Abstract
Background: There is a growing primary care workforce shortage. Primary care physicians (PCPs) already have a broad scope of work yet face increasing demands on their time and pressure to see more patients. This analysis aims to update PCP workforce shortage estimates by assessing the number of unique patients seen per year by each PCP in Virginia.
Methods: From Virginia’s all-payer claims database (APCD), we identified the unique number of patients seen annually between 2016 to 2021. We evaluated variation by physicians, practices, and communities by using licensure, claims, census, and Virginia Department of Health clinician survey data for PCPs across family medicine, internal medicine, pediatrics, and ob-gyn. We conducted iterative bootstrap simulation analyses to estimate the PCP workforce shortage, accounting for panel size and Full-Time Equivalent (FTE).
Results: There are 4,508 PCPs in Virginia, with 34.2% of PCPs over age 60 compared with 24.0% aged 30 to 49. The median number of unique patients seen by physicians in 2021 was 1,290. On average, physicians had 2,800 total visits annually, and patients were seen 2.12 times per year. PCPs saw more patients if they were aged 40 to 59, worked in health systems, practiced in rural communities, or served communities with lower social needs. Average FTE was 0.847 (SD: 0.215). In 2021, 1,305 additional PCPs were needed for each Virginian to have a PCP, preventing a 28.8% shortfall (95% CI: 27.9%–30.0%). This marks an increase from a 17.7% shortfall in 2016.
Conclusion: Primary workforce shortages are more dire than prior estimates when accounting for number of patients seen per year and FTE. The ideal number of patients a PCP can care for should be determined by clinicians and practices, with consideration of patient complexity, physician capacity, and practice resources.
- Access to Primary Care
- Family Medicine
- Health Policy
- Health Services Accessibility
- Health Workforce
- Practice-Based Research
- Primary Health Care
- Virginia
Introduction
There is a critical shortage of primary care physicians (PCPs) in the United States, contributing to reduced access to primary care and delays in care.1,2 Barriers to accessing primary care can worsen chronic disease, increase hospitalizations, and decrease life expectancy.1–4
Efforts to broaden access to primary care can lead to pressure for PCPs to see more patients.5,6 However, to provide sustainable levels of high-quality primary care, the number of patients that a physician cares for must reflect the capacity and resources of each practice and the unique needs of patients.7 Panel size, defined as the number of patients each clinician is responsible for, is a critical factor in understanding workforce shortages.8 This analysis focused specifically on the number of unique patients seen by each primary care physician annually in Virginia, a factor useful for informing panel size, and its implications for workforce shortages.
Primary care includes family medicine, internal medicine, pediatrics, internal medicine/pediatrics (med/peds), and obstetrics and gynecology (Ob-Gyn). PCPs are being asked to take on more responsibilities for each patient than ever before, often while seeing more patients in less time. PCPs manage chronic diseases for patients facing complex and layered medical issues and social drivers of poor health,9–11 while also addressing preventive care, behavioral health, and social needs.12–14 In addition to the broad scope of care they provide, PCPs are increasingly expected to deliver care beyond traditional office visits. Common tasks include responding to patient questions, coordinating care with specialists and pharmacies, responding to insurance prior authorizations, documenting care more extensively, and meeting performance metrics.15,16 To address all these new tasks, PCPs may have to reduce their panel sizes. While seeing more patients can create access for new patients, if a PCP’s panel size becomes too large, existing patients may not be able to schedule an appointment, and physicians may be pressured to see more people in less time, undermining high-quality, whole-person care.17,18 Accordingly, these rising demands and pressures on PCPs have corresponded with a decrease in panel size by 25% over the past decade.19
Despite the importance of understanding the impact of panel size on access to and quality of care, there is a paucity of literature on the variation of PCP panel size by practice, system, and community characteristics.19–21 Fewer studies use panel size to understand the primary care workforce shortage.21 Widely used panel size estimates rely on outdated assumptions rather than empirical evidence, leading to inaccurate projected workforce need estimations. For example, an often-cited ratio of 2,500 patients to 1 clinician is derived from an article that merely proposed this number as a feasible goal—it is not empirically-based or even based on lived experience.22,23 Studies using actual visits or physician-reported panel sizes suggest the real PCP panel size likely ranges between 1,200 and 1,900 patients.23,24
This study aims to better understand variations in the number of unique patients seen annually by PCPs in Virginia and to assess the potential implications on primary care workforce shortages.
