Community Relationships and Physician Retention in Clinician Shortage Areas

Key Points

Question

Is physicians’ retention in rural and urban underserved areas associated with how well communities meet physicians’ and their families’ needs and how well physicians integrate into their communities?

Findings

In this cross-sectional study using survey data from 1074 physicians working in rural and urban Health Professional Shortage Areas, community needs fulfillment and community integration were independently and comparably associated with anticipated retention in both settings, with no significant rural-urban differences.

Meaning

These findings suggest that strategies to help physicians have their community needs met and help them integrate into their communities will bolster retention in rural and urban underserved-area practices.

This cross-sectional study examines the associations of community needs fulfillment and community integration with anticipated retention among physicians in rural and urban underserved areas.

Abstract

Importance

Physicians’ retention in rural underserved areas is known to be affected by how well their communities meet their families’ needs and by how well physicians integrate into their communities, yet the relative importance of these 2 dimensions and whether they similarly affect retention in urban underserved practices are unknown.

Objective

To examine the associations of community needs fulfillment and community integration with anticipated retention among physicians in rural and urban underserved areas, and assess whether the relative importance of each dimension differs by setting.

Design, Setting, and Participants

This cross-sectional study used survey data collected May 2014 through January 2024, with analyses performed July 2025 to March 2026. Participants were physicians serving in rural and urban federally designated Health Professional Shortage Areas (HPSAs) in 33 US states participating in the Provider Retention and Information System Management collaborative.

Exposures

Community needs fulfillment (4-item scale; Cronbach α = 0.74) and community integration (2-item scale; α = 0.81).

Main Outcomes and Measures

The outcome of interest was anticipated retention beyond 2 years in the current practice. Separate rural and urban logistic regression models estimated independent associations with each exposure scale; z tests compared β coefficients across settings.

Results

Of 2163 physicians emailed invitations, 1174 (54.9%) responded; 1074 physicians (mean [SD] age, 40.6 [6.7] years; 648 [61.8%] female) met inclusion criteria and were analyzed (351 physicians [32.7%] in rural HPSAs, 723 physicians [67.3%] in urban HPSAs). Among these, 265 physicians (75.5%) in rural HPSAs and 515 physicians (71.2%) in urban HPSAs anticipated remaining beyond 2 years (P = .14). In both settings, higher community needs fulfillment and community integration scores were independently associated with anticipated retention. Findings were comparable for the 2 scales among rural HPSA physicians (community needs fulfillment: β = 0.612 [SE, 0.192]; community integration: β = 0.578 [SE, 0.156]; Nagelkerke R² = 0.19) and among urban HPSA physicians (community needs fulfillment: β = 0.565 [SE, 0.130]; community integration: β = 0.687 [SE, 0.110]; R² = 0.16). The z tests showed no significant rural-urban differences for either scale.

Conclusions and Relevance

In this cross-sectional study of 1074 physicians, community needs fulfillment and community integration were independently and similarly associated with anticipated retention among physicians in rural and urban HPSA practices. Practices and communities should pursue strategies to help physicians have their needs of their communities met and become integrated into their communities.

Introduction

Rural physicians’ relationships with their communities influence how long they remain in their practices.^1,2,3,4,5 A central dimension of this relationship is how well communities meet physicians’ and their families’ material and lifestyle needs for things like employment for spouses, good schools for children, social networks, recreational activities, physical safety, and proximity to extended family.^5,6,7,8,9,10 Studies show that meeting these needs is important to the retention of physicians working in rural areas generally, as well as for physicians working in rural areas that are underserved.^{1,5,6,8,9,10,11}

A second dimension of physicians’ relationships with their communities is their integration into their communities. Community integration—when an individual feels accepted by, involved in, and part of the community—is often cited by rural physicians as a key reason for remaining in their practices for decades or an entire career.^{1,2,3,4,9,11,12,13,14} Community integration includes the psychological and sociological concepts of sense of place and belonging, rootedness, membership, and community participation, reflecting the affective bonds people form with the social and physical aspects of a place.^3,4,12,15,16 Studies find that community integration is important for the retention of rural physicians generally and for physicians working in underserved rural areas.^{1,3,9,10,17,18}

It is unknown whether these 2 dimensions of the physician-community relationship—how well communities meet physicians’ needs and physicians’ community integration—similarly influence retention for physicians working in urban and urban underserved areas. Urban sociology has extensively documented how characteristics of communities, including the built environment, social capital, social networks, and civic engagement, affect urban residents’ well-being and behavior.¹⁹ Yet we are aware of no studies that have examined how community needs fulfillment and community integration relate to job retention for urban physicians.

