Variable financial rewards that are dependent on performance are used in many industries such as in banking, law, manufacturing, and retail. Recognition of work is crucial for motivating good practice, with tangible financial reward arguably one of the strongest of incentives. An understandable and fair system is needed to set goals, evaluate performance, and distribute financial rewards in order to incentivize professionals to meet these goals. A portion of salary must be variable, the size of which varies for meaningful impact.
Healthcare organizations have multiple roles to fulfil—meeting patients' expectations of access to and quality of care, meeting statutory obligations, maintaining accreditation by healthcare accreditation agencies, and maintaining financial prudence and viability. Academic healthcare organizations additionally train the next generation of professionals, and carry out research to improve future healthcare. Their medical professionals have to provide leadership and to do some administrative work. Their financial systems would need to recognize and incentivize clinical, teaching, research, and administration work. When pay is totally fixed, there is no monetary reward and less incentive to do more or to do better. Perhaps mindful of this, performance based compensation systems with variable rewards or bonuses are increasingly used by academic healthcare organizations.1–3
Such performance based compensation systems need to balance encouraging academic work and incentivizing clinical work. Doctors need to understand their department's and institution's goals, how the system encourages these goals, and how they will be rewarded. A system should be flexible enough to accommodate different departments' needs and different doctors' interests and strengths.4 A system needs to be fair and be seen to be fair. It can be difficult to get any such system entirely right first time, and adjustment will be needed over time.
Financial recognition if not well handled can demotivate staff, cause staff attrition, and cause mission failure especially in teaching and research. One issue is whether time spent on academic work will lead to a reduction in clinical income, and how to avoid this disincentive which may affect procedural specialties in particular.4 Another issue is whether clinical revenue should be used to support and reward non-clinical work. This has been described as clinical productivity cross-subsidizing academic pursuits, and can cause dissatisfaction when implemented within a department, let alone across departments.5 Should there be caps on incentives and salaries? Anaesthesia departments generally have less direct control over their case mix and workload than surgical and medical departments. In several countries, the primary source of anaesthetists' income is from working in the operating rooms (OR), and to maximize earnings they need to work in the OR every day. With academic medical centres being established in the UK, the report by Dr Sakai and colleagues6 in this issue about incentivizing clinical and academic work is timely and welcome.
Dr Sakai and colleagues6 report how the Department of Anesthesiology at University of Pittsburgh Medical Centre fared after implementing a performance based faculty compensation programme in 2004. In their institution, academic work included administration, teaching and research. Their doctors were classified as clinical or academic faculty, with academic faculty negotiating to spend as much as 80% of their working time on non-clinical work. They designed an original and comprehensive ‘merit matrix’, which quantified work and achievements in the administrative, teaching, and research domains in ‘credit hours’ or ‘matrix points’. They separated out a portion of total salary that was contingent on achievements, with the matrix points earned being translated to salary. This portion was clearly separated from base pay, and was termed ‘salary at risk’. Doctors had to ‘regain’ this ‘salary at risk’, perhaps a slightly unusual concept and term to use to motivate doctors. This ‘salary at risk’ was 30% of total salary, noticeable but not a huge proportion compared with other institutions.7 The other 70% was based on rank and years of service. Importantly, their doctors could by working more, earn more than they previously had before implementation of this system. Implicit in their system is the high degree of control that their managing clinicians had over their departmental finances and doctors' salaries, which they reviewed and adjusted annually.
Activity without achievement may mean little. Using their merit matrix, Dr Sakai's department had chosen to reward achievement and not just activity in teaching and research. Their merit matrix clearly defined and measured achievements, practicing the maxim ‘what gets measured gets rewarded’. Reassuringly, this included recognition of manuscript review work for journals, and differentiated service in national, regional, and institutional committees. The relative value of all types of non-clinical work is accounted for and is transparent. There did not seem to be a cap on merit matrix points, and doctors that earned beyond 150 points were even eligible for additional incentives.
Clinical performance was essentially measured by total hours of work and compared with a historical ‘site mean’. Those working more hours earned more. Recognition of clinical work should also take into account clinical complexity and subspecialty expertise. Their system had additional annual payments for those working in cardiac and transplant anaesthesia. Their system using total clinical hours worked is a simpler alternative to the American Society of Anesthesiologists Relative Value Unit (ASA RVU) system. The RVU system is very comprehensive, covering OR, and non-OR clinical work. Total units for an anaesthesia episode comprise: basic units (which takes into account complexity and type of operation), modifier units (such as for patients of poor ASA status and extremes of age), and time units (length of anaesthesia).8 It is possible for their system to be used for distribution of clinical revenue to incentivize all types of work, while fees can be charged using the ASA RVU system.
What thoughts has this report of performance based compensation system provoked? First, while the hours of clinical work is a simple primary measure, perhaps we can also recognize the quality of clinical work. We could measure outcomes that matter to patients (and perhaps also their lawyers in this era), such as regional anaesthesia success rates, pain and nausea on waking, complication rates, and dental trauma. The quality should arguably be measured in multiple dimensions, rather than simple rates.9 Their system does not address anesthesiologist efficiency, itself a marker of quality, vs simply clocking up more hours. Quality of care and efficiency can be difficult to define, sensitive to measure, and more measurements mean yet more work.
