A Multiattribute Decision Analysis for the Level of Frozen Blood Utilization

The technology for freezing red blood cells, available in many blood banks, appears to provide higher quality, more disease free blood with reduced inventory costs, but at a higher cost per unit. A diverse range of operational, therapeutic, and economic effects and great uncertainty concerning its performance make a decision on its level of use difficult. Multiattribute decision analysis is applied to assist a blood bank administrator in determining the most desirable utilization level. Use of the model is demonstrated for a hypothetical individual hospital blood bank and for a hypothetical regional system of interacting blood banks. In these cases our significant results include the specification of the most critical uncertainties and information needs identified by a sensitivity analysis. The methodology focuses on workable procedures for the establishment and systematic aggregation of objectives and measures of effectiveness and for practical problems in probability estimation and utility assessment. Statistical and utility independence properties are considered and used, whenever possible, to simplify assessment and calculations. Actual utility assessments and the resulting decision analyses are performed for two decisionmakers. We conclude with a critique of the methodology and a discussion of experience gained in its use. The illustration of the methodology itself should be useful for other decisions concerning the introduction of new technology with its inherent risks in both economic and social dimensions.