What can O.R. do for you?

Why There Needs to be More O.R. Inside Performance Measures

We're all familiar with performance measures, whether in the shape of school league tables or hospital ratings. But are they a good or a bad thing? The answer is, 'it depends'. Well designed performance measures are generally a good thing - they help top management to identify poorly performing units, help the public make informed choices, and tell local management whether they are doing as well as can reasonably be expected. On the other hand, badly designed performance measures can make matters worse, leading to distortions of service and diversion of resources away from where they could be most effectively employed.

Unfortunately, there are all too many instances of poorly designed performance measures that have produced unintended consequences. Well known examples include school league tables which damage good schools by failing to take into account the level of attainment of the pupils when they entered the school. These league tables have also led to some schools excluding weaker pupils from exams in order to improve the school's pass rates. Hospital waiting list targets have led, in some cases, to patients needing minor surgery being given higher priority than those who are seriously ill, in an attempt to reduce waiting lists as quickly as possible. There are many other examples of crude performance measures which, lacking the benefit of O.R. inside, lead to distortions in performance rather than the improvements they are intended to bring about.

The best way of ensuring that performance measures do the job they're supposed to is to put some O.R. inside them. This enables much more sophisticated measures to be developed. To get technical for a moment, there is a battery of techniques, including data envelopment analysis, data mining, and clustering techniques, which enable things like differences in the intakes of different schools to be adjusted for, enabling fair comparisons to be made. Other techniques, such as multicriteria analysis, can be used in cases where there are conflicting multiple objectives, as in the case of a hospital trying to reconcile the need to cut waiting lists with the need to give priority to the most seriously ill patients. With the benefit of O.R. inside, performance measures can be designed which ensure that resources are properly allocated by giving an appropriate weight to each of the objectives.

Making performance measures much more sophisticated by putting O.R. inside them can remove incentives to distort performance and enable the performance measures to do reliably the job they're intended to do - to lead to improved performance.

Banner Watch our 'What is O.R. movie' Download our Guide to O.R.