Table 1.

Factors Influencing Providers’ Geographic Choice of Practice Location

Predictor variable No. of studies GRADE	Effect and direction Results and supporting statement
	Provider aspects
Rural background 22 studies GRADE: High*	Positive association • Rural hometown was a predictor in a multivariate analysis of West Virginia medical student graduates (N = 1517; OR 4.02; CI 2.17–7.74)24 • Significant association with being raised in rural area in multivariate model of Oklahoma State University graduates (N = 190, p < 0.05) and graduates of the University of Minnesota (N = 3365; OR 2.82; CI 2.1–3.79)25 , 26 • Rural origin was a significant predictor in a multivariate analysis of Michigan State University College of Human Medicine graduates (N = 2382; OR 2.80; CI 2.09–3.74)27 • Significant association with rural high school in multivariate analysis of West Virginia physician assistants (N = 168; p < 0.01)28 • Being raised in a rural area was associated with practicing in a less populated county in a multivariate analysis (N = 683; p < 0.05)29 • Significant correlation with non-urban high school or college30 • Respondents who graduated from a rural high school were significantly more likely to practice in rural settings31 • Significant association with population of hometown32 • Qualitative analysis suggested rural exposure via upbringing33 • Significant difference due to rural childhood34 , 35 • 70% of rural providers had a rural background36 • 60% of rural providers had lived in a rural community44 • Birthplace in rural county increased odds23 • A combination of growing up in a rural area, plans to practice in rural area, and plans for family medicine showed a positive association37 • Higher proportion attending rural high school in rural vs. urban providers38 • Significant association with having a rural upbringing39, 47 • Significant relationship with rural background40 No association • Majority of rural providers did not grow up in small town41 , 42
Family 12 studies GRADE: Very low†‡	Association • Family ties reported as major reason43 • Family/spouse reported to be a very important factor34 • Significant association with location partner grew up in30 • Proximity to family listed as motivation36 • Significant association with having a child during or before medical school30 • Conclusion that support of and for significant other was most important factor31 • Many interviewees had sought out life partners who were willing to live in a rural community44 No association • Having children was not associated with practice location30 • Family obligation did not influence decision36 • Job of spouse was rated as very important by only 28% of participants38 • Spouse’s job location was cited by only 30%36 • Proximity to relatives was not a particularly influential factor40
Gender 11 studies GRADE: Very low‡	Association • Male gender was a significant predictor in a multivariate analysis of Michigan State University College of Human Medicine graduates (N = 2382; OR 1.39; CI 1.10–1.75)27 • Being male increased odds23 • Slightly smaller number of female rural practitioners than in overall population45 • Female physicians were less likely to practice in rural areas46 No association • Gender was not associated in a multivariate analysis of 1120 University of Louisville medical school graduates47 • No significant association with gender in multivariate analysis28 , 29 , 47, 48 • No difference by gender group30 , 31 , 34
Age 4 studies GRADE: Moderate	No association • Age was not associated with rural practice location in multivariate analysis of 1120 University of Louisville medical school graduates47 • Age was not associated with practicing in small town48 • Age at graduation was not associated with rural setting for first practice31 • Age at graduation, OR 1.0323
Marital status 4 studies GRADE: Very low†‡	Positive association • Being married increased odds (OR 1.47)23 Negative association • Those who were single were significantly more likely to practice in a rural setting as first employment31 No association • Being married was not associated30 , 48
International medical graduate (IMG) 4 studies GRADE: Very low‡	Positive association • Odds of South Asian IMGs working in a rural community 1.6 times the odds of US medical graduates in a multivariate analysis (N = 3862)49 (Slight) negative association • IMGs constituted 22% of the clinically active workforce but 19% of rural PCP workforce50 • 15.1% of IMGs work in rural areas compared to 17% of non-IMGs (p < 0.001)51 No association • 13% of IMGs compared to 18% DOs and 11% MDs were practicing in a rural location45
Race, ethnicity 3 studies GRADE: Moderate	No association • Race was not associated with rural practice location in a multivariate analysis of 1120 University of Louisville medical school graduates47 • Practicing in small town not associated with race48 • Rural setting for first practice not associated with race31
Exposure 2 studies GRADE: Low†	Positive association • Qualitative analysis suggested exposure via recreation facilitated future rural practice33 • Previous time spent in similar area was an important factor34
	Training
