Assessing researchers’ capabilities, opportunities, and motivation to conduct equity-oriented dissemination and implementation research, an exploratory cross-sectional study

We used the Capability, Opportunity, and Motivation Model of Behavior Change (COM-B) to guide data collection, situate and organize results, and interpret findings [15]. To be able to engage in, develop, and conduct equity-oriented D&I research, researchers and their teams must have: 1) the Capability to psychologically and physically integrate equity-oriented in D&I research (i.e., knowledge, skills); 2) the Opportunity through external factors to make this change (i.e., environment, time, resources); and 3) the Motivation to do so, including reflective processes (e.g., making plans, evaluating changes already tried) and automatic processes (e.g., emotions). Because education and training are interventions that target behavior, and because COM-B has been used to design training (e.g., we believed this framework would be a good fit to help us understand potential behavior change mechanisms that could facilitate or hinder the conduct of equity-oriented D&I research [16, 17]. We hypothesized that the interaction of each factor is theoretically necessary to influence our decision-making and behaviors in learning, developing, and conducting equity-oriented D&I research.

We conducted an exploratory cross-sectional survey that was distributed online from December 2020 to April 2021. This was an online survey, collected via Qualtrics survey platform [18]. This study was approved by the Institutional Review Board at Washington University of St. Louis (HRPO #202,012,103).

We recruited participants at conferences (e.g., the 2020 D&I conference, hosted by the AcademyHealth), meetings, or listservs on either D&I or health disparities research (e.g., Division 45 [Society for the Psychological Study of Culture, Ethnicity, and Race] from the American Psychological Association, the Implementation Research Institute, National Implementation Research Network, Implementation Science Centers in Cancer Control). We also sent tweets and emails from our personal accounts to recruit participants. While we did not have eligibility inclusion criteria for recruiting, our emails and tweets explicitly invited participants to give us input on training needs with regards to health equity and implementation science, with the goal of understanding existing gaps and to develop future training opportunities in this space. The emails and tweets contained a link to the survey developed using the Qualtrics survey platform [18]. The consent form was provided at the start of the survey and described the purpose of the study and study team. Once consented, participants were able to answer the survey. To minimize bias in reporting, we did not collect any identifiable information about participants.

All analyses were conducted using SPSS Version 27.0. We calculated descriptive statistics including frequencies, means, and standard deviations and reported on all data from questions answered by participants. To assess differences in training opportunities to conduct equity-oriented D&I research by participants’ training level, we conducted an Analysis of Variance (ANOVA) with training level as the independent variable (i.e., trainees, faculty, other) and individual survey items assessing motivation, information, and skills as dependent variables. We used ANOVA as the diagnostic test given that we wished to compare means of the three groups and our outcome variables are continuous (i.e., based on aggregate responses to the survey questions). Fischer’s Chi-square test would not be appropriate for this analysis because it is used to compare group means when the outcome variables are categorical. After assessing overall effects using ANOVA, we conducted post hoc analyses to determine specific between-group differences using Tukey’s honestly significant difference test.

Given our focus on reporting only descriptive statistics and comparing only completed responses, Little’s Test of Missing Completely at Random [20] was not conducted and data were not imputed.

A total of 180 people answered the online survey. Most participants identified as women (81.7%), white (66.1%), academics (78.9%), and faculty members (53.9%). Additional participant characteristics are reported in Table 1. With regards to competencies, many participants reported they were ‘Advanced’ (36.7%) or ‘Advanced Beginners’ (27.8%) in the D&I field. Participants reported varying levels of expertise in health disparities research, community based participatory research (CBPR), and health equity. Many participants (37.8%) reported being novice to D&I research that focused on health equity. Table 2 provides the detailed descriptive statistics for respondents’ self-reported expertise in the five different categories of research (i.e., D&I Research, Health Disparities Research, Community Based Participatory Research, Health Equity Related Research, D&I Research with a Focus on Health Equity).

Below we report results of the discrete sets of questions. For each table, we provide the question and descriptive statistics based on the respective COM-B domain that informed the question.

Table 3 includes descriptive statistics, means, and standard deviations for questions related to capability and motivation for conducting equity-oriented D&I research. Regarding questions about knowledge or information, the overall mean of the four specific questions was 4.2, with participants responding to individual items assessing knowledge with a range of means from 3.99 (SD = 1.56) to 4.62 (SD = 1.51). The overall mean for questions that assessed necessary skills was 4.4, with participants responding to individual items with a range of means from 4.25 (SD = 1.76) = to 4.67 (SD = 1.17). Finally, the overall mean for questions assessing motivation was 6.2, with participants responding to individual items assessing motivation with a range of means from 5.97 (SD = 1.32) to 6.31 (SD = 1.03).

