A Brief Survey to Identify Priorities for Improving Clinician Recruitment and Retention: Results from Hawai‘i Island Physicians

Abstract

Background

In light of the documented physician shortage on Hawai‘i Island, the Hawai‘i Clinician Recruitment and Retention survey was designed and implemented to assess perceptions of quality of life and the work environment among clinicians on Hawai‘i Island and to identify aspects of the environment on Hawai‘i Island that predict responses to questions regarding recruitment and retention.

Methods

The respondents were 127 Hawai‘i Island clinicians, specifically 96 physicians, 15 nurses, five pharmacists, four physician assistants, two social workers, and five “other” healthcare workers. The internal reliability of the survey was high (alpha=.91) and its convergent validity was supported by the significant correlation of item total scores with anchor items that measured overall ratings of the environment and likelihood of recruitment and retention. Given the small number of non-physician clinicians responding, descriptive analyses included only physicians. Physicians who indicated they plan to retire within 5 years were excluded from the correlation analyses to focus on patterns within the target group for retention.

Results

Overall, results indicate that, while the majority of physicians who relocated to Hawai‘i Island did so primarily for the quality of life, the best predictors of retention are financial sustainability, professional opportunities, community support, and access to good K-12 schools. Survey results also indicate that Hawai‘i Island will lose 32% of its current physicians within the next five years due to retirement or other causes.

Discussion

These findings indicate that increased urgency to find solutions is warranted.

Introduction

Physician shortages in rural areas continue to be well documented.¹ Interventions have focused on strategies to recruit clinicians/trainees from rural areas and/or expose them to rural practice during their training.² Using results from a qualitative study using a semi-structured interview design and incorporating previous research findings, Hancock and colleagues have proposed a model of physician recruitment and retention.³ This 2009 model begins with the importance of exposure to rural environments and, absent that, factors that increase the propensity to practice in rural settings. Once practicing in a rural setting, according to this model, physicians are likely to stay to the extent that they have a sense of place, comfort, outdoor opportunities, community, efficacy, and freedom. In contrast, those who experience greater financial trouble, lack of services, limited social network, or desire a career change are more likely to leave. The model provides a useful framework for designing and testing recruitment and retention interventions.

On a local level, projections indicate that the physician shortage in Hawai‘i is significant and growing.⁴ The physician shortage is more severe in rural counties. Applying Hancock's model to this unique setting, the geographic isolation and relatively high cost of living are likely to be significant barriers to effective recruitment and retention. The unique quality of life factors in rural Hawai‘i may offset these barriers. On rural Hawai‘i Island, the largest county in Hawai‘i with the lowest population density, physician shortages have been well-documented⁵ and significant efforts are underway to better understand and address this problem. In particular, the launch of the University of Hawai‘i School of Medicine Rural Family Practice Residency Program at Hilo Medical Center is a critical, evidence-based approach to improving recruitment of physicians to Hawai‘i Island. This approach is supported by research indicating that exposure to rural settings during clinical training, particularly longitudinal rotations, are good predictors of retention in rural areas.^3,6,7

While there is strong local support for the residency program, there is also significant interest in finding other solutions to the shortage. The Hawai‘i Island Healthcare Alliance is a diverse group of stakeholders focused on achieving a vision where Hawai‘i Island residents lead healthy and productive lives, make healthy choices, and have access to quality healthcare. The Alliance Recruitment and Retention Committee has been focused on addressing the broad range of workforce issues that impact the achievement of this vision. Discussion among committee members led to the decision to design and administer a survey to facilitate the prioritization of efforts to improve clinician recruitment and retention. Therefore, the purpose of this research was to assess perceptions of quality of life and the work environment among clinicians on Hawai‘i Island. The goal was to identify aspects of the environment on Hawai‘i Island that predict retention and that could be used to recruit clinicians to work there. The Hawai‘i Clinician Recruitment and Retention survey was designed to measure clinician ratings of the general environment as well as aspects of the environment specific to healthcare and to shed light on the applicability of Hancock's model of recruitment and retention to the unique factors of practicing on Hawai‘i Island.

