Kathy, in my practice, i threw out that O'Sullivan's test in 1988 and replaced it with a 50g post-prandial.  I have had much better results with eliminating false positives, which as you may know means the woman is then subjected to a 100g load of (unnatural) glucose and blood draws are then done once an hour for 3 hours.  Those women who tested high after the PP were then doing their 3hour GTTs and I almost always got a borderline or "true" gestational diabetes.  I find that most of the women who then went on to do their 3hour GTT just had a fluke in their tests and can repeat it and get normal values, which is why I dislike this way of testing so much.  That leaves only a few who are at risk and benefit greatly from this diet.  But alas, I do have to transfer them to an OB and they just end up on insulin and your typical ADA diabetes diet.
Anyway, I do encourage the paleo diet with my clients and those who participate have lowered their blood pressures, lowered their glucose tests, and had "normal" sized babies, etc.  I truly believe this is the way they were meant to eat. I wish I could get all my mamas to eat this way, lol. THen I could perform a "study".  WOuld be mostly anecdotal since my practice is small.
 I must say though,that healthy vegetarians also have healthy pregnancies and healthy outcomes.  <shrug>
> In my practice caring for pregnant women, it is proper protocol to do a
> diabetes screen at 28 weeks. The woman is given a 50g load of glucose to
> drink, then her blood sugar is monitored for elevation 1 hour later. We
> subject a person to a very unnatural situation, then see if she has a
> reaction. Quite absurd, really. If they have gestational diabetes, what
> do we do? We put her on the type of diet that she should have been on
> anyway!
>
> Kathy
>