Retaining clinicians in practices in underserved areas is particularly important. This study examines how the anticipated retention of physicians within federally designated Health Professional Shortage Areas (HPSAs) varies with how well their communities meet their personal and families’ needs and how well physicians integrate into their communities.²⁰ This study extends prior work by assessing the relative importance of these 2 dimensions of the physician-community relationship to anticipated retention in rural vs urban HPSAs and the relative importance of the 2 dimensions to the anticipated retention of physicians in each setting.

Methods

Participants

This cross-sectional study analyzed questionnaire data from physicians working in rural and urban HPSAs who had received loan repayment from the National Health Service Corps (NHSC). As physicians completed their initial term of loan repayment support, they completed questionnaires as part of their states’ participation in the Provider Retention and Information System Management (PRISM) collaborative of the National Rural Recruitment and Retention Network, which provides feedback to states from clinicians participating in public loan repayment and related support-for-service programs.^21,22,23 Use of these administrative data was determined to be exempt from human participants review and informed consent requirements by the University of North Carolina Office of Human Research Ethics (Study 12-0626) because they were existing data initially collected for administrative purposes. Study results were reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

The Bureau of Health Workforce provides information from its Bureau of Health Workforce Management Information System Solution to the PRISM collaborative on clinicians participating in the NHSC, including the dates of their loan repayment contracts, which guide the timing of the end-of-contract surveys used for this study’s data.²⁴ Eligible participants were allopathic and osteopathic physicians in PRISM’s 33 states whose support from the NHSC Loan Repayment Program (LRP), Rural Community LRP, or Substance Use Disorder Workforce LRP ended between May 2014 and January 2024 and who had worked in their current practices for at least 6 months and worked at least 20 hours per week.^25,26

Data Collection

As part of PRISM’s ongoing clinician feedback efforts, a month before physicians completed their loan repayment terms, they received email invitations from their state primary care office or other health workforce office to complete questionnaires. They were informed that the primary purpose of questionnaires was program feedback and improvement, that participation was voluntary, and that responses would not be shared with employers or affect their loan repayment support.

During the study period, a total of 2163 physicians in 33 states were emailed invitations, of whom 1174 (54.9%) completed questionnaires. Analyses excluded 15 respondents who had worked in their practices for less than 6 months, and 85 others who were missing information to determine their rural vs urban practice location, anticipated retention duration, physician-community relationship scales, or assessments of their current work and jobs.

Measures

In questionnaires, physicians reported their demographics, including their race and ethnicity and gender, and confirmed their current practice’s name, location, the date they began working there, and type of practice. Race and ethnicity were classified as Asian, Black or African American, multiple, non-Hispanic White, and other (eg, American Indian or Alaska Native, Hispanic, or Native Hawaiian or Other Pacific Islander). Race and ethnicity are included in these analyses because they are often important factors in individuals’ relationships with communities.

We constructed a scale of how well communities met physicians’ and their families’ needs by calculating means of responses to 4 survey items, each rated on a 5-point Likert-type scale from strongly agree to strongly disagree: “I/We enjoy the activities the community offers,” “I/We have access to most of the things we like to do,” “I/We feel safe in the community,” and “I/We live close enough to family” (Cronbach α = 0.74). Items were based on concepts and measures from prior literature.^{5,7,8,11,27,28,29} A second, community integration scale was constructed by calculating the mean level of agreement with statements of 2 items also drawn from the literature: “I feel a sense of belonging to the community” and “I am involved in community issues important to me” (α = 0.81).^{10,12,13,14,15,16,30}

