EQUIP Emergency: study protocol for an organizational intervention to promote equity in health care

EQUIP Emergency is a study of an organizational-level intervention to improve care quality at the point of care for those who face health inequities. This study is a three-way collaboration among health researchers, health care staff and Indigenous/community leaders aimed at developing an evidence-based intervention framework to promote equity for Indigenous and non-Indigenous people in diverse EDs. The study is part of a broader program of research entitled EQUIP Health Care that aims to reduce health inequities at the point of care in pursuit of the quadruple aims of health system optimization: improving the health of populations, enhancing patient experiences and outcomes, reducing per capita cost of care, and improving the work life of staff [29, 30]. A key goal is to create intervention processes that are, if evaluated to be effective, scalable and “ready to implement.” As such, our approach to knowledge mobilization, broadly defined as “a wide range of activities relating to the production and use of research results, including knowledge synthesis, dissemination, transfer, exchange, and co-creation or co-production by researchers and knowledge users,” [31] is both integrated and grounded in key principles of implementation science [32, 33].

Approaching EDs as complex adaptive systems [34], this mixed-methods multisite study will:

A tripartite model of Indigenous/community, practice and research team members is used throughout. The research leadership team is comprised of an Indigenous Elder with extensive research experience (author RP), a health care leader with wide-ranging expertise in practice leadership and research (author DB), and four academic researchers (authors CV, VB, AJB, EW) as co-Principal Investigators. Additional co-investigators are diverse Indigenous, practice and research leaders. The project is guided by an advisory panel, comprised of Indigenous community, practice and research leaders. For all intervention activities, the research team is supporting the engagement of Indigenous and community leaders, health care leaders and all staff in each ED.

We are working with 3 EDs in one Canadian province, purposely selected to capitalize on our extant research relationships and maximize diversity of settings: an inner-city, a suburban area, and a northern region serving rural, remote and small urban communities. We are developing, implementing, testing and refining the framework to optimize care for Indigenous people and a wider range of people who commonly face stigma and discrimination. We have designed an overall process that will be tailored to the unique characteristics, context and strengths of each ED. For example, one ED has extensive expertise providing care to patients with substance use problems, another has expertise caring for diverse groups of new immigrants, and one is located in a region with relatively high proportions of Indigenous populations. Collaborating with organizational leadership in these diverse health authorities promotes understanding of the potential scale up to other EDs. Table 1 provides some details regarding each.

Our understanding of equity and EOHC is informed by critical theoretical understandings of social justice, and the structures that perpetuate health and social inequities. The overall study is guided by complexity theory using an integrated approach to implementing and mobilizing interventions. These approaches and our tripartite leadership model are complemented by a change leadership approach known as Front Line Ownership (FLO), all of which align with understanding health care systems and EDs as complex adaptive systems.

Intervention research to bring about change within health care settings has historically had low success rates. Lack of success has been associated with failing to attend to the complexity of environments in which interventions are carried out, or factors associated with capacity and readiness for change [37‐39]. Consequently, the intervention detailed in this protocol was influenced by central tenets of complexity theory. As noted, EDs are complex adaptive systems with many component parts that interact and influence one another [39]. These interactions include and extend beyond the patient level to the provider, organization, and policy levels of health care [40]. Complexity theory directs us to analyze the diverse and complex interactions among these components.

In addition to attending to the complexity of the ED settings, our intervention design utilizes the evidence-based and theoretically-informed change management approach of Front Line Ownership. With its roots in complexity theory, FLO is an intentional departure from prevalent hierarchical health care change management practices whereby leaders promote ideas to staff and try to get buy-in. Instead, FLO involves those most actively engaged in direct patient care developing the ideas, making the decisions and designing and acting on the plans they create [41, 42]. FLO includes change processes that support participatory, peer-based learning and recognizes that learning and change are most effective and sustainable when the solutions (in this case, to the problem of inequities in health and health care) are generated by the health care staff themselves [42]. FLO does not infer absence of formal change leadership; it requires that leaders support the development of staff capacity to both identify and lead change initiatives. Supportive strategies, often referred to as “liberating structures,” [43] include quick-to-learn tools for the staff to enhance their capacity for facilitation among their peers including active listening, engagement and inclusion of ideas and strategies among the diversity of staff working in the setting (e.g., nurses, physicians, security personnel, housekeeping, unit coordinators or admitting clerks). FLO supports staff to both identify and think about problems differently and to take a leadership role to generate solutions to problems that occur in their settings [43]. FLO achieves increased change sustainability by: identifying local solutions that reflect the unique context in which issues arise; increasing organizational interconnectivity and the development of new action pathways via a process of point of care, worker-directed collaboration; and building the capacity of these workers for real-time problem-solving over time [42]. FLO is a contextual and non-linear approach to change management and change leadership ideally suited to interventions occurring in complex settings.

