Background

Low- and middle-income countries (LMICs) face a growing burden of cancer and a pressing need to strengthen their cancer care delivery systems. Predictions suggest that by 2030, 13 million people will die from cancer annually, and three quarters of the deaths will occur in LMICs [1]. The overall case fatality from cancer in low-income countries is approximately 75%, compared with 46% in high-income countries (HICs) [2]. This outcome gap, largely attributable to disparities in access to early detection and standard treatment, translates into millions of preventable deaths.

Effective delivery of evidence-based practice is a critical component of addressing global disparities in cancer outcomes. Evidence-based clinical practice guidelines are widely used in oncology for clinical decision-making, healthcare quality assessment, payment decisions, and training. In recent years, several international organizations have developed resource-stratified clinical practice guidelines for use in LMICs. This began with the Breast Health Global Initiative (BHGI) in 2006 [3] and was followed by the National Comprehensive Cancer Network (NCCN) in 2015 [4] and the American Society for Clinical Oncology (ASCO) in 2016 [5]. In November 2017, the African Cancer Coalition and partners announced the new NCCN Harmonized Guidelines™ for sub-Saharan Africa for prevalent cancers and supportive care categories [6].

Despite these highly publicized international efforts to develop resource-stratified clinical practice guidelines for cancer, there has been little research to evaluate the best strategies for dissemination and implementation in LMICs. Generally speaking, guideline publication alone is insufficient to result in widespread adoption into routine practice [7]. This may be especially true in LMICs, which are the target of increasing numbers of guidelines from international and national health authorities. Well-intended guidelines frequently prove ineffective in LMICs due to inadequate resources to support successful dissemination and implementation [8]. Barriers to clinical guideline implementation in general have been well described and include provider factors such as knowledge and attitude, guideline factors such as format and content, and external factors such as lack of resources, organizational constraints, heavy workload, and cultural norms [9]. Recent surveys of oncology providers in LMICs identify numerous barriers to the successful implementation of international cancer treatment guidelines, namely inadequate infrastructure and inclusion of an overwhelming amount of complex information in the guidelines [10, 11].

Considerable research from the field of dissemination and implementation science (D&IS) has shown that structured, multifaceted implementation strategies designed to target barriers to guideline use are most likely to improve guideline adherence [9, 12, 13]. Theories, models, and frameworks are increasingly used to identify the determinants of guideline use in a specific context and design interventions tailored to overcome barriers and leverage facilitators [14, 15]. Notably, the vast majority of guideline implementation research has been conducted in HICs [16]. There is growing recognition of the urgent need to investigate how to adapt proven implementation strategies to LMIC settings, as well as to develop and evaluate novel approaches for LMICs [17, 18]. In cancer care and control broadly, implementation interventions have been characterized by uneven quality and questionable impact even in HICs, likely due to the unique complexity of the field and failure of researchers to consistently embrace high-quality D&IS standards, such as adequately describing all aspects of the interventions under investigation [19]. The global oncology community has acknowledged the need for D&IS [20,21,22], but the development and evaluation of guideline implementation strategies for cancer management in LMICs remains an unmet need.

In preparation for the launch of Tanzania’s first-ever National Cancer Treatment Guidelines, we developed a theory-driven implementation strategy for guideline-based practice at Ocean Road Cancer Institute (ORCI) in Dar es Salaam, Tanzania. We aim to respond to calls for detailed description of how implementation interventions are developed by using the Intervention Mapping framework to describe our stepwise process in accordance with the Template for Intervention Description and Replication (TIDieR) checklist [23,24,25].

Methods

Setting

The United Republic of Tanzania is an East African country of nearly 60 million people, and Dar es Salaam is the largest city and leading commercial center. Tanzania is classified as a low-income country by the World Bank [26]. GLOBOCAN 2018 estimated 42,000 new cancer cases and over 28,000 cancer deaths per year in Tanzania [27]. In 1996, Ocean Road Cancer Institute (ORCI) was established as the national referral center for cancer in Dar es Salaam, with a mission to provide equitable, accessible, affordable, and high quality services of early detection and cancer care to the public [28]. The government of Tanzania sponsors free care to 5400 new cancer patients per year at ORCI, including radiotherapy, chemotherapy, and palliative care. Other services such as diagnostic pathology and surgery are provided at affiliated Muhimbili National Hospital and other referring hospitals and clinics throughout the country. Beginning in 2017, Tanzania’s Ministry of Health, Community Development, Gender, Elderly and Children (MoHCDGEC) commissioned development of the country’s first comprehensive National Cancer Treatment Guidelines. The process for the development of the national guidelines is described elsewhere [29]. In preparation for the planned guideline launch in 2019, a team embedded within our broader institutional collaboration between Muhimbili University of Health and Allied Sciences (MUHAS), ORCI, and the University of California, San Francisco (UCSF) (“The MUHAS-ORCI-UCSF Cancer Collaboration”) developed a dissemination and implementation strategy for ORCI using a theory-driven approach.