The study’s primary outcome was physicians’ anticipated retention, a widely used proxy for actual turnover, given its predictive validity and expediency.^31,32 Anticipated retention was measured with a validated survey item: “Looking ahead from now, about how many years do you anticipate remaining in your current practice?”³³ For regression analyses, anticipated retention was dichotomized as 2 years or less vs more than 2 years. Practice zip codes were mapped to counties. We then used the Area Health Resource File to pull in 2013 Rural-Urban Continuum Codes to identify rural (nonmetro; codes 4-9) vs urban (metro; codes 1-3) counties, as well as county non-Hispanic White racial and ethnic compositions and poverty rates.^34,35

Statistical Analysis

Analyses, conducted July 2025 through March 2026, initially described and compared the rural and urban HPSA physician groups on their demographics, practices, and county population measures. The 2 groups were then compared on proportions agreeing with each item of the community needs fulfillment and community integration scales and then the proportions for whom their mean responses in each scale fell within the agree vs neutral and disagree ranges using χ² tests. A Pearson correlation coefficient assessed the strength and direction of correlation between the 2 physician-community relationship scales as continuous measures. The features of physicians, practices, and counties associated with each scale were then assessed for physicians working in rural HPSAs and in urban HPSAs, using 1-way analysis of variance.

For rural and urban HPSA physicians whose responses on each community relationship scale fell in the strongly agree or agree vs neutral, disagree, or strongly disagree ranges, we plotted proportions who anticipated remaining in their practices over the next 4 years. Differences in proportions anticipating retention were assessed at each time point (1, 2, 3, and 4 years) with χ² tests.

To assess the relative strength of association of the 2 community relationship scales with anticipated retention for the rural and urban HPSA cohorts, 2 logit models of anticipated retention of 2 years or less vs more than 2 years were constructed. To avoid falsely attenuating the measured associations of the 2 scales with anticipated retention, no variables indicating physician, practice, or county characteristics were included as controls in the logit models because any association with retention for these variables is either known or seems likely to be mediated, at least in part, through physicians’ relationships with their communities.^30,36

Missing values were not imputed. A statistical significance level of 2-sided P ≤ .05 was set for this study’s principal analyses addressing the associations between measures of the physician-community relationship and anticipated retention. For the multiple tests of comparisons of the characteristics of the rural and urban cohorts and identifying correlates of physicians’ community needs fulfillment and community integration, a P < .05 for significance was set with the Benjamini-Hochberg correction for a 10% false discovery rate.³⁷ Analyses were conducted using SPSS software version 29.0 (IBM Corporation).

Results

Physician Demographics and Types of Practice

Of 2163 physicians emailed invitations, 1174 (54.9%) responded; 1074 physicians (mean [SD] age, 40.6 [6.7] years; 648 [61.8%] female) met inclusion criteria and were analyzed (351 [32.7%] in rural HPSAs, 723 [67.3%] in urban HPSAs). Most physicians were married (826 physicians [78.4%]) and had children (710 physicians [68.4%]) (Table 1). By race and ethnicity, there were 69 Asian physicians (6.6%), 78 Black or African American physicians (7.4%), 100 physicians (9.5%) who identified as multiple races or ethnicities, 748 non-Hispanic White physicians (71.3%), and 54 physicians (5.1%) who identified as other race or ethnicity. Physicians in rural HPSAs were more likely than those in urban HPSAs to be male, married, and non-Hispanic White (Table 1). In both settings, the most common specialty was family medicine, followed by pediatrics (Table 1).

Table 1. Physician, Practice, and County Characteristics in Rural vs Urban HPSAs.