Rural rotation in training or residency 15 studies GRADE: Moderate*‡	Positive association • A rural campus was a significant predictor in a multivariate analysis of Michigan State University College of Human Medicine graduates (N = 2382; OR 2.80; CI 2.09–3.74)27 • Graduates from the University of Louisville medical rural campus were more likely to choose a rural practice location according to a multivariate analysis (N = 1120; OR 5.46)47 • Rural programs increased odds in addition to being raised in a rural community in a multivariate analysis (N = 3365; OR 4.62; CI 3.01–7.07)25 • Difference in rural practice between rural- and traditional-track graduates remained significant in a multivariate analysis (N = 106; OR 7.54; CI 1.5–37.9)52 • Rural residency trainees were 3 times as likely to practice in rural areas45 • Interviews suggested that exposure via education facilitated rural practice33 • Rural clerkship and rural residency training were associated with rural practice30 • Optional summer rural externship increased probability26 • Association with medical school in rural area, (OR 2.65); rural elective, (RR 1.53–1.93)23 • Significant relationship with rural clerkship40 • Many interviewees had developed an interest in rural medicine before or during medical school44 No association • University of Mississippi graduates were not more likely to practice in rural areas than physicians who graduated elsewhere48 • Medical school had discouraged rural practice for 40% of practitioners36 • No association with medical school location29 • No difference in rural rotation between rural and urban practitioners34 • Study showing a significant relationship with rural clerkship also reported that respondents indicated that participation in rural training was not particularly influential40
Primary care and family medicine focus 7 studies GRADE: Moderate	Positive association • Choosing a family medicine residency increased the odds in a multivariate analysis of University of Louisville medical school graduates (N = 1120; OR 5.46)47 • Primary care specialty was a significant predictor in a multivariate analysis in Michigan State University College of Human Medicine graduates (N = 2382; OR 1.65; CI 1.31–2.08)27 • Primary care physicians were 2.4 times as likely as specialists to practice in small towns in a multivariate analysis (N = 927; p < 0.001)48 • Rural family medicine residency graduates were 3 times as likely to practice in rural care45 • Specialty distribution (primary care, specialty) was significantly different between rural and urban groups31 • Association with career in family medicine, (OR 2.65); family medicine clerkship, (RR 1.26–1.44)23 • Association with primary care residency (RR 1.22–1.79)23 No association No significant association with primary care specialty35 Career in primary care OR 1.0623
Osteopathic medicine degree 2 studies GRADE: Low†	Positive association • 6% of workforce were DOs but 18% practiced in rural care45 • 4.9% of the workforce but contributed 10.4% to rural primary care50
	Financial aspects
Student loan or scholarship 9 studies GRADE: Very low†‡	Positive association • Second major reason was a loan or scholarship obligation43 • Medical school loan repayment correlated with rural practice32 • NHSC loan repayment, NHSC scholarship, and debt increased odds23 • Loan repayment program had an important influence on community providers’ choice to practice for 42%38 No association • Student loan debt was not a predictor of practicing in small towns48 , 53 • The amount of loan debt was a less important factor38 • For 71%, education debt had no influence on location of initial job54 • A loan forgiveness/repayment program was not rated as a particularly influential factor40 • Loan repayment was rated an important factor by only 11%34
Salary 5 studies GRADE: Very low†‡	Association • Importance of income as a factor in practice location differed between rural and urban groups55 • 58% found salary to be an important factor38 • Pay correlated with selecting rural care32 No association • Salary was not a predictor of practicing in small towns in a multivariate analysis48 • Salary/signing bonus was rated as very important by only 24–28%34
	Setting
Scope of practice 6 studies GRADE: Very low†‡	Positive association • Broad scope of practice was cited as an important reason of general surgeons30 • Scope of practice was important to 71% for healthcare providers38 • Most participants had chosen to practice in a rural community, in part, because they could maintain a broad scope of practice44 • High agreement with serving the health needs of the community, type of practice, supervising physician characteristics40 No association • Scope of practice was rated very important only by 30% of emergency department physicians34 • Full scope of practice was important to only 10% of female physicians36
Recreational activities 4 studies GRADE: Very low†‡	Positive association • Access to amenities/recreation was rated as important for choosing practice location34 • Recreational activities were rated as important by 58%38 • Hunting of birds and large game was associated with rural practice30 No association • Currently hunting or fishing, fishing, and hunting of small game showed no difference30 • Cultural and recreational activities, educational facilities in the community, and community recruitment efforts were not a particularly influential factor40
Lifestyle, small town life 2 studies GRADE: Low†	Positive association • Lifestyle was rated as very important34 • Qualitative interviews identified desire for small town life as important41

*Upgraded due to size of effect (see text). ^†Downgraded due to study limitations. ^‡Downgraded due to inconsistency; for full study details see evidence table in the Appendix

CI confidence interval; IMG international medical graduate; OR odds ratio; NHSC, National Health Service Corps; PCP, primary care physician