Stages of change: Strategies to promote use of a Pediatric Early Warning System in resource‐limited pediatric oncology centers

Abstract Background Pediatric Early Warning Systems (PEWS) assist early detection of clinical deterioration in hospitalized children with cancer. Relevant to successful PEWS implementation, the “stages of change” model characterizes stakeholder support for PEWS based on willingness and effort to adopt the new practice. Methods At five resource‐limited pediatric oncology centers in Latin America, semi‐structured interviews were conducted with 71 hospital staff involved in PEWS implementation. Purposive sampling was used to select centers requiring variable time to complete PEWS implementation, with low‐barrier centers (3–4 months) and high‐barrier centers (10–11 months). Interviews were conducted in Spanish, professionally transcribed, and translated into English. Thematic content analysis explored “stage of change” with constant comparative analysis across stakeholder types and study sites. Results Participants identified six interventions (training, incentives, participation, evidence, persuasion, and modeling) and two policies (environmental planning and mandates) as effective strategies used by implementation leaders to promote stakeholder progression through stages of change. Key approaches involved presentation of evidence demonstrating PEWS effectiveness, persuasion and incentives addressing specific stakeholder interests, enthusiastic individuals serving as models for others, and policies enforced by hospital directors facilitating habitual PEWS use. Effective engagement targeted hospital directors during early implementation phases to provide programmatic legitimacy for clinical staff. Conclusion This study identifies strategies to promote adoption and maintained use of PEWS, highlighting the importance of tailoring implementation strategies to the motivations of each stakeholder type. These findings can guide efforts to implement PEWS and other evidence‐based practices that improve childhood cancer outcomes in resource‐limited hospitals.


| BACKGROUND
Approximately 40% of pediatric oncology patients require critical care during their cancer treatment. 1 Pediatric oncology patients admitted to the pediatric intensive care unit (PICU) experience a higher rate of mortality compared to the general PICU population, particularly in resourcelimited hospitals. 1,2 Pediatric Early Warning Systems (PEWS) are quality improvement tools used to facilitate early detection of critical illness. 3 PEWS consist of a scoring tool and action algorithm to identify and monitor hospitalized children at risk of clinical deterioration. 4 Bedside nurses evaluate vital signs, behavioral indicators, and relative concern to calculate a standardized score. 4 This score is associated with an action algorithm that guides the clinical team's response. 4 PEWS have been shown to reduce clinical deterioration events and PICU utilization, improve interdisciplinary and provider-family communication, enhance perceptions of healthcare quality, support clinician emotions, and result in cost-savings. [4][5][6][7][8][9][10] However, PEWS are rarely used in resources-limited hospitals, in part due to challenges during implementation. 3 Prior work identified stakeholder "stage of change" as integral to successful PEWS implementation. 11 Stakeholders are individuals with interest and influence to affect implementation of evidence-based practices like PEWS. 12 The stages of change are phases of readiness that describe stakeholder willingness to adopt and use a new practice: precontemplation (resistant to behavior change), contemplation (ambivalent towards behavior change), preparation (expressing interest in a plan of action to change behavior), adoption (changing behavior), and maintenance (independently sustaining behavior change). 13 The Capability, Opportunity, Motivation-Behavior (COM-B) model is a behavior change model that explains how stakeholders progress through the stages of change. 14 By explaining what a stakeholder needs for their perceptions towards PEWS to change, this model can guide implementation planning to tailor strategies to unmet stakeholder needs and promote adoption of evidencebased practices like PEWS. This study uses behavior change theory and the COM-B model to explore strategies used by implementation leaders to successfully promote stakeholder adoption and maintained use of PEWS in resource-limited pediatric oncology centers.

| METHODS
We conducted a secondary analysis of PEWS implementation in resource-limited pediatric oncology centers. 11 The primary analysis identified barriers and enablers in the domains of characteristics of individuals (clinical staff), inner (hospital) setting, outer setting (external factors), the PEWS intervention, and implementation process. 11 The stages of change were a major theme identified in the inner setting domain as an important barrier or enabler of implementation success. 11 This secondary analysis evaluated the process by which implementation leaders successfully converted implementation barriers into enablers for clinical staff using a stages of change framework. We followed Consolidated Criteria for Reporting Qualitative Research (COREQ) guidelines. 15 Methodology for the primary study was previously described 11 and is summarized below.

| Site and participant sampling
Escala de Valoración de Alerta Temprana (EVAT) is a Spanish-language PEWS validated in pediatric oncology patients. 16 Proyecto EVAT is a multicenter quality improvement collaborative to scale up PEWS in Latin America. 11 We recruited Proyecto EVAT centers who completed PEWS implementation prior to March 2020. From 23 centers meeting these criteria, we used purposive sampling to select low-barrier (requiring 3-4 months from pilot initiation to successful implementation) and highbarrier sites (10-11 months). A local study lead identified 10-15 clinical and administrative hospital staff involved in PEWS implementation at each center to participate in a semi-structured interview (estimated number required for thematic saturation). 17 Research team members then contacted identified staff for recruitment to the study; all identified participants agreed to participate and completed interviews.

