An Evaluation of the Decline in Primary Care Physician Visits,2010 to 2021

Abstract

Introduction:

Primary care is the backbone of the United States healthcare system, yet it continues to be underfunded and inaccessible to many. Previous studies explored primary care visit patterns until 2016, leaving gaps in our understanding of how recent factors like state Medicaid expansions, the Coronavirus Disease 2019 (COVID-19) pandemic, and population growth have influenced these patterns. Hence, the objective of the current study was to analyze the trends in primary care visits provided by outpatient clinicians over time and by visit type and contextualize study findings within the changing healthcare landscape.

Methods:

The Medical Expenditure Panel Survey data (2010-2021) were used to examine trends in the total number and share of ambulatory, preventive, acute, and chronic care visits and investigate sociodemographic factors associated with a given clinician visit. The outcome variable was the clinician type: Primary Care Physicians (PCPs), Internal Medicine (IM) subspecialists, and Nurse Practitioners or Physician Assistants or Registered Nurses (NP/PA/RNs). Explanatory variables included gender, age, race/ethnicity, education, region of respondent’s residence, income-to-poverty ratio, insurance coverage, number of chronic conditions, and survey year. Univariate, bivariate, and multinomial logistic regression analyses were performed.

Results:

The visits led by PCPs and IM subspecialists declined by 43% and 23% between 2010 and 2021, respectively. However, visits led by NP/PA/RNs increased by 98%. From 2010 to 2021, the proportion of preventive care visits provided by PCPs, IMs, and NP/PA/RNs increased by 25%, 7%, and 4%, respectively. PCPs provided fewer acute and chronic care visits in 2021 than in 2010. Regression analyses illustrated that relative to Non-Hispanic White patients, non-White patients had a higher likelihood of seeing PCPs. Patients reporting 1 chronic condition were more likely to obtain care from an IM or NP/PA/RN than a PCP. In contrast, those with 2 or more chronic conditions had a greater propensity to see PCP than NP/PA/RN.

Conclusion:

Despite a growing need for primary care services, the decrease in visits to PCPs is concerning and requires further examination. The declining trends in acute and chronic care visits raise questions as to whether primary care, in its current form, can continue to provide its essential attributes and services.

Keywords

primary care primary care physicians workforce chronic disease preventive medicine practice patterns

Introduction/Objective(s)

The benefit of high-quality primary care to patients and populations has been demonstrated repeatedly.^1-6 Access to a usual source of primary care for patients is associated with lower overall healthcare costs and more appropriate utilization of healthcare, including fewer visits to the emergency room.⁷ Given their comprehensive training, primary care clinicians are uniquely equipped to manage a wide range of disease processes, making them indispensable in treating chronic conditions within an aging population increasingly affected by multimorbidity. Unfortunately, primary care is currently in crisis, limiting access for many patients in the US.⁸ The primary care system is overburdened by various factors, including access issues to clinician burnout.^8,9 Despite an increasing need for primary care services in the United States,^10-12 the number of primary care physician (PCP) visits has continuously decreased in the last 2 decades.^13-17 Between 2008 and 2016, these visits declined by 6% to 25% based on multiple data sources.¹⁰ This drop surprisingly persisted after the enactment of the Affordable Care Act (ACA) in 2010, which intended to make care more accessible to patients.

Prior studies explore patterns of primary care visits until 2016. Since then, there have been many changes that would increase the need for primary care visits primary care practice patterns, including state Medicaid expansion, the Coronavirus Disease 2019 (COVID-19) pandemic, and population growth, in addition to the downstream effects of the ACA.^18
-21 Legislation from 40 states and the District of Columbia has improved insurance coverage for millions of Americans by expanding Medicaid benefits.¹⁸ This expansion has likely improved the individual’s ability to receive care, yet the primary care workforce has not seen a significant rise in training or active clinicians.^8,18,21 How primary care interacts with patients underwent a considerable disruption during the COVID-19 pandemic. Clinical care went from in-person to virtual; clinicians left the workforce or decreased their patient-facing time due to burnout, and mistrust in the medical community changed the day-to-day interaction with patients. These factors have likely impacted the visit trends for primary care.

Using the Medical Expenditure Panel Survey (MEPS), this study aimed to explore trends in outpatient visits before the enactment of the Affordable Care Act through the COVID-19 pandemic.

Methods

Data Source

We used the 2010 to 2021 Medical Expenditure Panel Survey (MEPS) data for this analysis. Administered by the Agency for Healthcare Research and Quality (AHRQ), MEPS surveys the civilian and non-institutionalized population in the United States to provide national estimates of healthcare use and costs.²² Participants answered detailed questions about their demographics, health conditions, healthcare use, medications, spending, and insurance coverage over 5 rounds.^23
-25 For a sample of MEPS respondents, MEPS collects data on dates of visits, healthcare service use, expenditures, payment sources, diagnoses, and procedure codes from their clinicians to verify and supplement the information provided by the respondents.²⁶ From 2010 to 2021, the response rates varied between 27.6% and 56.3%.²⁵ The COVID-19 pandemic reduced the response rates. Data was collected using phone interviews instead of in-person interviews in 2020 and 2021.²⁷ MEPS extended panels 23 and 24 to 9 rounds instead of the historical 5 rounds to overcome challenges with lower response rates in 2020 and 2021.^25,27

We derived data from 2010 to 2021 consolidated, office-based, outpatient, and medical condition files. Outpatient visits included visits to a clinic, acute care center, retail clinic, and telehealth visits. Although MEPS collected data on telehealth visits through a separate module, it reclassified them as outpatient or office-based visits. The information about the visit, including the reason for the visit and the provider’s specialty, was included in the office-based and outpatient files. The medical conditions file included the International Classification of Diseases (ICD)-9 or ICD-10 diagnosis codes. The patient-level sociodemographic characteristics were included in the consolidated files. We linked office-based and outpatient visit-level files for each survey year with the medical condition files and consolidated files, and then combined data from 2010 to 2021.

Study Population

Our sample included all the respondents in the 2010 to 2021 pooled data who had at least 1 office-based visit to examine trends in the distribution of ambulatory care visits by clinician type. For the bivariate and multivariate regression analyses, we restricted our sample to MEPS survey participants aged 18 years and over with non-missing values for all the study measures.

