| Description |
Routine testing of the urine of diabetic patients for the presence and amount of sugar and acetone is a standard procedure on hospital wards. However, studies have demonstrated a significant human error factor in the results of these tests. The purpose of this study was to measure the presence and degree of bias on urine sugar and acetone tests as performed by hospital nursing personnel. The two types of bias considered were (1) the influence of the concentration of sugar in the urine on the interpretation of acetone concentration, and (2) the influence of prior knowledge of a patient's past urine sugar and acetone on subsequent urine tests. The sample 122 included 51 registered nurses, 47 licensed practical nurses, and 24 nurse's aids or orderlies in staff positions in two general hospitals. Each subject test six urine specimens previously prepared to exact concentrations by the investigator. A questionnaire designed to measure the subject's knowledge of diabetic urine test, his educational preparation, and the frequency with which the subject performed the test on hospital wards was completed by each subject following completion of the urine testing. The influence of these variables on the amount of bias was examined. Variation in sugar concentration from high to low did not influence (or bias) the interpretation of acetone concentration by hospital nursing personnel. This lack of bias on acetone was maintained when the actual concentration of acetone was varied from moderate to negative. Prior knowledge of a patient's past acetone values was found to bias subsequent acetone readings by nursing personnel. The bias was in the direction of earlier tests. Prior knowledge of the sugar values did not, however, show a influence on subsequent readings. Neither the subject's knowledge about diabetic urine tests nor his educational preparation was found to have any relation to the degree of bias he possessed. However, the frequency with which the subject performed urine test on the hospital wards had a positive correlation to the amount of bias. A mean error score of 34 percent was shown in answer to test items measuring the subject's knowledge of basic facts about diabetic urine tests. Licensed practical nurses had a statistically greater error score than did registered nurses. No statistical difference between any of the groups on accuracy in performance of the urine tests was found. However, the overall mean error score for all 122 subjects on performance of the 1 2 urine test was 4.3, meaning 36 percent of all reported readings were in error. This study thus demonstrated that significant errors in urine testing for sugar and acetone are made by all groups of nursing personnel employed in an adult medical-surgical hospital service. These errors bear no relation to the educational level attained by the individual tested, and they appear to be increased with those individuals who perform the test most frequently. Furthermore, the reliability of these tests is impaired when the subject has knowledge of the patient's previous urine tests. Routine testing of urine for sugar and acetone on the average hospital ward must, therefore, be considered inaccurate until it clearly established that the testing individual have a reliable, reproducible technique. Such inadequacies constitute nursing care which is less than ideal. Means must be south to improve this particular facet of day to day patient care. One might suggest that testing errors could be reduced by such techniques as intensive inservice training, by frequent reassignment of individuals responsible for urine testing, be periodical testing of the individuals responsible for reporting diabetic urines, or by two or more of these approaches. Such techniques, of course, must be subject to careful scrutiny. Accordingly, studies which are similar in design to the present study might be employed to evaluate ward personnel performance before and after such training techniques. |