The survey included an item response format that has been shown by Ware and Hays to produce more sensitive measures of satisfaction.⁸ Many surveys attempting to assess perceptions of quality yield highly skewed response distributions with low variability.⁶ In their study, Ware and Hays compared two response formats: a five-point scale offering the response options of “excellent, very good, good, fair, and poor” and a six-point scale with response options of “extremely satisfied, very satisfied, somewhat satisfied, neither satisfied nor dissatisfi ed, somewhat dissatisfied, and very dissatisfied.” The former format yielded less skewed results with greater variability across settings, producing a more sensitive instrument for measuring satisfaction. Furthermore, the “excellent, very good, good, fair, poor” response format better predicted behaviors and behavioral intent. Thus, the recruitment and retention survey used this optimal response format (5 = excellent, 4 = very good, 3 = good, 2 = fair, 1 = poor) to minimize positive response bias and optimize variability and predictive validity.

Methods

The Hawai‘i Clinician Recruitment and Retention survey was designed to be brief enough to be practical and user-friendly but thorough enough to cover key environmental, community, and work-related factors that might be important to practicing clinicians based on a review of the literature. The draft survey was then distributed to Hawai‘i Island Healthcare Alliance leaders for feedback and suggestions were incorporated. The final survey was transferred to surveymonkey.com format for administration.

In addition to identifying clinician type, the survey included basic information about the clinician's practice setting, location of clinical training, location of previous clinical experience, reason for relocating to Hawai‘i Island, years in practice, and plans to retire (shown in Table 1). Clinicians were asked to indicate their ratings of specific aspects of living and working on Hawai‘i Island (shown in Table 2), as well as their overall ratings and their intentions to continue practicing and refer colleagues to practice there. These overall ratings and behavioral intent questions were the “anchor” items included to identify the best predictors of overall satisfaction, recruitment (as measured by the question “Would you recommend Hawai‘i Island to your healthcare colleagues who are interested in relocating?”), and retention (as measured by the question “Do you plan to continue practicing on Hawai‘i Island for the next 5 years?”). The recruitment and retention questions had a response format of “4 = Yes, definitely, 3 = yes probably, 2 = no probably not, and 1=no definitely not” as used in the Ware and Hays study.⁸

Table 1.

Physician sample characteristics

Question (n)	Response	Percent
Primary clinical area (95)	Primary care Specialty care	52.6% 47.4%
Primary practice setting (95)	Outpatient, owner Outpatient, employee Hospital Other	32.6% 23.2% 24.2% 20.0%
Plan to retire in the next 5 years (96)	Yes No	19.8% 80.2%
No. of years practicing on Hawai‘i Island (96)	Less than 1 year 1–5 years 6–10 years 11–20 years More than 20 years	5.2% 32.3% 19.8% 21.9% 20.8%
Location of previous clinical experience (excluding training) (96)	none (all experience on Hawai‘i Island) relocated from another island in Hawai‘i relocated from another state relocated from another country	28.1% 12.5% 56.3% 3.1%
Setting of previous clinical experience for those who relocated to Hawai‘i Island (69)	Primarily urban area Primarily rural area	72.5% 27.5%
For those who relocated to Hawai‘i Island, primary reason (69)	Closer to family Improved earnings Quality of life Other	14.5% 2.9% 62.3% 20.3%
Primary location of advanced clinical training (96)	Hawai‘i Island Another island in Hawai‘i Another state Another country	3.1% 11.5% 85.4% 0%
If not Hawai‘i Island, setting of advanced clinical training (93)	Primarily urban area Primarily rural area	92.5% 7.5%
District where you currently live (94)	Hamakua Hilo Kau North Kohala South Kohala Kona Puna	3.2% 45.7% 2.1% 5.3% 13.8% 21.3% 8.5%
District where you currently work (94)	Hamakua Hilo Kau North Kohala South Kohala Kona Puna	1.1% 51.1% 2.1% 3.2% 17.0% 20.2% 5.3%

Table 2.

Survey item means (standard deviations) and correlations with anchor items (and 2-tailed significance), excluding physicians planning to retire in the next 5 years; strongest predictors of recruitment and retention are highlighted