Methods
We conducted a retrospective secondary data analysis using licensure, claims, census, and Virginia Department of Health (VDH) data. We calculated the number of unique patients seen annually and examined variation by physician, practice, and community. We then completed a novel iterative bootstrap simulation analysis to estimate the primary care workforce shortage based on these findings. The Virginia Commonwealth University Institutional Review Board deemed our study as not human subjects research.
Participants and Setting
Our sample included all PCPs and all census tracts in Virginia between 2016 to 2021. We did not include Advanced Practice Providers (APPs), such as Nurse Practitioners (NPs) or Physician Assistants (PAs) (implications addressed in the discussion).
Data Sources
Our approach to identifying primary care physicians and practices has been described in prior publications.25–27 We broadly include all potential PCPs (family medicine, internal medicine, pediatrics, and ob-gyn); and we include most potential primary care practice types (private, hospital-owned, insurance-owned, direct primary care, urgent care).25,26 From the all-payer claims database (APCD), we identify active physicians who have at least 100 claims in a year. The APCD reports claims on 6.2 million of Virginia’s 8.8 million residents, including all Medicare, Medicaid, and most commercial claims data.28 It does not include patients who are uninsured, military, federal employees, or from Employee Retirement Income Security Act (ERISA) covered plans.28 We update our PCP list annually using the National Plan and Provider Enumeration System (NPPES), Virginia Department of Health Professions (VDHP) licensure data, APCD, manual reviews, and surveys. We triangulate addresses in the data sources with manual review to nest each PCP into a practice, health system, group, and/or Accountable Care Organization.25
We used 2023 VDH survey data for self-reported full-time equivalent (FTE) for PCPs, 2024 VDHP licensure dataset for gender and year of medical school graduation, and 2023 NPPES for National Provider Identifier, enumeration dates, and specialty types. We linked datasets by National Provider Identifier (NPI) and physician licensure numbers. We used the 2021 American Community Survey (ACS) and HealthLandscape Virginia for data on community characteristics and demographics to estimate physician shortages.29,30
Outcomes
Measurements included the number of unique patients seen annually by each PCP between 2016 and 2021, and the average number of visits per patient annually. Practice and physician-level characteristics included ownership type, physician-reported gender, approximate age, and specialty. We approximated age from VDHP licensure data on graduation year, and subsequently calculated the percentage of physicians over the age of 60 years as a proxy for physicians likely to retire or reduce their panel size in the subsequent 5 years.31–33
Community characteristics included the rurality of the census tract, calculated using the National Center for Education Statistics locale classification codes,34 and Social Deprivation Index (SDI), a composite measure of area-level deprivation based on data from the ACS.35 FTE was obtained from self-reported VDH physician survey data on the approximate number of physician hours (of 40) spent on direct patient care. As detailed below, our team calculated potential workforce shortages and the number of physicians needed to address the shortage based on the number of unique patients seen annually versus the number of residents in each community.
Statistical Analysis
To account for missing commercial insurance claims, we used an adjustment factor based on the percentage of Virginia’s population covered through employer and nongroup markets when calculating the number of unique patients seen per year.36 We calculated differences in mean and median unique number of patients seen annually based on physician and practice-level characteristics. We evaluated the average annual number of total patient visits per physician per year, as well as the average number of visits per patient per year. We assessed variations in total visits and average number of visits per patient, based on the number of unique patients seen annually. Comparisons of annual number of unique patients seen were made between levels of various physician- and practice-level characteristics using Kruskal-Wallis tests, employing a 5% significance level to detect differences.
