Conceptual framework for task shifting and task sharing: an international Delphi study. — TEI-tagged view

Aaron M. OrkinpersName, Sampreeth RaopersName, Jeyasakthi VenugopalpersName, Natasha KithulegodapersName, Pete WegierpersName, Stephen D. RitchiepersName, David VanderBurghpersName, Alexandra MartiniukpersName, Fabio Salamanca-BuentellopersName, Ross UpshurpersName (2021date). “Conceptual framework for task shifting and task sharing: an international Delphi study..” Human Resources for Healthtitle·j.

Object of study — task shiftingterm·dom and task sharingterm·dom in health workforcesterm·key, and the contexts suited to it

Research question — what differentiates task shiftingterm·dom, and which criteria identify contexts suited to it

Contribution — the COATS Frameworkterm: a definition, purpose statement, five opportunities and eight implementation criteria for task shifting and sharingterm·dom

Full text — tagged

The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.

Conceptual framework for task shiftingterm·dom and task sharingterm·dom: an international Delphiterm·mth study.

Abstract Task shifting and sharingterm·dom (TS/Sterm·dom) involves the redistribution of health tasks within workforces and communities. Conceptual frameworks lay out the key factors, constructs, and variables involved in a given phenomenon, as well as the relationships between those factors. Though TS/Sterm·dom is a leading strategy to address health worker shortages and improve access to services worldwide, a conceptual framework for this approach is lacking.

We used an online Delphiterm·mth process to engage an international panel of scholars with experience in knowledge synthesis concerning TS/Sterm·dom and develop a conceptual framework for TS/Sterm·dom. We invited 55 prospective panelists to participate in a series of questionnaires exploring the purpose of TS/Sterm·dom and the characteristics of contexts amenable to TS/Sterm·dom programmes. Panelist responses were analysed and integrated through an iterative process to achieve consensusterm·key on the elements included in the conceptual framework.

The panel achieved consensusterm·key concerning the included concepts after three Delphiterm·mth rounds among 15 panelists. The COATS Frameworkterm (Concepts and Opportunities to Advance Task Shifting and Task Sharingterm) offers a refined definition of TS/Sterm·dom and a general purpose statement to guide TS/Sterm·dom programmes. COATSterm describes that opportunities for health system improvement arising from TS/Sterm·dom programmes depending on the implementation context, and enumerates eight necessary conditionsterm·dom and important considerationsterm·dom for implementing TS/Sterm·dom programmes.

The COATS Frameworkterm offers a conceptual model for TS/Sterm·dom programmes. The COATS Frameworkterm is comprehensive and adaptable, and can guide refinements in policy, programme development, evaluation, and research to improve TS/Sterm·dom globally.

The online version contains supplementary material available at 10.1186/s12960-021-00605-z.

Background Shortages of trained and accessible health professionals are a threat to health and health equity worldwide. Health systems and professionals face substantial burdens to respond to new pressures, such as the COVID-19 pandemic, while maintaining the operation of routine services and care. Task shiftingterm·dom and task sharingterm·dom (TS/Sterm·dom) involve the strategic redistribution of tasks among health workforceterm·key teams and personnel. Specific tasks are moved, shared or delegated, usually from highly trained health workers to those with shorter training or fewer qualifications, including laypeople [ 1 ].

Task shiftingterm·dom and task sharingterm·dom is a promising way to address global health workforceterm·key shortages and insufficient access to care for critical health problems, and has been used effectively to improve health in a range of contexts and conditions [ 1 , 2 ]. Task shiftingterm·dom and task sharingterm·dom strategies allow for more efficient use of health human resources, especially as health systems worldwide struggle to maintain essential services while responding to the COVID-19 pandemic [ 2 ].

Conceptual frameworks lay out the key factors, constructs, and variables involved in a given phenomenon, as well as the relationships between those factors [ 3 ]. Conceptual frameworks can guide programme design and implementation, and assessments of programme performance, including effectiveness, equity, efficiency and quality [ 3 , 4 ]. Conceptual frameworks offer a theoretical synthesis of a given phenomenon, however, there are no conceptual frameworks to identify commonalities among the contexts and conditions suited to TS/Sterm·dom programmes [ 5 , 6 ]. A conceptual framework for the conditions and contexts suited to TS/Sterm·dom programmes could refine the development and implementation of TS/Sterm·dom programmes and research globally.

The goal of this study was to develop a conceptual framework for TS/Sterm·dom, using an online Delphiterm·mth process to engage an international panel of scholars with experience in knowledge synthesis around TS/Sterm·dom.

Methods Delphiterm·mth design We used a Delphiterm·mth process to guide the iterative development of the Concepts and Opportunities to Advance Task Shifting and Task Sharingterm (COATSterm) Framework [ 7 ]. We used a modified online Delphiterm·mth process to engage and gather opinions from scholarly experts on TS/Sterm·dom, and identify consensusterm·key responses to two research questions: What differentiates task shiftingterm·dom from other circumstances where workers with limited training or qualifications are involved in health care? What criteria can be used to identify conditions and contexts suited to task shiftingterm·dom?

