PharmPK Discussion - Huge standardized residuals

PharmPK Discussion List Archive Index page

On 6 Dec 2005 at 12:41:26, "Nollevaux, Fabrice (Wavre)" (Fabrice.Nollevaux.-at-.sgs.com) sent the message


The following message was posted to: PharmPK

Dear all,

We are currently doing some compartmental modeling using WinNonlin 4.0.
Here is the result of a fit obtained using a 3-compartment model for
a drug infused over 0.5 h (data are weighted according to the inverse
of the observed value):

    X         OBSERVED   PREDICTED   RESIDUAL    WEIGHT    SE-PRED
STANDARDIZED
              Y
Y                                              RESIDUAL
     0.000      0.000      0.000      0.000      0.000
    0.5000     0.7263     0.7374    -0.1107E-01 0.1129     0.4696E-01
-0.5613E+05
    0.7500     0.4438     0.4536    -0.9786E-02 0.1847
0.3671E-01  -91.89
     1.000     0.3846     0.3902    -0.5610E-02 0.2131     0.3062E-01
-0.3695
     1.500     0.3195     0.2936     0.2595E-01 0.2566
0.1674E-01  0.9877
     2.033     0.2153     0.2215    -0.6235E-02 0.3807     0.1543E-01
-0.3058
     3.000     0.1173     0.1418    -0.2452E-01 0.6988
0.1251E-01  -1.735
     4.000     0.1099     0.9763E-01 0.1227E-01 0.7459
0.1003E-01  0.8037
     6.017     0.6370E-01 0.5718E-01 0.6524E-02  1.287
0.9178E-02  0.6243
     8.000     0.3150E-01 0.3963E-01-0.8125E-02  2.602
0.6785E-02  -1.154
     12.07     0.2330E-01 0.2153E-01 0.1766E-02  3.518
0.7922E-02  0.6247

    CORRECTED SUM OF SQUARED OBSERVATIONS = 0.459232
    WEIGHTED CORRECTED SUM OF SQUARED OBSERVATIONS = 0.132424
    SUM OF SQUARED RESIDUALS =          0.182519E-02
    SUM OF WEIGHTED SQUARED RESIDUALS = 0.995667E-03
    S = 0.157771E-01 WITH     4 DEGREES OF FREEDOM
    CORRELATION (OBSERVED,PREDICTED) = 0.9981

As you may observe, the overall goodness-of-fit is quite good but we
get very high standardized residuals at 0.5 h (-56130) and (in a
lesser extenta at 0.75 h (-91.89).

Please can anybody give us some advice on how to interpret these huge
standardized residuals?

Thanks,

- -

Fabrice Nollevaux, M.Sc.
Pharmacokinetic Analysis and Outcomes Research Manager
SGS Life Science Services,
Biometrics Department,
Vieux Chemin du Poete 10, 1301 Wavre (Belgium)
Tel.: ++ 32 10 421 149  -  Fax: ++ 32 10 421 106
www.sgs.com/life_sciences

Back to the Top

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "Huge standardized residuals" as the subject

PharmPK Discussion List Archive Index page