We examine the targeting effects of increased scrutiny in the screening of Disability Insurance (DI) applications using exogenous variation in screening induced by a policy reform. The reform raised DI application costs and revealed more information about the true disability status of applicants at the point of the award decision. We use administrative data on DI claims and awards and merge these with other administrative data on hospitalization, mortality and labor market outcomes. Regression Discontinuity in Time (RDiT) regressions show substantial declines in DI application rates and changes in the composition of the pool of applicants. We find that the health of those who are not discouraged from applying is worse than those who are. This suggests that the pool of applicants becomes more deserving. At the same time, compared with those who did not apply under the old system of more lax screening, those who are discouraged from applying are in worse health, have substantially lower earnings and are more often unemployed. This indicates that there are spillovers of the DI reform to other social insurance programs. As we do not find additional screening effects on health at the point of the award decision, we conclude that changes in the health condition of the pool of awarded applicants are fully driven by self-screening of (potential) applicants.