We excluded outlier physicians from this analysis, including 19 physicians seeing greater than 7,500 patients and 442 seeing fewer than 100 patients in a year. Physicians seeing greater than 7,500 patients saw a median of 10,780 unique patients annually and were largely older physicians in health systems, who had lower-than-average FTE. We anticipate that these are senior physicians under whom other clinicians (eg, residents, APPs) are being billed. Physicians seeing fewer than 100 patients had a median of 16 unique patients seen annually and had similar characteristics to the study sample. Characteristics of outlier physicians are reported in Appendix 1.
To estimate the number of additional PCPs needed to ensure every Virginia resident has access to a PCP, we used an iterative bootstrap approach. We assumed that future physicians would have a similar distribution of patient volume and FTEs as current physicians. To simulate the number of unique patients seen by these additional physicians, we sampled with replacement from the existing PCP population until the total number of patients covered, across both current and simulated physicians, exceeded Virginia’s actual population, representing full PCP coverage. We incorporated population growth over time into our analysis. We repeated this process 10,000 times, and for each iteration, we determined the number of additional PCPs needed. We then calculated the mean and 95% confidence interval for the simulations of additional PCPs needed to achieve statewide coverage.
The R statistical software version 4.2.0 was used for all analyses.
Results
There were 4,508 PCPs in Virginia in 2021. The primary care workforce in Virginia has a high proportion of older physicians, with 1,541 (34.2%) PCPs over the age of 60 years versus 1,084 (24.0%) PCPs between the ages of 30 and 49 years, as shown in Figure 1.
2024 age distribution of active primary care physicians in Virginia.
There was broad variation in the number of patients seen by physicians, with a right skew to the distribution due to a small number of physicians seeing 2,500 to 7,500 unique patients per year (online Appendix 2, histogram of number of patients seen by physicians in 2021). After adjusting for the rate of missing commercial claims, we estimate the median number of patients seen by a PCP annually ranged between 1,196 and 1,290 patients between 2016 to 2021 (Table 1). Median number of unique patients seen by PCPs decreased between 2016 to 2020 (−90.0, P < .001) but then increased in 2021 to 1,290, such that there was no significant difference in unique patients seen between 2016 and 2021 (−4.0, P = .583).
Unique Number of Patients Seen by Physician by Year (2016–2021), by Mean, Median, and Quartile
The number of unique patients seen per year varied based on physician and practice characteristics (Table 2). PCPs within health systems saw more patients compared with others (1,347 vs 1,257, P < .001). Rural PCPs saw more patients than their urban counterparts (1,312 vs 1,251, P = .034). There were variations in patients seen by specialty (Internal Medicine = 1,127, Ob-gyn = 1,305, Family Medicine = 1,280, and Pediatrics = 1,592, P < .001). PCPs practicing in communities with a high SDI (higher levels of deprivation) saw fewer patients than those in communities with a low SDI (1,214 vs 1,365, P < .001). Median number of patients seen was larger for male versus female physicians (1,367 vs 1,241, P < .001) and varied by age (age <40 = 1,224, 40 to 49 = 1,262, 50 to 59 = 1,456, >60 = 1,269, P < .001).
Number of Unique Patients Seen per Year by Primary Care Physicians Based on Practice and Community-Level Characteristics (2021)
The average number of total patient visits per year per physician was 2,800, and varied by number of unique patients a physician saw per year, with an average of 435 total visits by physicians with a low numbers of unique patients seen annually (<500 patients), and an average of 7,176 total visits by physicians with higher numbers of unique patients annually (>2,500 patients). Patients had an average of 2.12 visits per year to their PCP. Patients seeing physicians with fewer patients (<500) had more visits per year than those seeing physicians with more patients (>2,500) (2.07 vs 1.93, P < .001).
