| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Chapter 6.4 |
Key words: practice plans, dental care access, evaluation, quasi-experimental research design, data triangulation
This chapter considers the practice plans of graduating dental school seniors in the Pipeline, Profession, and Practice: Community-Based Dental Education program, drawing on data collected and analyzed from multiple data sources and stakeholder groups. To address the evaluation questions, we analyzed data collected exclusively in the Pipeline program schools (multiple stakeholder interviews and a faculty survey) and data collected in all accredited dental schools in the United States (using the American Dental Education Association [ADEA] annual survey of dental school seniors as well as contextual variables). Chapter 4 of this report describes the methods used to collect and analyze each of these data sources.1 Since multiple sources of data were used to address the evaluation questions, we triangulated the evaluation results when possible.
In comparing the dental schools, we will sometimes distinguish among three groups: 1) the National Pipeline schools, which are the ten dental schools across the United States that received funding to conduct Pipeline programs from the Robert Wood Johnson Foundation; 2) the California Pipeline schools, which are the additional four dental schools in California that received funding to conduct Pipeline programs from The California Endowment and the University of California, San Francisco, which received funding from both foundations; and 3) non-Pipeline schools, the remaining thirty-seven U.S. dental schools (out of a total fifty-two accredited schools included in the study) that were not part of the Pipeline program. References to "Pipeline schools" in general include all fifteen dental schools that conducted programs.
This analysis is organized around three major evaluation questions. After discussing those, we examine the policy, delivery system, and university/dental school recommendations that can influence the practice plans of dental school seniors toward providing care to greater percentages of underserved patients.
 |
1. What are facilitators and barriers for students to provide care for underserved patients? |
|---|
 TopAuthor information 1. What are facilitators... 2. Are students planning...3. What factors influence...Discussion and Study limitationsReferences |
|---|
We analyzed a series of responses to site visit interview questions about facilitators and barriers for graduating seniors to provide care to underserved subgroups. The first question was the following: "To what extent does the school of dentistry help graduating seniors learn about opportunities to practice in settings that provide care to underserved populations (e.g., diverse racial/ethnic groups, special needs populations)?" Forty-one percent of academic and 33 percent of clinical administrators/faculty (versus only 14 percent of fourth-year students) reported the school communicated opportunities "well or very well." In contrast, about 43 percent of fourth-year students (compared to 24 percent of academic and 17 percent of clinical administrators/faculty) reported the school communicated opportunities "not well or not very well." Essentially, the students interviewed were far less convinced about the school efforts to communicate information to their students regarding opportunities to practice in underserved areas.
A second question was about sources of information for students. The stakeholders reported that students obtained information from a wide variety of sources that heightened their awareness about providing care to underserved patients. Table 6.4.1
shows "dental school" sources were mentioned most often by stakeholders (60 percent) as a source of information about practice opportunities. These included scholarships, announcements, noon presentations, and management courses. "Community rotations" were a close second, with 59 percent of respondents mentioning internships, extramural rotations, and general practice residency (GPR) programs. Fifty-one percent of the interviewees mentioned various dental and other organizations such as the U.S. Air Force, the military in general, Indian Health Service, state primary care associations, community health centers (CHCs), Special Care Dentistry, National Institute of Dental and Craniofacial Research (NIDCR), Rural/Underserved Opportunities Program (RUOP), and Regional Initiatives in Dental Education Program (RIDE). None of the respondents mentioned either the state or local health departments. Thirty-nine percent mentioned programs providing financial incentives including tuition reimbursement, scholarships, and loan repayment in exchange for service, and 11 percent mentioned career coaching opportunities such as alumni networks, career placement, and mentoring programs. About 9 percent of the respondents in the fourteen schools reported their school did not have an organized system for students to learn about opportunities to practice in settings that provide care to predominantly underserved patients.
