Abstract
For the purpose of assessing the accuracy of the reading of birth weight, delivery records of normal live births in O-public hospital (6627 cases from 1966 to 1975) and in A-private clinic (9080 cases from 1955 to 1974) are used. In distributions of the lowest digit, the lower two digits and the lower three digits of the birth weights noted in the records were analysed. Scale used in O-hospital is a counter scale (capacity of 10 kg, minimum graduation of 1 g) and that used in A-clinic is a slanted dial spring scale (capacity of 10 kg, minimum graduation of 50 g). In both establishments the same scales were used consistently. In O-hospital the values having [0] or [5] at the lowest digit were 71.4%. In A-clinic the values having [00] or [50] at the lower two digits were 93.9%. In both establishments the rounding of the reading value were evidently recognized. When the differences of roundings by sex, parity, weight grade and years were tested, only in O-hospital did they gradually tend to increase significantly by years. The reading seems to be influenced by type, capacity and minimum graduation of a scale and criteria for reading the weight. The following were discussed: the reading of the weight of neonate was significant from the point of view of community health and the technical development of the scale appropriate for that purpose was important.