Study Measures

For the trend analyses (Table 1), the primary outcomes included the total number of ambulatory, preventive, acute, and chronic care visits by clinician type. Based on the responses to the question regarding which clinicians respondents saw for their office-base care, the clinicians were categorized into (1) primary care physicians (PCPs – family physicians, general practitioners, internists, pediatricians, and geriatricians); (2) internal medicine subspecialists (IM subspecialist—allergists or immunologists, cardiologists, endocrinologists, gastrologists, hematologists, nephrologists, oncologists, pulmonologists, and rheumatologists); (3) Nurse Practitioners, Physician Assistants, and Registered Nurse (NP/PA/RN); (4) surgical specialists (general surgeons, ophthalmologists, orthopedic surgeons, otolaryngologists, plastic surgeons, thoracic surgeons, and urologists); (5) obstetricians and gynecologists (Ob/Gyn); (6) psychiatrists (7) other physicians (anesthesiologists, dermatologists, neurologists, osteopaths, pathologists, physical medicine/rehabilitation doctors, proctologists, radiologists, and physicians unclassified in MEPS); (8) behavioral health clinicians (clinical psychologists, clinical social workers, marriage counselors and family therapists (latter 2 clinicians were included since 2018) and (9) Other non-physician clinicians (chiropractors, dentists/dental hygienists, clinical nurse midwives, optometrists, podiatrists, physical therapists, occupational therapists, massage therapists, homeopaths, naturopaths, herbalists, other complementary, and alternative therapists and non-physician clinicians in MEPS).

Table 1.

Study Measures.

Trend analyses
Variable type	Measure
Exposure	Clinician type
Outcome	Total ambulatory care visits (#)
	Preventive care visits (#)
	Acute care visits (ICD-9 or 10) (#)
	Chronic care visits (ICD-9 or 10) (#)
Regression analyses
Variable type	Measure
Outcome	Three-Category Clinician Type (1. PCP, 2. IM—Subspecialist, and 3. NP/PA/RN)
Predictor variables	Gender (female, male)
	Age (18-29, 30-39, 40-49, 50-59, 60-64, and ≥65 years)
	Race-Ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic Other, and Hispanic)
	Education (completion of high school/GED, <12 years and ≥12 years)
	Income (≤99% FPL, 100%-124% FPL, 125%-199% FPL, 200%-399% FPL, and ≥400% FPL or greater)
	Region (South, Northeast, Midwest, and West)
	Insurance coverage (private, public, and uninsured)
	Number of chronic conditions (0, 1, and 2 or more conditions)
	Survey Year

The preventive care measure was created by combining responses to questions related to routine check-ups, well-child exams, and immunizations or shots. We determined the acute and chronic care visits based on the ICD-9 or ICD-10 diagnosis codes available in the MEPS pooled data. Two physician researchers on the study team reviewed each diagnosis code to identify whether the visit was acute, chronic, or both acute and chronic. In the event of a disagreement, the researchers discussed the type of visit and came to a shared decision.

In the regression analysis (Table 1), the clinician type was the outcome variable. It had 3 categories (1. PCP, 2. IM subspecialist, and 3. NP/PA/RN). NP/PA/RNs here are not categorized as either IM subspecialists or PCP as their practice type is not specified in MEPS, despite the likelihood that some are likely practicing in primary care, some in subspecialty care and some in urgent or other forms of care. The explanatory variables included patient-level socio-demographics (age, gender, race-ethnicity, education, income-to-poverty ratio), census region of respondent’s residence, insurance coverage, number of chronic conditions, and survey year . We included age as a 6-category measure (18-29, 30-39, 40-49, 50-59, 60-64, and 65 or more) and binary variable for gender (male or female). We grouped respondents by race/ethnicity into 4 categories (non-Hispanic White, non-Hispanic Black, Hispanic, and non-Hispanic Other). There were 4 categories for income-to-poverty ratio—1. ≤99% Federal Poverty Level (FPL), 2. About 100% to 124% FPL, 3.125% to 199% FPL, 4. 200% to 399% FPL and 5. ≥400% FPL. We included 3 categories for respondents’ educational attainment (less than a high school diploma or GED, high school diploma, and more than high school education). Based on the respondents’ residence, we had 4 categories of census region (South, Northeast, Midwest, West); We created 3-category measures for insurance coverage (private, public, or uninsured) and the number of chronic conditions ((1) 0 condition, (2) 1 condition, (3) 2 or more conditions; and survey year. Based on the ICD-9 and ICD-10 codes, we identified the 15 most common chronic conditions found in the data consistent with the HHS chronic conditions working group.²⁸ The Institutional Review Board, American Academy of Family Physicians deemed the study exempt (IRB Application #: 24-494, Approval Date: 04/25/2024) because the study used publicly available deidentified MEPS data.

First, we calculated the total number and share of ambulatory care visits for all clinician types to examine the trends in the distribution of office-based visits by clinician type. We then computed the total number and percentage of preventive, acute, and chronic care visits by 3 clinician types only (1. PCP, 2. IM subspecialist, and 3. NP/PA/RN). We conducted bivariate analyses to evaluate the differences in the distribution of socio-demographics, insurance type, census region, number of chronic conditions, survey year, and health characteristics among adults by the 3 clinician types mentioned earlier. We assessed the statistically significant differences in patient sociodemographics between the 3 clinician types using Chi-squared tests.

We performed multinomial logistic regression models to investigate the relative risk ratio of respondent factors associated with visiting a PCP versus an IM subspecialist or PCP versus an NP/PA/RN. The unit of analysis was the ambulatory visit for all the analyses except for socio-demographic characteristics, where we used individual respondent-level data. We set the alpha level at P < .05. We used survey design and weight variables to account for the complex survey design used in MEPS.^25,29 We conducted all analyses using Stata 18.0.³⁰

Results

The pooled sample consisted of 2.2 million (unweighted) ambulatory visits from 2010 to 2021 (the weighted sample size ranged from 1.6 billion in 2010 to 2.6 billion in 2021 nationally).