|
. Sixty-one percent of the respondents referred to financial concerns such as high educational debt, low compensation by safety net organizations, and limitation in government funding for loan forgiveness programs as major barriers for students to practice in underserved settings. The next most frequently mentioned barriers were personal preferences against providing care to underserved patients (23 percent) and challenges related to providing care to the patient population (17 percent).
|
shows several factors reported by stakeholders as facilitators for dental students to practice in underserved settings. Among these factors, sufficient financial support and enhanced dental school communication were most often mentioned (21 percent). Next, 19 percent stated that personal preferences (e.g., motivation to work with the underserved) and predispositions (e.g., students from rural/underserved areas willing to go back and serve their communities) were influential factors.
|
shows that the largest percentage of respondents (56 percent) suggested an increase in financial support for students (waived or reduced tuition, scholarships, loan forgiveness, and a better stipend or salary). Forty-five percent believed improved and enhanced communication in programs and curriculum should make a difference in encouraging more dental students to practice in underserved settings. About one-third suggested greater exposure to underserved patients (31 percent) and increasing diversity among students and recruiting students predisposed to care for under-served patients (28 percent). Finally, 3 percent recommended improving the delivery system and practice setting to deliver care to the underserved by creating a national system for tracking graduates working in underserved areas and providing transportation for underserved patients to receive dental care.
|
shows results separately for the five California Pipeline and nine National Pipeline schools. The chart shows the percentage of faculty reporting a "major" barrier (as opposed to "no, "minor," or "moderate" barrier). Confirming the findings from the site visits, 86 percent of faculty in California Pipeline schools and 75 percent of faculty in National Pipeline schools reported financial debt as the major barrier for graduating seniors to provide care to the underserved. The next major barrier was "low reimbursement from public insurance programs," with 75 percent of California Pipeline and 66 percent of National Pipeline school faculty perceiving this as a major barrier. Less than half of faculty thought the other barriers listed in Figure 6.4.1 were major barriers.
|
shows the percentage of faculty reporting "very often" or "often" for the listed sources (as opposed to "not very often or "never"). As the most important source, 80 percent of California Pipeline and 89 percent of National Pipeline school faculty reported students often/very often learn about opportunities during "extramural clinical rotations." Other sources of information described frequently were dental safety net organizations, such as the National Health Service Corps and Indian Health Service, faculty advisors, and job bank.
|
reports the dental schools’ targeted efforts to improve access or culture related to providing dental care, indicating the percentage of faculty who strongly agreed or agreed with the statements in this category (as opposed to those who were neutral, disagreed, or strongly disagreed). More than 90 percent of the faculty members in Pipeline dental schools strongly agreed/agreed that their dental school "has a targeted effort to provide care to URMs [underrepresented minorities]," as did 85–90 percent that the dental school "has a targeted effort to provide students with opportunities to care for URMs." Large percentages also agreed with the rest of the statements in Figure 6.4.3 regarding the culture of the Pipeline dental schools emphasizing technical skills, patient-centered care, and community service, although among these options a culture emphasizing community service showed diminishing agreement.
|
|
2. Are students planning to provide care to more underserved patients as a result of the Pipeline program? |
|---|
|
TopAuthor information1. What are facilitators...2. Are students planning...3. What factors influence...Discussion and Study limitationsReferences |
|---|
In collaboration with ADEA, the NET developed questionnaire items to measure components of community-based dental education (CBDE) and student characteristics. Fifty-two of the fifty-five accredited U.S. dental schools returned surveys in 2003; we selected these same schools from the 2007 survey for our impact analysis. In 2002–03, the overall student response among schools participating in the survey was 86 percent, and in 2006–07, the overall response rate was comparable at 85 percent. Because the University of Maryland/Baltimore College of Dental Surgery and the Boston University Goldman School of Dental Medicine had response rates of less than 5 percent in the 2007 survey, we used data from the previous year (2005–06) for these two schools.
