In the U.S., erythropoietin is typically delivered as 3x/week intravenous injections in the ESRD population. This turns out to be a fairly significant cost, as Medicare-related payments for EPO comprise about 25% of the ESRD budget, which itself is a significant chunk (about 6-7% as of 2002) of the overall Medicare budget.
This cost could probably be decreased by moving to subcutaneous dosing of erythropoietin. One study in France showed in a 1-year crossover study in which ESRD patients received subcutaneous dosing for 6 months, followed by intravenous dosing for 6 months, the EPO dose required to keep the Hgb within a usual target range was far less in the subcutaneous group (74 Units/kg/week) compared to the intravenous group (156 Units/kg/week).
I would imagine that the main barriers towards achieving this would be (a) financial–as it stands now, dialysis units make more money if they administer greater amounts of EPO, (b) legislative–the Social Security Act which has made the Medicare ESRD Program into what it is today specifies that self-administered medications are not covered by Medicare, (c) patient preference–most patients would prefer to avoid another needle stick each dialysis session, and (d) the status quo–that’s just the way we’re used to doing things.