The Characteristics of Physicians Who Primarily Practice in Nursing Homes

The number of clinicians who primarily practice in nursing homes (NHs), recently referred to as “SNFists,” increased by 33% between 2012 and 2015.¹ Recent research suggests that SNFists provide higher-quality care to NH patients than other clinicians,² and NH medicine has been proposed as a formal specialty,³ yet little is known about the characteristics of SNFists. Such information will be important in further linking attributes that may be associated with outcomes—for example, training and practice setting—and which may have policy implications. Using a nation-wide sample, we compared the characteristics of physicians who were SNFists with those of physicians who never visited NHs and those who occasionally visited NHs.

Methods

We merged Medicare Part B claims for a 20% national sample of beneficiaries with the Provider and Other Supplier Public Use File for the period 2014 through 2017 using National Provider Identifiers. After limiting the sample to physicians with ≥100 claims, 3 categories of physicians were identified: (1) SNFists with ≥90% of their evaluation and management (E&M) visits in NHs annually^1,4; (2) non-NH physicians with no E&M visits in NHs; and (3) occasional NH physicians with ≥1 E&M visit in a NH, but with <90% of their E&M visits in NHs annually.

Medicare claims were linked to the Medicare Data on Provider Practice and Specialty, Medicare Accountable Care Organization (ACO) Provider file, and IQVIA Physician Database⁵ to identify physician characteristics. We compared SNFists to (1) non-NH physicians and (2) occasional NH physicians. Comparisons were based on age, gender, primary care specialty (internal medicine, family practice, geriatrics, general practice), top 25 medical school graduate,⁶ foreign trained, ACO participation, practice size, and rural location. The statistical significance of differences was assessed using t tests for continuous variables and chi-square tests for categorical variables (2-sided, P < .05) (Supplementary Material).

Findings

We identified 6094 (1.1%) SNFists, 468,842 (87.4%) non-NH physicians, and 61,328 (11.4%) occasional NH physicians over the study period (Table 1). In 2017, SNFists and occasional-NH physicians billed 32.3% and 67.6%, of all E&M claims for NH visits by physicians, respectively.

Table 1.

Characteristics of SNFists, Non—nursing Home Physicians, and Occasional Nursing Home Physicians

	SNFists*, n (%) (n = 6094; 1.1)	Non—Nursing Home Physicians*, n (%) (n = 468,842; 87.4)	Occasional Nursing Home Physicians*, n (%) (n = 61,328; 11.4)
The 5 most common specialties	Internal Medicine 2008 (33.0)	Family Practice 63,376 (13.5)	Internal Medicine 18,917 (30.9)
	Family Practice 1637 (26.9)	Internal Medicine57,453 (12.3)	Family Practice 17,475 (28.5)
	Physical Medicine and Rehabilitation 493 (8.1)	Emergency Medicine 43,222 (9.2)	Hospitalist 4385 (7.2)
	Psychiatry 447 (7.3)	Hospitalist 25,546 (5.5)	Cardiology 2190 (3.6)
	Geriatrics 395 (6.5)	Orthopedic Surgery 21,107 (4.5)	Psychiatry 2019 (3.3)
Age categories† <30 y	11 (0.2)	1438 (0.3)	82 (0.1)
30–39 y	846 (13.9)	99,054 (21.1)	8914 (14.5)
40–49 y	1454 (23.9)	127,401 (27.2)	15,247 (24.9)
50–59 y	1412 (23.2)	113,562 (24.2)	16,358 (26.7)
60–69 y	1482 (24.3)	93,849 (20.0)	15,792 (25.8)
≥70 y	889 (14.6)	33,505 (7.2)	4933 (8.0)
Female†	2309 (37.9)	141,338 (30.2)	15,168 (24.7)
Primary care†	4259 (69.9)	125,763 (26.8)	38,097 (62.1)
Attended a top 25 medical school‡,§	437 (12.2)	61,184 (17.9)	4272 (11.3)
Foreign trained‡	2038 (36.3)	111,412 (24.6)	22,022 (36.9)
Participation in an ACO||	1656 (27.2)	163,654 (34.9)	25,375 (41.4)
Practice size†,**
Solo	1352 (22.2)	62,396 (13.3)	13,864 (22.6)
2–9 physicians	1040 (17.1)	77,563 (16.5)	13,276 (21.7)
10–49 physicians	1588 (26.1)	85,586 (18.3)	11,139 (18.2)
50–99 physicians	819 (13.4)	41,549 (8.9)	5359 (8.7)
≥100 physicians	1295 (21.3)	201,748 (43.0)	17,690 (28.8)
Rural	666 (10.9)	37,820 (8.1)	12,333 (20.1)