The logic model used in our baseline study was applied again in this chapter, this time to study the impact of the Pipeline program.3 The model suggests the contextual environment, CBDE, and student characteristics influence a dental student’s decision to care for URM patients. Our baseline article details variables in the logic model, along with their definitions and distributions.3 For the dependent variable, seniors responded to the following ADEA question (Q30a): "When you enter practice, about what percent of your underserved patients do you expect will be from underserved racial/ethnic minority populations?" Response categories were 0 percent; 1–10 percent; 11–24 percent; 25–50 percent; greater than 50 percent. For our analysis we constructed a dichotomous dependent variable with "0" indicating 0 percent to 24 percent and "1" indicating 25 percent or greater underserved patients.
We constructed an intervention variable to measure Pipeline program status, using the three categories described in the introduction (National Pipeline schools, California Pipeline schools, and non-Pipeline schools). The University of California, San Francisco (UCSF) was funded by both foundations; we included UCSF in the California Pipeline category because all schools in the state received Pipeline program funding and engaged in collaborative statewide recruitment and health policy initiatives. Appendix 1 of this report provides the variables, variable definitions, and distribution for each variable.4 In addition to describing the data collection methods, Chapter 4 of this report describes the statistical methods used to conduct the impact analysis reported here.1
First, we examined data reported on the dependent variable by student cohorts over time (2003 through 2007), by Pipeline and non-Pipeline schools. Table 6.4.5 shows the distribution of the dichotomous dependent variable for 2003 and 2007, stratified by type of school (California Pipeline, National Pipeline, all Pipeline, non-Pipeline, and total). In general, 20–27 percent of the students expected to serve 25 percent or more underserved patients, regardless of school type or year. There were no significant changes from 2003 to 2007 regardless of school type, suggesting no impact of the Pipeline program on practice decisions. Table 6.4.6 shows bivariate associations among candidate independent variables categorized as contextual, CBDE, and student characteristics by the dichotomous dependent variable measuring the percentages of underserved patients seniors expected to serve upon graduation (less than 25 percent or 25 percent or more).
|
|
). In order to produce a more parsimonious model, a two-step method was used to derive the final multiple variable model. To conduct the multiple variable analysis, we examined the correlation matrix of our outcome with candidate variables and selected those that were correlated at the p=0.20 significance level or less. Then, we regressed the outcome on these selected predictors using a best-subset regression approach. The best-subset regression is a method used to select the predictor variables that are used in a final multivariate model.*
|
To examine the impact of the Pipeline program, Table 6.4.7
shows the five intervention indicators measuring the following: 1) school variable—odds that Pipeline (California, National, or both) differed from non-Pipeline schools in 2003; 2) year variable—odds of change within non-Pipeline schools from 2003 to 2007; 3) school x year—indicates the difference in the rate of change between non-Pipeline and Pipeline schools from 2003 through 2007 (not interpretable as an odds ratio); 4) year + (school x year)—a constructed variable that indicates the odds of change within Pipeline schools from 2003 to 2007; and 5) school + (school x year)—a constructed variable that indicates the odds that Pipeline schools differed from non-Pipeline schools in 2007.
The results of the impact analysis reported on Table 6.4.7 show that, for California Pipeline schools, the odds of a student serving 25 percent or more underserved patients is greater in 2003 than for a non-Pipeline school student that same year (school, OR=1.49). We also see that, for California Pipeline schools, the odds of serving 25 percent or more underserved patients is significantly lower in 2007 compared to 2003 (year + school x year, OR=0.62). For the National Pipeline schools and Pipeline schools as a whole, however, no significant differences were found between Pipeline vs. non-Pipeline schools, nor were there any changes within these schools from 2003 to 2007.