| Data collection
An interview guide assessing barriers and enablers to PEWS implementation ( Figure S1) was developed in English, translated into Spanish, and iteratively revised. 18 We piloted the interview guide with three individuals from a Proyecto EVAT center not recruited for this study. From June to August 2020, bilingual interviewers (PE, SG) conducted virtual semi-structured interviews using WebEx. These interviewers did not have a prior relationship with participants, did not work at the centers, and were not involved in Proyecto EVAT. Interview duration was approximately 1 h. Bilingual interviewers (PE, SG) conducted interviews in Spanish and audio recorded. A professional translation company transcribed, deidentified, and translated the interviews into English for analysis.

| Analysis
Two analysts (AA, GF) created a codebook (Table S1) with codes developed a priori and inductively based on iterative review of nine transcripts. 18 The analysts then independently coded all transcripts using MAXQDA software, achieving a kappa of 0.8-0.9, with discrepancies resolved by a third analyst (DG).
For this study, thematic content analysis was used to explore segments coded as "stage of change", defined as "willingness, or lack of willingness of individuals or authorities in the hospital to gain new skills, accept change, or show interest/enthusiasm for PEWS" (Table S1). Relevant to successful implementation, the stages of change describe a stakeholder's evolving perception towards PEWS. We chose the COM-B model for its description of factors affecting an individual's willingness to change their behavior and move through the stages of change: physical and psychological capability to understand how to do the behavior, physical and social opportunity to provide the necessary resources, and reflective and automatic motivation to create inner drive. 14 Through iterative review of transcripts, we identified strategies used by PEWS implementation leaders to promote stakeholder movement through each stage of change and categorized these based on the behavior change wheel, a framework that describes behavior change strategies. 14 The behavior change wheel defines strategical actions used to address a stakeholder's needs and thus change a stakeholder's perception of PEWS. We used constant comparative analysis to explore how strategies related to stakeholder current stage of change, COM-B factors, and different stakeholder types.

| Ethical considerations
The St. Jude Children's Research Hospital Institutional Review Board approved this study as exempt with minimal risk; study participants provided verbal consent at the beginning of each interview. We obtained additional ethics approval from participating centers as necessary.

| RESULTS
The 71 participants included nurses (45%), physicians (45%), and hospital administrators and data managers (10%) from three low-barrier centers (San Luis Potosi (SLP), Mexico; Cuenca, Ecuador; San Salvador, El Salvador) and two high-barrier centers (Xapala, Mexico; Lima, Peru; Tables S2 and S3). Participants identified two stakeholder types important to PEWS adoption: clinical staff responsible for using PEWS in patient care and hospital directors who approved policies supporting hospitalwide PEWS implementation.
Participants identified six interventions (strategic actions that did not require formal approval)-training, incentives, participation, evidence, persuasion, and modeling-and two policies (directives needing leadership approval)-environmental planning and mandatesthat promoted stakeholder adoption and maintained use of PEWS. Implementation leaders used different strategies to address each stage of change and COM-B factor.

| Strategies across stages of change
Stakeholders presented in various stages of change throughout the implementation process (Table 1). Implementation leaders, therefore, employed distinctive strategies to address different concerns associated with each stage of change (Table 2, Figure 1).
To move stakeholders from the precontemplation to contemplation stage, implementation leaders explained why they should use PEWS: "We knew [ Implementation leaders demonstrated PEWS' feasibility specifically in the local hospital setting, and as stakeholders learned how to use PEWS, stakeholders moved from preparation to adoption: "The pilot showed that PEWS worked, and this maybe opened the doors to create the conditions to run it as a quality project from the hospital" (physician director, San Salvador). Additional environmental planning provided physical resources necessary to use PEWS: "When we had all the material, the boards updated… the first thing [physicians] do is check the sheet and see if someone has a red or yellow [elevated] PEWS, so they can start to work on that patient" (nurse director, Lima).
To move stakeholders from adoption to maintenance, implementation leaders encouraged stakeholders to feel pride as leaders in the program: "When [clinical staff] talk about PEWS, you can see their enthusiasm about the project. They own the project, and I think that was essential. Because the director can be excited, but if the people in charge of applying it are not involved and convinced, it would be more difficult" (nurse director, San Salvador).  [PEWS] will improve the attention for the patient and will also help them in their job because they will no longer have a patient who is in critical condition" (implementation leader, Lima). Training, incentives, participation, and environmental planning specifically targeted clinical staff. PEWS implementation required clinical staff training to develop skills and knowledge, and environmental planning to provide physical resources to use PEWS: "[PEWS] is applied by the nurses. They had to be trained and learn a new way of taking vital signs" (implementation leader, Xapala). Incentives and participation recognized and encouraged the active role of clinical staff in PEWS: "I think the positive reinforcement that nurses get from the nurse leaders [is] that the work is well done, that they have a good attitude. That we are a team, and we must stay together for the children, like the pediatrician with the residents. That big motivation [is] to listen to the nurses" (physician director, Cuenca).