Ambulatory Care Visits Patterns

The absolute number of ambulatory visits progressively increased from 2010 to 2019, decreased in 2020, and increased again in 2021. Specifically, the absolute number of PCP visits increased between 2010 and 2018 and then decreased from 2019 to 2021 (Figure 1, Table A1). While absolute number of ambulatory visits decreased for all other physician types in 2020 and increased in 2021, visits to PCPs continued to decline in 2021. The proportion of ambulatory visits to PCPs decreased by 43% compared to a decrease of 22% for IM subspecialists (Figure 1, Table A1). The proportion of visits provided by NP/PA/RNs increased by 98%. Notably, behavioral health clinicians also experienced a 241% rise in the percentage of outpatient visits from 2010 to 2021.

Figure 1.

Trends in percentage of ambulatory care visits by clinician type.

Primary Care Visit Patterns

The ambulatory care visits were subsequently analyzed by encounter type, including preventive care, acute care, and chronic care visits. There was an increase in the absolute number of preventive care visits. The percentage of preventive care visits as a share of total visits provided by PCPs, IM subspecialists, and NP/PA/RNs also increased by 25%, 7%, and 4%, respectively (Figure 2, Table A2).

Figure 2.

Trends in percentage of preventive care visits by clinician type.

From 2010 to 2021, total acute care visits for all clinician types increased by 17%. While PCP and IM subspecialists experienced a decrease in acute care visits (−30% and −38%, respectively), NP/PA/RN experienced a 35% increase (Figure 3, Table A3) during this period.

Figure 3.

Trends in acute care visits by clinician type.

Chronic care visits decreased 41% for all clinicians from 2010 to 2021. This translated to a 20% decrease in chronic care visits for PCPs, a 17% decrease for IM subspecialist physicians, and a 46% decrease for NP/PA/RNs (Figure 4, Table A4).

Figure 4.

Trends in chronic care visits by clinician type.

Associations Between Outpatient Provider and Patient Characteristics

Bivariate analysis

Of the 76 423 adults in the analysis sample who had at least 1 visit with either a PCP, IM subspecialist, or NP/PA/RN, 70% reported seeing a PCP at least once, 11% seeing an IM subspecialist at least once, and 19% seeing NP/PA/RN at least once (Table 2). In comparison to males, a higher proportion of females saw an NP/PA/RN than a PCP or IM-subspecialist (NP/PA/RN—56%, PCP—50%, IM-subspecialist—52%, P < .001). Adults under 40 years were more likely to receive care from an NP/PA/RN than a PCP or an IM subspecialist (40% vs 32% vs 16%). In contrast, adults 65 years or older were more likely to have a visit with an IM subspecialist (40%) than a PCP (23%) or NP/PA/RN (19%; P < .001). Among the three clinician types, the respondents with a post-high school education were more likely to receive care from an IM subspecialist (56% versus 46% seeing PCP and 52% seeing NP/PA/RN). Similarly, a higher percentage of adults with income greater than 400% FPL or more saw IM subspecialists (46% versus 38% PCPs and 42% NP/PA/RN, P < .001). In contrast, those receiving care from PCPs were less educated and poor (<200% FPL). Patients who reported no chronic conditions predominantly saw an NP/PA/RN or PCP rather than an IM subspecialist (66% vs 60% vs 42%, respectively). Patients who reported two or more chronic conditions increasingly saw an IM subspecialist or PCP than an NP/PA/RN (28% vs 21% vs 15%, respectively).

Table 2.

Distribution of Socio-Demographic and Health Characteristics by Clinician Type.

Characteristics	Clinician type			P-value
Characteristics	Primary care physician (n = 56 138 (70.1%))	IM subspecialist (n = 7748 (11.0%))	Nurse practitioner/physician assistant/registered nurse (n = 12 537 (18.9%))	P-value
Gender
Male	26 790 (49.9)	3612 (48.4)	5279 (43.8)	<.001
Female	29 348 (50.1)	4136 (51.6)	7258 (56.2)
Age in years
18-29	6523 (17.6)	469 (7.8)	2120 (22.5)	<.001
30-39	5853 (14.7)	581 (8.1)	1760 (17.7)
40-49	6567 (16.4)	875 (11.9)	1610 (15.6)
50-59	7581 (18.9)	1470 (21.0)	1756 (17.2)
60-64	3478 (9.1)	806 (11.5)	840 (7.8)
≥ 65	9547 (23.4)	2853 (39.7)	2156 (19.2)
Race-Ethnicity
NH, White	22 542 (58.5)	4291 (70.2)	7681 (72.0)	<.001
NH, Black	10 783 (13.3)	1429 (11.5)	1745 (9.5)
NH, Others	5704 (9.5)	640 (7.1)	1056 (7.7)
Hispanic	17 109 (18.7)	1388 (11.2)	2055 (10.9)
Educational attainment
Education 12 years	12 352 (22.1)	2237 (27.3)	3133 (23.2)	<.001
Education <12 years	25 282 (37.0)	1769 (17.5)	3603 (25.1)
Education >12 years	18 504 (40.8)	3742 (55.2)	5801 (51.7)
Income to Poverty Ratio
≤99 % FPL	12 121 (13.5)	1239 (10.3)	2054 (11.1)	<.001
100-124% FPL	3572 (4.7)	446 (4.4)	660 (4.1)
125-199% FPL	9299 (13.9)	1137 (12.3)	1755 (11.8)
200-399% FPL	16 015 (30.0)	2106 (27.4)	3815 (31.3)
≥400% FPL	15 131 (37.9)	2820 (45.8)	4253 (41.7)
Census region
South	8840 (17.6)	1490 (21.5)	1708 (15.2)	<.001
Northeast	10 573 (20.6)	1491 (19.9)	3100 (24.5)
Midwest	21 936 (39.3)	3193 (40.8)	4565 (37.4)
West	14 789 (22.5)	1574 (17.8)	3164 (22.9)
Insurance coverage
Private	30 563 (66.3)	4701 (69.1)	8013 (72.2)	<.001
Public	21 221 (28.0)	2779 (28.4)	3576 (21.7)
Uninsured	4354 (5.8)	268 (2.5)	948 (6.0)
Number of chronic conditions
Zero	23 007 (59.8)	2860 (41.5)	6541 (65.5)	<.001
One	7659 (18.9)	2145 (30.7)	2011 (19.2)
Two or more	8883 (21.3)	2049 (27.8)	1690 (15.3)
Survey year
2010	4225 (7.4)	518 (7.1)	536 (4.3)	<.001
2011	5375 (9.1)	652 (8.6)	623 (4.7)
2012	6185 (9.8)	693 (8.2)	735 (5.5)
2013	5260 (8.2)	618 (7.6)	856 (6.1)
2014	4981 (8.2)	570 (7.0)	941 (7.3)
2015	5158 (8.2)	592 (7.7)	990 (7.8)
2016	4430 (7.5)	520 (6.2)	857 (6.8)
2017	6013 (11.1)	909 (11.2)	1460 (11.4)
2018	5400 (10.6)	932 (12.5)	1494 (11.7)
2019	3549 (7.6)	602 (8.2)	1045 (8.4)
2020	2484 (5.8)	475 (7.2)	956 (8.8)
2021	3078 (6.5)	667 (8.5)	2044 (17.3)

