Current operator volumes of invasive coronary procedures in medicare patients: Implications for future manpower needs in the catheterization laboratory

Abstract

Objectives

We seek to assess the per-operator volume of diagnostic catheterizations and percutaneous coronary interventions (PCI) among US cardiologists, and its implication for future manpower needs in the catheterization laboratory.

Background

The number of annual Medicare PCIs peaked in 2004 and has trended downward since, however the total number of catheterization laboratories nationwide has increased. It is unknown whether these trends have resulted in a dilution of per-operator volumes, and whether the current supply of interventional cardiologists is appropriate to meet future needs.

Methods

We analyzed the Centers for Medicare and Medicaid Services 2008 Medicare 5% sample file, and extracted the total number of Medicare fee-for-service (Medicare FFS) diagnostic catheterizations and PCIs performed in 2008. We then determined per-physician procedure volumes using National Provider Identifier numbers.

Results

There were 1,198,610 Medicare FFS diagnostic catheterizations performed by 11,029 diagnostic cardiologists, and there were 378,372 Medicare FFS PCIs performed by 6,443 interventional cardiologists in 2008. The data reveal a marked difference in the 2008 distribution of diagnostic catheterizations and PCIs among operators. Just over 10% of diagnostic catheterizations were performed by operators performing 40 or fewer Medicare FFS diagnostic catheterizations, contrasted with almost 30% of PCIs performed by operators with 40 of fewer Medicare FFS PCIs. A significant majority of interventional cardiologists (61%) performed 40 or fewer Medicare FFS PCIs in 2008.

Conclusions

There is a high percentage of low-volume operators performing PCI, raising questions regarding annual volume recommendations for procedural skill maintenance, and the future manpower requirements in the catheterization laboratory. © 2012 Wiley Periodicals, Inc.

INTRODUCTION

Over the past decade, progressive technical improvements have led to improved outcomes and expanded indications for both diagnostic catheterizations and percutaneous coronary intervention (PCI). A rational plan for the future of interventional cardiology [1] requires a firm comprehension of the number of invasive and interventional cardiologists currently practicing, accurately ascertaining the number of interventions currently being performed by these physicians, and then extrapolating how many procedures will be performed at a future date based on an assessment of population growth and advancement in the field.

According to the Agency for Healthcare Research and Quality (AHRQ), there were 1,475,602 inpatient diagnostic catheterizations and 693,315 inpatient PCI procedures performed in the United States in 2009 [2]. Most recently, data from the Centers for Medicare and Medicaid Services (CMS) has shown that yearly Medicare PCI volumes peaked in 2004 and have trended downward since [3]. The cause for this trend is likely multi-factorial, including cardiologists' response to the negative results of several recent clinical trials (COURAGE, OAT etc.) [4, 5], as well as reduced rates of in-stent restenosis with the increasing usage of drug-eluting stents, and longer periods of dual antiplatelet drug therapy after PCI. Despite the downward trend in PCI volumes, the number of hospitals performing PCI has increased substantially, estimated to have increased from 1,223 in 2001 to 1,637 by 2008 [6]. The number of active invasive and interventional cardiologists in the United States is more difficult to ascertain with accuracy. According to the American Board of Internal Medicine there are currently 23,696 active board certified cardiologists, but there are no accurate estimates of how many perform PCI. There are 5,196 individuals with active interventional cardiology certificates and as many as an additional 1,000 individuals who have not recertified [7].

In this study, we use the CMS 2008 Medicare 5% sample file to estimate aggregate volumes and per-physician yearly averages of diagnostic cardiac catheterizations and PCIs among US cardiologists. We evaluate the 2008 volume distribution among practitioners and identify its implications regarding future manpower needs.

METHODS

To obtain accurate information for operator volume, this study analyzes the CMS procedure billing data from 2008 to report physician practice trends in diagnostic cardiac catheterization and PCI (both angioplasty and stenting). Data obtained from the CMS for the year 2008 through the Medicare utilization statistics data set, Part B (Medicare Standard Analytic Files) were analyzed. The Standard Analytic Files contain all inpatient, outpatient, skilled nursing, physician, and supplier claims for a 5% random sample of fee-for-service Medicare beneficiaries. Beneficiaries enter the sample on the basis of their beneficiary identification number and remain in the sample until they leave the Medicare program. The 2008 Medicare 5% sample data was summarized to show the count of all services and the count of interventional and diagnostic catheterizations. Using billed American Medical Association Current Procedural Terminology (CPT) codes from 2008, the total number of diagnostic catheterizations (CPT codes 93508–93529), and PCIs were compiled. PCI was broken down into stenting (CPT code 92980) and angioplasty (CPT code 92982) used as a primary treatment modality, not in combination with other therapies.