The Delphi methodterm·mth is a structured approach to develop consensusterm·key among experts through a series of questionnaires and structured feedback. It involves collecting and distilling knowledge from a dispersed panel of experts to build reliable group consensusterm·key on a specific issue. This approach has been used to forecast trends across a range of policy and research domains, and to drive decision-making and resource allocation [ 8 ]. There are numerous Delphiterm·mth variants, but most share design features such as purposive sampling, participant anonymity, iterative design, structured communication between participants and researchers, thematic analysis of expert responses, and statistical aggregation of group feedback [ 7 , 9 , 10 ]. Delphi methodsterm·mth add structure to the process of building group consensusterm·key and reduce biases that can arise in face-to-face meetings by limiting the social, psychological, and political aspects of conventional group interactions, improving the quality, efficiency, long-term accuracy and validity of panelists’ predictions and judgements [ 10 , 11 ].

The Delphiterm·mth panel and method were not intended to deliver consensusterm·key or final approval for the specific language, structure, and format of the conceptual framework. The Delphiterm·mth process was intended to solicit and synthesize a broad range of knowledge and experience concerning TS/Sterm·dom and provide consensusterm·key on the elements included in the COATS Frameworkterm.

Panel selection, recruitment and retention We determined that three Delphiterm·mth rounds using email for correspondence would be sufficient to achieve consensusterm·key and stability within our expert panel [ 7 , 9 , 10 ]. We aimed to retain a minimum of 12 experts after three rounds of Delphiterm·mth participation, and based on our experience with previous Delphiterm·mth studies, we planned for 40% attrition in each round. Therefore, we estimated that 40 invitees would be required in the first round of the Delphiterm·mth.

We used two techniques to convene the panel of scholarly experts. The first set of panelists recruited to the Delphiterm·mth were the corresponding authors from published, international systematic reviews on TS/Sterm·dom (n = 36). Rather than approaching researchers engaged in the evaluation of individual TS/Sterm·dom programmes, we chose scholars who had demonstrated experience in knowledge synthesis around TS/Sterm·dom and could contribute to the development of a conceptual framework. We deliberately sought scholars who had published systematic reviews because these experts would have high-level, global knowledge of TS/Sterm·dom and the concepts needed to synthesize multiple TS/Sterm·dom programmes. They were identified from a separate systematic review underway by members of our research team [ 12 ]. Given that these invitees were corresponding authors on global systematic reviews concerning TS/Sterm·dom, this sample offered a panel with global and representative knowledge of the scholarly literature and research concerning TS/Sterm·dom. Second, panelists in our initial set of 36 contacts were invited to recommend colleagues who, in their opinion, might be interested in participating in this Delphiterm·mth on the basis of their work or expertise.

Throughout the Delphiterm·mth process, panelists were blinded to the identity of the other panelists, except for the individuals who referred us to subsequent panelists. Survey content was never associated with a panelist identifier; only the investigators could associate panelists with responses. All the included questions from the survey concerned the respondents’ area of professional expertise.

We sent email reminders to panelists on a weekly basis to increase the response rate. As an incentive to complete the study and to express appreciation for panelists’ participation, we offered to donate in the participant’s name to a medical humanitarian organization of the participants’ choice. We gathered information about panelists’ institutional affiliations and scholarly work from their publicly available academic profiles and publications.

Delphiterm·mth rounds, data collection and analysis To ensure a systematic and meaningful synthesis of responses, we drafted and refined the questions asked of the panelists in every round. We piloted the questionnaires among members of our authorship team who were not directly involved in designing the Delphiterm·mth. We communicated with the panelists in English and used Google Forms to conduct our surveys.

We developed a prespecified definition of consensusterm·key based on two criteria [ 13 ]. First, the panel was eligible to have achieved consensusterm·key around a given survey item if at least 70% of respondents agreed with that item. When using a seven-point Likert scale, we defined “disagreement” as a score of three or less. This criterion ensured that a strong majority of respondents agreed with any included survey item. Second, an item was said to have achieved consensusterm·key only if none of the dissenting respondents raised concerns that were fundamentally incompatible with the inclusion of that survey item. This criterion aligns with approaches from formal consensusterm·key decision-making, where a structured discussion is used to understand and resolve the merits and drawbacks of a given proposal [ 14 ]. This approach recognizes that essential insights can be tendered by a minority of decision-makers, and attends to the substance of minority opinions. Procedurally, these minority opinions were gathered by requiring that panelists offer free-text comments if they disapproved of a survey item. We analysed these free-text responses and incorporated that feedback into subsequent rounds of the Delphiterm·mth and into the final COATS frameworkterm. As the analysis advanced, an emphasis on and reiteration of certain issues above others became more apparent. These elements coalesced into the final categories and themes.

Round I was designed to elicit broad and general concepts from the panelists using unstructured, open-ended, questions: What is, in your opinion, the purpose of task shiftingterm·dom? What are the three to five characteristics of a health problem that make it amenable to task shiftingterm·dom? What are three to five examples of delegation of responsibilities from highly qualified health workers to individuals with fewer qualifications and shorter training that are not task shiftingterm·dom?