Source: Authors Analyses of Medical Expenditure Panel Survey (2010-2021; N = 76 423); Unit of analysis: Individual Respondent Level.

Abbreviations: FPL, federal poverty level; IM, internal medicine; NH, non-Hispanic; NP/PA/RN, nurse practitioner/physician assistant/registered nurse; PCP, primary care physicians.

Primary care physicians include family physicians, general practitioners, general internal medicine, general pediatrics, and geriatrics. IM subspecialists include allergists or immunologists, cardiologists, endocrinologists, gastrologists, hematologists, nephrologists, oncologists, pulmonologists, and rheumatologists. NP/PA/RN includes nurse practitioners, and physician assistants or registered nurses.

Multinomial regression analysis

Factors associated with an IM subspecialist relative to a PCP visit

Compared to 18 to 29 years, adults 65 years and older had 2.2 times the relative risk of having an IM subspecialist visit than a PCP visit. (Table 3, Table A5) Relative to respondents with high school education, those with less than high school education were more likely to visit a PCP than an IM subspecialist (Relative Risk Ratio (RRR) 0.9, 95% Confidence Interval (CI) [0.8-1.0]). In contrast, those with post-high school education, as compared to those with high school education, were less likely to visit PCP (RRR = 1.1, 95% CI [1.0-1.2]). Patients with income of 400 percent of the FPL or more than those income less than 100% FPL had a lower likelihood of seeing a PCP versus an IM subspecialist (RRR = 1.2, 95% CI [1.1-1.3]). Relative to privately insured, uninsured were 50% more likely to have a PCP visit than an IM-subspecialist (RRR = 0.5, 95% CI [0.4-0.6]). Patients who self-report having no chronic condition than those who report 1 chronic condition had a higher propensity of obtaining care from a PCP than an IM subspecialist.

Table 3.

Multinomial Logistic Regression: Respondent Characteristics Associated with Having a Visit to a Given Clinician Type.

Characteristics	IM subspecialist	Nurse practitioner/physician assistant/registered nurse
	RRR (95% CI)	RRR (95% CI)
Gender
Male	1.00 (1.00-1.00)	1.00 (1.00-1.00)
Female	0.98 (0.93-1.04)	1.16** (1.06-1.26)
Age in years
18-29	1.00 (1.00-1.00)	1.00 (1.00-1.00)
30-39	1.00 (0.84-1.21)	0.89 (0.75-1.05)
40-49	1.26** (1.07-1.49)	0.69** (0.58-0.83)
50-59	1.62** (1.38-1.91)	0.61** (0.52-0.72)
60-64	1.84** (1.55-2.19)	0.62** (0.52-0.74)
≥65	2.20** (1.88-2.58)	0.54** (0.46-0.64)
Race-ethnicity
NH, White	1.00 (1.00-1.00)	1.00 (1.00-1.00)
NH, Black	1.08 (0.98-1.19)	0.58** (0.51-0.67)
NH, other	0.77** (0.68-0.87)	0.62** (0.52-0.73)
Hispanic	0.97 (0.86-1.08)	0.45** (0.39-0.53)
Educational attainment
Education 12 years	1.00 (1.00-1.00)	1.00 (1.00-1.00)
Education less than 12 years	0.88* (0.79-0.99)	0.90 (0.74-1.10)
Education greater than 12 years	1.09* (1.02-1.17)	1.10 (0.99-1.22)
Income-to-poverty ratio
≤99 % FPL	1.00 (1.00-1.00)	1.00 (1.00-1.00)
100-124% FPL	1.00 (0.88-1.14)	0.93 (0.77-1.13)
125-199% FPL	1.00 (0.89-1.12)	0.89 (0.74-1.07)
200-399% FPL	1.07 (0.96-1.18)	0.85* (0.73-1.00)
≥400% FPL	1.21** (1.08-1.34)	0.87 (0.73-1.03)
Census region
South	1.00 (1.00-1.00)	1.00 (1.00-1.00)
Northeast	0.79** (0.71-0.87)	1.13 (0.99-1.30)
Midwest	0.86** (0.78-0.95)	0.92 (0.81-1.05)
West	0.76** (0.68-0.85)	1.12 (0.96-1.31)
Insurance coverage
Private	1.00 (1.00-1.00)	1.00 (1.00-1.00)
Public	0.94 (0.87-1.01)	1.10 (0.99-1.23)
Uninsured	0.47** (0.38-0.57)	1.00 (0.81-1.23)
Number of chronic conditions
Zero	1.00 (1.00-1.00)	1.00 (1.00-1.00)
One	1.48** (1.39-1.58)	1.24** (1.14-1.36)
Two or more	1.00 (0.94-1.07)	0.85** (0.78-0.92)
Survey year
2010	1.00 (1.00-1.00)	1.00 (1.00-1.00)
2011	0.98 (0.87-1.09)	1.04 (0.90-1.20)
2012	1.00 (0.88-1.13)	1.08 (0.90-1.30)
2013	1.10 (0.97-1.24)	1.56** (1.31-1.87)
2014	1.09 (0.97-1.22)	2.04** (1.63-2.55)
2015	1.00 (0.88-1.15)	1.89** (1.56-2.30)
2016	0.96 (0.84-1.10)	1.91** (1.58-2.31)
2017	1.03 (0.91-1.18)	1.90** (1.61-2.23)
2018	1.15* (1.02-1.29)	2.15** (1.80-2.56)
2019	1.20** (1.06-1.37)	2.33** (1.99-2.72)
2020	1.31** (1.14-1.51)	2.99** (2.45-3.65)
2021	1.34** (1.18-1.52)	3.90** (3.34-4.55)
Constant	0.20** (0.16-0.24)	0.37** (0.29-0.47)