Paying for performance at hospital level has been tried to drive improvements in healthcare, and manage costs. Jha and colleagues10 evaluated the Medicare Premier Hospital Quality Incentive Demonstrative, comparing 252 hospitals in the programme and 3363 control hospitals. They found little difference in mortality rates for myocardial infarction, heart failure, pneumonia, and coronary artery bypass surgery. They found little to suggest that financial incentives improved patient care. An earlier study also found little improvement in cervical cancer screening, mammography, and haemoglobin A1c testing practices with paying for performance.11 Thus, it seems unclear if more pay improves clinical quality.
Secondly, to incentivize work in the different domains, the relativity of achievements and attendant financial rewards need inclusive negotiation and collective agreement. For example, Dr Sakai's merit matrix awards 10 credit hours for a lecture that is repeated annually, while achieving a first author publication of original research in this journal might achieve 400 credit hours. Achieving a publication or grant requires arguably more effort than lecturing, not to mention coping with the uncertainty and heartaches inherent in research. Conversely, we can regard teaching students and residents and ensuring that they become competent, to have more immediate value. Leadership work was clearly recognized and valued in their system, this is crucial for harnessing talent and delegating responsibility.
Thirdly, clinical output, often termed ‘productivity’, can dramatically increase with concurrent supervision of multiple OR. While we can encourage concurrency, some subspecialty work is unsuitable for multi-OR supervision, and may disadvantage their practitioners. In their department, the additional annual pay for those managing complex work may have addressed this disadvantage. One consideration is that in US institutions accredited for residency training by the Accreditation Council for Graduate Medical Education, physician faculty can supervise anaesthesia at not more than two locations simultaneously.12 Multi-OR supervision entails its own stressors. Perhaps other measures of individual clinical output can be considered, such as clinical days worked, time that the doctor is available for clinical work, total ASA RVU units billed.13
Finally, activity and achievements are hollow without teamwork, respect for patients and colleagues, integrity, and compassion for patients. While we would like to take these values as given for doctors, their absence can have drastic effects on patient care and department morale. Communicating clearly, staying calm during critical situations, interpersonal skills, being responsible and proactive are also attributes of good anaesthetists.14 Clinical training takes time, may involve additional risks, and often it is easier just to do a procedure yourself than to take a trainee through it. Encouraging a willingness to teach can be difficult. Can and should these values be addressed in a performance based compensation system? While values can be defined, their measurement may not be easy, and communication of the results to individuals may be even less easy.
Can such a performance based compensation system work? Dr Sakai and colleagues reported improvements in clinical work output, amount of original research, and teaching parameters, after implementing their system which integrated clinical and academic performance. The academic faculty all ‘regained’ their ‘salary at risk’. Their hypothesis and conclusion that a performance based system with variable salaries can better incentivize doctors compared with fixed salaries are plausible. What is unclear is whether the improvements were due only to the performance based compensation system they implemented, and whether the system could on its own have caused the improvements. Scheduling or ‘protecting’ time for academic work conveys recognition and may have been a strong motivator too.15 Other factors such as financial pressure, the prevailing economy, the need for job security, departmental culture, and departmental leadership all can contribute to how well a department does. Personal interest and commitment are key factors that lead people to work beyond the clinical domain and to excel in teaching, research and leadership. Academic medical centres may attract different doctors from those who chose to work in non-academic or private practice settings.
Is their system applicable elsewhere? The outcomes data they used to measure performance is the type of data that all departments should regularly review, regardless of salary systems. The relative value of rewards for different achievements in the clinical and academic domains can be adjusted to drive a department's specific needs. Performance based systems focusing on clinical ‘productivity’ can increase clinical output with fewer staff, and without adversely affecting academic output.7,15 However, financial control may not be vested within clinical departments in other countries. The Merit Awards in the NHS, which can substantially increase a consultant's salary, are not within a department's control. The managing clinicians need not only training in and understanding of human resource and finance work, but also commitment to and time to do the work.
‘Buy in’ from the clinical faculty is crucial, if any clinical revenue is to be used to fund academic work. A survey of academic surgeons in the US cautioned against relying on clinical revenue to subsidize academic work, as many surgeons were disaffected by this.5 While ‘buy in’ was not specifically addressed in Dr Sakai and colleagues' report, their doctors were able to choose their track and type of work, and their department did not seem to have problems with staff retention even as workload increased.
A last thought. It takes strategic and strong leadership to commit manpower and time to non-clinical work, in the face of demands for clinical service and clinical revenue. What Dr Sakai and colleagues have modestly left unwritten about are the vision, leadership and political will of their department to implement their system and to make it work for 9 years. That they have their performance based compensation system is an achievement in its own right and we should congratulate them for this.