わが国の平均出生時体重の年次推移

保健所別,単産・複産合計

抄録

We investigated the secular trends in the mean birth weight of total births (including the single and multiple births) at health centers, using the vital statistics of four prefectures: Iwate, Ishikawa, Yamanashi and Nara. We also compared the trends in the mean birth weight at the health center level with those at the prefectural and regional levels, which was reported in the previous paper (Ref. 1).
1. Only in these four prefectures are the tables of frequency distribution of the birth weight of total births by sex and by health center available annually since 1961, and these tables are exactly the same or can be used as tables employed in the vital statistics of the Japanese government.
The mean birth weights and their standard deviations of each available year (1961-72) were computed by sex and by health center (Table 1). Although there are tolerable fluctuations annually at the health center level, a remarkable increase of the mean birth weight is clearly noted (Fig. 1).
2. The regression coefficients of the straight lines were computed from the time series of the mean birth weight. The mean birth weight in 1960 was then estimated from the regression lines. In three other prefectures besides Nara, the statistically significant negative associations are recognized between the linear regression coefficient (which corresponds to the annual increase rate of the mean birth weight of each health center between 1961 and 1972) and the mean birth weight of 1960-61 (Table 2). As we mentioned in the previous report, this negative association suggests that the annual increase rate during this period is closely related to the levels of the birth weight of each health center in 1960-61.
Then, two kinds of the regression equations and the residual mean squares were computed from the two variables: the estimated mean birth weight in 1960 and the linear regression coefficient pointed above (Table 3).
3. In order to investigate the areal differences of the mean birth weight in each of the four prefectures, the coefficient of the correlation in rank among the mean birth weight of each year (1961-72) was computed by sex. The frequency of the statistically significant positive association of the coefficient of the rank correlation is low in each prefecture. However, this frequency of the statistical association increases markedly when the summations are made at the health center level between two sets of prefectures (Iwate and Ishikawa, Yamanashi and Nara) in which the mean birth weight is nearly at the same level (Table 5). This fact suggests that it would also be possible to determine the areal differences of the mean birth weight at the health center level, if the choice is made of a prefecture with rather more health centers and/or, treating a few adjacent prefectures or one region all together.
In each prefecture the coefficient of the variation of the mean birth weight was computed by sex in each year (Table 6). While the mean birth weight increases gradually, the areal differences tend to decrease.
4. In conclusion, we would like to emphasize that it is advisable to apply the “mean birth weight” as a community health indicator at the health center level as well as the prefectural and regional levels, taking into consideration the rather remarkable annual fluctuations of the mean birth weight and of the occasionally undetectable areal differences among prefectures.