The mean self-reported physician FTE was 0.847 (SD 0.215) (Table 3). FTE did not differ significantly based on ownership, specialty, or age. Mean FTE was lower for female versus male physicians (0.830 vs 0.868, P < .001), rural versus urban physicians (0.818 vs 0.853, P < .001), and physicians practicing in high SDI versus low SDI communities (0.810 vs 0.878, P < .001).
Primary Care Physician Full Time Equivalent (FTE) Based on Practice and Community-Level Characteristics (2021)
Using the bootstrap simulation approach, we estimated that 1,305 more PCPs were needed for each Virginia resident to have a PCP in 2021 (see online Appendix 3 for results by year), up from 845 more PCPs needed in 2016. Therefore, a 28.9% increase in the PCP workforce was needed in 2021 for each Virginia resident to have a PCP, higher than the estimated 17.7% increase in the workforce needed in 2016.
Discussion
Understanding how many patients a physician can see in a year is critical for ensuring the delivery of high-quality primary care and understanding workforce shortages. Our data show that the median number of unique patients seen annually by a PCP in Virginia is 1,290, with substantial variation based on physician, practice, and patient population characteristics. This is much lower than traditional estimates of the PCP panel size, which means that the primary care workforce shortage is more dire than projected.
Ensuring PCPs are responsible for the right number of patients is crucial to promoting access to care while safeguarding the delivery of the core functions of primary care known as the 4Cs: first contact, continuity, comprehensiveness, and coordination.37 Seeing too few patients will limit initial access for new patients,38–40 while seeing too many will limit access for existing patients, decrease continuity, and compromise a physician’s ability to provide comprehensive care and coordination–issues that are increasingly reported by patients.38–40
Not every patient is seen by their PCP every year, and PCPs remain responsible for patients outside of office visits. The full collection of patients a PCP is responsible for is considered their true patient panel. Patients may be seen less often because they are younger or healthier and do not have needs, but also because of systematic barriers preventing access to care (eg, lack of insurance, no appointment availability, no transportation, work/life conflicts). In general, patients with chronic conditions and in need of prevention should be seen at least annually,40,41 as managing patients outside of visits represents true work for physicians.
Our measure of unique patients seen annually could underestimate PCPs’ true panel size. Others have used unique patients seen over 2 or 3 years to represent patient panels.41 However, doing this can overestimate panel size as patients seen more remotely may have moved, transitioned to another PCP, or died. Conversely, including urgent care and same-day physicians in our measure could inflate our estimates of panel size, as visits are often shorter and focused on one acute need that is less complex than typical comprehensive primary care visits. In addition, our sample likely included physicians with other clinicians billing under their NPI (eg, APPs or residents), which could account for the right-skewed distribution of the number of unique patients seen annually (see Appendix 2) and may further inflate our panel size estimate.
We found that PCPs saw decreasing numbers of patients from 2016 to 2020. There was an increase in 2021, possibly related to a postpandemic surge in accessing delayed care, but collectively, PCPs are seeing fewer patients than in decades prior.19 Decreasing panel sizes may reflect rising demand for asynchronous care outside of visits, administrative burdens, increasing nonclinical workload, and burdens from greater numbers of prior authorizations.17,19 This is likely compounded by an increase in patient complexity and a broad scope of practice as primary care strives to provide continuous and comprehensive care.17,19 As a result, burnout from these factors has increased in recent years.42,43 PCPs may try to mitigate burnout by seeing fewer patients, spending fewer hours in direct patient care, or even leaving the workforce. Collectively, these factors are likely driving reductions in panel size and increases in the PCP workforce shortage.25,44
A report by the American Association of Medical Colleges (AAMC) estimated a shortage of 20,800 PCPs nationally in 2021.45 This is approximately a 7% PCP workforce shortfall, which is substantially lower than what we predict using our bootstrap simulation and panel size assumptions. The AAMC calculations are based on the number of PCPs required to remove a Health Professional Shortage Area (HPSA) designation, which assumes 1 physician per 3,000 residents and assumes each physician is 1.0 FTE.46,47 A panel size of 3,000 is higher than most PCPs can appropriately care for, and our use of unique patients seen annually may reflect a more realistic PCP panel size. For comparison, using Virginia’s HPSA designations, we would only need an additional 210 PCPs,48 while our panel size and boot strap simulation estimate is 6 times larger (n = 1,305).49
Assuming an average reported FTE of 0.847 among Virginia PCPs and a median panel size of 1,290 patients, a 1.0 FTE PCP would be estimated to have a panel size of approximately 1,523 patients. While this is a tempting assumption, FTE data from VDH is limited as it only reports hours of direct patient care and assumes a work week of 40 hours. This likely underestimates the true work done by PCPs. Other studies report an average outpatient PCP FTE of 0.67,49 and working an average of 54 hours per week.50 Given this, it is more likely that the 0.847 FTE reported by Virginia PCPs truly represents full-time work. In which case, the real panel size for PCPs would be 1,290 and not 1,523 patients.
