Addressing the Physician Shortage in Hawai‘i: Recruiting Medical Students Who Meet the Needs of Hawai‘i's Rural Communities

Abstract

Background

Past studies in the continental US have demonstrated that students from rural areas and those who go into primary care are more likely to practice in rural areas than urban-born and specialty physicians.

Methods

This study uses two separate data sets to examine whether medical students and young physicians in Hawai‘i follow the same pattern. A retrospective study of graduates of the University of Hawai‘i John A. Burns School of Medicine from 1993–2006 was performed examining the relationship between practice location and high school attended for those practicing in Hawai‘i. In addition, a survey was conducted with the first, second and third year medical students examining their practice intentions as related to where they grew up. Both data sets were analyzed using Chi Squared tests to determine the significance of associations between individuals from rural backgrounds practicing or intending to practice in rural areas.

Results

The relationship in both cases showed that students and physicians from rural areas were more likely to practice in rural areas. However, 81% of all respondents reported being willing to consider practicing in rural area, especially if lifestyle, work environment, and employment opportunities were favorable.

Discussion

If the State of Hawai‘i wants to expand the physician workforce in the rural areas of Hawai‘i, recruiting more students from rural areas and increasing desirability of rural practice settings are excellent paths to take.

Background

The problem of physician shortage in rural areas is not new¹ and given current supply and demand trends, it is expected to worsen with time.^2–4 Within rural sections of the country, there is a higher prevalence of poverty and chronic disease, the population is typically older, and a greater proportion of residents are without health insurance or receive Medicaid or Medicare.⁵ Rural areas of the United States are home to approximately 20% of the overall population (60 million people) but less than 8% of the nation's physicians,⁶ resulting in inadequate numbers of physicians to meet the health needs of rural populations.^7,8 Primary care physicians (PCPs) are traditionally the focal providers of health services in rural areas but their supply is insufficient to meet the ever-growing demands of these communities.⁹ Physician supply in shortage areas is resistant to even the strongest market pressures — during the period between 1980 and 1999, a time of rapidly increasing aggregate physician numbers in the United States, primary care was the only practice arena to show a steady decline in practitioners.¹⁰

Several factors have been shown to be predictive of whether or not a medical student will ultimately choose to work in a rural environment. There is ample evidence to support the claim that medical school graduates who were raised in rural environments are more likely to choose to practice in rural areas.¹¹ The literature also supports the assertion that interest in rural family medicine is stronger among students who completed high school in a small town (population less than 50,000), and have strong family ties in rural communities.¹² Due to financial constraints and the weak political feasibility of making major changes to the US health care system in the near future, a careful policy approach to altering existing programs affecting physician supply may be the most realistic means of satisfying the nation's current need for physicians in rural areas.¹³ One cost-effective approach to addressing physician shortage is for medical schools to recruit and select more students who are inclined to seek careers in rural medicine. A number of US medical schools have chosen this model and through rural program tracks have made significant contributions to the physician workforce in their state's rural regions.¹⁴

Methods

Two distinct research projects were conducted to analyze rural origins and practice locations. The first, a retrospective study of rural origins and practice location, was conducted using data on practice location of past University of Hawai‘i John A. Burns School of Medicine (JABSOM) graduates working in Hawai‘i, and the second, a survey of rural practice intention, was conducted among medical students.

1. Retrospective Study of Rural Origins and Practice Location

Information on high school attended was obtained for the graduates of John A. Burns School of Medicine for the years 1993 to 2009 from the JABSOM Office of Student Affairs. Information on classes prior to 1993 is stored in paper format, and location of high school could not be easily isolated from other confi- dential information, and therefore was not reviewed for this study. The end date of 2006 was selected to allow graduates at least five years for residency training, as longer residencies indicate specialties more likely to be found in urban areas. High school attended was used as a proxy measure for community of residence. The zip code of the high school was identified as either rural or non-rural based on the island where it was located (O‘ahu was considered urban; all other islands rural). Practice location information was obtained through physician surveys, internet searches, direct calling, and provider lists as described by Withy, et al. in their healthcare workforce assessment article in this edition. Primary care was defined as: Internal Medicine, Med-Peds, Family Medicine, General Practice, Pediatrics, and Geriatrics.

