The 2003 Institute of Medicine (recently renamed the National Academy of Medicine [NAM]) report “Who Will Keep the Public Healthy?” highlighted the importance of academic–public health practice collaboration for a well-functioning public health system and well-educated workforce.1 Previous research has demonstrated that relatively few public health graduates pursue careers in governmental public health agencies and that relatively few governmental public health employees have formal academic public health training.2 Academic-practice collaboration represents a strategy for addressing both of these challenges.
Public health departments and academic institutions engage in a range of cooperative activities. Such cooperation can greatly benefit a public health department,3 can be mutually beneficial for a university,4 and can provide benefits to the individual faculty members, students, and practitioners involved in the engagement.5 Previous research has demonstrated that academic-practice collaboration can strengthen core public health functions,6 emergency preparedness capabilities,7 , 8 and overall scope of services offered.9 Recognizing the importance of collaborating with academia, longer-term, formalized collaborations can result in the formation of an academic health department, an arrangement that can provide “mutual benefits in teaching, research, and service, with academia informing the practice of public health, and the governmental public health agency informing the academic program.”10 Recognizing the importance of academic-practice collaboration, the Public Health Accreditation Board requires a health department to document engagement with academia, especially with respect to evidence-based practices, although there are no requirements that this engagement takes place within an academic health department model. In short, we know that academic-practice collaboration can be an effective strategy for public health departments to pursue. Yet, much of what we know about how or why these collaborations achieve their goals comes from case reports or analyses of smaller or nonrepresentative samples.
While academic-practice collaboration can be mutually beneficial, these affiliations are not without limitations. Barriers to academic-practice collaboration reported in most previous research have focused on factors that inhibit initiation of any partnership. Such factors include lack of time, lack of funding, and lack of interest among health department employees.4 One academic-practice partnership struggled with the perception from some that the collaboration was “just another obligation” added to an already full suite of activities.11
Establishing an evidence base of successful strategies for health department–academic cooperation may help overcome some of these barriers and promote a more valuable experience for both academics and practitioners. At their core, academic-practice collaborations bring together individuals working within and across organizations to achieve mutually beneficial results.
Yet, little nationally representative research has been undertaken in this area to date. To our knowledge, this study is the first to specifically differentiate academic-practice collaborations perceived by public health practitioners as successful versus those perceived as unsuccessful. For example, could a partnership that struggled with the perception of being “just another obligation”11 have flourished if created or administered under different conditions or with different responsible staff members? Are certain types of positions, skills, or characteristics of an individual involved in a partnership associated with collaborations that prove to be successful?
By exploring practitioners' views of a successful academic collaboration, we strove to define correlates that are beneficial to a state health department. These promising practices could then be tested at public health departments. Specifically, practitioners can use these findings to help guide future collaborations toward programs and personnel within their departments that may be good fits. The purpose of this study was to explore predictors and correlates of beneficial academic collaborations from the perspective of those on the front lines—the practitioners constituting the public health workforce.
Data and sample
Data for this analysis come from the Public Health Workforce Interests and Needs Survey (PH WINS), a survey of over 23 000 public health workers from state health agencies in 37 states. The survey methodology is described in detail in the accompanying article by Leider et al12 in this supplement. Briefly, PH WINS is a nationally representative survey of state health department central office employees that was fielded in 2014 and covered a range of topics, including academic collaboration. The overall response rate was approximately 46%.
All variables and data used in the analyses came from PH WINS. A total of 10 246 responses were available for this analysis, constituting a nationally representative sample of the state health department workforce. The PH WINS sample was limited to individuals working in the state health department (N = 10 246) and excluded state health officials working in local/regional health departments (n = 7229), in other agencies (n = 890), and those who were not permanent central office staff (n = 552). PH WINS local health department pilot study data were not used for this study, as those data are not nationally generalizable.
Outcome measure: Successful collaboration
This study focused on 2 measures related to academic–public health practice collaboration: first, whether such a partnership occurred, and, second, whether that partnership was successful.