Intervention Mapping

Intervention Mapping is a framework for intervention development that maps out a stepwise process from needs assessment to evaluation [30]. The six steps in Intervention Mapping integrate theory and evidence into the major program planning activities of conducting a needs and capacity assessment, developing and implementing a program, and evaluating a program’s effectiveness [24]. We used Intervention Mapping as a basis for developing a guideline dissemination and implementation strategy at ORCI in Tanzania.

Step 1: Needs assessment

ORCI leaders and staff previously identified a need for improved translation of evidence to practice and standardization in clinical care, which is substantiated by available data indicating that significant numbers of patients at ORCI have not received standard treatment despite resource availability [31,32,33]. The proposed launch of National Cancer Treatment Guidelines in Tanzania presented an opportunity to improve evidence-based practice; however, ORCI leaders recognized that a dedicated implementation effort would be necessary to ensure that the guidelines are adopted into routine practice rather than relegated to collect dust on the shelves. The goal of our needs assessment was therefore to identify what would be required in order to implement the new guidelines at ORCI. We began with brainstorming sessions [34] among members of our research team, which includes oncologists, oncology nurses, clinical and qualitative researchers, and an implementation scientist. We then held meetings with key stakeholders, including ORCI leaders, clinical managers, oncology trainees (“residents”), and patient advocates. Finally, we conducted three focus groups with ORCI oncologists, residents, radiotherapists, and nurses (Luhar et al., unpublished data, 2019). Through this formative evaluation, we identified barriers and facilitators to guideline-based practice at ORCI. We crosschecked our findings with the literature on determinants of guideline implementation.

Step 2: Program objectives

The main objective of our program is to develop an implementation strategy that will effectively lead to the adoption of guideline-concordant practice at ORCI. Based on the needs assessment in Step 1, we identified proximal program objectives such as expanding access to treatment guidelines, increasing familiarity with guideline content, and improving attitudes toward guideline-based practice among providers. Following the proximal objectives, we identified both behavioral and environmental performance objectives, which include increasing guideline-based decision-making and rates of guideline-concordant treatment plans made and completed, and establishing clinical systems that promote guideline-concordant practice. The long-term objectives are to reduce inappropriate variability in clinical practice and improve quality of care, patient outcomes, and resource utilization.

Step 3: Select theory-based methods and practical strategies

Successful implementation of clinical practice guidelines depends on uptake by care providers, which requires sustained behavior change. In order to design an intervention that would optimally target the behavior of guideline-based clinical practice, we used the Capability, Opportunity, Motivation and Behavior/Behavior Change Wheel (COM-B/BCW) framework [35]. COM-B/BCW was developed through a systematic review and synthesis of 19 existing behavior change frameworks and provides a coherent, systematic method for identifying and organizing all potential barriers to behavior change, selecting the barriers that, if modified, are most likely to lead to behavior change in a given context, and choosing evidence-based behavior change techniques most likely to be effective in overcoming targeted barriers. We categorized the key organizational-level and individual-level barriers identified in our needs assessment into the COM-B domains of Capability, Opportunity, and Motivation (Table 1). Through iterative consultation with oncology providers and clinical leaders at ORCI, we used the BCW framework to (1) select intervention functions to address each key barrier, (2) select behavior change techniques likely to help enact each intervention function, and (3) select a feasible mode of delivery for each technique (Table 2).

Table 1 COM-B Theoretical Domains Framework for barriers to adoption of guideline-based clinical practice at ORCI
Full size table
Table 2 Behavior Change Wheel (BCW) framework for adoption of guideline-based clinical practice at ORCI
Full size table

Step 4: Program plan

We organized the behavior change techniques and modes of delivery derived in Step 3 into a phased implementation strategy, summarized in Table 3. The focus of phase 1 is guideline dissemination, with distribution of hard and soft copies and a publicity campaign. Phase 2 includes dedicated knowledge and skills training at a National Cancer Treatment Guideline Summit, and phase 3 encompasses ongoing reinforcement through clinical systems restructuring, point-of-care clinical forms, and behavior modeling and promotion of guideline adherence by Implementation Champions (“Champions”).

