CMIXCommercial Materials Dispersion Apparatus ITA Experiment
Copyright 1988-2018, All rights reserved.
References in periodicals archive ?
Much more dramatic, however, is the highly statistically significant relationship (p < 0.00001) between centre, PI, FX, CMIX and duration of stay.
Birmingham and Rhymney data set, n = 342 (deaths excluded): percentage of variation of (In) duration of stay which was `explained' by multiple regression model Percentage of variation explained ([R.sup.2] x 100) Variables Model+ Model + centre in the model Model centre +age + sex CMIX 19.5 25.2 25.2 PI + FX 19.2 25.0 25.0 PI 13.0 19.6 20.1 FX 14.1 19.3 19.4 Centre 6.4 6.4 7.4 Sex 1.6 7.0 7.4 Age 2.3 7.1 7.4
As the coefficients associated with FX and PI in the above equation are practically identical (0.48 and 0.51) it is not surprising that the CMIX variable (which assigns equal weighting to FX and PI scores) also explained 25% of the variation in duration of stay when combined with the `centre' variable (see Table IV) and 19.5% when it was entered alone.
Figure 2 presents the duration of stay of patients in CMIX categories 0, 1, and 2.
However, the three subcategories of CMIX had CVs of 0.82, 0.93, and 0.85 suggesting that the subgroups were more homogeneous than the original sample.
By combining these two basic aspects of functional status with a very simple classification of presenting illness we produced a three-point casemix score (CMIX) that was correlated with length of stay and place of discharge.
The CMIX variable (or the two variables PI and FX from which it was derived) explained about 20% of the variability in duration of stay, but the addition of age and sex to the model did not improve it further.