Data for the first measure came from a PH WINS question asking: “In the past year, have you worked with members of the academic community (faculty/staff/students) on public health practice issues?” (Response options included “Yes” and “No.”) Data for the second measure came from the question asking: “To what extent was this collaboration helpful to you in your work?” (Response options included “not at all helpful,” “not very helpful,” “somewhat helpful,” and “very helpful.”) For this study, a partnership was classified as “successful” if the respondent answered “very helpful” and classified as “unsuccessful” if the respondent's answer was any of the other 3 options.
Previous studies have relied on theory-driven or deductive approaches to defining positive or sustainable collaboration for public health.13 , 14 This study relied on an inductive approach that defined a successful academic-practice collaboration from the perspective of the public health workforce.
These outcomes were explored in association with 4 sets of predictors: (1) respondent's demographic characteristics; (2) respondent's position characteristics; (3) respondent's skills and positional duties; and (4) departmental characteristics.
Respondent's demographic characteristics included respondent age (categorized as <40, between 41 and 60, and >60 years), gender, race/ethnicity (categorized as American Indian/Alaska Native, Asian, black, Hispanic, Native Hawaiian/Pacific Islander, white, ≥2 races), and plans to retire prior to 2020. Two educational background measures were also included: receipt of a bachelor's degree or higher, and receipt of a public health degree. Length of time in current position and length of time with the current organization were included (each dichotomized to over vs under 5 years).
Characteristics of a respondent's position included supervisory status (nonsupervisory, team leader, supervisor, or manager/executive), number of direct reports (categorized as 0, 1-5, 6-10, or >10), and the position's classification (administrative, clinical and laboratory, public health science, or social services).
Respondent's skills and positional duties were self-reported to PH WINS through a series of questions covering (a) how important a given item is for an individual's day-to-day work (“positional duties”) and (b) the individual's current skill level for the item (“skills”). A total of 18 items were surveyed (eg, “Assessing the broad array of factors that influence specific public health problems,” “Applying quality improvement concepts in my work,” “Influencing policy development,” “Preparing a program budget with justification”). Because of the large number of items, we performed exploratory factor analyses to identify related items.15 Results from these analyses were used to create separate subscales for positional duties and individual skills as described in the Supplemental Digital Content Appendix (available at: http://links.lww.com/JPHMP/A164). A list of items included in each of the 3 scales is shown in Table 1. The 3 scales created were virtually identical for both positional duties and respondent's self-reported skills and are referred to throughout as: Community Collaboration and Data, Management and Communication, and Budgeting. See the Supplemental Digital Content Appendix (available at: http://links.lww.com/JPHMP/A164) for full description of scale creation process. Scores for the 3 duties scales ranged from 1 (not important) to 4 (very important); scores for the 3 skills scales included 1 (not applicable) and ranged from 2 (unable to perform) to 5 (expert).
A fourth set of predictors hypothesized to be salient to the relationship was organizational characteristics. PH WINS data do not enable identification of a respondent's state of residence/employment. However, anonymized state dummy variables were available and were included in models to account for otherwise unobservable state-level characteristics (eg, departmental structure, governance).
Analytic sample creation
Of the 10 246 observations in the data set, complete data were available for a total of 8718 respondents (85.1% of total). A total of 1528 observations (14.9%) were excluded because of missing data for at least 1 variable: 138 were missing data on the presence of academic partnership, 613 were missing at least 1 skills or positional duties scale, 103 were missing race/ethnicity, 83 were missing age, and an additional 591 were missing at least 1 additional predictor variable.
We computed univariate statistics to assess variable distributions and variation. Bivariate analyses were then used to assess correlation among variables. Sample weights were used in all analyses to account for survey design and nonresponse.16
Multivariate logistic models were used to explore associations between outcomes of interest and predictor variables. Separate models were estimated for (a) whether a PH WINS respondent collaborated with an academic institution, and (b) for those who did participate, whether the collaboration was successful (as defined earlier). In both models, state-level characteristics were estimated through the use of fixed effects via inclusion of dummy variables for respondent's state.17
Analyses to assess the sensitivity of results to variable coding, survey weighting, use of state-level fixed effects, and other model specifications revealed that overall findings were not sensitive to variable coding, use of survey weighting, or model alternative model specifications. All analyses were performed using Stata (version 13.1). The Arizona State University institutional review board approved the research plan.