Table 3 Summary of phased implementation strategy derived from the BCW/COM-B framework
Full size table

Step 5: Program implementation

We developed a logic model to guide the planning, implementation, and evaluation of our intervention (Table 4). To operationalize the components, we developed a project management spreadsheet, divided the responsibilities among our team, and have held biweekly videoconference calls to review progress, discuss problems, and plan next steps. ORCI team leaders have worked closely with the Tanzanian MoHCDGEC to coordinate the publication of hard copies of the guidelines and to plan a National Summit for Guideline Training. We established a linkage system, or mechanism to involve program adopters and implementers, through engendering program ownership among ORCI-based team leaders and clinical managers and training Champions.

Table 4 Logic model for implementation of guideline-based clinical practice at ORCI
Full size table

Step 6: Evaluation plan

Based on the logic model, we identified the relevant indicators of the process, outcome, and impact of our intervention. We classified and refined these indicators using the RE-AIM framework [36], orienting our evaluation toward issues relevant to program adoption, implementation, and sustainability in order to strengthen its external validity. For each indicator, we determined an appropriate research methodology for measurement among the categories of direct observation, questionnaire administration, clinical chart reviews, and cost-effectiveness analysis (Table 5). We then developed research protocols employing these methods for a selection of indicators. The first study, “aim 1,” will evaluate the activities and outputs of the implementation strategy itself in order to assess its feasibility, acceptability, fidelity versus adaptation, and sustainability using direct observation and pre-post questionnaires. The second study, “aim 2,” will evaluate the effectiveness of the implementation strategy using a pre-post design focused on guideline-concordant treatment completion, healthcare quality metrics, and survival outcomes for breast cancer and colorectal cancer. The third study, “aim 3,” will evaluate the impact of guideline-based practice on the cost of cancer treatment and resource utilization at ORCI. The results of these studies will be reported separately.

Table 5 Indicators of process, outcome, and impact classified using the RE-AIM framework
Full size table

Discussion

In Tanzania and many other LMICs, shifting to guideline-based oncology practice represents a change in clinical culture and behavior, and guideline publication alone is unlikely to result in sustained adoption or measurable impact on clinical care delivery. Using the stepwise process outlined by the Intervention Mapping framework, our team successfully developed a guideline implementation strategy derived from the prominent behavior change theory COM-B/BCW. To our knowledge, this is the first report of a multifaceted theory-driven implementation strategy designed to promote the uptake of cancer treatment guidelines in sub-Saharan Africa.

Given the recent surge in resource-stratified cancer treatment guidelines internationally, there is a widespread need for dedicated guideline implementation efforts. While the intervention described here is tailored to ORCI, we surmise that there are many commonalities between the needs at ORCI and other centers in other centers in sub-Saharan Africa and in other LMICs. Indeed, the barriers to guideline implementation that we identified at ORCI are consistent with those found in a scoping review of the general literature, including lack of familiarity and agreement with guideline-based practice among physicians, insufficient access to guidelines, and constraints within clinical systems and resources [9]. Additional barriers reported in low-resource settings emerged in our formative evaluation as well, including lack of technical capacity, tradition of using expert opinion-based approaches, lack of training on guideline use, and competing priorities [8]. Moreover, the components of our intervention derived through COM-B/BCW map onto proven guideline implementation strategies, such as the distribution of educational materials (e.g., hard and soft copies of the guidelines) and media (e.g., publicity campaign) in phase 1 of our intervention, educational meetings and marketing in the form of interactive trainings at the summit in phase 2, local opinion leaders (e.g., Implementation Champions) and reminders (e.g., clinical forms) in phase 3, and audit and feedback in the evaluation plan [37]. These consistencies with well-established barriers and strategies enhance the validity of our process and results and predict a degree of generalizability to other settings.

Notably, however, these “proven” guideline implementation strategies have largely been tested in HICs. In a 2017 Cochrane overview of 18 systematic reviews of implementation strategies to change health worker behavior, only 1.6% of 820 primary studies took place in a low-income country, and 10% in a middle-income country [16]. While it seems plausible that guideline implementation strategies may be similarly effective in LMICs, this cannot be assumed. Our project responds to calls for “urgently needed” investigation of the transferability of evidence on implementation strategies generated in resource-rich countries, including research to learn how to best adapt strategies for LMICs as well as the discovery and evaluation of novel approaches [17].

Our rigorous evaluation plan will measure the process, outcomes, and impact of our intervention at ORCI. Importantly, we will also systematically document the modifications and adaptations made to the originally planned intervention using the Framework for Reporting Adaptations and Modifications-Expanded (FRAME) [38]. If our intervention ultimately proves effective, the next step will be to validate the approach at other sites in Tanzania and beyond, ideally using a quasi-experimental design. We hope that by providing a detailed, stepwise description of our intervention development process, others may endeavor to replicate the process in their settings.