Search PubMed⌕ Search

Biomedical subjects

K Uberla

Publications and source records attributed to K Uberla.

At least 91 records · Page 5Linked to original sources
PubMed1978

[Experimental design in phase I (AUTHOR'S TRANSL)].

Clinical pharmacology has its domain during phase I. The formalisation of phase I is not very advanced. Sometimes it is argued that the design of experiments during that phase is unnecessary. Just the opposite is true. Trying a new substance for the first time in man poses a special necessity for the design of experiments in a scientific context. Phase I is described in a formal way. The design of experiments with a single subject -- the first one to be exposed to a new substance -- is described. Finally Latin squares are dealt with as a useful tool for phase I.

Clinical Trials as Topic↗
PubMed1976

[Preventive screening in two factories. I. Methods and results (author's transl)].

A screening study for coronary disease, chronical bronchitis, diabetes mellitus, hypertension, peripheral circulatory disturbance and overweight is described. 2429 persons aged over 40 years and working in two factories were studied. Typical laboratory tests, a short standardised examination by a physician and a questionnaire were used. In a 10 per cent sample the questionnaire was repeated by an interview and the serum was sent to the laboratory not only by mail, but also by a special car transport in a cooled transport box. The results of the laboratory tests are presend according to age, sex and factory. The family doctor had to be informed in nearly 70 per cent of the men and about 60 per cent of the women because of at least one suspicious symptom or sign. There was a pathological value of glucose in the urine in 14.7 per cent, a rise of glucose in the blood (above 113 mg per cent) in 5.7 per cent, of triglicerides (above 181 mg per cent) in 12.6 per cent, of cholesterol (above 264 mg per cent) in 15.4 per cent, of uric acid (male above 7.7 mg per cent, female above 7.1 mg per cent) in 6.8 per cent, of creatinine (above 1.3 mg per cent) in 6.4 per cent and the presence of albumin in urine in 2.2 per cent of the cases.

Adult↗
PubMed1976

[Preventive screening in two factories. II. Exactitude of laboratory tests and questionnaire (author's transl)].

The exactitude of some widely used laboratory tests (triglycerids, cholesterol, glucose, uric acid, creatinine) is examined in a situation relevant for practising physicians. Different statistical methods for reporting errors are compared. Accuracy and precision are useful measures for the quality of analytic procedures. They are not sufficient for medical judgement for a single patient, since sampling procedures, type of sample or transport and storage of sample are not considered in accuracy and precision. Such sources of error can largely devaluate the quality of accuracy and precision and of the analytic procedures. Error coefficients are proposed as a new method of reporting laboratory errors. The mathematical model starts with an analysis of variance and total error, technical error and transport error are defined. The study shows that the error introduced by transporting serum via postal service compared to that conveyed in a cooled transport box by a special car is considerable. It is nearly impossible for a physican to use single laboratory value for a rational decision "normal-not normal" if the values are based on samples which are sent by postal service and if the values lie in the broad borderline between the normal and pathological area. This is especially true for creatinine (transport error 47.0 per cent) and uric acid (transport error 38.7 per cent), but not so much for cholesterol (22.9 per cent), triglicerids (14.3 per cent) and glucose (13.3 per cent). Variables with high transport error like creatinine should not be used in screening programs, as long as the transport is made by postal service and the method is not improved. The agreements between the answers of a questionnaire and of an interview are analysed with the same patients (n=235). These agreements lie between 98.7 per cent and 61.7 per cent depending on the single question. The exactitude of medical questionnaires is in the same size order as the exactitude of laboratory tests-at least when the sample is sent by postal service.

Clinical Laboratory Techniques↗
PubMed1976

[Preventive screening in two factories. III. Relations between quantitative variables and questionnaire-items (author's transl)].

In the preceding articles a preventive screening model in two factories was described, the results were given and the exactitude of laboratory tests presented. In this paper the relations are studied between all the observed variables. Contingency-tables and the x2-statistic are used. For the 57 observed variables-(laboratory variables, e.g. triglicerides, cholesterol, glucose, uric acid, creatinine and questionnaire-items designed to screen for angina pectoris, myocardial infarction, claudicatio intermittens and bronchitis)-all possible contingency tables between two variables were calculated (1596 tables). The significant connections are reported in form of a hypothesis-matrix. There exists no statistical correlation between laboratory tests and questionnaire-items. The anamnestic questionnaire cannot be replaced by laboratory tests and vice versa. Nearly all items of the questionnaire are correlated. There are some correlations between laboratory tests, blood pressure and smoking habits.

Clinical Laboratory Techniques↗