2. Medical Student Survey of Rural Practice Intention

In a separate 2010 study, a survey was administered to all 186 first through third year medical students present at mandatory colloquia in the Spring of 2010. A total of 166 JABSOM students completed the survey that asked them to identify their intended location of practice (89% response rate). The study asked participants if they would consider rural practice and factors that might influence their choice to practice in a rural setting. The participants were asked their ideas for recruiting and retaining physicians in rural areas of Hawai‘i. The purpose, risks, and benefits of participating in the survey were explained before it was distributed to class members in compliance with the University of Hawai‘i Committee on Human Subjects approval of the study.

Analysis: For both studies “Rural” was defined as Hawaiian islands other than O‘ahu and “Urban” as any location on O‘ahu.

1. Retrospective Study of Rural Origins and Practice Location

An Excel table was created with high school coded as 0 for out of state, 1 for rural, and 2 for urban. Practice location was similarly coded. Because of the limitations of the data, only the graduates who were both from Hawai‘i and practicing in Hawai‘i could be included. The final data set included 177 individuals meeting the criteria. A 2×2 table was developed and chi square analysis performed to test for an association between rural origin and rural practice location.

2. Medical Student Survey of Rural Practice Intention

A dataset was created with codes for rural origin and rural intentions with respect to practice location. If a student listed multiple practice locations under consideration, then the response was counted as “rural” if any of the locations met the definition of rural. Responses from 61 students were excluded from the analysis because they were either undecided, from outside Hawai‘i, or intended to practice outside Hawai‘i leading to a sample size of 105. Again, a chi square analysis was conducted to compare rural origin to rural intention practice location.

Descriptive responses were included from all respondents. They were reviewed by two researchers for common themes and counted in a category upon concordance by the two researchers. The most common thematic categories are reported in the results section.

Results

Of physicians practicing in Hawai‘i who graduated from JABSOM during a 9-year period (1997–2006), 20 were identified as practicing in rural Hawai‘i (Table 1). Nearly half (46%) of physicians from rural backgrounds who stayed in Hawai‘i chose to practice in a rural setting, while only 5% of physicians from non-rural backgrounds chose to practice outside of O‘ahu. Thus, Hawai‘i-schooled physicians from rural settings are nine times more likely to practice in a rural location than those who did not go to high school on a neighbor island. After conducting a chi square test, the correlation between rural background and rural practice was found to be statistically significant (P <.0001). Of the subset of JABSOM physicians in primary care, 15% chose to practice in rural Hawai‘i compared to 8% of non-primary care physicians. After conducting a chi square test there was weak evidence (P=0.09) for an association between rural practice and primary care specialty (Table 2). The low statistical significance of this test is likely linked to the small number of 1993–2009 JABSOM graduates practicing in rural Hawai‘i (12).

Table 1.

Comparison of JABSOM graduates rural origin vs. rural practice location, P < 0.0001^*

	Rural Practice	Non-Rural Practice	Total
Rural Background	12	14	26
Non-Rural Background	8	143	151
Total	20	157	177

^*chi square analysis

Table 2.

Comparison of JABSOM graduates type of practice and rural practice location, P = 0.09^*

	Rural Practice	Non-Rural Practice	Total
Non-Primary Care Specialty	7	86	93
Primary Care	13	71	84
Total	20	157	177

^*chi square analysis

The results of the prospective survey of JABSOM students, after excluding those from or intending to practice outside Hawai‘i, 17% responded that they intended to practice in a rural setting (Table 3). Students from rural backgrounds were 11 times (75% vs 6.7%) more likely to state a rural location as one of their intended places of practice. While students from rural backgrounds made up only 15% of the total respondents, they represented 67% of those with stated intent to practice in rural settings. The correlation between students' rural background and rural intended practice location was found to be statistically significant (P <.0001). Additionally, though less than one fifth of the 105 included respondents intended rural practice, the majority (81%) of the total 166 students polled said they would consider practicing in a rural location. Factors they described as possibly influencing their decision included lifestyle, work environment, and employment opportunities for both themselves and their spouses. Survey participant responses to the question, “What suggestions do you have for recruiting and retaining physicians in Hawai‘i's rural areas?” are listed in Table 4. Participants were not given multiple choice options for this question and were instead asked to give free-form answers. The most common responses were increased compensation and loan repayment opportunities.

Table 3.

Comparison of current JABSOM students rural origin vs. rural practice location intention based on a 2010 survey of JABSOM medical students regarding practice intention, P < 0.0001

	Rural Practice	Non-Rural Practice	Total
Rural Background	12	4	16
Non-Rural Background	6	83	89
Total	18	87	105

^*chi square analysis

Table 4.