Authors Analyses of Medical Expenditure Panel Survey (2010-2021); Reference Category: PCP; Unit of Analysis: Visit; Dependent Variable: Three-Category Clinician Type (1. PCP, 2. IM Subspecialist, and 3. NP/PA/RN). Adjusted for gender, age, race-ethnicity, education, income-to-poverty ratio, census region, insurance coverage, number of chronic conditions, and survey year.

Abbreviations: CI, confidence interval; IM, internal medicine; NH, non-Hispanic; NP/PA/RN, nurse practitioner/physician assistant/registered nurse; PCP, primary care physicians; RRR, relative risk ratio.

P < .05. **P < .01.

Factors associated with an NP/PA/RN relative to a PCP visit

The relative risk of women compared to men visiting an NP/PA/RN versus visiting a PCP was 1.2 (RRR = 1.2, 95% CI [1.1-1.3]). Consistent with bivariate results, compared to respondents between 18 and 29 years, respondents in the age groups 40 to 49, 50 to 59, 60 to 64, and 65 and older were more likely to see a PCP than an NP/PA/RN. In comparison with non-Hispanic White adults, non-White adults (non-Hispanic Black (RRR = 0.6, 95% CI [0.5-0.7]; non-Hispanic Other (RRR = 0.6, 95% CI [0.5-0.7]; and Hispanic (RRR = 0.5, 95% CI [0.4-0.6]) were more likely to receive care from a PCP than an NP/PA/RN. Adults with 2 or more chronic conditions had higher likelihood of obtaining care from a PCP than an NP/PA/RN.

Discussion

General Findings

Primary care is the only specialty shown to extend life expectancy, yet the percentage of visits with primary care physicians continues to decrease between 2010 to 2021.³¹ Although the proportion of PCP visits decreased, overall, the proportion of patients seeing PCPs was the highest. Primary care physicians continue to lag in the percentage of chronic care and acute care visits as compared to care provided by subspecialists and NP/PA/RNs. These findings are not groundbreaking, but disheartening given the total investment in healthcare in the US and the need for improved individual and population health outcomes.³²

As the US population has grown older, patients are more likely to see subspecialists. Sicker patients are seen by primary care physicians or subspecialists. Acute care visits are being completed by NP/PA/RNs. Individually, these trends are not staggering. Put together, these trends demonstrate the significant fragmentation of care and lack of continuity.

Inequities also arose in our evaluation; individuals with higher income and with more education are more likely to visit a subspecialist. Less educated, uninsured and poorer individuals are more likely to receive care from a PCP. Our findings demonstrate that access to subspecialty care remains tiered, and not accessible to all individuals who may benefit from subspecialty care equitably.

An encouraging finding of this study is the increasing number of preventive visits provided by primary care physicians, which is consistent with previous studies.^7,8 Surprisingly, this trend persisted through the COVID-19 pandemic despite decreasing total visits to primary care physicians from 2020 to 2021. Given that primary care providers are uniquely responsible for providing full-spectrum chronic disease and cancer screening, increasing access to preventive services is essential to improving population health outcomes in the United States.

Workforce Trends and Continuity of Care

Between 2010 to 2020, there was a 7.4% in the total population in the United States.³³ This population growth alone should have increased the demand for primary care. Despite this, our study shows that PCPs provided 44% fewer visits between 2010 and 2021. Nurse practitioners, PAs, and advanced practice registered nurses have augmented the primary care and subspecialty workforce with an increasing number of program graduates in the last decade. As such, the number of visits to these clinicians should increase, consistent with our study’s findings.^8,34 The number of NP/PA/RN visits increased from 108 million in 2010 to 326 million in 2021. It was notable for a significant increase between 2020 and 2021, many of which were for COVID-19 vaccination visits. The proportion of NPs and PAs that enter primary care is unclear, but perhaps between 32% to 34% of NPs and 27% to 30% of PAs work in primary care.⁷ Studies comparing NP/PA/RN to physician-led visits have demonstrated similar quality of care metrics, such as chronic disease control measures, patient safety outcomes, and patient satisfaction scores.^35,36 However, NPs and PAs tend to see fewer and less medically complex patients than physicians. Our study confirms that individuals with more chronic comorbidities are seen in primary care rather than by an NP/PA/RN. As such, NPs and PAs do not provide a one-to-one replacement for physicians, underscoring the need for investment in a diverse clinical workforce.³⁷

Urgent care and retail settings have taken on a large proportion of visits that likely were previously seen in primary care. Patients may prefer to be seen in a retail clinic due to walk-in hours and shorter wait times than emergency or primary care. When used appropriately, such as for uncomplicated pharyngitis or urinary tract infections, retail clinics have been shown to provide fast, high-quality care at a lower total cost.³⁸ However, patients who visit retail clinics are less likely to have continuity of care with a provider.³⁹ While an external provider can appropriately manage simple, acute diagnoses, the urgent care or retail clinic visit may become a missed opportunity to connect with the primary care physician, discuss chronic medical conditions, ensure medications are adequately managed, and receive preventive screening.

Policy Implications

The drivers of change in visit patterns cannot be elucidated by the proportion of visits to different clinicians alone; there are likely several factors that change a patient’s interaction with the health system. Alternative methods of communication, such as patient-provider messaging on the electronic medical record, the accessibility of artificial intelligence and online resources, and a shift toward less trust in the health system may also play a role. Based on the data available, however, policies can strengthen primary care’s role in the health system to improve patient outcomes.