^*SNFists defined as physicians with ≥90% of their evaluation and management visits in nursing homes (NHs) (Health Common Procedure Coding System Codes 99304–99310, 99315, 99316, 99318); non-NH physicians defined as physicians with no evaluation and management visits in NHs; and occasional NH physicians were defined as physicians with ≥1 E&M visit in a NH, but with <90% of their evaluation and management visits in NHs. For statistical significance of comparisons, t tests were used for continuous variables and chi-square tests for categorical variables (2-sided, P < .05). All differences were statistically significant, except for having attended a top 25 medical school (P = .12) and being foreign trained (P = .37), in comparisons of SNFists and occasional NH physicians.

^†The Medicare Data on Provider Practice and Specialty file was used to identify, age, gender, specialty, and practice size.

^‡The IQVIA Physician Database was used to identify medical schools attended by physicians.

^§Top 25 medical school rankings were drawn from the US News and World Report 2017, America’s Best Graduate Schools, Schools of Medicine.

^||The Medicare Accountable Care Organization Provider file was used to identify physicians participating in Accountable Care Organizations.

^**Practice size was defined as the number of unique National Provider Identifiers in the practice.

SNFists vs Non-NH Physicians

Compared to non-NH physicians, SNFists were more likely to be aged ≥70 years (14.6% vs 7.2%), primary care physicians (69.9% vs 26.8%), foreign trained (36.3% vs 24.6%), and in solo practice (22.2% vs 13.3%) (Table 1). SNFists were less likely to be top 25 medical school graduates (12.2% vs 17.9%) or ACO participants (27.2 vs 34.9%). The majority of SNFists, 59.9%, specialized in internal medicine or family practice compared to 25.8% of non-NH clinicians. The 5 most common specialties among SNFists were internal medicine (33.0%), family practice (26.9%), physical medicine and rehabilitation (8.1%), and psychiatry (7.3%). Only 6.5% of SNFists specialized in geriatrics.

SNFists vs Occasional NH Physicians

In comparison to occasional NH physicians, SNFists were more likely to be aged ≥70 years (14.6% vs 8.0%) and female (37.9% vs 24.7%), and were less likely to be ACO participants (27.2% vs 41.4%) or in rural locations (10.9% vs 20.1) (Table 1).

In all comparisons above, differences were statistically significant.

Discussion

Only a small percentage of physicians (12.5%) who billed Medicare had any claims for NH visits. SNFists provided a disproportionate amount of primary care in NHs, accounting for 1.1% of all physicians but billing for 32.3% of all NH E&M visits. SNFists differed from other physicians in distinct ways; they were older, more likely be foreign trained, more likely to be in solo practice, but less likely to have attended a top 25 medical school or to have participated in an ACO. SNFists were mostly primary care physicians; only a small percentage were geriatricians.

Individuals receiving care in NHs are among the most vulnerable in the US health care system, as exemplified by the outsized impact of the COVID-19 pandemic on this population. Low ACO participation rates among SNFists are concerning if they reflect a reluctance to take part in innovative delivery models that may improve the quality of care. This is particularly important given the disproportionate share of primary care that SNFists provide to NH patients. Further study is needed to determine whether SNFist characteristics, practice type, or low participation rates in innovative delivery models are associated with the quality of care they provide, and also to learn more about how SNFists, compared to non-SNFists, provide care to NH patients. Determining the “value proposition” for physicians practicing in NHs will also require broader use of medical provider—specific quality metrics and measures that accurately reflect medical staff organization rather than identifying SNFists solely on the volume of NH visits.^7,8