|
3. What factors influence a student’s decision to provide care to underserved patients? |
|---|
|
TopAuthor information1. What are facilitators...2. Are students planning...3. What factors influence...Discussion and Study limitationsReferences |
|---|
shows bivariate associations of students’ decisions to provide care to greater percentages of underserved patients with contextual, CBDE, and student characteristics in 2003 and 2007. Regarding contextual variables, students who expected to serve more than 25 percent underserved patients after graduation were significantly more likely to be in states with a higher mean percentage of URM state legislators (2003 only), states with no adult Medicaid coverage (2007 only), counties with a larger URM population, counties with greater percentages of the population below 200 percent of the federal poverty level (2007 only), counties with a larger number of federally qualified health centers (FQHCs) (2003 only), California Pipeline schools (2003 only), private schools (2003 only), schools with a larger percentage of URM students (2007), and schools with lower first-year expenses (2007 only). Regarding CBDE, students who expected to serve more underserved patients after graduation were significantly more likely to report the importance of the following: high school counselor in pursuing dentistry, recruitment by a dental school, and participating in a pre- or postbaccalaureate dental program. Additionally, students who perceived the time devoted to cultural competence in school was inadequate (2003 only), who felt prepared to provide care for racially and culturally diverse groups (2007 only), who felt extramural clinical experiences influenced practice location plans, who expected to spend more weeks in extramural rotations (2003 only), and who planned to work in a community clinic or government service were more likely to report plans to provide care to underserved minorities.
For student characteristics, students who expected to provide care for more underserved patients after graduation were significantly more likely to be female, twenty-nine years of age and older, URM and Asian/PI, not married, from a lower-income family, participating in a loan repayment program (2003 only), having a father who did not attend college, and having a higher service orientation score, lower entrepreneurial score, and higher social consciousness score. Student characteristics strongly reflect the differences in student attitudes towards providing care to underserved patients, regardless of year.
Table 6.4.7 provides a more stringent test for the significant factors associated with student decisions by controlling for all the other variables in the multivariate model. Significant contextual predictors associated with the odds that students plan to provide care to more underserved patients included residing in states with no adult Medicaid coverage, counties with a greater percentage of the population below 200 percent of the federal poverty level, and counties with a greater dentist per 10,000 population ratio (California Pipeline only).
Regarding CBDE characteristics, the following predictors were significantly associated with the odds of providing more care to underserved patients: importance of workforce supply and demand in selecting dentistry as a career (California Pipeline only), inadequacy of time devoted to cultural competence (National Pipeline and all-Pipeline only), preparedness to provide care to diverse groups (all-Pipeline only), extramural clinical experiences that influenced practice plan locations, and plans to serve in a community clinic or in government service after graduation.
Significant student characteristics associated with the odds of providing care to more underserved patients included the following: female gender, age twenty-nine and up (National Pipeline only), URMs, Asian/PIs, unmarried (California Pipeline only and all-Pipeline only), parental income $50,000 or less, educational debt upon graduation greater than $168,000 (National Pipeline and all-Pipeline only), a higher service orientation score, a lower entrepreneurial score, and a higher socially conscious score. Additionally, we conducted a similar analysis taking out the three minority schools: Meharry Medical College, Howard University, and the University of Puerto Rico. The results (not shown) were identical in terms of significant intervention and student characteristics. The significant contextual and CBDE predictors were similar as well, providing evidence of the robustness of the final model.
|
Discussion and Study Limitations |
|---|
|
TopAuthor information1. What are facilitators...2. Are students planning...3. What factors influence...Discussion and Study limitationsReferences |
|---|
Regarding the first evaluation question, we investigated the facilitating factors and barriers to providing care to underserved patients upon graduation. We used site visit interviews and faculty survey data to address this question. From the site visit interviews, we found discordant views among the dental school seniors, who were far less convinced about the dental schools’ efforts to communicate opportunities to practice in underserved areas, compared to administrators and faculty members. From the faculty perspective in both California Pipeline and National Pipeline schools, students learned about opportunities while in dental school primarily through the extramural clinical rotations and safety net organizations and to a lesser extent (listed in descending order) from faculty advisors, job bank, didactic courses, mentor program, and alumni network.
