The role of remuneration in clinical productivity of paediatric physicians

Increased attention devoted to patient safety, quality of care and accountability has heightened interest in physician performance. Studies, to date, have largely investigated dimensions of performance that are similar across physician specialties, such as professionalism, administrative leadership and academic activity (ie, research and teaching). There is a paucity of work that delineates performance measures for different specialties. An examination of the limited literature on clinical productivity of paediatric physicians yields some interesting and, at times, contradictory findings regarding ways to measure, encourage and reward clinical productivity. There are generally two types of clinical productivity measures: those based on financial data (for example, fee-for-service earnings per physician) and those based on clinical activity (for example, the number and length of patient visits) (Table 1) (1,2). Neither of these measures account for the quality of patient care. It is thus prudent to examine the underlying clinical circumstances (including the context) of practice when evaluating clinical productivity, especially in academic clinical settings where many factors are beyond the control of individual clinicians (1). Examples of measures used to report clinical productivity of paediatric physicians Examples of measures used to report clinical productivity of paediatric physicians Most of the literature pertaining to the clinical productivity of paediatric physicians has been generated in the United States. The most commonly reported measure has been the resource-based relative value scale (RBRVS), which was developed to assign resource values to various services provided by physicians. The RBRVS has designated relative value units (RVUs) for more than 7000 services and encompasses physician work, practice expenses and cost of malpractice expenses (1,3). RVUs are used extensively in the United States Medicare system. Several articles and editorials have suggested that the RVUs established in the United States for paediatric specialties are not comprehensive, and may underestimate the effort required to provide paediatric care. The American Academy of Pediatrics (AAP) states that the RBRVS is incomplete because some services unique to paediatric care, such as child abuse services, are absent. The AAP has also argued that the original studies conducted to establish the RBRVS did not collect sufficient data on paediatric work, resulting in the unique characteristics of paediatric care not being well reflected in the values established (4). RVUs that were first established were intended to reflect typical patients being treated by typical doctors, with the expectation that over the course of practice, inequities in patient characteristics would be addressed. However, this assumption may not hold true in paediatric populations (5,6). A recent evaluation of paediatric ambulatory diabetes care at the Seattle Children's Hospital (Washington, USA) demonstrated that payment rates assigned to paediatric endocrinology procedures using the RBRVS covered less than 50% of all practice expenses (3). One major factor that accounted for this shortfall was underestimation of the cost of providing multidisciplinary care. The under-representation of paediatric providers in the establishment of RBRVS was cited as a key contributor to the reimbursement shortfall. With the growth of managed care plans in the United States, the AAP has developed guiding principles for managed care arrangements in paediatrics (7). The AAP's recommendations address many of the gaps identified in costing paediatric services, including consideration of age and establishment of modifiers in capitation rates for children with special health care needs. The extent of uptake of these principles is unknown. In the Canadian context, there has been limited evaluation of the reliability of resource intensity weights (RIWs) as a measure of clinical productivity in the provision of paediatric or other specialty services. A 2001 study (8) conducted in Ontario (not specific to paediatric care) revealed that the expected cost of surgical care for trauma patients based on the use of RIWs was significantly lower than the actual cost of care. This is partly attributable to variability in the intensity of trauma care that was often invariant to length of stay. While it may be inappropriate to extrapolate these findings to other specialties, further consideration of the use of RIWs as measures of clinical productivity is warranted, including their use in paediatric care, because RIWs are readily available and may be modified to reflect sub-specialty practices. Several studies (5,9,10) on the clinical productivity of paediatric physicians highlight the important differences between adult and paediatric care. Unique characteristics of the paediatric population include small size, young age, presence of congenital disease, greater complexity of presenting anatomical factors, and higher anxiety levels during examination and procedures (8,11). Such differences may increase the technical complexity and length of some procedures, complications and/or the need for sedation when absent in many adults (4,9,10). While it has been argued that caring for children is more complex and time intensive, there are some circumstances in which care may take less time – for example, when an otherwise healthy child is recovering from surgery and may require fewer follow-up visits with surgeons (10). For procedures performed on both children and adults, physician time and effort may vary according to the age of the patient, so the use of a single service code for reimbursement purposes would be inappropriate. It is rare for fee-for-service codes in the United States (or elsewhere for that matter) to be tailored to the age of the patient (10). This contrasts, somewhat, to the use of RIWs to measure the intensity of hospital services because RIWs are stratified by the age of the patient, albeit by only three large age groups. It has been suggested that age-specific modifiers be introduced to augment existing codes, noting however that this solution may not work for all specialties, and that some paediatric subspecialties may require an additional range of codes (5). Studies have been conducted in cardiology and nephrology on the extent of variation between the clinical effort required to provide similar services to adults and children. The American College of Cardiology and the AAP engaged a panel of paediatric and adult cardiologists in a modified Delphi process to ascertain the work effort associated with the provision of cardiac care to paediatric populations. The panel quantified the effort required to perform 20 cardiology procedures listed in the Medicare RBRVS in paediatric populations. Work effort was deemed to be similar for adults and children for 25% of the codes that were rated. For the remaining codes, work effort was found to be consistently higher for paediatric populations (averaging 44% higher) than the associated physician work RVUs for typical adult cases (5). A study of the work effort required to provide end-stage renal disease care, ie, dialysis to children, showed that delivery of care to infants