Characteristic	Physicians, No. (%)		P value
	Rural HPSAs (n = 351)	Urban HPSAs (n = 723)
Period when surveyed
Before COVID-19 (2016-2019)	191 (54.4)	380 (52.6)	.57
During COVID-19 (2020-2022)	104 (29.6)	209 (28.9)
After COVID-19 (2023-2024)	56 (16.0)	134 (18.5)
Physician characteristics
Age group, y (mean [SD], 40.6 [6.7] y)
31-34	56 (16.7)	124 (17.8)	.91
35-39	122 (36.4)	240 (34.4)
40-49	120 (35.8)	259 (37.2)
50-68	37 (11.0)	74 (10.6)
Gender
Female	196 (57.0)	452 (64.2)	.02
Male	148 (43.0)	252 (35.8)
Third gender	0	0
Marital status
Married	283 (82.0)	543 (76.7)	.05
Not married	62 (18.0)	165 (23.3)
Parenthood
Has children	243 (72.1)	467 (66.6)	.08
No children	94 (27.9)	234 (33.4)
Race and ethnicity
Asian	8 (2.3)	61 (8.6)	<.001
Black or African American	16 (4.7)	62 (8.8)
Multiple	25 (7.3)	75 (10.6)
Non-Hispanic White	281 (82.2)	467 (66.1)
Othera	12 (3.5)	42 (5.9)
Medical specialty
Family medicine	220 (62.7)	366 (50.6)	.003
Pediatrics	48 (13.7)	155 (21.4)
Psychiatry	30 (8.5)	69 (9.5)
Obstetrics and gynecology	24 (6.8)	76 (10.5)
Internal medicine	28 (8.0)	56 (7.7)
Remaining education debt, $ (mean [SD], $126 762 [$128 179])
0-19 000	86 (25.3)	178 (25.4)	.35
20 000-99 999	89 (26.2)	191 (27.3)
100 000-199 999	83 (24.4)	139 (19.9)
≥200 000	82 (24.1)	192 (27.4)
Practice variables
Type of practice
FQHC and Look-Alike Health Centers	97 (27.6)	470 (65.0)	<.001
Rural Health Clinic	96 (27.4)	20 (2.8)
Hospital-based site	34 (9.7)	53 (7.3)
American Indian health facility	44 (12.5)	36 (5.0)
Mental health facility	10 (2.8)	19 (2.6)
Otherb	70 (19.9)	125 (17.3)
Tenure in current practice, y (mean [SD], 4.66 [3.12] y)
0.50-2.49	50 (15.7)	154 (22.9)	.002
2.50-3.99	133 (41.8)	213 (31.6)
4.00-6.99	92 (28.9)	185 (27.5)
7.00-22.00	43 (13.5)	121 (18.0)
County population characteristics
Persons below poverty level, %
<12.0	49 (14.0)	233 (32.2)	<.001
12.0-16.9	117 (33.3)	319 (44.1)
≥17.0	185 (52.7)	171 (23.7)
Minoritized race and ethnicity, %c
<30.0	225 (64.1)	199 (27.5)	<.001
30.0-49.9	65 (18.5)	257 (35.5)
≥50.0	61 (17.4)	267 (36.9)

Abbreviations: FQHC, Federally Qualified Health Center; HPSA, Health Professional Shortage Area.

^a Includes American Indian or Alaska Native, Hispanic or Latino, and Native Hawaiian or Other Pacific Islander.

^b Includes health departments, substance use disorder facilities, correctional facilities, emergency response sites, and others.

^c Includes American Indian or Alaska Native, Asian, Black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander, 2 or more races, and some other race.

Two-thirds of physicians in urban HPSAs worked in Federally Qualified Health Centers (FQHCs) or Look-Alike Health Centers.³⁸ One-quarter of rural HPSA physicians worked in Federally Qualified Health Centers or Look-Alike Health Centers, and another one-quarter worked in Certified Rural Health Clinics (Table 1). Physicians in rural and urban HPSAs combined began working in their practices a mean (SD) of 4.66 (3.12) years earlier, which was generally more than 2 years before they began receiving NHSC loan repayment support. Physicians in rural HPSAs practiced in counties with higher poverty rates and lower percentages of residents from minoritized groups than counties where urban HPSA physicians worked.

Community Needs Fulfillment and Community Integration

A total of 238 rural HPSA physicians (67.8%) and 536 urban HPSA physicians (74.1%) reported that their communities met their and their families’ needs overall (P = .03), that is, their mean level of agreement on the 4 survey items of the community needs fulfillment scale fell within the agree and strongly agree range (Table 2). Among the scale’s 4 component items, more urban HPSA physicians agreed that they and their families enjoyed the activities offered in the community and had access to most things they liked to do, whereas more rural HPSA physicians felt safe in their communities. Physicians in rural HPSAs were more likely than those in urban HPSAs to feel integrated into their communities (232 physicians [66.1%] vs 426 physicians [58.9%]; P = .02), more often reporting both that they felt a sense of community belonging and that they were involved in community issues (Table 2). Both scales of the physician-community relationship were strongly positively correlated with one another among rural HPSA physicians (r = 0.51; 95% CI, 0.42-0.58) and moderately positively correlated among urban HPSA physicians (r = 0.35; 95% CI, 0.28-0.41).