Results of a poll of JABSOM students (class of 2010–2013) on suggestions for recruiting and retaining physicians in Hawaii's rural areas. Listed in order of popularity. Of 134 survey participants, there were 91 respondents to this question, providing these 140 responses

Student Suggestions
Increased Compensation (40)
Increased Loan Repayment Opportunities (33)
More rural training experiences in medical school and residency (17)
Better facilities, infrastructure, and support on neighbor islands (16)
Increased exposure to the culture of rural areas via talks and visits from neighbor island physicians (12)
Increased recruitment of medical students from rural areas (11)
Improve schools (primary and secondary education) in rural areas (3)

Discussion

Although all counties of Hawai‘i currently experience physician shortages, the neighbor islands suffer more significant shortages of primary care doctors than O‘ahu, according to 2010 Hawai‘i Physician Workforce Data.¹⁵ A number of prior studies identified being from a rural area as a predictive factor for working in a rural area. Similarly, studies have shown that practicing family medicine is predictive of work in a rural area, particularly when combined with having a rural background.^{2, 16–21}

The results of this study affirm the same is true at JABSOM, as 46% of JABSOM-trained physicians from rural backgrounds have chosen to practice on the neighbor islands in comparison to only 5% of those raised on O‘ahu. Furthermore, both current JABSOM students and past graduates are statistically more likely to practice in rural areas if they are from an island other than O‘ahu. If JABSOM's mission is to train doctors for all of Hawai‘i, it is reasonable to conclude that in order to produce more doctors for the neighbor islands, we should increase recruitment and admission of medical students from locations with the greatest unmet needs for physicians.

A small number of undergraduate medical schools in the United States have published outcomes data on their comprehensive rural health programs designed specifically to increase the supply of rural primary care physicians. The most thoroughly studied of these undergraduate medical programs is the Physician Shortage Area Program (PSAP), started in 1974 at the Jefferson Medical College (JMC) in Philadelphia, Pennsylvania. Research within the PSAP has identified the selective admissions process as the most powerful factor influencing the program's success.⁵ Through admittance of students based largely on their rural background and commitment to family medicine (but who also meet academic and other admission criteria), the PSAP is composed of students dissimilar to those comprising the non-PSAP population at JMC (one study notes that more than two thirds of PSAP graduates were not accepted to any other medical school and probably would not have been accepted under regular admissions to JMC) but the attrition rate and undergraduate and post graduate academic performance of PSAP graduates are nonetheless similar.²² While other successful models differ slightly from the PSAP program in their curriculum and in other areas, they all share a focus on students from rural areas and on the practice of family medicine.^23–25

Studies have compared PSAP graduates to non-PSAP graduates of JMC, and to all medical school graduates in Pennsylvania. The most recent study was conducted in 1999, and evaluated the program over 22 years.⁵ The authors found that compared to non-PSAP graduates, PSAP graduates were three times as likely to practice in a rural area of the United States (34% vs 11%), more than three times as likely to practice in an underserved area (30% vs 9%), four times as likely to practice family medicine (52% vs 13%), and 8.5 times as likely to practice family medicine while also practicing in a rural area (21% vs 2%). When evaluated on a statewide scale, PSAP graduates represent 1% of all graduates from the state but they account for 21% of family physicians practicing in rural areas. Importantly, these physicians are not only more likely to initially practice in rural areas, they are also statistically more likely to stay.²⁶ Alternately, a 2001 study using a logistic model for predictive factors found that of the non-PSAP graduates who grew up in an urban area and did not express an early interest in family practice, only 1.8% went on to become rural primary care physicians. Of matriculants who did not participate in PSAP, non-PSAP graduates who had grown up in a rural area and had a freshman-year plan for family practice were 75% as likely to practice rural primary care as PSAP graduates, thus suggesting that even without a specially designed undergraduate program to support rural primary care, medical schools could increase rural physician numbers simply by admitting more students from rural areas with an interest in primary care.²⁷

Yearly, the PSAP program and five other rural health medical programs produce over 100 physicians likely to practice primary care and to enter rural residency programs, an extremely important resource for rural graduate medical education (GME) programs and rural communities.²⁸ A 2008 review of these programs in Academic Medicine revealed that on average, 64% of program graduates were practicing in populations of < 50,000 and 57% were in rural communities (Table 5).¹⁴ According to this study, if all US medical schools were to expand their class sizes by 30% and include these programs as part of their expansion, they would produce five times as many rural physicians as an expansion without these programs. Because JABSOM is currently expanding its incoming class size, it is an opportune time to consider a specific rural health program as it could increase the supply to Hawai‘i's rural areas much more than an increase in class size alone.