In order to increase ambulatory visits, an increase in the number of PCPs is needed. Congress took steps in 2021 to increase graduate medical education and training, with specific focus on underserved populations. However, a more targeted approach to increase primary care training slots would benefit public access.⁴⁰ Secondly, reimbursement must change to meet the demands of primary care. Cognitive disciplines like primary care receive lower reimbursement than procedural disciplines.⁴¹ With more equitable reimbursement, PCPs will be able to focus on long-term continuity, rather than quick, production-based visits.

Limitations

This study has important limitations that need to be addressed. First, the MEPS data is self-reported, which can introduce recall and reporting bias. To address this, MEPS collects data over multiple rounds across a 2-year period, reducing recall time and enhancing reporting accuracy.²⁴ Second, the data does not specify whether care was received from Registered Nurses or Advanced Practice Providers, such as Nurse Practitioners or PAs, and lacks details on their specialties. As cited in the methods, the total ambulatory, preventive, acute, and chronic care provided by NPs, PAs, and RNs may be lower than reported with only a proportion of those working in primary care. Vaccine visits are included in the data. Finally, the COVID-19 pandemic impacted the MEPS data collection efforts by switching from in-person to telephone interviews.²⁵ To address low response rates in 2020 and 2021, MEPS extended data collection to 9 rounds.^25,27 While the AHRQ recommends caution when interpreting trends from these years, their researchers found little evidence of bias.^25,27 Despite these limitations, MEPS remains a nationally representative survey of the civilian, non-institutionalized U.S. population, using a rigorous methodology using a rigorous methodology that provides valuable insights into healthcare utilization and costs.

Conclusion

Despite a growing need for primary care services, decreasing visits to primary care physicians (PCPs) is concerning and requires further examination. The declining trends in acute and chronic care visits raise questions as to whether primary care, in its current form, can continue to provide its essential attributes and services.

Outpatient clinicians, including primary care physicians, subspecialists, and NP/PA/RNs, each have a role in providing care for patients in a collaborative manner instead of as a disconnected and inefficient replacement for each other. This quality health care begins with comprehensive and coordinated primary care. It is essential to provide enhanced support for primary care to strengthen the workforce and retain clinicians in ambulatory care settings to promote individual and population health goals.

Footnotes

Appendix

Table A5.

Multinomial Logistic Regression: Respondent Characteristics Associated with Having a Visit to a Given Clinician Type.^a

	IM subspecialist		Nurse practitioner/physician assistant/registered nurse		IM subspecialist	Nurse Practitioner/physician assistant/registered nurse
Characteristics	Coefficient (std. error)	P-value	Coefficient (std. error)	P-value	RRR (95% CI)	RRR (95% CI)
Gender
Male
Female	−0.02 (0.03)	.568	0.14 (0.04)	.001	0.98 (0.93-1.04)	1.16 (1.06-1.26)**
Age in years
18-29
30-39	0.00 (0.09)	.961	−0.12 (0.09)	.165	1.00 (0.84-1.21)	0.89 (0.75-1.05)
40-49	0.23 (0.08)	.006	−0.37 (0.09)	.000	1.26 (1.07-1.49)**	0.69 (0.58-0.83)**
50-59	0.48 (0.08)	.000	−0.49 (0.09)	.000	1.62 (1.38-1.91)**	0.61 (0.52-0.72)**
60-64	0.61 (0.09)	.000	−0.48 (0.09)	.000	1.84 (1.55-2.19)**	0.62 (0.52-0.74)**
≥ 65	0.79 (0.08)	.000	−0.62 (0.09)	.000	2.2 (1.88-2.58)**	0.54 (0.46-0.64)**
Race-ethnicity
NH, White
NH, Black	0.07 (0.05)	.133	−0.54 (0.07)	.000	1.08 (0.98-1.19)	0.58 (0.51-0.67)**
NH, Other	−0.27 (0.06)	.000	−0.48 (0.09)	.000	0.77 (0.68-0.87)**	0.62 (0.52-0.73)**
Hispanic	−0.03 (0.06)	.571	−0.79 (0.08)	.000	0.97 (0.86-1.08)	0.45 (0.39-0.53)**
Educational attainment
Education 12 years
Education <12 years	−0.12 (0.06)	.027	−0.10 (0.10)	.300	0.88 (0.79-0.99)*	0.90 (0.74-1.10)
Education ≥12 years	0.09 (0.04)	.012	0.09 (0.05)	.079	1.09 (1.02-1.17)**	1.10 (0.99-1.22)
Income to poverty ratio
≤99 % FPL
100-124% FPL	0.00 (0.07)	.983	−0.07 (0.10)	.485	1.00 (0.88-1.14)	0.93 (0.77-1.13)
125-199% FPL	−0.00 (0.06)	.960	−0.12 (0.10)	.205	1.00 (0.89-1.12)	0.89 (0.74-1.07)
200-399% FPL	0.07 (0.05)	.208	−0.16 (0.08)	.045	1.07 (0.96-1.18)	0.85 (0.73-1.00)*
≥400% FPL	0.19 (0.05)	.001	−0.14 (0.09)	.107	1.21 (1.08-1.34)**	0.87 (0.73-1.03)
Census region
South
Northeast	−0.24 (0.05)	.000	0.13 (0.07)	.070	0.79 (0.71-0.87)**	1.13 (0.99-1.30)
Midwest	−0.15 (0.05)	.002	−0.08 (0.07)	.209	0.86 (0.78-0.95)**	0.92 (0.81-1.05)
West	−0.27 (0.05)	.000	0.11 (0.08)	.153	0.76 (0.68-0.85)**	1.12 (0.96-1.31)
Insurance coverage
Private
Public	−0.06 (0.04)	.100	0.09 (0.06)	.089	0.94 (0.87-1.01)	1.10 (0.99-1.23)
Uninsured	−0.77 (0.10)	.000	0.00 (0.11)	.985	0.47 (0.38-0.57)**	1.00 (0.81-1.23)
Number of chronic conditions
Zero
One	0.39 (0.03)	.000	0.22 (0.04)	.000	1.48 (1.39-1.58)**	1.24 (1.14-1.36)**
Two or more	0.00 (0.03)	.918	−0.17 (0.04)	.000	1.00 (0.94-1.07)	0.85 (0.78-0.92)**
Survey year
2010
2011	−0.02 (0.06)	.665	0.04 (0.07)	.626	0.98 (0.87-1.09)	1.04 (0.90-1.20)
2012	−0.00 (0.06)	.981	0.08 (0.09)	.395	1.00 (0.88-1.13)	1.08 (0.90-1.30)
2013	0.09 (0.06)	.137	0.45 (0.09)	.000	1.10 (0.97-1.24)	1.56 (1.31-1.87)**
2014	0.08 (0.06)	.154	0.71 (0.11)	.000	1.09 (0.97-1.22)	2.04 (1.63-2.55)**
2015	0.00 (0.07)	.947	0.64 (0.10)	.000	1.00 (0.88-1.15)	1.89 (1.56-2.30)**
2016	−0.04 (0.07)	.553	0.65 (0.10)	.000	0.96 (0.84-1.10)	1.91 (1.58-2.31)**
2017	0.03 (0.06)	.597	0.64 (0.08)	.000	1.03 (0.91-1.18)	1.9 (1.61-2.23)**
2018	0.14 (0.06)	.025	0.76 (0.09)	.000	1.15 (1.02-1.29)*	2.15 (1.80-2.56)**
2019	0.18 (0.07)	.005	0.84 (0.08)	.000	1.2 (1.06-1.37)**	2.33 (1.99-2.72)**
2020	0.27 (0.07)	.000	1.1 (0.10)	.000	1.31 (1.14-1.51)**	2.99 (2.45-3.65)**
2021	0.29 (0.06)	.000	1.36 (0.08)	.000	1.34 (1.18-1.52)**	3.9 (3.34-4.55)**
Constant	−1.63 (0.11)	.000	−1 (0.13)	.000	0.2 (0.16-0.24)*	0.37 (0.29-0.47)**
Observations	1 036 358		1 036 358		1 036 358	1 036 358

