OBJECTIVE: We sought to determine the relationship between relative value units (RVUs) and intended measures of work in catheterization for congenital heart disease.
METHODS: RVU was determined by matching RVU values to Current Procedural Terminology codes generated for cases performed at a single institution. Differences in median case duration, radiation exposure, adverse events, and RVU values by risk category and cases were assessed. Interventional case types were ranked from lowest to highest median RVU value, and correlations with case duration, radiation dose, and a cases-predicted probability of an adverse event were quantified with the Spearman rank correlation coefficient.
RESULTS: Between January 2008 and December 2010, 3557 of 4011 cases were identified with an RVU and risk category designation, of which 2982 were assigned a case type. Median RVU values, radiation dose, and case duration increased with procedure risk category. Although all diagnostic cases had similar RVU values (median 10), adverse event rates ranged from 6% to 21% by age group (P < .001). Median RVU values ranged from 9 to 54 with the lowest in diagnostic and biopsy cases and increasing with isolated and then multiple interventions. Among interventional cases, no correlation existed between ranked RVU value and case duration, radiation dose, or adverse event probability (P = .13, P = .62, and P = .43, respectively).
CONCLUSIONS: Time, skill, and stress inherent to performing catheterization procedures for congenital heart disease are not captured by measurement of RVU alone.
- AE —
- adverse event
- ASD —
- atrial septal defect
- CHARM —
- Congenital Heart Disease Adjustment for Risk Method
- CPT —
- Current Procedural Terminology
- PDA —
- patent ductus arteriosus
- RUC —
- Relative Value Update Committee
- RVU —
- relative value unit
What’s Known on This Subject:
Few reports have explored the measurement validity of the relative value unit (RVU) system, particularly in pediatrics. The RVU system, although broadly applied in health care settings, was developed for the adult population and thus may possess unique inadequacies in pediatrics.
What This Study Adds:
We found deficiencies in the ability of the RVU system to capture features of case mix complexity and differences related to age. Additional investigation may be warranted to determine the validity of RVU as a measurement tool in pediatrics.
In the 1980s, reform of the traditional physician payment system was prioritized by the US government, physicians, insurers, and patients. An alternative system developed at the Harvard School of Public Health, based on resource-input costs, was proposed, enacted into federal law (Omnibus Budget Reconciliation Act 1989), and subsequently implemented in 1992 by the Health Care Financing Administration, which changed the Medicare Fee Schedule to a system based on the concept of the relative value of work.1,2
Absent any previous attempts to quantify physician work except for the California relative value scale, investigators began by attempting to determine physician work associated with specific services. The specific services were identified, of convenience, by Current Procedural Terminology (CPT)-4. Physicians were surveyed as to relative work for CPT-defined services taking into account 4 components: time spent, skill required (technical and cognitive), and stress incurred. Of the 7000 CPT-coded services at the time, ∼400 were studied and results were extrapolated to the remainder. The results of the physician estimates of relative amounts of work, relative value units (RVUs), were found to be reasonable and consistent.
It was understood that this scale was provisional and subject to change; thus, by congressional mandate, the RVU system and scale are reassessed every 5 years. In addition, the Relative Value Update Committee (RUC) meets 3 or 4 times each year to consider requests to revalue procedures and assign values to new procedures.3,4 Soon after the development of the RVU system, most payers adopted the fee schedule established by the Health Care Financing Administration.5 Some health care institutions use the scale to determine the value of physician work and, in some cases, to determine reimbursement relative to other health care providers.6–8
The potential inaccuracy of applying adult-based RVUs to pediatric practice was immediately recognized by pediatric specialties and professional organizations. In the early 1990s, 81 congenital heart operations were assigned CPT codes and corresponding RVUs. These RVUs were found to be reasonable relative measures of physician work in a validation study by Jenkins et al.9 However, in these same patients, catheterization procedures involving an increasingly diverse set of therapeutic interventions have not systematically been assigned RVUs.
Since the inception of the RVU scale, few studies have questioned the relationships between the assigned RVU and the time, skill, and stress, which the RVU values are intended to capture. We suspected that the CPT-based RVU assignment method did not accurately reflect the unique physician work component among catheterization procedures for congenital heart disease. To test this proposition, we used available proxies for physician time, skill, and stress. Time was relatively straightforward to quantify and we simply chose case times. For skill, we considered the amount of radiation required as a surrogate for the amount of information obtained and processed during the procedure, and for technical skill, we used the number of interventions. Stress is clearly less quantifiable, and therefore we assumed that stress was higher in cases with a known higher incidence of adverse events (AEs). We considered that in the relatively young field of pediatric catheterization, where many of the procedures have only become common practice in the period since the inception of the RVU system, the CPT-based RVU system might be uniquely inadequate to measure the intended components of relative value.