Diagnostic cardiac catheterization and PCI volume was summarized by National Provider Identifier (NPI), which uniquely identifies individual physicians. All providers and all billed procedures in the 2008 Medicare 5% sample were identified. All diagnostic catheterizations and PCIs were then identified per physician provider. The NPI summary was tabulated by the volume of procedures reported for each NPI using the CPT codes.

A skewed distribution was observed in the 2008 Medicare 5% sample file, and using the “binomial distribution,” or “repeated Bernoulli trial,” the original distribution of procedure volumes in the year 2008 was inferred. Using the statistical correction, the result is a parametric estimate of the distribution of actual procedure volumes in the universe of beneficiaries, based on what was observed in the 2008 Medicare 5% sample data. When necessary, a simple linear interpolation between exact percentiles to obtain data for round-number percentiles for display purposes was utilized. Estimated 2008 volumes are presented, which were calculated using a multiplicative factor of 20.

It should be noted that the absolute volumes reported include only Medicare fee-for-service (Medicare FFS) procedures. The data presented is the estimated total 2008 Medicare FFS volume based on the 5% sampling data. The results show the distribution of these Medicare FFS diagnostic catheterization and PCI services among physicians performing any Medicare FFS procedures.

Furthermore, the underlying claims only reflect Medicare FFS data, and we did not convert these numbers into an estimated total volume of diagnostic catheterizations and PCIs for all payers. To convert the Medicare FFS volume to total volume for all payer types, a multiplicative factor of 2 would be a reasonable approximation based on the 2009 AHRQ data, where Medicare accounted for an estimated 51.4% of all diagnostic catheterizations, and 51.3% of all PCIs, respectively [2]. In this analysis, this conversion is not utilized in the results section but is referred to in the discussion section. The resulting procedure volume distributions presented here, based on the Medicare FFS data only, should be a reasonable proxy for the distribution of all diagnostic catheterization and PCI services among physicians.

RESULTS

In 2008, there were 1,198,610 Medicare FFS diagnostic catheterizations performed by 11,029 diagnostic-only and/or dual-diagnostic and interventional cardiologists. There were 378,372 Medicare FFS PCIs performed by 6,443 interventional cardiologists. The mean 2008 Medicare FFS procedural volume per-operator was 109 diagnostic cardiac catheterizations and 59 PCIs, respectively (see Table I).

Table I. 2008 Medicare fee-for-service Diagnostic Catheterization and PCIa Data

	Diagnostic catheterizations	PCI
a PCI, percutaneous coronary intervention.
Total procedure volume	1,198,610	378,372
Total cardiologists performing procedures	11,029	6,443
Average procedure volume per cardiologist	109	59

There was an uneven distribution of Medicare FFS diagnostic catheterizations and PCIs among cardiologists. For diagnostic catheterizations, 123,567 Medicare FFS procedures were performed by cardiologists whose total 2008 Medicare FFS diagnostic volume was 40 cases or fewer. Conversely, 187,741 Medicare FFS diagnostic catheterizations were performed by a very small group of high-volume cardiologists (359 total), each of whom performed >400 Medicare FFS procedures in 2008 (see Fig. 1).

Figure 1. Total Medicare fee-for-service diagnostic catheterizations by 2008 physician volume cohort.

Dividing diagnostic-only and/or dual-diagnostic and interventional cardiologists into three categories based on Medicare FFS procedural volume for 2008 (those performing 40 or fewer, 41–200, and >200 Medicare FFS diagnostic procedures in 2008), there were 4,009 cardiologists who performed 40 or fewer Medicare FFS diagnostic catheterizations in 2008. This cohort represents 36.3% of all diagnostic-only and/or dual-diagnostic and interventional cardiologists, and their aggregate volume represents 10.3% of the total Medicare FFS diagnostic catheterizations performed in 2008. There were 1,449 cardiologists who performed >200 Medicare FFS diagnostic catheterizations, representing 41.6% of the total Medicare FFS diagnostic procedures performed in 2008 (see Table II).