Table 2. Rural vs Urban HPSA Physicians Who Agree or Strongly Agree With Items of the Community Needs Fulfillment and Community Integration Scales^a.

Overall scales and component survey item	Physicians, No. (%)		P value
	Rural HPSAs (n = 351)	Urban HPSAs (n = 723)
Community needs fulfillment
Overall	238 (67.8)	536 (74.1)	.03
“I/We enjoy the activities the community offers”	268 (76.4)	621 (85.9)	<.001
“I/We have access to most of the things we like to do”	250 (71.2)	629 (87.0)	<.001
“I/We live close enough to family”	217 (61.8)	449 (62.1)	.93
“I/We feel safe in the community”	313 (89.2)	613 (84.8)	.05
Community integration
Overall	232 (66.1)	426 (58.9)	.02
“I feel a sense of belonging to the community”	280 (79.8)	526 (72.8)	.01
“I am involved in community issues important to me”	238 (67.8)	443 (61.3)	.04

Abbreviation: HPSA, Health Professional Shortage Area.

^a Compared with neutral, disagree and strongly disagree.

Factors Associated With Community Needs Fulfillment and Community Integration

In both rural and urban HPSAs, physicians with children vs those without children had higher mean community needs fulfillment score (mean [SD]: rural, 4.04 [0.77] vs 3.59 [0.78]; P < .001; urban, 4.09 [0.73] vs 3.97 [0.66]; P = .04) and community integration scale (mean [SD]: rural, 4.10 [0.90] vs 3.63 [1.01]; P < .001; urban, 3.88 [0.81] vs 3.64 [0.92]; P < .001) scores (Table 3). Physicians’ age, gender, and specialty were not associated with scale scores in either setting. Physicians who had worked longer in their practices had greater mean community integration scores in rural HPSAs and greater community needs fulfillment scores in urban HPSAs (Table 3).

Table 3. Associations Between Physicians and Their Practices and Counties With Their Community Needs Fulfillment and Community Integration.

