Who provides what care? An analysis of clinical focus among the national emergency care workforce

Abstract

Objective:

Clinician expertise has been associated with improved patient outcomes, yet ED clinicians often work in various clinical settings beyond the ED and, therefore, may risk expertise by having less clinical focus. We sought to describe clinical focus among the emergency care workforce nationally.

Methods:

Using the 2017 Medicare Public Use Files (PUF), we performed a cross-sectional analysis of clinicians receiving reimbursement for emergency care Evaluation & Management (E/M) services from Medicare fee-for-service Part B. Clinicians were categorized by type as EM physicians, non-EM physicians, and advanced practice providers (APPs). The primary outcome was the clinical focus of the individual clinician, defined as the proportion of E/M services within the ED setting relative to a clinician’s total E/M services across all practice settings.

Results:

Of 65,710 unique clinicians providing care to Medicare fee-for-service beneficiaries in the ED setting, 39,016 (59.4%) were classified as EM physicians, 8,123 (12.4%) as non-EM physicians, and 18,571 (28.5%) as APPs. The individual clinician median focus was 92.8% (interquartile range [IQR]: 87.0, 100.0) for EM physicians, 45.2% (IQR: 5.1, 97.0) for non-EM physicians, and 100.0% (IQR: 96.3, 100.0) for APPs.

Conclusion:

EM physicians have twice as much clinical focus in comparison to non-EM physicians providing emergency care to Medicare fee-for-service beneficiaries. These findings underscore the importance of diverse training and certification programs to ensure access to clinically focused ED clinicians.

1. Introduction

Of nearly 60,000 unique emergency care clinicians, an analysis of 2014 Centers for Medicare & Medicaid Services (CMS) data classified 61% as EM physicians, 14% as non-EM physicians, and 25% as advanced practice providers (APPs) [1]. Notably, non-EM physicians and APPs comprising the workforce provide a significant proportion of emergency care in rural areas [1,2]. Yet, outcomes for patients cared for by clinicians with less clinical focus in emergency care (i.e. working concurrently in other practice settings) may be worse. Prior work has shown that Internal Medicine-trained physicians, providing care in the ED setting in addition to office-based or inpatient care, were more likely to have a patient return to the ED in need of a hospitalization compared to their EM-trained colleagues [3]. Supporting the need for clinical focus, a literature review identified that clinical focus and residency-training in EM has been shown to contribute to effective, safe, efficient, and cost-effective emergency care [4].

Previous studies of the national emergency medicine workforce have been limited to descriptions of the count of, but not the practice of, clinicians providing care in the ED setting [5-9]. Prior work has not acknowledged that some clinicians work across multiple settings, a practice which may risk the loss of focus in providing care in the ED setting. Understanding variation in the clinical focus of clinicians in the ED setting can inform workforce development policies, training programs, and incentives to reduce potential maldistributions of expertise across the US. We sought to use a national administratve claims-based dataset of clinician care for Medicare beneficiaries to compare clinical focus of EM physicians, non-EM physicians, and APPs providing care in the ED setting.

2. Methods

2.1. Study Design and Dataset

We performed a cross-sectional analysis using the CMS 2017 Provider Utilization and Payment Data from the Physician and Other Supplier Public Use File (PUF). Because patient identifiers were not used, this study was deemed exempt by the Institutional Review Board.

The PUF provides information on services provided to Medicare beneficiaries by healthcare professionals, and is based on CMS administrative claims data. The data are available from the CMS Chronic Condition Data Warehouse (CCW), a database with 100% of Medicare enrollment and fee-for-service claims data. The PUF provides information on all clinicians with active national practitioner identification (NPI) numbers reimbursed greater than 10 times for a specific E/M service in 2017 by the Medicare fee-for-service Part B program. Clinician demographics (name, credentials, gender, address, entity type) within the PUF are included from the National Plan and Provider Enumeration System (NPPES) [10].

2.2. Definitions

We included all clinicians in the Medicare PUF with addresses registered to the 50 states and Washington D.C. in the analysis. Referred to in subsequent text, ‘clinicians providing care in the ED setting’ were defined as those who received greater than 10 reimbursements for any E/M (levels 1 to 5) service, based on HCPCS codes 99281, 99282, 99283, 99284, and 99285 denoting care in the ED setting. EM physicians were identified in the PUF based on the ‘Provider Type’ variable, identified from the clinician specialty code reported on the claim received from participating institutions. Non-EM physicians were defined as those with another medical specialty. For example, at the corresponding author’s institution, the medical biller uses the specialty associated with the highest level of training for each clinician. For clinicians identified as belonging to multiple specialties, the PUF reports the specialty for which the clinician billed the largest number of services that year. APPs were defined based on ‘Provider Type’ and included nurse practitioners (NPs) and physician assistants (PAs). NPs and PAs are able to provide services independent of direct physician involvement and bill Medicare under their own NPI.

