Recently it has been suggested (Massa 1992; Howarth & Reid 1993), from the C IV ultraviolet resonance line profile of HD 93521, that there is a high-speed component in the polar outflow from the star as well as a low-speed component in the equatorial regions. In this paper we present theoretical calculations of the line profiles that would be produced by such a model. We find from Hubble Space Telescope (HST) Goddard High Resolution Spectrograph (GHRS) observations of HD 93521 that the observed C IV profile can be produced if the star has an inclination angle very close to 90 deg and if the star is surrounded by a thin disk, whose half-width is approximately 3 deg in latitude. The geometry of this disk is similar to what one would expect from the wind-compressed disk model of Bjorkman & Cassinelli, so this star may provide an ideal observational test of that model. In addition to the C IV resonance line, we examine both the Si IV and N V resonance lines. The Si IV line exhibits low-velocity absorption that is similar to that seen in the C IV line, but the emission is much weaker. On the other hand, the N V line has weaker absorption and much stronger emission than the C IV line. N V is a higher ionization state than C IV, so it is likely that N V is one stage above the dominant state, N IV. Apart from fitting individual line profiles, we also examine the differences between the two Hubble Space Telescope observations of HD 93521. We find evidence for a pair of narrow absorption components, seen at low velocity, as well as evidence for a discrete emission feature in the blueshifted absorption cores of the lines. This blueshifted emission at low velocity can cause what instead appears to be an interstella absorption line. Without multiple observations that can reveal the temporary emission, one must be very careful when determining interstellar column densities to stars like HD 93521.