Characteristic	Rural HPSAs (n = 351)				Urban HPSAs (n = 723)
	Community needs fulfillment scale		Community integration scale		Community needs fulfillment scale		Community integration scale
	Mean (SD)	P value	Mean (SD)	P value	Mean (SD)	P value	Mean (SD)	P value
Physician demographics
Age, y
31-34	4.06 (0.68)	.19	4.13 (0.88)	.31	4.10 (0.71)	.14	3.74 (0.89)	.31
35-39	3.83 (0.87)		3.98 (1.01)		4.01 (0.75)		3.77 (0.86)
40-49	3.98 (0.81)		3.95 (0.95)		4.08 (0.66)		3.86 (0.82)
50-68	3.78 (0.71)		3.74 (0.93)		3.90 (0.80)		3.69 (0.91)
Gender
Female	3.91 (0.84)	.90	3.95 (0.93)	.64	4.05 (0.74)	.43	3.81 (0.88)	.86
Male	3.92 (0.74)		4.00 (0.98)		4.01 (0.67)		3.79 (0.82)
Marital status
Married	3.99 (0.78)a	<.001	4.03 (0.94)a	.02	4.05 (0.71)	.61	3.85 (0.83)a	.01
Not married	3.61 (0.82)		3.72 (0.97)		4.02 (0.73)		3.66 (0.93)
Parenthood
Has children at home	4.04 (0.77)a	<.001	4.10 (0.90)a	<.001	4.09 (0.73)a	.04	3.88 (0.81)a	<.001
No children at home	3.59 (0.78)		3.63 (1.01)		3.97 (0.66)		3.64 (0.92)
Race and ethnicity
Asian	3.47 (0.73)	.09	3.75 (1.31)	.94	3.81 (0.74)a	.006	3.75 (0.82)	.45
Black or African American	3.59 (0.66)		4.03 (0.99)		3.87 (0.79)		3.73 (0.87)
Multiple	3.65 (0.80)		3.96 (0.97)		4.23 (0.64)		3.87 (0.90)
Non-Hispanic White	3.98 (0.80)		3.97 (0.95)		4.07 (0.70)		3.79 (0.87)
Otherb	3.88 (0.73)		4.21 (0.86)		3.99 (0.77)		4.00 (0.64)
Medical specialty
Family medicine	4.00 (0.74)	.07	4.01 (0.90)	.33	4.07 (0.67)	.60	3.74 (0.82)	.11
Pediatrics	3.88 (0.82)		4.02 (0.98)		4.05 (0.67)		3.84 (0.78)
Psychiatry	3.68 (0.97)		3.60 (1.09)		4.01 (0.81)		3.96 (0.97)
Obstetrics and gynecology	3.80 (0.78)		4.06 (1.02)		3.91 (0.89)		3.80 (0.89)
Internal medicine	3.64 (0.90)		3.93 (1.06)		4.02 (0.74)		3.86 (1.08)
Practice variables
Type of practice
FQHC and Look-Alike Health Centers	3.99 (0.75)a	<.001	3.98 (0.94)a	.01	4.06 (0.69)	.12	3.79 (0.82)	.60
Rural Health Clinic	4.15 (0.68)		4.20 (0.89)		3.85 (0.74)		3.93 (0.89)
Hospital-based site	3.86 (0.88)		4.04 (1.08)		4.20 (0.67)		3.95 (0.79)
American Indian health facility	3.51 (0.70)		3.55 (0.85)		3.95 (0.75)		3.67 (0.90)
Mental health facility	3.58 (1.05)		3.90 (0.91)		4.14 (0.85)		3.92 (1.11)
Other types of practices	3.84 (0.86)		3.91 (0.97)		3.92 (0.77)		3.76 (0.95)
Tenure in current practice, y
0.50-2.49	3.77 (0.74)	.11	3.60 (0.91)a	<.001	3.98 (0.82)a	.001	3.75 (0.97)	.08
2.50-3.99	3.89 (0.79)		4.00 (0.91)		3.94 (0.70)		3.70 (0.79)
4.00-6.99	4.06 (0.85)		4.08 (0.96)		4.14 (0.70)		3.83 (0.81)
7.00-22.00	4.06 (0.71)		4.40 (0.74)		4.22 (0.55)		3.94 (0.85)
County characteristics
Persons below poverty level, %
<12.0	4.10 (0.70)a	<.001	4.15 (0.86)	.19	4.14 (0.69)a	.02	3.82 (0.79)	.76
12.0-16.9	4.10 (0.72)		4.03 (0.92)		4.02 (0.70)		3.80 (0.86)
≥17.0	3.76 (0.83)		3.90 (0.98)		3.93 (0.76)		3.75 (0.94)
Minoritized race and ethnicity, %c
<30.0	4.02 (0.79)a	<.001	4.06 (0.94)	.09	4.13 (0.64)	.04	3.89 (0.81)a	.03
30.0-49.9	3.88 (0.83)		3.90 (1.02)		4.03 (0.70)		3.83 (0.85)
≥50.0	3.57 (0.67)		3.77 (0.87)		3.97 (0.78)		3.69 (0.89)

Abbreviations: FQHC, Federally Qualified Health Center; HPSA, Health Professional Shortage Area.

^a Statistically significant at the Benjamini-Hochberg corrected significance rate of P < .05.

^b Includes Hispanic, Native American or Alaska Native, or Native Hawaiian or Other Pacific Islander.

^c Includes American Indian or Alaska Native, Asian, Black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander, 2 or more races, and some other race.

Among physicians in rural HPSAs, community needs fulfillment scores were lower in counties with higher poverty percentages and higher minoritized population percentages. Among physicians in urban HPSAs, community needs fulfillment scores were lower in counties with higher poverty percentages and community integration scores were lower in counties with higher minoritized population percentages.