The database maintained in the Children's Hospital Boston cardiac catheterization laboratory for reporting and quality assessment was queried for this study. This is a prospectively collected database that includes patient and procedural characteristics as well as a record of AEs.10 Patient characteristics collected included age, gender, diagnosis, and presence of a genetic or noncardiac comorbidity. Type of case, admission type, method of airway management, types of interventions, duration of case (defined as time sheaths were inserted until removed), fluoroscopy dose-area product (μGy ⋅ m2), and AEs were reported in this cohort. AEs were classified according to the International Pediatric Congenital Cardiac Code catheterization complication nomenclature 5 scale severity ranking11 in which severity level 3, 4, and 5 classified events are high severity and potentially life-threatening events.
In the catheterization database, all angiograms, interventions, and type of diagnostic procedure performed are recorded and linked to CPT codes. Angiograms are assigned the most appropriate CPT codes by site. Types of interventions and sites are chosen at the time the catheterization procedure according to an intervention list similar to the nomenclature developed by the International Pediatric Congenital Cardiac Code, published in 2011.12 In some cases, specific CPT codes exist for highly specific interventions such as atrial septal defect (ASD) closure, Fontan fenestration closure, and patent ductus arteriosus (PDA) closure. In other cases such as angioplasty and stent placement, interventions may be performed in very dissimilar patients in a variety of vascular sites but have common CPT codes. All cases were assigned an acquired or congenital diagnostic procedure code for right and/or left catheterization as appropriate.
RVU values available from the 2010 National Physician CMS Fee Schedule Relative Value File were matched to corresponding CPT codes (copyright 2009 by the American Medical Association). We calculated the RVU value for cases in the CHB catheterization database by matching the RVU value to the CPT codes generated for the case. Examples of cases, designated CPT codes, and corresponding RVU values are provided in Table 1. The database was queried for diagnostic, interventional, or biopsy cases performed from January 1, 2008, to December 31, 2010. Six cases with an RVU value of 0 and 329 cases with a CPT code of 93799 or 92992, which do not have an assigned RVU value, were excluded from the cohort of 4011 for a sample of 3676 cases.
Risk categories were originally created as a way of grouping a relatively large number of diverse and sometimes infrequently encountered procedures into categories with similar risk of AEs (Table 2).10,13 Categories were developed by using expert judgment and empirical methods in a derivation cohort and validated in a prospective cohort in the Congenital Cardiac Catheterization Project on Outcomes database. A case is assigned to 1 category based on the characteristics of the case; for diagnostic procedures the cases are stratified by age group. If 1 or more interventions are performed, the case is assigned to the highest category. The categories have a strong association with the occurrence of clinically important AEs; therefore, although this association has not been tested, we considered them a surrogate for physician stress associated with performing a procedure. Of the 3676 cases meeting inclusion criteria, 119 could not be assigned a risk category, so the final sample size for analysis was 3557.
Cases were also grouped for this study according to technical skill required and number of interventions. Diagnostic catheterization and myocardial biopsy without additional interventions were designated group 0. Group 1 included interventional procedures with a single intervention, and group 2, interventional procedures with more than 1 intervention. Among the 3557 cases in the Children’s Hospital Boston data set (2008–2010) with RVU values and designated risk category, 2982 case types (84%) could be assigned to 1 of these groups. The analysis by case type is limited to this subcohort (Table 3). Examples of case types performed infrequently and not captured in the subcohort (n = 575) include isolated distal pulmonary artery interventions, angioplasty or stenting in vessels or shunts other than the aorta, aorta stent redilation, intracardiac angioplasty or stenting, isolated coil occlusions, and mitral valvuloplasty.
Cohort characteristics were summarized by frequency and percentage for categorical variables and by median and interquartile range for continuous variables. Comparisons of median case duration, radiation exposure, AE rates, and RVU across procedure risk categories were performed by using the Kruskal-Wallis test. RVU values were summarized for 18 case types and by the 3 groups and reported as mean, SD, median, minimum, maximum, and interquartile range.
AE rates, case duration, radiation exposure, and RVU values were also calculated and compared among diagnostic catheterization cases stratified by age group (neonates <1 month, infants 1 month to 1 year, children >1 year, and adults >18 years).