Table II. 2008 Medicare fee-for-service Diagnostic Catheterizations by Physician Procedure Volume Cohort

Diagnostic catheterization volume	Physicians in cohort	% Total	Aggregate diagnostic catheterizations performed by cohort	% Total	Cohort average volume
1–40	4,009	36.3	123,567	10.3	31
41–200	5,571	50.5	576,145	48.1	103
>200	1,449	13.1	498,898	41.6	344
	11,029		1,198,610

For PCI, 112,679 Medicare FFS procedures were performed by interventional cardiologists whose total 2008 Medicare FFS interventional volume was 40 cases or fewer. Conversely, 3,911 Medicare FFS PCIs were performed by a very small group of high-volume interventional cardiologists (nine total), each of whom performed >400 Medicare FFS PCIs in 2008 (see Fig. 2).

Figure 2. Total Medicare fee-for-service PCI by 2008 physician volume cohort.

Dividing interventional cardiologists into three categories based on Medicare FFS procedural volume for 2008 (those performing 40 or fewer, 41–200, and >200 Medicare FFS PCIs in 2008), there were 3,929 interventional cardiologists who performed 40 or fewer Medicare FFS PCIs in 2008. This cohort represents 61.0% of all interventional cardiologists, and their aggregate volume represents 29.8% of the total Medicare FFS PCI procedures performed in 2008. There were 185 interventional cardiologists who performed >200 Medicare FFS PCIs, representing 13.8% of the total Medicare FFS PCI procedures performed in 2008 (see Table III). The low-volume cohort (interventional cardiologists performing 40 or fewer Medicare FFS PCIs in 2008) per-physician Medicare FFS procedure average is 29, whereas the mid- and high-volume cohorts averaged 92 and 283 Medicare FFS PCI procedures in 2008, respectively (see Table III). This “skew” in the distribution of procedure volume is substantial, and the distribution is clearly not a Normal or Gaussian distribution. Consequently, calculation of an overall average per-physician procedure volume is highly misleading. This explains why even though 61% of all physicians performed 40 or fewer Medicare FFS PCIs in 2008, the calculated average procedure volume for all physicians is 59 (see Table I).

Table III. 2008 Medicare fee-for-service PCIa by Physician Procedure Volume Cohort

PCI volume	Physicians in cohort	% Total	Aggregate PCI performed by cohort	% Total	Cohort average volume
a PCI, percutaneous coronary intervention.
1–40	3,929	61.0	112,679	29.8	29
41–200	2,329	36.1	213,345	56.4	92
>200	185	2.9	52,348	13.8	283
	6,443		378,372

DISCUSSION

In the United States over the last decade, the general direction of invasive procedures has been toward fewer procedures and revascularizations per capita [6]. In contrast, the supply of interventional cardiologists has increased over this same time frame, with significant implications for the manpower requirements in the field over the coming decade. In this study, we report the number of interventional cardiologists performing Medicare FFS procedures was 6,443 in 2008. Based on self-reported numbers from the last workforce assessment from ACC, the number of cardiologists performing PCI may be closer to 8,000–9,000 [8].

Current ACC/AHA/SCAI guidelines suggest that interventional cardiologists perform 75 or more PCIs yearly to maintain their procedural proficiency. With 693,315 inpatient PCIs performed in the US in 2009, the ACC/AHA/SCAI guidelines imply an upper bound for the optimal total number of practicing interventional cardiologists nationwide of 9,244. This crude estimate assumes a uniform distribution of procedures across interventional cardiologists, ignoring real-world dynamics like variations in operator volume relating to individual practice structure, population density, and the presence (or absence) of regional referral centers for advanced cardiac care. The skew in the distribution of volume per operator seen in the 2008 CMS data clearly demonstrates that anticipating future manpower needs is far more complex than a simple calculation of this type.

The 2008 Medicare FFS data for diagnostic catheterizations and PCIs presented here, when segregated by yearly operator volume, reveals a marked difference in the distribution of the two respective procedures among operators. Just over 10% of diagnostic catheterizations were performed by operators with 40 or fewer Medicare FFS diagnostic catheterizations in 2008, contrasted with almost 30% of PCIs performed by operators with 40 of fewer Medicare FFS PCIs that same year. Indeed, a significant majority of interventional cardiologists (61%) performed 40 or fewer Medicare FFS PCIs in 2008, compared to only 36.3% of diagnostic-only and/or dual-diagnostic and interventional cardiologists performing 40 or fewer Medicare FFS diagnostic catheterizations. Given the correlation between higher operator volume and improved patient outcomes, this data raises questions regarding whether the current distribution of PCIs across interventional cardiologists is rational [9].