Physicians’ Anticipated Retention and Its Associations With Community Needs Fulfillment and Community Integration

Approximately three-quarters of physicians working in rural (265 physicians [75.5%]) and urban (515 physicians [71.2%]) HPSAs anticipated remaining in their current practices for more than 2 years (P = .14). In both settings, physicians who agreed or strongly agreed overall with the items of the community needs fulfillment and community integration scales were significantly more likely to anticipate they would still be working in their practices in 1, 2, 3 and 4 years than physicians who were neutral or disagreed with these items (Figure).

Figure. Line Graphs of Physicians’ Agreement That Their Community Needs Were Met and That They Were Integrated Into the Community and the Proportions Anticipating Remaining in Rural and Urban Health Professional Shortage Area (HPSA) Practices^a.

^aGroups differed on both metrics and in both setting at 1, 2, 3, and 4 years (all P < .001).

Among physicians working in rural HPSAs, the community needs fulfillment score was independently associated with anticipated retention beyond 2 years (β = 0.612 [SE, 0.192]), as was the community integration score (β = 0.578 [SE, 0.156]) when the 2 were assessed together as continuous variables in a logistic regression model (Table 4). The 2 scales had similar β coefficients, indicating comparable strengths in their associations with anticipated retention. For physicians in rural HPSAs, the strength of the association between the combined dimensions of physician-community fit and anticipated retention was moderate (Nagelkerke R² = 0.19).³⁹

Table 4. Anticipating Retention Beyond 2 Years by Community Needs Fulfillment and Community-Integration Scale Scores: Comparing Physicians in Rural and Urban HPSAs.

Variable	Rural HPSAs (n = 351)		Urban HPSAs (n = 723)		z Test for differences in associations (2-tailed)
	β (SE)	P value	β (SE)	P value	z Score	P value
Community meets physician needsa	0.612 (0.192)	.001	0.565 (0.130)	<.001	0.203	.84
Physician-community integrationa	0.578 (0.156)	<.001	0.687 (0.110)	<.001	−0.571	.57
Model constant	−3.407 (0.732)	<.001	−3.864 (0.579)	<.001	NA	NA
Model χ2	47.65	<.001	87.34	<.001	NA	NA
Model −2 log likelihood	343.22	NA	780.36	NA	NA	NA
Model Nagelkerke R2	0.189	NA	0.163	NA	NA	NA

Abbreviations: HPSA, Health Professional Shortage Area; NA, not applicable.

^a Continuous scales, from 1 (strongly disagree) to 5 (strongly agree).

Similarly, among physicians in urban HPSAs, both scales were independently associated with anticipated retention at 2 years when included together in a logistic regression model and, once again, β coefficients for the 2 scales were similar (community needs fulfillment: β = 0.565 [SE, 0.130]; community integration: β = 0.687 [SE, 0.110]), indicating comparable strengths of association (Table 4). A model R² of 0.16 indicated that the strength of the association between the scales and anticipated retention for urban HPSA physicians was moderate.

The z tests comparing β coefficients between rural and urban HPSA models showed no difference between physician groups in the strength of association with anticipated retention for the community needs fulfillment scale (z = 0.20; P = .84) or for the community integration scale (z = −0.57; P = .57) (Table 4).⁴⁰

Discussion

In this cross-sectional study of 1074 physicians, community needs fulfillment and community integration were independently and similarly associated with anticipated retention among physicians in rural and urban HPSA practices. Physicians, like all workers, do not remain long in jobs where the terms of employment are poor or where the work is poorly organized or unfulfilling.^6,40 Physician turnover studies, principally based on the fields of organizational behavior and management, have focused on the effects of these job- and work-related factors.^41,42,43 But physician turnover cannot be fully understood if physicians are approached solely as workers.⁴⁴ As people, physicians and their families want and need things of their communities to be able to live and grow happily and successfully, and they need to feel accepted and at home.^3,4,6,15,16 As a workforce that freely crosses state lines for jobs and often has no extended family nearby, there is little keeping physicians in communities where they do not feel accepted or that do not meet their needs.