The 16 interventional case types were ranked from lowest to highest median RVU value, and the relationship between case ranking by median RVU value and the following characteristics were explored: median duration of case, median radiation dose, and median predicted probability of an AE. The predicted probability of an AE was calculated for each case using the Congenital Heart Disease Adjustment for Risk Method (CHARM) risk adjustment model, which is based on the assigned risk category, indicators of hemodynamic vulnerability, and age <1 year.14 Associations between ranked interventional case types and each measure were quantified by using the Spearman rank correlation coefficient.
Patient and procedure characteristics for this population are summarized in Table 4. Most of the population had complex congenital heart disease, with either 2 ventricles (39%) or single ventricle physiology (17%). Most cases were scheduled electively (80%). The largest proportion of cases were classified as interventional (55%). Among the remaining, 25% were classified as diagnostic, which may have involved hemodynamic and/or angiographic assessment, and 20% as myocardial biopsy after heart transplantation. There was a wide range in case duration. At least 1 AE of any severity was recorded in 14% of cases, and high severity level 3, 4, or 5 events occurred in 5% of cases.
The largest proportion of cases were ranked in the lowest risk category (n = 1354 [38%]) and were composed mostly of diagnostic cases in patients >1 year of age, as well as those undergoing myocardial biopsy. The remaining cases were diagnostic cases in patients <1 year of age and cases involving single or multiple interventions in decreasing frequency by risk category (risk category 2 = 25%, 3 = 23%, and 4 = 13%). Procedure time, radiation dose, AE rates, and median RVU value increased by risk category (all P < .001) (Fig 1).
The associated median RVU value for case group 0 (diagnostic catheterization and myocardial biopsy), 1 (interventional catheterization, single intervention), and 2 (interventional catheterization with >1 intervention) is shown in Fig 2. Median RVU increased according to groups with broader interquartile ranges among more complex and heterogeneous groups 2 and 3 (Table 5). Isolated interventions, such as closure of an ASD, PDA, or fenestration and aortic or pulmonary valvuloplasty, had similar median RVU values to one another compared with those involving multiple interventions (P < .001 for the 3-group comparison).
Diagnostic procedures composed 25% of the cohort in 885 patients over a wide age range from infants to adults (Table 6). Radiation dose per unit of time increases as a function of body surface area; thus, older patients had higher radiation doses during similar length procedures. But although case duration was similar among age groups, there were significant differences in AE rates, the surrogate used for stress in this study (P < .001). Despite this difference in risk associated with the procedure, all the diagnostic cases were assigned similar RVU values (Table 6, Fig 3).
The largest proportion of patients (55%) in this cohort underwent interventional procedures with varied case types and associated RVU values. Interventional cases were ordered from lowest to highest median RVU value, and procedure characteristics are summarized (Table 7). Among these different types of procedures, no correlation existed between ranked RVU value and median procedure time, radiation dose, or the predicted probability of an AE as calculated for each case type using the CHARM method (Fig 4).
The relative value of physician work is determined through consensus methods by considering the time associated with the encounter, the amount of related stress, and the amount of required technical and cognitive skill.3 In this study, we found that in the field of pediatric cardiac catheterization, RVU values did not capture intended elements of RVU (time, skill, and stress) among interventional cases or discriminate among important differences by patient age. Thus, the established RVU system does not reflect all the intended elements of work for patients undergoing catheterization for congenital heart disease. We propose that either new RVUs are generated or additional metrics be considered when measuring the complexity of work in this pediatric specialty care service.
After widespread adoption of the RVU scale, and to address inadequacies of the RVU system, the Society for Thoracic Surgeons recommended to the AMA CPT Editorial Panel several new codes to more correctly describe congenital heart surgery procedures. Once these were adopted for CPT, the Society for Thoracic Surgeons conducted surveys to establish work values for the entire group of congenital heart surgery codes and presented these recommended work values to the RUC. These codes were then given value by using standard methods based on consensus and relative value compared with other established procedures; later the utility of these codes was supported by work showing correlation with mortality, length of stay, and charges in congenital heart surgery.9
In catheterization for congenital heart disease, CPT codes (93530–93533) and associated RVU values for basic diagnostic cases without intervention have been established and were reviewed and revised by the RUC in 2010. Codes have also been established for a few common predictable and well-circumscribed procedures for placement in interatrial communications, including ASD, PFO, and Fontan fenestrations (CPT 93580). Nevertheless, due to the complexity and diversity of procedures performed in congenital cardiac catheterization and the sparse number of codes specific to the treatment of congenital heart disease, values for many of the interventional procedure codes have been extrapolated from codes developed for procedures performed in adults by cardiologists and interventional radiologists. For example, CPT codes 92997 and 92998 (transcutaneous pulmonary artery balloon angioplasty) are used for both angioplasty and stent redilation at multiple sites, including proximal pulmonary arteries, distal pulmonary arteries, aortopulmonary collaterals, and conduits. Despite a common procedural code, there is considerable variation among these interventions and potentially the expected complexity by site.