Data were available for a total of 8718 respondents in 37 states. Nationally, 27.3% of all state health department practitioners reported participating in at least 1 academic collaboration with members of the academic community on public health practice issues within the past year. All states had at least 1 respondent note that he or she participated in a collaboration within the past year.
As shown in Table 2, distinct patterns appeared in terms of respondents who did versus did not participate in academic-practice collaboration. Without adjusting for other factors, younger respondents, those with at least a bachelor's degree, those with a public health background, those without plans to retire before 2020, and those who had been in their current position for longer than 5 years were all more likely to report participating in an academic-practice collaboration. Likewise, supervisors, persons with 6 or more direct reports, and persons in public health science positions were also more likely to have participated in a collaboration. Job duties and self-reported skills also differed significantly between collaboration participants and nonparticipants.
After accounting for these factors in a multivariate model, several remained statistically significantly associated with the likelihood of collaboration participation. Table 3 shows that, of individual-level characteristics examined, only a respondent's educational background was significantly associated with the likelihood of collaborating. Having a public health degree (eg, MPH or DrPH) increased the odds of collaborating with an academic institution by 85% (P < .001). Several positional characteristics were significantly associated with likelihood of partnering. Team leaders were significantly more likely than nonsupervisors to report partnering (OR = 1.51; P < .001). Compared with administrative positions, all other position types were significantly more likely to collaborate (P < .05 for all 3 comparisons). Positional duties and skills of respondents were also significantly associated with the likelihood of partnering, with positions having duties related to budgeting, community partnerships, and data being associated with an increased likelihood of partnering.
The study's second objective was to examine factors associated with a successful academic-practice collaboration. A total of 2616 PH WINS respondents reported collaborating with at least 1 member of the academic community within the past year. Of these, 46.6% of collaborations were successful. All states had at least 1 respondent who noted that he or she participated in a successful collaboration within the past year.
Table 4 shows results from the multivariate logistic regression model predicting a successful academic-practice collaboration. All of the 3 skills scale variables were significantly associated with the likelihood of a successful collaboration. Specifically, having stronger skills in 1) community partnterships and data and 2) management and communication were positively associated with the likelihood of a successful collaboration, whereas having stronger budgeting skills was negatively associated with the likelihood of a successful collaboration. Positional duties were not significantly associated with the likelihood of collaboration success. Individual characteristics significantly associated with collaboration success were a respondent's public health background (OR = 1.31; P < .05) and respondent's race/ethnicity—with both black and Native Hawaiian/Pacific Islanders being more likely to report a successful collaboration. Positional characteristics were less strongly associated with collaboration success, with the only significant factor being that team leaders were less likely to report a successful collaboration experience.
This study expands upon a growing body of evidence that collaboration between academicians and public health practitioners can be beneficial. This study is the first to explore the extent of collaboration between faculty, staff, or students at academic institutions and state health agencies. Findings indicate that these collaborations are widespread at the organizational level, as all states had at least 1 reported collaboration perceived to be helpful. At the same time, successful collaborations are less common. Less than half of all collaborations (46.6%) were successful in the eyes of the health department practitioner. This means that only approximately 12% of the state public health workforce has participated in a successful collaboration within the past year. Results showed that predictors of whether a PH WINS respondent has participated in an academic collaboration are based in large part on the characteristics of the respondent's position (eg, supervisory status and position classification) and the duties required by an individual's position. This aligns with previous findings that posited collaboration to be especially relevant for specific public health functional areas such as emergency preparedness.7 , 8 Our findings suggest that there may be certain aspects of a position that enhance the likelihood of wanting or needing to engage in academic collaboration. For example, administrative positions were significantly less likely to report partnering with academics than other position types. This aligns with a view of academic-practice collaboration, whereby academia informs the practice of public health.10 An alternate view of this finding is that administrative public health positions may not be benefiting from academic collaboration to the same extent that other functional areas are.
