Incentivize Everything, Incentivize Nothing

Abstract

f everything in life is of equal value, people willchoose what they do based solely on personal pref-erence. If you want people to do something that isat odds with their personal preference, you must pro-vide incentives that outweigh the personal satisfactionof just doing whatever they please. That is why we arepaid to go to work. Our salary is a regular incentive toget up at 5:45 am and drive to work, instead of loung-ing late in the morning in a warm bed, and headingout even later to a soft white sand beach.In this issue, Abouleish et al. (1) have provided afascinating glimpse into how academic anesthesia de-partments induce their faculty to work harder in theclinical arena. They characterized the construct of clin-ical incentive plans among 83 responding academicprograms. Most (69%) did not pay incentives based onwork performed during normal working hours.Ninety percent did not alter total salary by more than25%. They note that many comments on the surveysuggested total departmental productivity did notchangewiththeintroductionofanincentiveplan.Thisis not surprising, in that anesthesiologists have little orno control over booking and performing surgeriesduring their working hours. Now that we know howacademic chairs are crafting clinical incentive pro-grams, can we analyze those comparative data to helpus all do it better? That remains to be seen. It isinteresting that 64% of respondents have had theirplans in place less than 5 years, so many of us are ona steep learning curve. This well-timed article mayallow us to learn from each other and better serve bothour faculty and facility. Furthermore, we may now beable to summarize these data and thus respond moreintelligently to our group practice plan managers whoinsist on productivity-based compensation.While we now know how everyone else is doing it,what more should we know about incentives in gen-eral to help gauge whether we are all doing it right?Incentivesmotivatepeopletoperforminacertainway(more, better, quicker). The key to a working incentivesystem requires the following: 1) The individual be-lievesthatthedesiredbehaviorwillleadtothedesiredoutcomes (i.e., the incentive is guaranteed to be paid;it is not contingent on profits or anything else outsidetheir control), 2) the individual believes that theseoutcomes have positive value for him or her (i.e., anoble purpose or a sufficient sum of money), and 3)the individual believes that he or she is able to per-form at the desired level (the lack of control over whatis booked into one’s room combined with a lack ofincentives for daily scheduled work in most plans area reflection of this reality).It is a basic truth that people always act in theirrational self-interest. Therefore, incentives almost al-ways work perfectly to provide the behavior that isrewarded,assumingtheincentiveislargeenough.Theonly problem is that we often fail to understand whatbehavior we really want, how to measure what wewant (a metric), and how to link the incentive clearlyto that metric without introducing new problems.There is a classic article, “The Folly of Rewarding Aand Hoping for B,” (2) that details the problems thatcan occur with poorly designed incentive systems thatcreate perverse results. An example is when an aca-demic department starts to reward clinical billings inan effort to “get everyone to bill more” and all of asudden finds there are no volunteers for off site cases,the IRB, the preoperative screening clinic, or the med-ical school Admission’s Committee. Reluctance to getinvolved in these nonbillable activities is even greaterwhen they require evening or weekend time.The ideal incentive has several characteristics. First,there should be a target of expected performance be-low which there is no incentive. A number of depart-ments described by Abouleish et al. base their incen-tives on total dollars or total work. This degrades thedifferences between those with exceptional productiv-ity and average productivity. If one wants marginalproductivity (productivity above that which already