				Correlation coefficient with anchor items (2-tailed significance)
Survey item (n)	Mean (SD)	Overall rating of the healthcare environment	Overall rating of the community environment	Recommend Hawai‘i Island to colleagues?	Plan to continue practicing on Hawai‘i Island?
Access to cultural activities (77)	2.83 (.979)	.091 (.353)	.238* (.016)	.128 (.202)	.172 (.090)
Opportunities to enjoy outdoor activities (77)	4.43 (.802)	.237* (.020)	.226* (.029)	.141 (.178)	.091 (.389)
Access to shopping and other services (77)	2.29 (.792)	.264** (.008)	.166 (.100)	.238* (.020)	.222* (.032)
Cost of living (77)	2.08 (.914)	.183 (.064)	.171 (.087)	.275** (.007)	.244* (.017)
Access to good K-12 schools (72)	2.15 (1.083)	.372** (.000)	.226* (.026)	.437** (.000)	.282** (.006)
Social support in the community (75)	2.81 (1.023)	.373** (.000)	.443** (.000)	.338** (.001)	.268** (.008)
Support from the business community (74)	2.59 (1.072)	.435** (.000)	.427** (.000)	.354** (.000)	.307** (.002)
Support from elected officials (75)	2.21 (1.017)	.433** (.000)	.294** (.003)	.246* (.015)	.122 (.232)
Amount of personal time away from work (75)	2.81 (1.182)	.199* (.039)	.307** (.002)	.097 (.333)	.027 (.786)
Earnings from main practice (77)	2.74 (1.093)	.473** (.000)	.323** (.001)	.424** (.000)	.398** (.000)
Stability of earnings (75)	3.13 (1.082)	.359** (.000)	.317** (.001)	.400** (.000)	.340** (.001)
Quality of healthcare facilities and equipment (76)	2.80 (.938)	.431** (.000)	.332** (.001)	.260* (.010)	.233* (.022)
Quality of health information technology (77)	2.81 (1.077)	.296** (.002)	.278** (.004)	.234* (.018)	.211* (.034)
Quality of healthcare workforce (76)	2.88 (1.070)	.429** (.000)	.437** (.000)	.298** (.003)	.231* (.021)
Access to Continuing Education opportunities (76)	2.59 (.941)	.305** (.002)	.259** (.010)	.230* (.023)	.086 (.400)
Opportunities to achieve professional goals (77)	2.96 (1.019)	.575** (.000)	.431** (.000)	.421** (.000)	.389** (.000)
Access to current scientific literature (77)	3.21 (1.207)	.371** (.000)	.348** (.000)	.286** (.004)	.259** (.009)
Access to consultants for referrals (77)	2.06 (.978)	.336** (.001)	.243* (.015)	.198 (.050)	.126 (.217)
Degree of acceptance by other clinicians in the community (77)	3.56 (1.019)	.370** (.000)	.460** (.000)	.359** (.000)	.247* (.014)
Quality of clinician-patient relationships (77)	3.88 (.888)	.294** (.003)	.360** (.000)	.147 (.150)	.226* (.027)
Cooperation among healthcare providers (76)	3.34 (.917)	.646** (.000)	.413** (.000)	.385** (.000)	.347** (.001)

^*Correlation is significant at the 0.05 level (2-tailed),

^**Correlation is significant at the 0.01 level (2-tailed).

This study was determined by the University of Hawai‘i Committee on Human Studies to be exempt from IRB review and all procedures with regard to informed consent and data confidentiality were followed. In July 2010, key healthcare leaders, including senior leaders of Hawai‘i Island hospitals, federally qualified health clinics and physician groups were contacted regarding the distribution of the survey on behalf of the Hawai‘i Island Healthcare Alliance. These leaders were asked to send a standardized email with link to the survey at surveymonkey.com to the clinicians working in their respective organizations with the request that the survey be completed within three weeks. Before the survey deadline, these leaders were sent two reminders, including an update on the number of clinicians that had responded to date. The intent was to obtain survey responses from all clinician types, including physicians, nurses, pharmacists and others. Because the survey invitation was sent to clinicians through these healthcare leaders, response rates based on those invited to complete the survey are not known.

From this convenience sample, descriptive statistics were obtained to identify key aspects of the clinician sample as well as mean scores on each item. Analyses of the reliability and preliminary validity of the Hawai‘i Clinician Recruitment and Retention survey were performed. Correlation analyses were performed to identify predictors of clinician intention to continue practicing on Hawai‘i Island and to recommend Hawai‘i Island as a place for other clinicians to practice. Given that the survey produced ordinal data, Kendall's tau-b was selected to obtain correlations using SPSS software (PASW Statistics 18).