Source: Authors Analyses of Medical Expenditure Panel Survey (2010-2021); Reference Category—PCP; Unit of Analysis—Visit; Dependent Variable -Three-Category Clinician Type (1. PCP, 2. IM Subspecialist, and 3. NP/PA/RN).

Includes both Coefficients (std. error) and RRR (95% CI).

Abbreviations: CI, confidence interval; FPL, federal poverty level; IM, internal medicine; NH, non-Hispanic; NP/PA/RN, nurse practitioner/physician assistant/registered nurse; PCP, primary care physicians; std, standard; RRR, relative risk ratio.

P < .05. **P < .01.

Authors’ Note

Preliminary findings were presented at the 2023 NAPCRG Conference, San Francisco, and, the conference abstract was published online in Annals of Family Medicine, 2023.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Anuradha Jetty

Supplemental Material

The Medical Expenditure Panel Survey data used in the current study are publicly available and downloadable from the Agency for Healthcare Research and Quality at .

References

Macinko

Starfield

Shi

Quantifying the health benefits of primary care physician supply in the United States. Int J Health Serv. 2007;37(1):111-126.

Shi

Starfield

The effect of primary care physician supply and income inequality on mortality among blacks and whites in US metropolitan areas. Am J Public Health. 2001;91(8):1246-1250.

Starfield

Shi

Grover

Macinko

. The effects of specialist supply on populations’ health: assessing the evidence: the evidence suggests that populations do not necessarily benefit from an overabundance of specialists in a geographic area. Health Affairs. 2005;24(Suppl1):W5-W97.

Macinko

Starfield

Shi

The contribution of primary care systems to health outcomes within Organization for Economic Cooperation and Development (OECD) countries, 1970–1998. Health Serv Res. 2003;38(3):831-865.

Starfield

Shi

Macinko

Contribution of primary care to health systems and health. Milbank Q. 2005;83(3):457-502.

Shi

Starfield

Politzer

Regan

Primary care, self-rated health, and reductions in social disparities in health. Health Serv Res. 2002;37(3):529-550.

Bazemore

Merenstein

Handler

, et al. The impact of interpersonal continuity of primary care on health care costs and use: a critical review. Ann Fam Med 2023;21(3):274-279.

Jabbarpour

Jetty

Byun

, et al. No One Can See You Now: Five Reasons Why Access to Primary Care Is Getting Worse (and What Needs to Change). Robert Graham Center, Milbank Memorial Fund, The Physicians Foundation. 2024. Accessed July 28, 2024. https://www.aafp.org/dam/rgc/documents/publications-reports/reports/2024-scorecard-final-report.pdf

Abraham

Zheng

Poghosyan

Predictors and outcomes of burnout among primary care providers in the United States: a systematic review. Med Care Res Rev. 2020;77(5):387-401. doi:10.1177/1077558719888427

10.

Bian

Carretta

Lee

Hogan

Shenkman

Charness

Prevalence of multiple chronic conditions among older adults in Florida and the United States: comparative analysis of the OneFlorida data trust and national inpatient sample. J Med Internet Res. 2018;20(4):e8961.

11.

Mental Health America. Prevalence of Mental Illness. 2024. Accessed July 1, 2024. https://mhanational.org/issues/2024/mental-health-america-prevalence-data

12.

The National Institute for Health Care Management (NIHCM) Foundation. Mental Health: Trends & Future Outlook. 2019. Accessed April 4, 2024.https://nihcm.org/publications/mental-health-trends-future-outlook

13.

Johansen

Richardson

CR.

The ecology of medical care before and after the affordable care act: trends from 2002 to 2016. Ann Fam Med. 2019;17(6):526-537. doi:10.1370/afm.2462

14.

Ganguli

Shi

Orav

Rao

Ray

Mehrotra

Declining use of primary care among commercially insured adults in the United States, 2008-2016. Ann Intern Med. 2020;172(4):240-247. doi:10.7326/M19-1834

15.

Ray

Shi

Ganguli

Rao

Orav

Mehrotra

Trends in pediatric primary care visits among commercially insured US children, 2008-2016. JAMA Pediatr. 2020;174(4):350-357. doi:10.1001/jamapediatrics.2019.5509

16.