In contrast with the aforementioned discussion, findings showed that while the presence of a collaboration was associated with factors that by and large relate to the position, the bulk of the relevant factors for determining whether the collaboration was successful relates to the individual involved in the collaboration.
A respondent's self-reported skills across 18 items measured in PH WINS were all highly significantly associated with the likelihood of a successful collaboration. Having higher skills in community relationships, management, and communication was positively associated with successful collaborations.
Budgeting skills, on the contrary, were negatively associated with a successful collaboration. Thus, administrative positions are not only less likely to collaborate, when collaboration does occur, but persons with higher levels of skills for budgeting and administering also report having less successful collaborations. This represents both a challenge and an opportunity for fostering a more frequent and more successful collaboration between academics and administrative public health practitioners. The PH WINS data set does not have information about the academic specialties involved in collaborations. It is possible that additional outreach to a broader range of academic partners may yield future opportunities for successful collaborations. For example, schools of management or public policy/service may be useful in establishing collaborations related to administrative or financial responsibilities.
A respondent's educational background was very highly correlated with both the presence and success of a collaboration. Not only were individuals with a public health degree more likely to report collaborating, but they were also more likely to report a successful collaboration. Public health agencies looking to establish new collaborations or to refine existing collaborations may do well to consider purposefully including individuals with public health degrees in these efforts. Likewise, academics looking for additional partners within public health agencies may want to pay particular attention to individuals with public health training, as these individuals are statistically more likely to engage in successful collaborations.
The correlates identified as significantly associated with the likelihood of collaborating and the likelihood of a successful collaboration represent only one side of a 2-sided collaboration. The purpose of the study was to both explore the perspective of the public health workforce and to identify correlates of what works based on the experiences of thousands of respondents across the nation. Without data on the import, impact, or circumstances of the actual collaboration(s), the study is limited in its ability to make conclusions regarding the true causal nature of the relationships observed. But given the expansive scope of the data used for this study and the major, substantive importance of the outcomes used (Did a collaboration take place and, if so, was it helpful to the department?), it is perhaps most useful to view these findings as a promising administrative practice for promoting effective collaboration with academia. Previous studies have noted the underdevelopment of literature surrounding administrative best practices for public health.18 Given that data from than 8700 public health practitioners report that having a public health degree was associated with an approximately 30% increase in the likelihood of having a successful outcome, public health departments may be well served by exploring how to utilize existing resources (eg, workers with a public health degree or relevant job skills) when establishing or sustaining future collaborations.
Study findings should be viewed in light of several limitations. The information used for this study comes from PH WINS, a self-administered survey of public health practitioners. As such, no data were available on academicians' perception of the collaborations evaluated. Moreover, no data exist on the structure of the collaboration. Rather, this study focuses exclusively on the processes associated with successful collaborations. A second issue is the self-reported nature of our data. Respondents' recollections and perceptions of the helpfulness of a given collaboration may vary, and no information was available about the import or impact of the circumstances of the collaboration(s). Yet, numerous studies have touted the importance of mutual benefit to both practice and academia as a necessary success factor for sustainable collaboration.2 , 3 , 6 , 7 , 19 So, understanding how public health practitioners themselves perceive collaborations and conditions associated with a greater likelihood of a successful outcome is itself a relevant and important measure.
Collaboration between the public health workforce and faculty, staff, or students from academic institutions represents an opportunity to strengthen current public health agency operations and improve training for future generations of public health practitioners. Yet, capacity for external collaboration varies across many health departments.14 At a national level, little is known about where collaborations are occurring and how successful they are from the perspective of the public health workforce. This study found clear evidence that collaboration may be widespread, but that several levels of factors underlie both their presence and their likelihood of success.
