Introduction: Chile has a low organ donation rate of 7.8 donors pmp. The three main reasons are low familial consent to donation (50%), lack of possible donor (PD) referral (87%), and high number of discarded PDs during the procurement process (80%). While there have been interventions to target the first two, the third one hasn't had any attention. What factors contribute to PD discards? Are there differences between Local Procurement Coordinations (LPC), hospitals, or socioeconomic level?
Methods: We obtained the national database of all patients who entered procurement follow-up between 2013-2022. We used multilevel multivariate logistic regressions for the individual, hospital, and LPC levels. The explanatory variables were patient’s gender (the only accessible individual-level variable), hospital variables (complexity), and socioeconomic variables of the municipality where the hospital is located (poverty, rurality, years of schooling, Municipality Development Index, and % of native population). The binary dependent variable was defined as discard or contraindication for donation (1), against effective donor or familial refusal to donation (0). Additionally, we grouped all discard and contraindication causes in 4 groups: patient, process, social, and other causes. We adjusted the same type of regressions for each of them separately. We excluded private institutions to reduce bias.
Results: For the general model we found that the only significant variables were gender (OR=1.21; p<0.0002), poverty (OR=28,160; p<0.0002), and rurality (OR=14.65; p=0.007). We found significant random effects for both higher levels (variance of 0.13 y 0.12 respectively for hospital and LCP). For the group submodels, we found significant random effects for the hospital level in all of them (variances of 0.13; 0.23; 0.78; 0.29 respectively). Nonetheless, the LPC level had significant random effects only at the patient and process groups (variance of 0.23 and 0.48 respectively). The patient’s gender was significant for all models except Other (OR=1.38-p<0.034; OR=1.31-p=0.008; OR=1.45-p<0.02), and poverty was significant for all of them (OR=2,575-p<0.01; OR=1,448,342-p=0.005; OR=663,187-p<0.002). Additionally, the model for Process showed years of schooling as significant (OR=0.64; p<0.003), and the one for Social and Other had rurality as significant (OR=231.8; p=0.0012 and OR=9.85; p=0.04 respectively).
Conclusion: Our results show that the Chilean organ donation system is unexpectedly impacted by the structural social inequities, which we can’t target from our field. But we also found a large variance of PD discards among hospitals and LPCs, evidencing a lack of standardization of the procurement process. We can improve the procurement process at the hospital and LPC levels to reduce this variation and thus, improve our system. We propose to standardize the procurement process by using technological tools for monitoring and evaluation.