Ideal primary care panel size should be determined by clinicians and practices, with consideration of the medical and social complexity and needs of patients, clinician availability, and clinic resources and support. Payment reform is needed to increase support for primary care, including compensating physicians for comprehensive care provided outside of the traditional visit, and reducing the gap in insurance reimbursement for primary and specialty services.51
Our analysis did not include APPs, such as NPs or PAs. Less than 30% of NPs and PAs work in primary care in Virginia and nationally.52,53 We were not able to differentiate which APPs were practicing in primary care. In addition, APPs in collaborative practice agreements with physicians will bill under the physician’s NPI. APPs are essential members of the primary care interprofessional team and contribute to alleviating workforce shortages. However, prior studies have not demonstrated that APPs alone can alleviate workforce shortages, and APPs will face similar financial and workforce pressures as physicians.54,55 We aim to include APPs in our future workforce analyses.
Conclusions
As complexity and scope of primary care practice increase, PCP panel sizes are decreasing. Appropriate panel size, adjusted for patient and physician needs, is essential for the delivery of comprehensive primary care. Currently, health systems and public health planners overestimate primary care panel sizes. More realistic estimations show that the primary care workforce shortage is more dire than estimated, and it is worsening.
Appendix 1. Variation in Physician Characteristics Based on Number of Unique Patients a Physician Sees Annually (Less than 100 vs 100-7,500 vs Greater than 7,500 Unique Patients*) – 2021 Data
Appendix 2. Histogram of unique number of patients seen by physicians in 2021
Appendix 3. Percent More Primary Care Physicians (PCPs) Needed for Each VA Resident to Have a PCP, 2016 to 2021
Bootstrap Mean and CI Calculated from 10,000 Simulations
Notes
This article was externally peer reviewed.
Funding: Funding for this study was provided by the Community Engagement Core and the Biostatistics, Epidemiology and Research Design (BERD) cores of the C. Kenneth and Dianne Wright Center for Clinical and Translational Science (CTSA award No. UM1TR004360 from the National Center for Advancing Translational Sciences) and the Agency for Healthcare Research and Quality (award No. R03 HS30016-01A1). Research support was provided through the American Board of Family Medicine (ABFM) Research and Editing Fellowship.
Conflict of interest: Jacqueline Britz is an Editor of the JABFM.
- Received for publication March 12, 2025.
- Revision received May 11, 2025.
- Revision received July 20, 2025.
- Accepted for publication August 4, 2025.
References
Jump to section
- Article
- Abstract
- Introduction
- Methods
- Results
- Discussion
- Conclusions
- Appendix 1. Variation in Physician Characteristics Based on Number of Unique Patients a Physician Sees Annually (Less than 100 vs 100-7,500 vs Greater than 7,500 Unique Patients*) – 2021 Data
- Appendix 2. Histogram of unique number of patients seen by physicians in 2021
- Appendix 3. Percent More Primary Care Physicians (PCPs) Needed for Each VA Resident to Have a PCP, 2016 to 2021
- Notes
- References
- Figures & Data
- References
- Info & Metrics