Results

The respondents were 127 Hawai‘i Island clinicians, including 96 physicians, 15 nurses, five pharmacists, four physician assistants, two social workers, and five “other” healthcare workers. The internal reliability of the Hawai‘i Clinician Recruitment and Retention survey, as measured by Cronbach's alpha, was high (alpha = .91). Preliminary convergent validity of the instrument was supported by the finding that total scores for the items were significantly correlated with the scores for the four anchor items, P <.001 (r = .77 for “overall rating of the healthcare environment,” r = .73 for “overall rating of the community environment,” r = .68 for “would you recommend Hawai‘i Island to your healthcare colleagues who are interested in relocating?” and r = .38 for “do you plan to continue practicing on Hawai‘i Island for the next 5 years?”). Given the small numbers of non-physician clinicians responding, the remaining descriptive analyses include only the 96 physicians who completed the survey to allow greater certainty about applying results specifically to physicians. Based on an estimate of 310 practicing physicians on Hawai‘i Island,⁵ 31% of physicians completed the survey. Descriptive statistics for this physician sample are shown in Table 1. Table 2 presents mean ratings for each survey item, standard deviations and correlation analyses designed to identify priorities for improving recruitment and retention of physicians. The physicians who reported that they planned to retire in the next five years were omitted from these analyses to maximize relevance to those who are retention targets, leaving a sample of 77 physicians. Correlation coefficients indicate the strength with which answers to specific item ratings predict answers regarding recruitment and retention. Two-tailed significance tests were selected due to the exploratory nature of this survey such that hypotheses about whether significant correlations would be positive or negative were not stated a priori. The strongest item predictors of recruitment and retention, defined as items that significantly predict both recruitment and retention questions at the P ≤ 0.01 level, were all positive correlations and are highlighted in Table 2.

Discussion

The majority of physicians who relocated to Hawai‘i Island from a different island in Hawai‘i, another state, or another country, did so primarily for the quality of life; very few relocated to improve earnings. However, the best predictors of retention were financial sustainability, professional opportunities, community support, and access to good K-12 schools. Not surprisingly, these were also the strongest predictors of physician recruitment of colleagues to Hawai‘i Island. This means that physicians who gave high ratings in these areas were significantly more likely to recommend Hawai‘i Island to colleagues interested in relocation and significantly more likely to report they plan to continue practicing there. Likewise, those who gave low ratings in these areas were significantly less likely to recommend Hawai‘i Island or to plan to continue practicing there. These findings support elements of Hancock's model. However, inconsistent with Hancock's model, ratings of opportunities to enjoy outdoor activities and amount of time away from work were not significant predictors of recruitment and retention.

While cause and effect were not tested, this simple, cost-efficient survey methodology allows the identification of priorities for interventions based on items that best predict recruitment and retention questions. For physicians on Hawai‘i Island, the survey results suggest that improvements in financial sustainability, professional opportunities, community support, and access to good K-12 schools are most likely to improve physician recruitment and retention and, thus, should be priorities. The reliability and preliminary validity of the Hawai‘i Clinician Recruitment and Retention survey are strong, suggesting that it could be used to track changes over time and adapted for use in other communities to identify and compare priorities for improvement.

It is noteworthy that twenty percent of physicians reported that they planned to retire within five years. Of those who were not planning to retire, 12 (16%) indicated that they do not plan to continue practicing on Hawai‘i Island for the next five years. Combined, this suggests that Hawai‘i Island will lose 32% of its current physicians within the next five years. These findings indicate that increased urgency to find solutions is warranted. While efforts to improve the pipeline of medical students and residents with extended training on Hawai‘i Island are greatly needed and should continue, it is unlikely that these efforts will be sufficient to address the current and projected shortage.

Models of care that result in physicians and other clinicians functioning “at the top of their license” are likely critical to addressing the problem of physician recruitment and retention. This means the formation of teams in which physicians delegate tasks to other healthcare workers who can perform the tasks more cost-effectively. Such models extend the capability of physicians to optimally manage a greater volume of patients. From an economic perspective, this allows a more efficient use of health care dollars and an opportunity for all clinicians to be paid at a license-appropriate level. From a professional perspective, this allows physicians and other clinicians to do the work they were uniquely trained to do. Analyses also indicate that community support and infrastructure are important predictors of recruitment and retention, so these avenues should be further explored.

A potential limitation of this study is that the survey was distributed via leaders of hospitals, clinics, and physician groups and associations, which could have resulted in under-representation among solo, unaffiliated practices. It is also possible that the responders differed from non-responders in some way that biased the results. Furthermore, while the administration of this survey produced a sufficient response from physicians, there were insufficient responses from other clinicians for meaningful analysis. Future research should attempt to identify recruitment and retention priorities for other clinicians. Additional research is also needed to causally test the predictors identified in this survey. Finally, follow up surveys should be implemented to track progress of interventions designed to improve recruitment and retention on Hawai‘i Island.

Conflict of Interest

The author has identified no conflicts of interest to disclose related to the content of this manuscript.