Based on prior literature [11-13], we classified the universe of common E/M services into nine setting categories defined as ED care, critical care, observation care, office-based care, inpatient care, nursing facility care, telehealth, home visit care, and assisted living facility care. The complete list of included E/M services and HCPCS codes for each setting category can be seen in Supplemental Table 1.

2.3. Primary Outcomes

At the level of the individual clinician, the primary outcome was the clinical focus, defined as the proportion of E/M services within the ED setting relative to a clinician’s total E/M services across all practice settings. We calculated the volume of E/M services within a practice setting as a ‘per clinician’ proportion. For example, individual non-EM physician may have billed 45% of his/her services within the ED setting (HCPCS 99281-99285), 5% as critical care, 30% as office-based care, 10% as inpatient care, and 10% as nursing facility care.

2.4. Statistical Analysis

Data preparation and analyses were performed in Stata 16 (StataCorp, College Station, TX). We report descriptive statistics for clinical focus. For data visualization, we used the open source Pandas library and the ‘seaborn’ module, a Python library based on ‘matplotlib’. Grouped by clinician type, we present violin plots showing the median, IQR, and distribution of the service for the clinician level data.

3. Results

A total of 1,088,687 clinicians were included in the 2017 CMS PUF data set, including 65,710 unique clinicians that provided care in the ED setting. Of those, 39,016 (59.4%) were EM physicians, 8,123 (12.4%) were non-EM physicians, and 18,571 (28.5%) were APPs. Among non-EM physicians who provided care in the ED setting, Family Practice and Internal Medicine specialty classification was most common, respectively 41.0% and 17.9% (Supplemental Table 2). Physician assistants and nurse practitioners respectively accounted for 64.3% and 35.7% of APPs providing care in the ED setting.

At the level of the clinician type, data is presented for billed E/M services (Table 1, Figure 1). In 2017, 77.0% of all emergency care claims (HCPCS codes 99281-99285) were by EM physicians. Care in the ED setting accounted for 87.5% of all E/M services billed by EM physicians in caring for Medicare beneficiaries. Critical care (8.5%), office-based care (2.4%), and observation care services (0.7%) were among the most common remaining E/M services billed by EM physicians. Of those providing care in the ED setting, non-EM physicians primarily billed for office-based care (35.3%) and inpatient care (28.7%) services, followed by care in the ED setting (28.0%). Of those providing care in the ED setting, APPs primarily billed for care in the ED setting, accounting for 81.0% of all E/M services billed (Table 1).

Table 1.

Evaluation & management (E/M) category summations by clinician type.

	EM Physician	Non-EM Physician	APP	Total
ED care	15,886,052 (87.5)	1,685,090 (28.0)	3,064,021 (81.0)	20,635,163
Critical care	1,546,900 (8.5)	174,741 (2.9)	31,332 (0.8)	1,752,973
Observation care	124,088 (0.7)	84,511 (1.4)	35,868 (1.0)	244,467
Office-based care	438,327 (2.4)	2,122,288 (35.3)	478,208 (12.6)	3,038,823
Inpatient care	109,952 (0.6)	1,725,650 (28.7)	97,278 (2.6)	1,932,880
Nursing facility care	44,131 (0.2)	201,460 (3.3)	58,366 (1.5)	304,057
Telehealth care	1,744 (<0.01)	509 (<0.01)	361 (<0.01)	2,614
Home Visit care	5,063 (0.03)	10,077 (0.2)	7,692 (0.2)	22,832
Assisted Living Facility care	5,133 (0.03)	9,671 (0.2)	9,502 (0.3)	24,306
Total	18,161,390	6,013,997	3,782,628

Note: % are column percentages

Abbreviations: APP, advanced practice provider; E/M, evaluation & management.

Figure 1. Evaluation & management (E/M) category representation by clinician type.

Abbreviations: ALF, assisted living facility care; APP, advanced practice provider; CC, critical care; Inpt, inpatient care; NF, nursing facility care; Obs, observation care; Office, office-based care.

3.1. Clinical Focus

The individual EM physician had a median focus of 92.8% (IQR: 87.0, 100), identifying that care in the ED setting accounted for the majority of the individual EM physician’s E/M service volume billed. Of those providing care in the ED setting, the individual non-EM physician had a median focus of 45.2% (IQR: 5.1, 97.0) and the individual APP had a median focus of 100% (IQR: 96.3, 100). Values for the remaining E/M service settings are presented alongside violin plots (Figure 2).