Not surprisingly, the vast majority of stakeholders concurred that financial concerns, such as high educational debt, low compensation by safety net organizations, and low reimbursement for public insurance programs, were the primary barriers to graduates’ providing dental care to the underserved. About 20 percent of the respondents believed that personal preferences of students were an important determinant of practice decisions, which is a predisposing factor worthy of heightened emphasis in dental school recruitment practices, over and above recruiting students based on personal characteristics, such as race/ethnicity and income. Finally, stakeholders interviewed at the Pipeline schools recommended various approaches for increasing dental care to disadvantaged populations; these included financial support, improved dental school communications and curriculum, improved recruitment practices, and a national tracking system for following dental school graduates’ career paths.
To respond to the second evaluation question, we used multiple variable analysis shown on Table 6.4.7 (without controls), which revealed the Pipeline program did not have a short-term impact on student practice plans in the five-year period (2002–07). More specifically, the results show successive cohorts of students were not planning to provide care to greater percentages of underserved patients as a result of the Pipeline program. Valid financial barriers remain as the primary reason why students fail to practice in community clinics and safety net organizations upon graduation (e.g., exorbitant educational debt, limited funding for loan repayment programs). This is disappointing but probably not completely unexpected. Additionally, according to students interviewed, most did not feel the dental school promoted opportunities for providing care to underserved populations, and the faculty survey responses show that, when it comes to school culture, technical skills and patient-centered care were more strongly valued than community service.
The final evaluation question investigated significant factors influencing the students’ practice plans upon graduation. Specific contextual, CBDE, and student characteristics were found to be significant determinants of practice plans. Among the mutable factors, several CBDE program components were shown to be major correlates of practice plans (adequacy of the time devoted to cultural competence, preparedness to provide care to diverse groups, extramural clinical experiences that influenced practice plan locations, and plans to serve in a community clinic or in government service after graduation). These program components will need to be carefully considered and possibly incorporated by dental schools to develop and sustain Pipeline program components that address the dental care access crisis for underserved patients.
Additionally, specific student characteristics associated with practice plans to provide care to underserved patients reach far beyond the original recruitment program emphasized by Pipeline schools (i.e., URM race/ethnicity and lower income). These include female gender, older age, Asian-Pacific Islander race/ethnicity, unmarried, higher educational debt, and scores on several attitude scales (higher service orientation, higher social consciousness, and lower entrepreneurial scores). Many of the Pipeline dental schools revised their recruitment protocols as part of the five-year demonstration project. Findings from this longitudinal evaluation can be used to inform recruitment and selection practices that will yield greater numbers of students predisposed to address dental care access among vulnerable patient populations.
In terms of study limitations, we applied a multiple case study design to analyze qualitative site visit interview and faculty survey data collected in the Pipeline (intervention) schools. This enabled us to identify some of the generalizable barriers and facilitating factors for students to provide greater volume of care to the underserved. This evaluation design is strengthened by examining data from multiple schools and using a uniform data collection method and coding of survey data using multiple raters. Additionally, we used a quasi-experimental design to investigate the impact of the Pipeline program on the practice plans of dental school seniors. Our results are strengthened by having baseline measures, longitudinal data, and a non-equivalent comparison group (non-Pipeline schools). The greatest threat to this quasi-experimental design is the selection threat to validity, which is always problematic when randomized control trials (RCTs) and experimental design are not feasible or practical. Our multiple variable analyses enabled us to statistically control for differences in intervention and comparison groups, but weaknesses still exist because variables we failed to measure might have a significant effect on the outcome and that would bias our analysis. Other weaknesses in the study should be acknowledged. The Pipeline program did not include an explicit intervention component aimed at directly changing policy, financing, and/or individual practice decisions of dental school seniors. Nevertheless, we viewed practice plans as a longer-term outcome with a great potential for improving dental care access. Lastly, evaluation critiques have suggested the ADEA data are limited because seniors might not be credible in their reports, but this is purely anecdotal information that has never been scientifically validated. Additionally, our study results are fortified by using data collected and triangulated from multiple data sources and stakeholder groups.
