Improving the Availability of Stem-Cell Donors - A Letter and Email Intervention

Abstract

Attrition of stem-cell donors at the confirmatory typing stage non-trivially reduces the supply of life-saving stem-cell transplants. Attrition varies worldwide from 20% to over 70% at the stage of confirmatory typing, which are the last screening steps of prospective donors before donation. In a natural field experiment with DKMS Germany (registered at AEARCTR-0008938), a major stem-cell donor registry, we randomly vary the framing of email and letter invitations in an existing initiative, aimed at reducing donor attrition at the confirmatory typing stage and sorting donors on future availability. The initiative asks donors to fill out a health questionnaire and to report any periods where a donor is not available to donate longer than three weeks in advance. In the three experimental conditions, each with 8,500 recipients who are registered donors, invitations are framed so that donors are asked to (i) join a team of “quickly available donors”, or (ii) join a team of “highly committed” donors. The baseline group receives a letter (and email) with neither framing. We also compare out treatments to a pure control that got no invitation and the previous letter DKMS used. Our results show that i) the quickly available letter halved donor attrition in the first year compared to the control letter after receiving the letter. Second, the baseline and committed letters sort donors on availability persistently. We find that the committed group allows one to identify a group of potential donors that persistently behave in a manner consistent with being more committed, as measured by advance unavailability reports, and completing another follow-up health screening two years after the original invitation. Further, in the long run, donor availability is very similar, despite the strong effect of the overall letter framing on participation. If anything, the letter that reduces participation has a positive effect on donor availability. Overall, this suggests that the causal effect of participation as such is not as important as the invitation, and that the screening of donors using letters may be useful in settings where there is already a large enough pool of registry members to choose from.