PREDICTORS OF TRANSFUSION TIMING AND PRIORITIZATION IN A REGIONAL KENYAN HOSPITAL: INSIGHTS TOWARD TRIAGE TOOL DEVELOPMENT FROM MATERNAL AND INPATIENT BLOOD BANK PRACTICES ekp17@pitt.edu

 

PREDICTORS OF TRANSFUSION TIMING AND PRIORITIZATION IN A REGIONAL KENYAN HOSPITAL: INSIGHTS TOWARD TRIAGE TOOL DEVELOPMENT FROM MATERNAL AND INPATIENT BLOOD BANK PRACTICES

7

 

 

 

 

United States

Oral or Poster

In low- and middle-income countries (LMICs), limited blood product supply necessitates real-time clinical triage of transfusion requests. Yet, few systems employ standardized prioritization tools. Understanding predictors of transfusion fulfillment and time-to-transfusion (TTT) can reveal implicit triage patterns and inform development of structured allocation strategies.

A 6-week prospective observational study was conducted at two hospital-based blood bank laboratories in Nakuru County, Kenya: maternity unit and general inpatient service. Patients with blood requisitions submitted between 00:00–11:59 were screened. Variables included age, sex, pregnancy status, hemoglobin (Hb), urgency, clinical indication, blood group, product type, and transfusion timing. Statistical analyses included chi-squared tests, t-tests, ANOVA, Cox proportional hazards models, and decision trees stratified by clinical ward.

Among 161 patients (68 maternity, 93 inpatient), 47 (29.2%) were transfused. Transfused patients had lower Hb (7.55 vs. 12.1 g/dL, p<0.001). Although maternity patients had significantly higher mean baseline Hb than inpatients (8.64 vs. 6.87 g/dL, p<0.01), TTT was significantly shorter in maternity patients (mean 6.2 hrs) vs. inpatients (17.8 hrs, p=0.021), despite similar urgency labels. Fewer units were dispatched than requested (1.18 vs. 2.36, p<0.01). Inpatients with Hb ≥7.6 g/dL and prior transfusions had the shortest TTT (4.0 hrs), while urgent obstetric indications (e.g., APH, pre-eclampsia) predicted rapid transfusion in maternity units (TTT 3.8 hrs). Crossmatch within 24 hours strongly predicted transfusion fulfillment.

Transfusion acts as de facto triage in LMICs. Predictors such as hemoglobin level, urgency, and transfusion history can guide equitable, data-driven blood allocation tools to optimize delivery under resource constraints.