This variability in intervention types and patient populations may partially explain why assigned RVU values did not correlate with case duration, radiation dose, and anticipated AE occurrence (Fig 4). As an example, pulmonary vein dilation cases, designated in risk category 4, include some of the most ill and vulnerable patients encountered in the field and require one of the most technically challenging procedures performed in congenital cardiac catheterization. However, ASD closure (risk category 2), which is a single isolated, short, and often predictable intervention in a relatively healthy patient population, has a similar RVU value. Despite these important shortcomings, some aspect of risk was captured by the RVU designations, which showed a median increase according to the 4 risk category designations (Fig 1).
A distinction in the complexity and work of performing a diagnostic procedure for congenital heart disease has been recognized and a relatively higher RVU applied compared with populations without congenital heart disease. In this study, we also considered potential age-related differences in the risk associated with performing diagnostic procedures and found statistically significant differences in AE occurrence according to age despite equivalent RVU designations. Thus, in pediatric catheterization and perhaps other pediatric disciplines, age adjustment may need to be considered in assigning the relative value of work.
Recently, the American College of Surgeons published a new risk adjustment model for the National Surgical Quality Improvement Program. This model considered new risk categories and found additional independent explanatory value beyond just RVU for procedural measures of stress, skill, and procedural complexity.15 This finding lends additional credence to the consideration of factors other than or in addition to RVU when comparing and evaluating physician work output and outcomes in pediatric and congenital catheterization. In this analysis, we have demonstrated how RVU does not capture all the intended elements; however, additional work is necessary to determine measures that better evaluate and measure the procedural time commitment, complexity, and stress associated with performing procedures in congenital cardiac catheterization.
The measures used in this study were from a sample of convenience and not variables specifically developed to capture the components of RVU. There are no published billing standards in catheterization for congenital heart diseases. Because this analysis is limited to the experience at a single center, it does not allow for variation in the interpretation and valuation of a subset of interventional procedures for which there are no specific CPT codes. Although the risk categories and CHARM risk model were developed and validated in a multicenter dataset, the opportunity to test the categories and models in other data sets remains to be achieved. In addition, in this analysis we make the assumption that a higher anticipated AE occurrence is associated with physician stress, whereas other personal or systems factors may be important.
RVUs are intended to assign value to the physical, technical, and cognitive skills possessed by a physician, as well as emotional stress endured while performing a service. In this analysis of catheterization procedures for congenital heart disease, we have demonstrated some of the weaknesses of the RVU system in capturing the complexity or risk involved with performing a diversity of procedure types in a population with a wide age range undergoing catheterization at a pediatric institution. It is therefore important to consider additional factors such as age and validated measures of case mix complexity. In addition, an expanded set of CPT codes for types of interventional cases to describe cardiac catheterization procedures in different age groups could be created and then valued through the RUC valuation process. Additional analyses are warranted to explore the explanatory value of RVU in other pediatric populations and validate measures that are intended to capture elements of work unique to these populations.
- Accepted October 4, 2012.
- Address correspondence to Lisa Bergersen, MD, MPH, Department of Cardiology, The Children’s Hospital, 300 Longwood Ave, Boston, MA . E-mail:
All of the authors made substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; participated in drafting the article or revising it critically for important intellectual content; and provided final approval of the version to be published.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: The John F. Keane Fund, which provides support for research in the interventional catheterization division of Children's Hospital Boston.
COMPANION PAPER: A companion to this article can be found on page 340, and online at www.pediatrics.org/cgi/doi/10.1542/peds.2012-1434.
- ↵Blue Cross and Blue Shield of Massachusetts I: changes in reimbursement policy and reporting structure. Blue Cross and Blue Shield; 1995
- Bergersen L,
- Giroud J,
- Jacobs J,
- et al
- Bergersen L,
- Gauvreau K,
- Marshall A,
- et al
- Copyright © 2013 by the American Academy of Pediatrics