ALTERNATIVE SPECIMEN SOURCES: Methods for confirming positives

An excerpt from a publication on the CDC website.

"Slide 16 lists data obtained when testing 281
subjects with non-HIV disease but with potentially
interfering substances. Subjects included
people with autoimmune diseases, kidney disease,
liver disease, STDs, urinary conditions,
neoplasms, and pregnant women. These urine
specimens tested by ELISA were problematic.
Assuming all of these people were uninfected,
several categories were very reactive by urine
EIA. It was demonstrated that 32/50 people with
kidney/liver conditions, 22/47 with urinary conditions,
25/63 pregnant women, and 17/35 with
neoplasms were repeatedly reactive by the urine
ELISA. Overall, in this group 113/281 (40.2%)
were repeatedly reactive by ELISA and would
advance to Western blot. Two subjects were
urine Western blot false positive (specificity
99.3%) and 11/281 (3.9%) were urine Western
blot indeterminate. The remaining 268 were
urine Western blot negative."

So, this would seem to indicate that the ELISA test is problematic for testing pregnant women. That's almost half the pregnant women giving false positives. What does that do to a 30% statistic arrived at by ELISA testing blood from pregnant women? According to this one study anyway, about 39% might have expected a positive, HIV or not.