Weight Restricted (Multiplier) Models

When considering various inputs and outputs in the basic DEA models discussed in earlier chapters, we made no judgment about the importance of one input versus another, and we assumed that all outputs had the same importance. In fact, in our example data, we assumed outpatient visits would consume the resources at the same level as inpatient admissions. Similarly, in producing the patient outputs, we valued the contribution of nursing hours the same as the contribution of medical supplies. Beside these assumptions, DMUs in a DEA can become efficient by simply taking advantage of a particular input or output variable. Simply, a hospital can become efficient by emphasizing a favorable input or output. For instance, observing from the example data, Hospital 9 has relatively low nursing hours but a high amount of medical supplies. The low nursing hours may be the reason this hospital is at the efficiency frontier (see Fig. 2.7). In the DEA literature, these DMUs that take advantage of these weak assumptions are sometimes called maverick DMUs.