Table 4 provides an overview of participants' perceptions regarding the opportunities available to engage in, propose, and conduct equity-oriented D&I projects. When queried about their use of theories, models, or frameworks with an equity lens in current D&I projects, 68 (37.8%) reported using the RE-AIM extension for health equity and sustainability [21], 35 (19.4%) reported Health Equity Implementation Framework [6], and 25 (13.9%) reported using Baumann and Cabassa considerations for the Proctor Model [1]. Some examples of other frameworks participants reported using in the open text included the Conceptual Framework of the Multilevel Influence of Racism on Rural Cancer Disparities Across the Continuum [22], Consolidated Framework for Implementation Research [23], RE-AIM (without extension) [24], the Exploration, Preparation, Implementation, Sustainment framework [25], the Interactive Systems Framework [26], Theoretical Domains Framework [27] Black Feminist Thought [28], The World Health Organization’s Social Determinant of Health Framework [29], and Proctor’s Outcomes for Implementation Research framework [30] Interestingly, participants also mentioned research approaches (i.e., Community Based Participatory Research and Decolonizing Methodologies) as well as the Expert Recommendations for Implementing Change (ERIC) [31] as answers for this question.

When asked about measures used, only 28 (15.6%) participants answered they used any measure that incorporated an equity lens (e.g., measures for community engagement, racism, or discrimination) in their current projects. Forty-one percent reported using individual-level measures to examine issues of equity (e.g., stigma measures, implicit attitudes tests), followed by 31.1% using community-level measures (e.g., perceived structural racism scale, etc.) and 23.9% using healthcare setting measures (e.g., major experiences of discriminations in life domains, such as at school, job, housing). When asked for more information about measures used to incorporate an equity lens in current D&I projects, participants reported using qualitative approaches or not using any validated or existing measures.

We also asked participants what training opportunities they would need to help them incorporate a focus on health equity in D&I research. Most (76.7%) needed training to help select and utilize implementation strategies to promote equity, 66.1% needed training about how to operationalize health equity outcomes or determinants, and 61.7% needed training to guide assessment of context with a focus on health equity. When asked whether they would need other types of trainings, participants mentioned in open text that they would like training on: “everything health equity, D&I and intersectionality”, on “the global application of equity and D&I research”, on “how to navigate D&I and health equity where there is not an applicable evidence-based intervention”, and on “how to implement trauma informed care practices in healthcare settings.” Participants reported wanting more training on tools for how to do equity work, and training about how to, as early career faculty, advocate for equity work with more senior scholars.

Table 5 shows ANOVA results with the training answers by different levels of participants’ training. Graduate students and postdoctoral students were combined into a “trainee” category (n = 45), “faculty” includes people who responded they were in faculty positions (n = 97), and the “other” category is comprised of leadership, research staff, and people who selected the other category (n = 39). Results indicated mean differences by participants’ training level (i.e., trainees, faculty, other). There were main effects on three items. There was a significant overall effect by participants’ training level on the item “I have the skills necessary to identify implementation strategies to promote health equity in D&I research”, although post hoc analysis showed no meaningful differences between groups. There was a significant overall effect by participants’ training level on the item, “I have the information needed to apply theories, models, frameworks for promoting health equity in D&I research” such that faculty reported meaningfully higher scores than trainees (p = 0.027) and other participants (p = 0.021). Third, there was a significant overall effect by participants’ training level on the item, “I have the information needed to identify evidence-based interventions to promote health equity in D&I research,” such that faculty indicated higher levels of information than trainees (p = 0.004).

Table 6 includes responses to questions about factors that could influence participants’ ability to conduct equity-oriented D&I research. Forty-four percent of participants reported the challenge of not having skills needed to conduct this research, 32.2% reported challenges in receiving funding to support this work, and 30% reported challenges with finding collaborators.

Table 7 outlines specific challenges to funding (written into open text), including lack of “funding that honors local partnership and time to appropriately tailor/adapt and implement interventions to address long standing structural factors that contribute to health inequities.” Other reported barriers for conducting this type of research included “lack of mentorship, hierarchy (e.g., staff cannot control research agenda), concerns that existing tools are created by white people, novelty or newness of the field, and lack of infrastructure or experience.” All participants agreed that equity-oriented research is needed for advancing D&I research.

This study has some limitations. This is a cross-sectional, online survey that used newly developed questions not psychometrically validated. Given the paucity of empirical work in this area, we consider this to be an exploratory study in reaching a sample of researchers interested in these topic areas through meetings and social media. This exploratory attitude also allowed us to develop questions that were aimed at understanding researchers’ engagement with and conduct of equity-oriented D&I research. The sample of 180 participants is small and there may be sampling bias introduced through our sampling methods (e.g. underrepresenting those not active on social media or in professional organizations). Despite these limitations, this is, from our perspective, an innovative study in empirically exploring the perspective of participants from different backgrounds about factors that influence their ability to conduct equity-focused D&I research. The sample size for our study is comparable, if not larger, to the sample size of other survey studies used to inform D&I trainings (e.g., [69‐71]. This survey did not assess faculty ranks (i.e., Assistant, Associate, Full Professor) which may have given further insight into how these challenges could perhaps be different for early career compared to more established faculty. Our questions about expertise on the different fields was “select all”, making it difficult to disentangle unique challenges per discipline. While we had informal conversations with international collaborators who stated that they answered our survey, we did not capture the geographical location of our participants and hence, are unable to highlight global participation. Our question about the use of frameworks was very specific to our own work. As shown by the open responses from our participants, there is a large literature about health equity in public health, global health and other social science literatures. The development of training in equity-oriented D&I research should bring the broader literature and examine how these different disciplines could strengthen the work around equity through collaboration. Finally, participants could have also represented several topic areas, such as cancer or mental health, and therefore, have differential exposure to equity or D&I research based on topic interest.