At the definitional core of health equity is the critical analysis of power and the workings of discrimination dynamics in the pursuit of social justice [4, 44, 45]. Our conceptualization of EOHC is grounded in understanding discrimination as a socially structured and often implicitly sanctioned phenomenon, justified by ideology and expressed in interactions among and between individuals and institutions in ways that maintain privileges for members of dominant groups, and contribute to inequities for others [13, 46], with profound impacts [47]. Experiences of discrimination are often amplified when issues of poverty, substance use, or stigmatizing conditions such as HIV or chronic pain syndromes intersect with people's sense of being treated differentially on the basis of their ethno-cultural and other identities [48]. From this vantage point, interventions aimed at improving organizational cultural competencies or providing diversity education for service providers to reduce discrimination and increase access to services are often good starting points, but are insufficient to achieving equity-oriented practices. For example, such interventions often lack a focus on how social systems produce and sustain health inequities, thus allowing the fundamental causes of inequities to remain unidentified and unchallenged. In comparison, equity-oriented approaches seek to address the continuities between structural and interpersonal discrimination. Such approaches go beyond providing education to fill knowledge gaps, and aim to support critical reflection and action by providers and other staff on the interconnected barriers to health care across personal (e.g., stereotypes and assumptions about patients), interpersonal (e.g., interactions with patients that may be discriminatory), organizational (e.g., security protocols that create access barriers for patients who experience poverty, mental health or substance use issues) and structural levels (e.g., inadequate access to housing and social welfare supports).

EQUIP PHC, the precursor to the current study, showed that in PHC settings EOHC was associated with positive health outcomes [5], and that intervening at the organizational level improved staff confidence and comfort with providing such care [4]. We do not, however, know whether such an intervention can improve health outcomes, or how emergency department staff can be supported to provide EOHC. This project pursues three key dimensions for enacting equity-oriented health care derived from our prior research [6, 7, 20]: Culturally-Safe Care; Trauma- and Violence-Informed Care (TVIC); and Harm Reduction, each of which must be Contextually Tailored. These key dimensions, described in greater detail in previous publications [4], are briefly summarized below.

Culturally-Safe Care locates the culture of health care as the site for transformation and seeks to challenge discriminatory values and assumptions held within health care across interpersonal, historical, and structural levels (e.g., legacy of colonization and racism that shapes discrimination as a barrier to health care) [4, 7, 20]. It shifts the focus away from attempting to identify the unique cultural characteristics of individuals or groups as the potential barriers to care, towards critiquing structural conditions and inequitable power relations, requiring changes to dominant norms to enhance care for everyone.

Trauma- and Violence- Informed Care (TVIC) involves understanding past and ongoing interpersonal and structural violence as causes of trauma, and developing policies and practices to minimize harm, prevent re-traumatizing people during care, and promote healing. The primary goal of TVIC is to create safe and trusting environments [49].

Harm Reduction involves practices that mitigate harms, not only of substance use but also harms associated with the historical, socio-cultural and political determinants of substance use and related social and health policy [50].

Contextually-Tailored Care explicitly and continually adapts services, practices, organizational policies and clinical guidelines to optimally address the populations served and to local social and community contexts, thus expanding beyond the individually-focused concept of patient-centred care [4].

A mixed methods multisite design will examine changes in key outcomes specified in the intervention theory (Fig. 1) including: a) a longitudinal analysis of change over time based on structured assessments of patients and staff, b) an interrupted time series design (ITS) of administrative data (i.e., staff sick leave, patients who leave without care being completed, including those who register but leave before being seen by a care provider, before being assessed, before being treated, or against care provider advice), c) a process evaluation to assess how the intervention was implemented and the contextual features of the environment and the process that are influential for successful implementation, and d) a cost-benefit analysis. Table 2 provides a summary of our hypotheses and research questions and the related data sources and analytic approaches.

Changes in patients’ overall ratings of care (primary outcome), and self-reported experiences of discrimination in the emergency department (secondary outcome) will be examined using survey data collected at 5 points in time: 2 times pre-intervention, immediately post-intervention, and at 6 months and 12 months post-intervention. Staff perceptions of patient care (primary outcome) and staff satisfaction and engagement with their job and workplace (secondary outcome) will be examined using survey data collected at 4 points in time: pre-intervention, immediately post-intervention and 6 and 12 months later. Table 3 provides a summary of the patient and staff survey items and their sources.

Data will be collected at equally spaced intervals (monthly) on selected variables hypothesized to be sensitive to EOHC, and which have sufficient stability over time. Variables include the number of people who leave the ED prior to the completion of care (primary outcome) as a percentage of the total number of people seen, and the rate of staff sick time taken as a proportion of productive hours (primary outcome). Using retrospective data, we will collect monthly data from hospital administrative sources for 24 months prior to the start of the intervention, 12 months during the implementation of the intervention and 12 months post-intervention for each variable.

From a complex systems perspective, a research intervention is seen as an event that unfolds within and occurs in interaction with a system [38, 39, 61]. In order to accurately document and assess the implementation of an intervention and the relevance of the implementation processes for intervention outcomes, it is necessary to monitor the contextual influences shaping intervention activities. A process evaluation will be conducted to examine how contextual influences shape EQUIP Emergency implementation and ultimately, outcomes, including the readiness of the EDs to engage in activities to foster EOHC. Table 4 provides a summary of the proposed process evaluation phases mapped across EQUIP intervention activities and key data sources.