Figure 2. Distribution of services billed by individual clinician for E/M service categories.

Note: ED care (panel A), critical care (panel B), observation care (panel C), office-based care (panel D), inpatient care (panel E), nursing facility care (panel F), telehealth (panel G), home visit care (panel H), and assisted living facility care (panel I).

Note: White dot – median; thick black line – interquartile range; thin black line – 1.5x interquartile range; colored area – kernel density estimate.

Note: Denominator is the number of clinicians of a specific clinician type providing emergency care.

Abbreviations: APP, advanced practice provider; E/M, evaluation & management; IQR, interquartile range.

4. Discussion

In our study of a nationwide 2017 Medicare reimbursement database, we identified 65,710 unique clinicians providing care in the ED setting, over 7,000 more clinicians than a recent study assessing the 2014 database [1]. This is the first emergency medicine analysis assessing the individual clinician level volume of E/M services across practice settings. There are two key findings from our analysis. First, 77% of all emergency care claims were by EM physicians, and 87.5% of all E/M services billed by EM physicians were related to care in the ED setting. Comparatively, care in the ED setting accounted for a smaller proportion of all E/M services billed for non-EM physicians and APPs of those that provide care in the ED setting. Second, individual EM physicians had a high clinical focus of care in the ED setting, with non-EM physicians providing care in the ED setting being more likely to spend a majority of clinical care services outside the ED, and instead focused on care in office-based or inpatient settings. These findings provide fundamental descriptive data, which can inform future studies or serve as a comparison between clincians providing care in the ED setting.

Non-EM physicians and APPs are an important part of the emergency workforce equation. Compared to prior work [1], our work shows that APPs comprise an increasing proportion of clinicians providing care in the ED setting, and that the proportion of EM physicians and non-EM physicians are decreasing. Over the three-year timeframe between study analyses [1], the number of clinicians classified as EM physicians increased by 3,160 (8.8%), the number classified as APPs increased by 4,211 (29.3%), and the number classified as non-EM physicians decreased by 274 (3.3%). Despite controversy regarding their inclusion in specialty organizations, non-EM physicians frequently provide care in rural ED settings that are less likely to be staffed by EM-trained physicians [1,14,15]. Additionally, the number of emergency care NP/PA postgraduate training programs is increasing [16,17]. The supply of NPs and PAs providing care in the ED setting is projected to continue growing over the coming decade [18], and in rural ED settings the role of the APP is becoming more prominent. As aging non-EM physicians retire and the APP supply increases, future investigations will need to ensure that clinicians who spend a significant portion of their clinical time outside the ED setting (i.e. those less focused) are able to safely provide emergency care.

The volume-outcome adage “practice makes perfect” [19] identified in prior literature [20-26] may similarly apply to clinical focus at the ED clinician level, with our work suggesting the need for outcome comparisons between clinically focused and non-clinically focused clinicians providing care in the ED setting. Future analyses including emergency physicians with other non-clinical responsibilities (i.e. research, administration) may also prove worthwhile when investigating clinical focus.

There are several limitations of the present study that deserve mention. First, the PUF data may not be representative of the clinician’s entire practice as it only includes information on Medicare fee-for-service beneficiaries. As a result of the dataset not including pediatric or non-Medicare adult patients, we are unlikely to affect assessments of clinician focus which is based on the proportional distribution of a clinician’s work. Second, all aggregate calculations from the PUF are underestimates, as CMS excluded all line items with 10 or fewer beneficiaries. If included, medians and IQRs would be expected to be reduced slightly, however this reduction would be expected to be comparable across clinician types and not alter the direction of our findings. Also, institutional variation in reporting ‘Provider Type’ may exist for clinicians reporting multiple specialties. CMS’s selection of the ‘Provider Type’ specialty code associated with the largest number of services, if a clinician belongs to multiple specialties, is not expected to systematically redistribute clinicians in a biased fasion, thereby maintaining our findings. Finally, services billed for APPs may be underreported as this only captures APP services in which no attending physician supervision was present to sufficiently meet physician billing requirements. However, many groups bill directly under the APP, since 85-100% of the service fee can be captured.

In summary, EM physicians comprise 3 out of 5 clinicians in the emergency workforce. Individual EM physicians are highly focused on emergency care, billing over 90% of E/M services in the ED setting, while non-EM physicians providing care in the ED setting exhibited less focus and billing less than 50% of E/M services related to emergency care. This greater understanding of focus at the clinician level can inform workforce development policies, training programs, and incentives to ensure access to focused clinicians. Future work should explore the association of these differences in clinical focus in the ED setting with patient outcomes and costs.