Frost

Hargraves

HCCI Brief: Trends in Primary Care Visits. Health Care Cost Institute. 2018. Accessed April 4, 2024. https://healthcostinstitute.org/hcci-originals-dropdown/all-hcci-reports/trends-in-primary-care-visits

17.

Ganguli

Lee

Mehrotra

Evidence and implications behind a national decline in primary care visits. J Gen Intern Med. 2019;34(10):2260-2263. doi:10.1007/s11606-019-05104-5

18.

Assistant Secretary for Planning and Evaluation. National Uninsured Rate Reaches an All-time Low in Early 2023 After the Close of the ACA Open Enrollment Period. U.S. Department of Health and Human Services. 2023. Accessed April 15, 2024. https://aspe.hhs.gov/sites/default/files/documents/e06a66dfc6f62afc8bb809038dfaebe4/Uninsured-Record-Low-Q12023.pdf

19.

Lee

Monuteaux

Galbraith

AA.

Women’s affordability, access, and preventive care after the affordable care act. Am J Prev Med. 2019;56(5):631-638. doi:10.1016/j.amepre.2018.11.028

20.

Seiler

Malcarney

M-B

Horton

Dafflitto

Coverage of clinical preventive services under the Affordable Care Act: from law to access. Public Health Rep. 2014;129(6):526-532. doi:10.1177/003335491412900611

21.

Petterson

Liaw

Tran

Bazemore

AW.

Estimating the residency expansion required to avoid projected primary care physician shortages by 2035. Ann Fam Med. 2015;13(2):107-114. doi:10.1370/afm.1760

22.

Survey Background. Medical Expenditure Panel Survey. 2019. Accessed July 15, 2024. https://meps.ahrq.gov/mepsweb/about_meps/survey_back.jsp

23.

Agency for Healthcare Research and Quality. MEPS HC-224 2020 Full Year Consolidated Data File. 2022. Accessed July 14, 2024. https://meps.ahrq.gov/data_stats/download_data/pufs/h224/h224doc.pdf

24.

Chowdhury

Machlin

Gwet

KL.

Sample Designs of the Medical Expenditure Panel Survey Household Component, 1996–2006 and 2007–2016. Methodology Report #33. 2019. Accessed July 14, 2024. https://meps.ahrq.gov/data_files/publications/mr33/mr33.shtml

25.

Agency for Healthcare Research and Quality. MEPS HC-224 2020 Full Year Consolidated Data File. 2022. Accessed July 14, 2024. https://meps.ahrq.gov/data_stats/download_data/pufs/h224/h224doc.pdf

26.

Agency for Healthcare Research and Quality. MEPS-MPC Sample Design and Data Collection Process. Agency for Healthcare Research and Quality. 2004. Accessed July 14, 2024. http://www.meps.ahrq.gov/survey_comp/mpc_ data_collection.jsp

27.

Zuvekas

Kashihara

The impacts of the COVID-19 pandemic on the medical expenditure panel survey. Am J Public Health. 2021;111(12):2157-2166. doi:10.2105/AJPH.2021.306534

28.

Goodman

Posner

Huang

Parekh

Koh

HK.

Defining and measuring chronic conditions: imperatives for research, policy, program, and practice. Prev Chronic Dis. 2013;10:E66. doi:10.5888/pcd10.120239

29.

Medical Expenditure Panel Survey. Using Statistical Software Packages to Produce Estimates from MEPS Data File. 2021. Accessed July 13, 2024. https://meps.ahrq.gov//survey_comp/hc_samplecodes_se.pdf

30.

StataCorp. 2019. Stata Statistical Software: Release 16. StataCorp LLC.

31.

Basu

Phillips

Berkowitz

Landon

Bitton

Phillips

RL.

Estimated effect on life expectancy of alleviating primary care shortages in the United States. Ann Intern Med. 2021;174(7):920-926. doi:10.7326/M20-7381

32.

Bazemore

Petterson

McCulloch

KK.

US primary care workforce growth: a decade of limited progress, and projected needs through 2040. J Gen Intern Med. 2024;40:339-346. doi:10.1007/s11606-024-09121-x

33.

Ogunwole

Rabe

Roberts

Caplan

. Population Under Age 18 Declined Last Decade. US Census Bureau; 2021:12.

34.

Auerbach

Staiger

Buerhaus

PI.

Growing ranks of advanced practice clinicians - implications for the physician workforce. N Engl J Med. 2018;378(25):2358-2360. doi:10.1056/NEJMp1801869

35.

Barnett

Balkissoon

Sandhu

The level of quality care nurse practitioners provide compared with their physician colleagues in the primary care setting: a systematic review. J Am Assoc Nurse Pract. 2022;34(3):457-464. doi:10.1097/JXX.0000000000000660

36.

Stanik-Hutt

Newhouse

White

, et al. The quality and effectiveness of care provided by nurse practitioners. J Nurse Pract. 2013;9(8):492-500.e13. doi:10.1016/j.nurpra.2013.07.004

37.

Rotenstein

Apathy

Edgman-Levitan

Landon

Comparison of work patterns between physicians and advanced practice practitioners in primary care and specialty practice settings. JAMA Netw open. 2023;6(6):e2318061. doi:10.1001/jamanetworkopen.2023.18061

38.

Mehrotra

Liu

Adams

, et al. Comparing costs and quality of care at retail clinics with that of other medical settings for 3 common illnesses. Ann Intern Med. 2009;151(5):321-328. doi:10.7326/0003-4819-151-5-200909010-00006

39.

Reid

Ashwood

Friedberg

Weber

Setodji

Mehrotra

Retail clinic visits and receipt of primary care. J Gen Intern Med. 2013;28(4):504-512. doi:10.1007/s11606-012-2243-x

40.

Rains

Holmes

Pathak

Hawes

EM.

The distribution of additional residency slots to rural and underserved areas. JAMA. 2023;330(10):968-969.

41.

Sinsky

Dugdale

DC.

Medicare payment for cognitive vs procedural care: minding the gap. JAMA Intern Med. 2013;173(18):1733-1737. doi:10.1001/jamainternmed.2013.9257

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB

0.03 MB