and children took longer and was more intensive than care provided to adults (9). The study demonstrated that dialysis service times fall with age, with children's care being the most time intensive, followed by that of adolescents, and finally, adults. Differences in the volumes of specialized procedures performed in paediatric and adult patient populations may affect resource needs and productivity. For example, only 7% of all solid-organ transplant recipients in the United States in 2005 were children, and paediatric transplant procedures are widely distributed across a large number of institutions, with each performing a relatively small number of procedures (12). Minimum staffing requirements and standby costs for programs such as transplantation, which are already resource intensive, become even more expensive when customized for paediatric populations. Remuneration of physicians in academic and group practice settings has gradually moved away from fee-for-service toward capitation and blended models that better account for the multifaceted roles of physicians. In Ontario, several articles (13–15) have reported on the shift toward alternative funding of physicians, although there has been no evaluation of its impact on the quality of care. Studies focused on remuneration of paediatric physicians have reported mixed results. In the Netherlands, where most paediatricians moved to fixed salaries in the mid-1990s, there were no notable differences in practice patterns of fee-for-service and salaried paediatricians (16). Salaried paediatricians spent more time with their patients providing advice and information, but there were no differences in patient satisfaction and diagnostic testing. A paediatric primary care case study (17) in the United States showed that moving from fee-for-service billings to a prepaid group practice improved efficiency, continuity of care and education of medical trainees, but the outcomes of this study may have been confounded by other factors, such as the avoidance of cumbersome billings to multiple public and private insurers in the United States. Finally, an experiment to monitor physician practice behaviour under different reimbursement schemes randomly assigned paediatric residents to either fee-for-service or fixed salary methods of reimbursement. The study found that residents assigned to the fee-for-service group saw their patients more often than residents assigned to the salary group (18). Pay-for-performance models have become increasingly popular, although the jury is still out on the correlation between pay-for-performance and quality of care (19). There are several gaps in the existing body of evidence on pay-for-performance in paediatrics, the most significant being the dearth of studies regarding the provision of chronic or acute inpatient care for children. The majority of paediatric pay-for-performance studies have focused on immunization programs in primary care settings. Another limitation of the literature is that unintended consequences, such as patient selection (or cherry picking), have not been assessed. Linking incentives to the quality of care can be tricky for certain types of paediatric care. In the case of end-stage renal disease for children, limited access and compliance issues are well documented (20). The American Society of Paediatric Nephrology has expressed concern about the over application of pay-for-performance, because it may unduly penalize paediatric nephrologists for longer care times and poorer outcomes attributable to patient non-compliance. As such, there is the possibility that pay-for-performance may adversely affect access to care for marginalized patient populations. Despite the potential for unintended adverse consequences, there are some documented examples of successful pay-for-performance models in paediatric care. A large pay-for-performance program in the United States for paediatric asthma care established incentives associated with group participation and achievement of outcomes (21). All participating group practices earned a portion of the allotted financial incentive for their involvement, with 25% of practices earning the maximum 7% fee schedule increase based on demonstrated results. The program was successful in driving desired physician behaviour with predictable (albeit slight) increases in costs. While the authors support the use of financial incentives to influence clinical practice, they do not recommend individualized incentives. Instead, they promote emulation of their group-level pay-for-performance model to promote collaboration and shared learning (21). The University of Michigan Health System (Michigan, USA) implemented a productivity-based incentive system for academic paediatricians that resulted in a 22% increase in productivity and a 3% increase in costs (2). The success of this system was attributed to a well-designed compensation model. In addition to base, academic and administrative components, the model incorporated clinical incentives. Clinical incentives were earned on the basis of resource value units above the base salary, which was linked to a minimum performance threshold for all physicians that was also related to resource value units. Physicians earn clinical incentives by extending work hours, maximizing coding accuracy and treating more complex patients. This broad examination of paediatric physician productivity demonstrates that while some measures of clinical productivity exist, they do not fully delineate between care of children and adults. Pay-for-performance has proven to be successful when married with other remuneration models, such as fee-for-service and salaries. Much of the present review has presented clinical productivity in isolation from other dimensions. The complicated reality is that physicians, particularly in academic medical centres, are rarely involved exclusively in delivering clinical care; they carry out research, teaching and administrative duties as well. Remuneration models for physicians often begin by assigning relative time or effort to each area of responsibility, and then measure productivity in each of these areas (22,23). For example, research productivity may be measured through the number of publications and research grants; productivity in teaching may be measured by the number of medical residents supervised and hours of classroom time (23,24). However, productivity in each of these dimensions cannot be evaluated in isolation given the interplay between them. Physicians who supervise medical students have been reported to be less clinically productive than those without teaching responsibilities (25). In assigning value – explicitly in remuneration models or implicitly through organizational culture and values – to the different responsibilities of physicians, organizations may set up incentives for physicians to focus on one area of their responsibilities over others. Flexible career planning and remuneration models that account for individual physicians' interests as well as organizational goals may succeed in optimizing productivity in clinical and other areas of responsibility. The authors thank Dr James G Wright for his helpful comments and suggestions on an earlier version of this paper.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.