Evaluating Interventions to Improve Stem Cell Donor Availability

Behavioral narrative interventions and one video treatment designed to increase stem cell donor availability for confirmatory typing.

2026-01-22

2027-12-31

Stated CT availability, Actual CT availability

Survey completion rates (1), knowledge about extraction methods (2), contact information updates (3), absence reporting (4), actual CT availability (5)

This RCT tests whether behavioral narrative interventions increase stem cell donors’ availability for confirmatory typing (CT). The study uses five arms: three narrative video treatments, one video control, and one plain (no-video) control, and evaluates effects on both stated and actual CT availability. The narratives target reciprocity, community, and “zero cost.” The primary outcomes are (i) stated CT availability (self-reported) and (ii) actual CT availability.

Not available

Computer-generated random assignment

Individual donor-level

No

NA

Planned sample sizes follow the invitation split of 1.8M total invitations: Plain control (no video): 40,000 invitations Each of the four video arms (3 narrative treatments + 1 video control): 440,000 invitations each (total 1,760,000) Assuming 8% survey completion, the planned number of observations for stated CT availability (survey completers) is: Plain control: 40,000 × 0.08 = 3,200 Each video arm (T1, T2, T3, video control): 440,000 × 0.08 = 35,200 Total stated-outcome N: 144,000 Assuming 2% CT requests among completers over two years, the planned number of observations for actual CT availability (among CT requests) is: Plain control: 3,200 × 0.02 = 64 Each video arm (T1, T2, T3, video control): 35,200 × 0.02 = 704 Total CT-request N: 2,880

The experiment has five arms: three narrative video treatments (T1–T3), one video control, and one plain (no-video) control. We will send 1,800,000 invitations in total. The plain control receives 40,000 invitations; the remaining 1,760,000 invitations are allocated equally across the four video arms (440,000 each).
Assuming an 8% survey completion rate, the expected number of observations for survey-based outcomes (stated CT availability) is 3,200 in the plain control and 35,200 in each video arm, for a total of 144,000 completed surveys.
Assuming a 2% CT request rate among survey completers over two years, the expected number of observations for behavioral outcomes (actual CT availability among CT requests) is 64 in the plain control and 704 in each video arm, for a total of 2,880 CT requests.

(1) Stated CT availability (1.1) Planned comparisons: 3 treatments vs. video control Using a Bonferroni-adjusted α = 0.0167 for the family of three comparisons and 80% power, with 35,200 participants per group, the minimum detectable standardised mean difference is Cohen’s d = 0.0244. (1.2) Video control vs. plain (without video) control Assuming a two-sided t-test with α = 0.05 and 80% power, and n = 35,200 (video control) vs. n = 3,200 (plain no-video control), the minimum detectable effect is Cohen’s d = 0.0517. (2) Actual CT availability (binary; among CT requests) (2.1) Planned comparisons: 3 treatments vs. video control Using a Bonferroni-adjusted α = 0.0167 for the family of three comparisons and 80% power, with 704 observations per group and a baseline rate of 80%, the minimum detectable difference in proportions is 0.0642 (6.42 pp.).