The ADEA data are limited in that the questions address intended practice decisions rather than actual career choices made after graduation. However, our analyses yielded findings consistent with prior research. For example, we found URM students are more likely to report plans to provide care to underserved patients upon graduation. Prior studies have observed that Latino dental graduates are more than twice as likely to practice in high Latino zip codes, compared to non-Latino dental graduates.5 African American students, in comparison to whites, are also more likely to practice in underserved communities, provide care to uninsured and Medicaid beneficiaries, and continue caring for underserved populations after participation in the National Health Service Corps.6,7 Therefore, the practice intentions of dental school seniors seem to be an immediate antecedent of behavior—in our case, the decision to serve underserved patients. Decades of social-psychological studies by Fishbein and Ajzen have shown that intentions and attitudes have a high predictive validity with regard to future choices and behaviors.8
In the final analysis, although recommendations were made by stakeholders about how to improve communications in the dental schools to heighten awareness about opportunities to provide care to disadvantaged populations, the more challenging work lies in the health policy arenas, within both federal and state legislatures. Questions remain: Should dental school deans and faculty members become more involved in solving the dental care access crisis? Will the U.S. Health Resources and Services Administration (HRSA) and the FQHCs partner with dental schools across the country to improve access to dental care for low-income subgroups? Will professional associations and advocacy groups be able to muster enough political clout to influence health policy? These are but a few of the salient policy questions emerging out of the Pipeline program.
|
Author Information |
|---|
|
TopAuthor information1. What are facilitators...2. Are students planning...3. What factors influence...Discussion and Study limitationsReferences |
|---|
* The best subset involves examining all models created from all possible combinations of predictor variables and finding a "best" model(s) based on a selection criterion, such as a residual sum of squares. Several different measures are available to select the best model; we used the smallest AIC (Akaike information criteria) and the smallest Mallows C (p) criteria (Afifi AA, Clark VA, May S. Computer-aided multivariate analysis. Boca Raton, FL: Chapman & Hall, 2004). The results of the best-subset model were then used in a final random effects multivariate logistic regression model. This introduces a random effect into the linear model that allows each school to have its own influence (or intercept) that deviates from the population intercept. All tests of significance reported in Tables 6.4.5, 6.4.6, and 6.4.7 were adjusted for clustering amongst students within the same school, as these students are more likely to be similar on a variety of characteristics compared to students in other schools. 
|
REFERENCES |
|---|
|
TopAuthor information1. What are facilitators...2. Are students planning...3. What factors influence...Discussion and Study limitationsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. M. Andersen and P. L. Davidson Introduction to the Evaluating the Dental Pipeline Program Report J Dent Educ., February 1, 2009; 73(2_suppl): S10 - 14. [Full Text] [PDF]
|
|
|
|
|
|
P. L. Davidson, T. T. Nakazono, A. Afifi, and J. J. Gutierrez Methods for Evaluating Change in Community-Based Dental Education J Dent Educ., February 1, 2009; 73(2_suppl): S37 - 51. [Full Text] [PDF]
|
|
|
|
|
|
J. J. Gutierrez, T. T. Nakazono, D. C. Carreon, and R. M. Andersen Introduction to Case Studies of the Pipeline Programs at Fourteen U.S. Dental Schools J Dent Educ., February 1, 2009; 73(2_suppl): S52 - 57. [Full Text] [PDF]
|
|
|
|
|
|
J. J. Gutierrez, T. T. Nakazono, D. C. Carreon, and R. M. Andersen Introduction to the Cross-Site Comparisons and Multivariable Analyses of the Dental Pipeline Program J Dent Educ., February 1, 2009; 73(2_suppl): S236 - 237. [Full Text] [PDF]
|
|
|
|
|
|
R. M. Andersen, P. L. Davidson, K. A. Atchison, J. J. Crall, J.-A. Friedman, E. R. Hewlett, and A. Thind Summary and Implications of the Dental Pipeline Program Evaluation J Dent Educ., February 1, 2009; 73(2_suppl): S319 - 330. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