| Strategies across high-and lowbarrier centers
High-and low-barrier centers differed in the amount of time required to move from pilot initiation to successful implementation. Implementation leaders at high-and low-barrier centers differed in their application of some strategies, impacting time required for implementation.
While all low-barrier centers and one high-barrier center used incentives, one high-barrier center did not describe the use of any incentives. Without use of incentives, one high-barrier center waited on outcomes data to motivate staff, thus delaying implementation: "You'll have to spend some time to do measurements… I think in this case it took the doctor 18 months or 24 months. [The doctor] presents the results, and that's when everybody is convinced with the importance of the implementation" (quality director, SLP).
Similarly, PEWS adoption by hospital directors allowed for institutional policies, like mandates, that facilitated dissemination of PEWS: "Our directors and the people chosen to implement the project have credibility, leadership… without those, nobody would have paid attention" (data director, San Salvador). Implementation leaders from low-barrier centers involved hospital directors during the pre-implementation planning phase. In contrast, high-barrier centers waited longer to involve hospital leaders, delaying implementation: "The [nurse director] would put barriers. If [the nurse director] was doing that, the rest of the nurses would never feel like [PEWS] was something that they should do" (physician director, SLP). Early engagement of hospital directors in the PEWS implementation process allowed for earlier use of these effective strategies, thus reducing barriers to implementation.

| DISCUSSION
This study identified six interventions and two policies used by implementation leaders to successfully guide clinical staff and hospital directors through the stages of change to promote adoption and maintained use of PEWS in resource-limited pediatric oncology centers. Collectively, the identified strategies addressed all six COM-B factors required for behavior change, supporting the relevance of the COM-B model in these settings. High-and low-barrier centers used strategies differently, offering potential explanation for different length of time required for implementation. Recognizing different stakeholders' priorities and readiness to accept a new practice, implementation leaders tailored strategies to each stakeholder's stage of change, unmet needs, and specific interests. Stakeholder stage of change, influencing the decision to adopt and continue using a new behavior, is important for successful implementation of any evidence-based practice. 13 In previous work, an important determinant of sustainability and ongoing use of PEWS was hospital staff's perceptions of PEWS' relative importance and impact in patient care. 10,19,20 By using a stages of change framework to understand how stakeholders' perceptions towards PEWS change, this analysis provides further insight into how stakeholders choose to adopt new clinical practices, and how adoption can be promoted through targeted strategies. Identified differences between strategies used at low-and high-barrier centers suggests timing of stakeholder engagement influences implementation time, an outcome particularly relevant to resource-limited settings where longer implementation may require more resources and result in premature implementation abandonment. 21 Based on these findings, we recommend early engagement of hospital directors during planning of any program to implement a new clinical practice. Similarly, incentives and mandates can effectively overcome initial staff resistance to change, promoting early adoption. We also recommend active application of strategies informed by behavior change theory to facilitate implementation in resource-limited hospitals, where challenges with implementation may perpetuate health disparities by delaying uptake of effective clinical practices. 22 Implementation strategies that utilize behavior change theory are less frequently applied to plan implementation in these settings. 22,23 By leveraging behavior change theory to identify strategies for adoption, this study demonstrates how such work can be used to promote equitable implementation of effective clinical practices across variable resource-levels.

| Limitations
This study has several limitations. The study only included centers that completed PEWS implementation as all Proyecto EVAT centers which started eventually successfully implemented PEWS. 24 Thus, these findings have limited generalizability to other evidence-based practices that may have poor uptake or more frequent implementation failure. It is likely, however, that other factors, such as public health crises, political unrest, and resource challenges, rather than stakeholder buy-in, are more relevant to these early implementation challenges. 24 This study's focus on stakeholder stage of change is most relevant to hospitals actively planning and implementing new interventions. This study also focused on pediatric oncology centers implementing one evidence-based practice; thus, our findings may not be generalizable to centers implementing interventions for other populations. However, this study integrated multiple behavior-change theories with empiric evidence about implementation, thus strengthening our findings' validity to understand and contextualize participants' experiences. This study provides insight into strategies to move stakeholders from the precontemplation to maintenance stage for clinical practices in resource-limited hospitals.