Rural physician studies have long identified the importance of both community needs fulfillment and community integration to turnover.^2,3,6,11 By assessing both dimensions simultaneously in a rural HPSA physician cohort, these findings suggest that these are distinct aspects of the physician-community relationship, that each plays a role in anticipated turnover independent of the other, and that the estimated impact of each on anticipated turnover is comparable. Then further asking whether these relationships also hold for physicians in urban settings, this study found that both community needs fulfillment and community integration were no less important to the anticipated retention of physicians in urban HPSAs.

Communities and practices are in the position to help physicians find what they need from their communities and feel accepted and at home. Information packets given to prospective physicians help them know up front whether the community is likely to fit their needs.¹² Community liaison committees can help new physicians find homes in well-fitting neighborhoods, fulfilling employment for spouses, and the best childcare and education options for children and help families discover the area’s recreation offerings.^{7,12,13,44,45,46} Local families can introduce physicians’ families to community events, churches, civic organizations, and sports leagues.^12,13,45 Communities can make physicians feel embraced by publicly expressing their gratitude through pieces in the local media, municipal honors, and publicly including them in community events. Practices can shape physicians’ jobs to promote a healthy work-life balance to give physicians ample time to enjoy communities with their families.^12,44 These interventions are more likely to happen if they are made explicit parts of practices’ and communities’ recruitment and retention initiatives. Furthermore, loan repayment programs can help educate and guide physician-applicants to practices in well-fitting and welcoming communities, preferentially approve loan repayment applications where there is likely to be a strong physician-community fit, and decline to make loan repayment awards in communities that have proven unwelcoming to physicians.

Medical schools and residencies can also play a role by teaching the importance of the physician-community relationship, how to identify a community that will meet one’s needs, and how to become part of a community.^3,47 Curricula should include concepts from rural and urban sociology and culture, community engagement, and leadership development.¹³ Extended rotations in the types of community settings where physicians may later work give learners the opportunity to learn about their people, institutions, customs, and lifestyles.^3,4,48 Current Accreditation Council for Graduate Medical Education program requirements in family medicine, internal medicine, and pediatrics list competencies in understanding the social and community factors affecting patients (cultural awareness) and the “ability to recognize and develop a plan for one’s own personal and professional well-being”⁴⁹ but list no competencies for residents to understand how the community can affect them personally or to know how to build sustaining relationships with their communities. Few primary care physicians in the 1990s were taught how to become accepted in their communities and other aspects of community medicine but, promisingly, those who were taught these skills later reported greater participation and assimilation into their communities as practicing physicians.^50,51

Limitations

This study has some limitations. As an observational study, correlations identified may not represent causal relationships. Also, this study identified associations between aspects of the physician-community relationship and anticipated retention for broad rural and urban HPSA cohorts, and associations for some subpopulations likely differ, such as for those at different points in their careers or life course, and those new vs long established in their practices. Furthermore, anticipated retention is a measure of behavioral intention, which, although predictive of retention, will not perfectly reflect all who will actually remain in their jobs.^52,53

This study’s survey measures assessing how well communities fulfill physicians’ needs were worded broadly for relevance to all physician-participants. Narrower and more specifically worded items may have revealed even stronger associations between community needs fulfillment in specific areas and anticipated retention for specific subgroups, such as the item, “My spouse is happily employed” for the anticipated retention of married physicians.^6,27,54

A person’s community integration and sense of place are conceptualized in many ways and assessed with different survey measures in the literatures of various fields.^15,55 It is not known how well the survey measures used in this study captured the breadth or most important elements affecting physicians’ retention or whether weighting items within the scales would be more predictive. Future studies using further refined scales may find the community integration-retention connection to be even stronger.

The notion of community was not defined for survey respondents, allowing each, guided by their context and background, to envision its nature and breadth. It is likely that physicians in rural and urban areas saw the nature and breadth of their communities differently.

Conclusions

This cross-sectional study of 1074 physicians working in HPSA practices found that 2 dimensions of the physician-community relationship—community needs fulfillment and community integration—were important to physicians’ anticipated retention in both rural and urban underserved areas. Future studies should further explore the role of these and other aspects in the physician-community relationship in retaining physicians within underserved as well as nonunderserved areas and examine and verify the ways that practices, communities and educators can strengthen these relationships for the benefit of physicians, patients, practices, and communities.

Supplement.

Data Sharing Statement