BrianD

When we do the cross-match for the hospital, we issue the blood as "suitable", rather than "compatible", and issue a warning that, under such circumstances, the observations during the transfusion are of paramount importance - not that they aren't anyway.

You might consider keeping O Neg Emergency Issue RBC's ( 2 to 4 or other; your choice) that are pretagged for Emergency Patient (in your case, would cover ER trauma, if needed, and/ or L&D, or separate emergency issue units available for each); segments prepulled and labled with unit# lable for crossmatch after issue; and cooler/ ice and/or other means of tranport predetermined.

Tom, I agree your decision was practical and timely.  Given the situation, there are 3 questions that come to my mind:
Who is paying for the IRL testing?  I hope not the patient because this testing seems to have been done in hopes of an EX-supervisor "scoring" points against a newcomer. Has it been established that the patient is Kell negative?  Have you talked with your director to ascertain why he or she felt your decision couldn't be supported?  Your director has established a potentially bad precedent for your transfusion service.  best of luck.

Is your department a separate cost center? Are you being held directly responsible or is the Clinical Laboratory Administrator?
Look at the numbers of units returned to the donor center as a percentage of units received. If it is high, you need to re-think your inventory control as each unit returned is a unit you spent time/money confirming.
How do you confirm your donor units? Can your testing requirements be reduced?
Are you doing IgG crossmatches on all patients or just those with positive screen or history of clinically significant antibody? Electronic?
Are you doing a 3-cell screen versus 2-cell antibody screen?
Does your hospital host mobile blood donations? We negotiated a price reduction on rbc units if we met mutually agreed donation goals.
Do you have more than one blood supplier?
Have you optimized your reagent red cell shipments with your usage? Do you monitor them monthly?
Have you optimized your gel card shipments with your usage?
Do you make your own ProVue controls or purchase them?
Do you use both tube and gel methodologies? If so, you must purchase proficiency surveys accordingly. I purchase JAT only and no longer do J.
Just some thoughts, hope they help.
Dan

If this policy is defined in your SOPs you should follow them.  In my opinion, what you did was entirely logical and safe.  It may have even prevented a presumably K neg patient that is known to make antibodies from being exposed to the K antigen and making anti-K as well.  If you had a patient with multiple antibodies even the IRL may not be able to rule out everything and may suggest using antigen negative units for specificities that they can't rule out.  Lastly, you give blood every day to patients that have had anti-K ruled out using only one single dose K pos cell, because this is the case with every patient with a negative antibody screen.  They don't even get the advantage of an AHG xm (most places) which your patient got.
 
You could also check with your IRL and see how many K+ cells they require to rule it out.  They may require only one.
 
Afterthought: be careful of turf wars with ex-supervisors.    Look for ways for everyone to save face.   

Gel won't pick up weak rouleaux but it will give false positives with really strong rouleaux and there is no way to fix it except to do tube testing. Also, what we call "gel junk" (usually antibodies to the antibiotics in the gel diluent--made worse by improper storage) must be resolved by tube. You can also run into problems with warm autos and cold antibodies as described above. Anti-M's come up really nicely in gel--often showing dosage. You almost never find a Lewis antibody in gel. Small hospitals that choose to have no backup method must realize that they will not be able to do AHG xms at all for patients with the above problems that require AHG xms. You may need a policy for defining these things as clinically insignificant so IS xm is allowed but that doesn't work if the patient has an anti-K also.
Many reference labs are loathe to do gel and don't really trust it. You may find that you can't send out problems because your reference lab won't find the reactions in tube. There are real antibodies that react in gel and not tube so they aren't always completely right to rely only on tube.
Remember there are new instruments/vendors entering the US market in 2012 due to Ortho's patent rights expiration. Bio Rad will be able to market the European gel system in the US and Ortho is rolling out an improved automated instrument. I saw a European vendor at the AABB meeting with gel cards with 8 tubules instead of 6. There may be some new ways to skin this cat after next year.