合併症の少ない低出生体重児の出生体重からみた入院日数予測 第2報 : 入院予測日数を指数関数回帰式から算出する方法の検証と臨床応用 <原著>

本論文第一報にて我々はNICUに入院した低出生体重児の入院予測日数を2病院の出生体重と入院日数のデータをもとに指数関数の曲線回帰式から算出する方法を報告した.

本論文は①第一報で報告した回帰曲線の入院日数予測は, その後に入院した低出生体重の入院日数予測に実用可能だったか ②両病院NICUは移設・移転しており, そのことは入院予測日数に影響したか ③入院予測日数は両施設でどう異なるかを示し, 指数式回帰曲線からの入院日数予測の臨床応用において考慮する条件を検討するとともに, 施設毎に入院予測日数を算出する必要性を述べることを目的とした.

先行研究以後にNICUに入退院した低出生体重児N病院670人, R病院280人の出生体重と入院日数のデータからPASW.18により指数方程式による回帰曲線を出力させ入院予測日数をもとめた.

N病院では先行研究の入院予測日数の回帰曲線は今研究のNICUに入院した低出生体重児の指数モデルの回帰曲線とほとんど変わらず, 先行研究の入院日数予測はその後入院した低出生体重児の入院予測として実用可能であった. また先行研究以後の入院予測日数の回帰曲線は移設前後でほぼ等しく, 移設の影響はほとんど見られなかった.

一方, R病院では先行研究と移転後の回帰曲線は大きく異なり, 先行研究の入院予測は移転後の入院日数予測には適応できなかった. また, 移転後も入院予測日数の回帰曲線はまだ変動がみられ, 臨床応用するには今後も入院予測日数の見直しが必要であることを示唆した.

N病院とR病院の回帰曲線は異なっていたことから, 各施設で回帰曲線を作成し入院予測日数を求めることが正確なクリティカルパス作成につながる.

We reported a method to calculate and predict hospitalization period of the low birth weight infants in two NICU using a curve regression of exponential equation based on the data of birth weight and the hospitalization days by the first report.

The purpose of this study is whether the estimated hospitalization period of the regression curve was practical and possible to calculate the hospitalization days of the LBWI who had been hospitalized afterwards.

The subjects were 670 and 280 infants with a birth weight less than 2500 g and admitted to NICU within 0-2 days after birth, in N hospital and R hospital.

Data of hospital stay, birth weight and gestational age at birth were collected and conducted to statistical analyses using PASW Statistics 18.

NICU stay days' plots against birth weight fitted the exponential function well, and the obtained regression curve formula and the parameters showed the estimated hospitalization period by birth weight.

In N hospital, the regression curve of the estimated hospitalization period by the previous work was similar to the regression curve of this research. It was suggested the estimated hospitalization period of the previous work be adjusted to the hospitalization of LBWI who had been hospitalized afterwards. Moreover, the regression curve was almost equal before and after the transfer of hospital, and it means the transfer was little influence in hospitalization days.

In R hospital, regression curve of the previous work and the present study is fairly different, and it was suggested the estimated hospitalization period of the previous work has clearly indicated no adjustment after transfer hospital.

Since regression curve of estimated hospitalization period of LBWI was different between N and R hospitals; and request for making the regression curve in each facility could make an accurate critical path.

The calculation hospitalization period by the regression curve may alleviate the anxiety of the infant's mothers and family. Moreover it may be utilized for evaluation of nursing care.