The costs of the intervention will be calculated from budget monitoring data for the research budget and for the costs of staff time. The cost-benefit ratio will be computed for each of the primary outcomes. These cost-benefit ratios will reflect the cost to improve patient and staff ratings of care by 1 unit, cost to have one less patient leave without care being completed, and cost to reduce each hour of staff sick time.

The power analysis is based on the within-site analyses. For the longitudinal patient survey data, we estimated the number of patients needed per site to detect a small to moderate effect size of 0.25. Means and standard deviations from the EQUIP PHC study for patients’ experiences with care informed the power analysis. With power of 0.80 and alpha of 0.05, we can detect significant changes with an effect size of 0.25 with a sample size of 250 patients. We used methods suggested by Zhang et al. [62], to estimate the power for the ITS design. Based on the trend over time in the primary ITS outcomes extracted from the administrative data for the past 24 months, we estimate that the time series will have an autocorrelation of − 0.20 and will require an autoregressive model with 1 lag (AR1). Power is 0.85 to detect a moderate effect size with alpha of 0.05, AR1 model, and an autocorrelation of − 0.20 with 24 time points prior and 12 time points post the intervention.

Longitudinal analysis will be conducted within site and for all three sites combined. For each wave of patient and staff survey data, descriptive statistics appropriate to the level of measurement will be computed. Analysis of change across time will use General Estimating Equations (GEE) to account for the repeated measures over time. Time (5 levels for patient survey data and 4 levels for staff survey data) will be the independent variable. We hypothesize increases in patients’ mean ratings of care and in staff perceptions of care (primary outcomes), and decreases in patient experiences of discrimination in the ED and increases in staff engagement and team effectiveness (secondary outcomes).

For the analysis of the ITS data, segmented regression with autoregressive models will be used [64]. Interrupted time series are tested by comparing the level of the outcome of interest and the rate of change over time between the time period prior to the system change and the time period after the system change [65]. The model will yield an estimate of the level and slope across time of the outcomes prior to the intervention and changes in the level and slope after the intervention. The change in level provides an estimate of the immediate effect of the intervention, and the change in slope provides an estimate of the ongoing effect of the intervention across time after implementation. We hypothesize that there will be a significant change immediately after the intervention with a continued improving trend.

A cost per unit of improvement will be calculated for each of the primary outcomes. Improvement will be the change in the outcome from the time point immediately prior to the start of the intervention implementation to the 12 months after the intervention. Total cost will be divided by the change in outcome and scaled to reflect the cost per 1 unit improvement in patient and staff perceptions of care. For the number of patients who leave without care being completed, the difference in the total number of patients who left in the 12 months prior to and post-intervention will be computed. Total cost to deliver the intervention will be divided by this number to reflect the cost to achieve one fewer patient per year leaving without care being completed. This same approach will be used with respect to staff sick time (total staff sick time in the 12 months prior compared to the 12 months post-intervention).

Process evaluation analysis will be conducted for each site and then combined to assess the contextual features influencing intervention implementation at organizational (e.g., leadership within the health authority and specific health care facilities) and unit (e.g., emergency department staff and structures) levels. Each phase detailed in Table 4 will be analyzed first at each site to identify the degree to which the activities have achieved the intended purpose using the predetermined performance indicators for enablers assumed to foster successful implementation. Thematic analysis will be conducted using data from meeting minutes, field notes, and qualitative interviews to explain the outcomes of this analysis. Results from the analyses at each site will be combined and analyzed to identify similarities and differences across sites including potential rationale for any discrepancies. The analysis will be refined, advancing theoretical coding to assess the barriers and enablers related to the core theoretical tenets underpinning the intervention including front-line ownership as a change management approach and equity-oriented health care and its related dimensions. This analysis will identify the processes that facilitated or hindered the implementation of EQUIP Emergency and the circumstances under which EQUIP is likely to be successful, thereby providing important information about scale-up potential.

Key ethical issues are the prioritizing of care processes over research processes, the protection of privacy and ensuring informed consent for patients and staff, and ensuring that research processes do not compromise staff employment relationships. The emergency departments serving as study sites are very busy and often operate at over-capacity. In this context, patient care will be the top priority and no research activities will interfere with care. Efforts to ensure privacy will include a) using tablets for patient surveys and online surveys for staff, b) supporting patients to enter their own data directly, or at least read the questions (without them being read out loud so that others can overhear), c) using private spaces to complete surveys when this is not possible, and conducting all qualitative interviews with staff and leaders in private spaces. The EDs serve diverse populations, thus efforts to ensure informed consent for patients will include a) reading aloud study information and consent letters unless the patient indicates this is not needed (not assuming literacy), b) offering materials translated into the top two languages at each site (for a total of 4 languages: English, Hindi, Punjabi and Cantonese). All study processes have been reviewed and approved by the Research Ethics Boards at the University of British Columbia and each Health Authority.