The following regression equation was used for

 prediction:



 y = 3.43 + [(-0.02) age] + [(0.25) in situ + (1.03) local +
 (0.01) regional/direct + (0.26) regional/nodes +
 (0.64) direct/nodes + (-0.06) regional, NOS +
 (0.66) non-local] + [(-1.30) none + (-0.35) surgery +
 (-0.85) radiation + (-0.61) chemotherapy +
 (-0.36) surgery, radiation + (-0.21) surgery,
 chemotherapy + (-1.04) surgery, chemotherapy, radiation
 therapy + (-0.23) surgery, radiation, chemotherapy +
 (-1.16) radiation, chemotherapy + (-0.46) chemotherapy,
 radiation + (-0.59) chemotherapy, hormonal + (-1.13)
 radiation, hormonal] + [(-0.15) none + (0.31)
 radiation + (0.93) chemotherapy]

 The variable age was continuous in this equation. The

 levels of the other three variables were coded as ones and

 zeros in the analysis. For illustrative purposes, the values

 of b for the levels of each independent variable are con-

 tained in brackets in the statement of the equation. In

 order to predict an individual's survival, the appropriate

 values of b for their age, stage of cancer, initial treat-

 ment, and subsequent treatment were summed and added to the

 value for the intercept. For example, the prediction

 equation for a 65 year old individual with local disease who

 received surgery initially and chemotherapy at a later time

 would be as follows:



y = 3.43 + (-0.02) (65) + (1.03) + (-0.35) + (0.93)