The percentage of interventional cardiologists performing 40 or fewer Medicare FFS PCIs in 2008 may reflect the increasing number of facilities offering PCI nationwide, the expanding geographic coverage of PCI capability, and the increasing necessity to staff catheterization laboratories with 24 hour-per-day, 7 day-per-week coverage to meet guideline door-to-balloon times. From a heuristic standpoint, “24/7” catheterization laboratory staffing requires approximately five operators per laboratory. With 1,637 laboratories operating nationwide in 2008, simple arithmetic suggests that 8,185 interventional cardiologists would be an appropriate “equilibrium” level, independent of individual physician yearly volume. With acute myocardial infarction representing a growing percentage of PCI volume, it should be expected that increasing geographic coverage for PCI capability will necessarily dilute operator volumes as the procedure follows a natural migration from population-dense, high-volume centers into lower population-density exurban and rural communities.

There are a number of limitations to the data analysis presented. The data presented captures only Medicare FFS diagnostic catheterizations and PCIs performed in 2008. This excludes the entire universe of non-Medicare procedures. While Medicare patients accounted for roughly 51.3% of total PCIs performed based on the 2009 AHRQ data, we cannot be certain that simply “grossing up” the yearly Medicare FFS volume by a factor of 2 for each of the cohorts listed in Tables II and III represent an accurate estimate of total yearly volumes [2]. In addition, the data presented represents a nationwide estimate of procedure volumes. Any conclusions regarding per-operator volumes would result in an over-estimate of “true” per-operator volumes in states with older average populations (e.g., Florida), and an under-estimate in younger states (e.g., Colorado). Similarly, areas with differing rates of Medicare patients in the Medicare FFS program (as opposed to managed care) also skews the data.

Furthermore, with regard to the PCI data, fewer PCIs performed does not necessarily correlate with fewer advanced interventional procedures performed. The Medicare FFS data presented captures only coronary interventions, it does not include procedures where the use of intravascular ultrasound or fractional flow reserve technologies result in foregoing an intervention. Nor does it include transcatheter valvular procedures, percutaneous closure device placements, or peripheral interventions such as lower extremity or carotid artery percutaneous interventions. As the role of the interventional cardiologist evolves to include more non-coronary interventions it will be appropriate to include these procedures as part of their total annual volumes [1].

Finally, the data presented reflects the most recently available CMS Medicare 5% sample file, that for 2008. We are unable to present year-over-year procedure volume trends since CMS releases the Medicare 5% sample file data only periodically, with the next most recent 5% sample dating back to 2005.

The 2008 Medicare FFS data reveals a significant difference in the distribution of diagnostic catheterizations versus PCIs across physicians, with lower-volume operators representing over 60% of the universe of interventional cardiologists, versus lower-volume diagnostic-only and/or dual-diagnostic and interventional cardiologists representing just over 36% of the diagnostic universe. This analysis, however, only scratches the surface of current diagnostic and PCI volumes nationwide, and the implications for future manpower needs in the catheterization laboratory. Recent reports suggest that as the obesity epidemic increases, there will be a 30% increase in the number of coronary artery disease diagnoses, many of whom will require invasive and interventional procedures [10]. This data suggests that the number of interventional cardiologists currently being trained is adequate to meet that challenge, as long as the per-physician procedure distribution narrows. If the current variation in per-physician procedure distribution continues, or even widens, then many more interventional cardiologists will need to be trained, with many of those not performing an acceptable minimum of cases.

The main problem identified by this analysis is that the per-physician volume distribution of PCI, based on the 2008 Medicare FFS data, is concerning because there is a very high percentage of low-volume operators. Certainly this finding raises questions in connection with the necessary operator volume for technical and cognitive skill maintenance. Moreover, the fact that the incidence of coronary artery disease is likely to increase will offset some of the recent trends in regard to the volume and appropriateness of invasive procedures. A key factor in determining the overall manpower requirements for PCI in the future is whether or not the per-physician procedure distribution observed in this study, with its skew toward low-volume operators, will persist or normalize. The data presented suggests that clinical cardiologists who perform limited numbers of diagnostic and interventional procedures may have an increasingly difficult time maintaining their skills. The skew in the 2008 Medicare FFS PCI per-physician procedure volume distribution suggests that many interventional and invasive cardiologists in the US do not currently meet the minimum criteria traditionally set by the AHA/ACC/SCAI guidelines [11]. Notably, the most recent quality statements of the SCAI suggest moving away from a minimum volume criterion [12].

This study highlights the need for better data aggregation and surveillance regarding the total number of catheterization laboratories operating nationwide, the total number of physicians staffing these laboratories, the demographics, population density, and geographic areas they serve, and the clinical scenarios under which both diagnostic catheterizations and PCI are performed, to more accurately predict future manpower needs in the catheterization laboratory.