ultrasound department of a London teaching hospital. SUBJECTS: The fetuses of 663 women seen in the routine antenatal booking clinic whose ultrasound and menstrual dates agreed within 10 days. METHODS: Fetuses were scanned once only for the purpose of the study at gestations between 12 and 42 weeks, when up to 20 dimensions were measured. For each measurement separate regression models were fitted to estimate the mean and standard deviation at each gestational age. Centiles were derived by combining these two regression models, assuming that the measurements have a normal distribution at each gestational age. RESULTS: A total of 594 fetuses had their biparietal diameter measured and their head circumference measured directly. Both head diameters were recorded for 587 fetuses and the circumference was also derived from these, as was the cephalic area. New charts are presented for biparietal diameter (both outer-outer and outer-inner), head circumference (directly measured and derived from diameters). The directly measured head circumferences were consistently (by about 1%) greater than those derived from measurement of the head diameters. The new charts are compared with previously published charts that are in wide use. Charts for occipitofrontal diameter, cephalic index and cephalic area are also presented. CONCLUSIONS: We have constructed new size charts for the fetal biparietal diameter and for head circumference, both measured directly and derived from head diameters. We have demonstrated the difference between the size charts constructed from these two sets of values and hence the importance of using the appropriately derived chart when assessing the head circumference. The differences between the new charts for biparietal diameter and head circumference and previous ones may be largely due to methodological differences.

Anthropometry↗
PubMed1994

Charts of fetal size: 3. Abdominal measurements.

OBJECTIVE: To construct new size charts for fetal abdominal circumference and area. DESIGN: A prospective, cross sectional study. SETTING: The routine ultrasound department of a London teaching hospital. SUBJECTS: The fetuses of 663 women seen in the routine antenatal booking clinic whose ultrasound and menstrual dates agreed within 10 days. METHODS: Fetuses were scanned once only for the purpose of the study at gestations between 12 and 42 weeks, when up to 20 dimensions were measured. Separate regression models were fitted to estimate the mean and standard deviation as functions of gestational age. Centiles were derived by combining these two regression models, assuming that the measurements have a Normal distribution at each gestation. RESULTS: A total of 610 fetuses had their abdominal circumference measured directly. Abdominal diameters were recorded for 425 fetuses and the circumference was also derived from these, as was the abdominal area. New charts for abdominal circumference (directly measured and derived from diameters) are presented. The directly measured circumferences were consistently (by about 3.5%) greater than those derived from measurement of the abdominal diameters. The new charts are compared with previously published charts that are in wide use. A chart for abdominal area is also presented. CONCLUSIONS: We have constructed new size charts for the fetal abdominal circumference, both measured directly and derived from abdominal diameters. We have demonstrated the difference between the size charts constructed from these two sets of values and hence the importance of using the appropriately derived chart when assessing the abdominal circumference. The differences between the new charts and previous ones may be largely due to methodological differences.

Abdomen↗
PubMed1994

Charts of fetal size: 4. Femur length.

OBJECTIVE: To construct a new size chart for fetal femur length. DESIGN: A prospective, cross sectional study of fetuses scanned once only for the purpose of the study at gestations between 12 and 42 weeks. SETTING: The routine ultrasound department of a London teaching hospital. SUBJECTS: The fetuses of 663 women seen in the routine antenatal booking clinic whose ultrasound and menstrual dates agreed within 10 days. RESULTS: Femur length was measured on 649 of the 663 fetuses. A linear-cubic regression model was fitted to estimate the mean and a separate linear regression to estimate the standard deviation. Centiles were derived by combining these two regression models, assuming that the measurements have a normal distribution at each gestational age. A new chart for femur size is presented and compared with previously published data. CONCLUSIONS: We have constructed a new size chart for fetal femur length taking into consideration the increasing variability with increasing gestational age. We have compared our chart with other published data, and believe that the differences seen may be largely due to methodological differences.

Cross-Sectional Studies↗
PubMed1993

Construction of age-related reference centiles using absolute residuals.

This paper proposes a simple approach to the parametric derivation of age-related reference ranges which avoids the creation of arbitrary age groups, copes easily with a non-linear relation between variability and age, and is computationally simple. After the mean is modelled as a function of age, the age-specific standard deviation is estimated by regressing the absolute residuals on age. The method assumes that the data are, or can be transformed to be, normal at each age, and uses the properties of the half normal distribution. An example is given using 450 measurements of fetal foot length.

Age Factors↗