Table 5.

Specialty and Rural Practice Outcomes of Medical School Programs With a Mission to Increase Rural Physicians¹⁴

				Study Outcomes: no (%) of program graduates working in rural communities
Program	Medical School	Year Started	Sample Size	Rural	Population <50,000	Population <25,000
Rural Physician Associate Program (RPAP)	University of Minnesota (UM) Medical School	1971	284 (practicing in MN) 869 (practicing in the United States.)	167 (59) 521 (60)	223 (79)	192 (68)
University of Minnesota Medical School, Duluth (UMD)	University of Minnesota Medical School, Duluth	1972	1264		NA (62)	NA (47)
Upper Peninsula Program (UPP)	Michigan State University College of Human Medicine	1974	28		NA (50)
Physician Shortage Area Program	Jefferson Medical College, Thomas Jefferson University	1974	NA		NA (76)	NA (68)
Rural Medical Education Program (RMED)	State University of New York (SUNY), Upstate Medical University	1989	86		22 (26)
Rural Medical Education Program (RMED)	University of Illinois College of Medicine at Rockford	1993	61	56 (92)
Weighted Average				(57)	(64)	(53)

There are potential weaknesses of the current study that need to be addressed. It should be noted that this analysis does not take into account the number of UH residency graduates who ultimately practice in rural areas of the state, many of whom are graduates of other medical institutions. Furthermore, students currently practicing outside of Hawai‘i were eliminated from the data set, since their practice location is not documented as it is for Hawai‘i physicians. Thus it is possible that there are a large number of rural practicing physicians outside of Hawai‘i. Also, the rural origin of students included in this study was determined by the location of high school zip code, which may not be an accurate lifetime measure. For example, a small number of students from neighbor islands attend high school on O‘ahu but do not lose their ties to the neighbor islands and may be therefore miscategorized in this analysis. Additionally, students may have lived in different areas before attending high school. Finally, in the survey of current students, the response rate was 89% and only included those students present at the colloquia, which may provide a skewed description of student practice location intentions if the students who attended the colloquia were different from the students who did not attend.

Future research should address the location of Hawai‘i-trained medical graduates who practice outside of Hawai‘i as well as continue to study intentions and ultimate practice location of JABSOM students. An outstanding JABSOM program that provides quality education and medical school preparation for disadvantaged students (many of them from rural areas) is the Imi Ho‘ola (those who seek) Program, a one year post-baccalaureate program with automatic admission to JABSOM, started in 1973 and now institutionalized at the medical school. This program has proven to be extremely successful in supporting students from underrepresented backgrounds who might otherwise not attend medical school.

The implications of these findings are significant. Admission policies that do not include students raised in rural areas are unlikely to be successful at mitigating rural physician shortages. However, increasing the number of matriculants from rural areas or with freshman-year plans for family medicine will be both effective and cost-saving, particularly considering the high cost of physician recruitment and retention further along the physician pipeline.²⁹ Besides rural background and stated interest in family medicine, applicants' demonstrated commitment to primary care and underserved populations can be used as a measure of likelihood to practice in rural communities. In addition, rural training increases rate of rural practice and history of working in community-based volunteer programs such as AmeriCorps, Peace Corps, Teach For America, correlate with retention in rural areas.³⁰ Studies need to be carried out to determine if there is a positive correlation between experience prior to medical school and eventual choice of specialty.

In conclusion, although monies are scarce in these financially difficult times, reallocation of funds toward restructuring, supporting, and strengthening programs already in existence can have profound impacts on health care access for Hawai‘i's most underserved populations. The decreased burden of a program that operates within an existing medical school, coupled with the ease with which admission criteria can be adjusted to achieve program objectives, indicates a robustness that is necessary when approaching the physician supply and distribution problem. Furthermore, implementation of a rural health track could be facilitated through the Hawai‘i/Pacific Basin Area Health Education Center (AHEC), the University of Hawai‘i Department of Family Medicine, the JABSOM Office of Medical Education, other existing Title VII programs, and the model education programs already in place that could be used as guides for smooth implementation. Due to its low cost, high effectiveness, lack of need for regulatory overhaul, and ease of implementation, this policy option is politically feasible. While these changes may not provide an immediate resolution of Hawai‘i's physician shortages, they are likely to create cost-effective and sustainable solutions for the health care needs of this state.

Conflict of Interest

None of the authors report any conflict of interest.