Broadly, while characteristics of an individual's position are most significant in predicting whether a collaboration will occur, characteristics of the individual are more relevant in predicting whether a collaboration will be successful. An individual's public health background and the skills related to community partnerships and data and management and communication are more relevant to a collaboration's success than, for example, the number of direct reports an individual has or whether a person is in a supervisory role. Likewise, no evidence was found to suggest that older versus younger individuals, males versus females, or those with long tenures at their current agencies/positions are more likely to collaborate or to have a successful collaboration.
Public health leaders interested in fostering an environment that promotes a successful collaboration with academia may benefit from ensuring that the public health practitioners involved in such collaborations are themselves trained in public health and that they have the requisite relevant skills.
1. Institute of Medicine. Who Will Keep the Public Healthy? Educating Public Health Professionals for the 21st Century. Washington, DC: National Academies Press; 2003.
2. Conte C, Chang CS, Malcolm J, Russo PG. Academic health departments: from theory to practice. J Public Health Manag Pract. 2006;12(1):6–14.
3. Livingood WC, Goldhagen J, Little WL, Gornto J, Hou T. Assessing the status of partnerships between academic institutions and public health agencies. Am J Public Health. 2007;97(4):659.
4. Caron RM, Hiller MD, Wyman WJ. Public health system partnerships: role for local boards of health in preparing the future public health workforce
. J Community Health. 2014;39(1):29–34.
5. Erwin PC, McNeely CS, Grubaugh JH, Valentine J, Miller MD, Buchanan M. A logic model for evaluating the academic health [published online ahead of print February 26, 2015]. J Public Health Manag Pract. doi:10.1097/PHH.0000000000000236.
6. Neri EM, Ballman MR, Lu H, Greenlund KJ, Grunbaum JA. Academic-health department collaborative relationships are reciprocal and strengthen public health practice: results from a study of academic research centers. J Public Health Manag Pract. 2014;20(3):342–348.
7. Covich JR, Parker CL, White VA. The practice community meets the ivory tower: a health department/academic partnership to improve public health preparedness. Public Health Rep. 2005;120(suppl 1):84.
8. Dunlop AL, Logue KM, Vaidyanathan L, Isakov AP. Facilitators and barriers for effective academic-community collaboration for disaster preparedness and response [published online ahead of print December 12, 2012]. J Public Health Manag Pract. doi:10.1097/PHH.0b013e3182205087.
9. Downey LH, Thomas WA, Gaddam R, Scutchfield FD. The relationship between local public health agency characteristics and performance of partnership-related essential public health services. Health Promot Pract. 2013;14(2):284–292.
10. Erwin PC, Keck CW. The academic health department: the process of maturation. J Public Health Manag Pract. 2014;20(3):270–277.
11. Hamilton CB, Buchanan ML, Grubaugh JH, Erwin PC. Forming an academic health department: a case example. J Public Health Manag Pract. 2014;20(3):304–309.
12. Leider JP, Bharthapudi K, Pineau V, Liu L, Harper E. The methods behind PH WINS. J Public Health Manag Pract. 2015;21(suppl 6):S28–S35.
13. Alexander JA, Weiner BJ, Metzger ME, et al. Sustainability of collaborative capacity in community health partnerships. Med Care Res Rev. 2003;60(4)(suppl):130S–160S.
14. Rodriguez HP, McCullough JM, Hsuan C. Local health department collaborative capacity to improve population health. Front Public Health Serv Syst Res. 2014;3(4):7.
15. Pett MA, Lackey NR, Sullivan JJ. Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research. Thousand Oaks, CA: Sage; 2003.
16. Winship C, Radbill L. Sampling weights and regression analysis. Sociol Methods Res. 1994;23(2):230–257.
17. Allison PD. Fixed Effects Regression Models. Vol 160. Thousand Oaks, CA: Sage; 2009.
18. Brownson RC, Allen P, Duggan K, Stamatakis KA, Erwin PC. Fostering more-effective public health by identifying administrative evidence-based practices. Am J Prev Med. 2012;43(3):309–319.
19. Lee AF, Quade T, Dwinnells R. Evolution of the academic health department through public health academic and practice collaborations. J Public Health Manag Pract. 2014;20(3):290–296.