Following Round I, investigators combined and analysed the panelists’ responses in a taxonomy according to common themes and categories. We attempted to make the items on each list mutually exclusive and comprehensive. We synthesized these findings in a survey to elicit participants’ level of agreement with each of the themes and categories on a seven-point Likert scale for Round II. This survey also offered free-text response options for panelists to add additional comments or categories as required.

Our study initially used the term "task shiftingterm·dom" in isolation, rather than the broader concepts of both task shiftingterm·dom and task sharingterm·dom. In Round I, panelists raised conceptual distinctions between task shiftingterm·dom and task sharingterm·dom. Recognizing that this had emerged organically from the Delphiterm·mth process, we added questions to refine the distinction between task shiftingterm·dom and task sharingterm·dom and integrated this distinction into our conceptual model.

Once we received all responses from Round II, these data were again reviewed by the investigators. Concepts were eliminated and retained on the basis of the panelists’ scores and collapsed into more general categories, including a definition of TS/Sterm·dom, the purpose of TS/Sterm·dom, opportunities arising from TS/S programmesterm·dom, and conditions required for the implementation of TS/Sterm·dom. These results were sent back to the panelists as a survey for Round III. Panelists were asked to review the final list of items, state whether they agreed or disagreed with each item, and voice concerns or comments in free text. Following Round III, the investigators integrated the experts’ consensusterm·key responses into a reasonable and manageable set of concepts and sub-concepts to form the conceptual framework [ 3 ].

Results The Delphiterm·mth process was initiated on August 26, 2019 and completed on June 20, 2020. The process was paused in March through May 2020 given the COVID-19 pandemic.

Figure 1 provides a flowchart of panelist recruitment and retention in each of the Delphiterm·mth phases. Although the response rate for the first Delphiterm·mth round was 31% (17/55), attrition thereafter was low, with one non-responding panelist in each of the subsequent rounds. Fifteen (15) panelists participated in the full Delphiterm·mth. All panelists were university faculty or research institute appointees with expertise related to TS/Sterm·dom in the United States (7), South Africa (3), the United Kingdom (2), Nigeria (1), India (1), and Australia (1). Panelists included qualitative and quantitative researchers with primary appointments in public health and epidemiology (6), medicine (3), medical anthropology (3), clinical psychology (2), and health economics (1). Twelve panelists conducted all or most of their work concerning TS/Sterm·dom in low- or middle-income countries. Fig. 1 Flow chart of panelist recruitment and retention The panel arrived at a consensusterm·key concerning the themes and content of the conceptual model following three Delphiterm·mth rounds. Additional file 1 provides the questionnaires for each round of the Delphiterm·mth with summary statistics of panelist responses to each item. In Round I, 17 concepts were identified to describe conditions and contexts suited to TS/Sterm·dom. In Round II, these features were distilled into eight conditions by combining similarly ranked survey items and removing lower ranked features. In Round III, the eight conditions were split into four necessary conditionsterm·dom and four important considerationsterm·dom based on dissenting participants’ reasoning. Additional file 1 provides an analysis of dissenting views from Round III and how these were interpreted and incorporated into the COATS Frameworkterm. Dissenting views did not tend to oppose the concepts in the emerging framework. Dissenting comments generally offered feedback to soften or adapt the emerging framework so that it would be applicable to a wider range of contexts, communities of practice, and specific circumstances. Round III achieved consensusterm·key around all survey items.

The COATS Frameworkterm consists of three elements (Fig. 2 ): Definition and purpose of TS/Sterm·dom programmes (Yellow Box); Opportunities arising from TS/S programmesterm·dom (Blue Box); and Criteria for the implementation of a TS/Sterm·dom programme (Red Box). The analysis below describes how these elements emerged from the Delphiterm·mth process and how the framework can be implemented and adapted for various contexts. Throughout the COATS frameworkterm, the term “intervention” is used to refer to the task that is shifted or shared in the context of a TS/Sterm·dom programme. Interventions in this context may be preventive, curative, therapeutic, diagnostic, or another health action. For linguistic clarity, we recommend that the overall TS/Sterm·dom initiative be referred to as a programme rather than an intervention. Definition and purpose of TS/Sterm·dom sharing programmes (Yellow Box) The COATS Frameworkterm expands on the dominant WHO definition of task shiftingterm·dom with a revised general definition and purpose statement for TS/Sterm·dom programmes. Where the WHO defines task shiftingterm·dom as the “rational redistribution of tasks among health workforceterm·key teams”, the COATSterm definition offers a broader concept involving the redistribution or delegation of health care tasks within workforces or communities. This reference to communities underscores the role of lay and informal providers in TS/Sterm·dom initiatives. Delphiterm·mth panelists offered feedback concerning the distinction between task shiftingterm·dom and task sharingterm·dom. This led to refinements in the COATS Frameworkterm to offer distinctions between task shiftingterm·dom and task sharingterm·dom along with an overarching definition of TS/Sterm·dom. This distinction is also reflected in the TS/Sterm·dom literature [ 15 – 17 ]. In Round II, seven panelists identified task shiftingterm·dom and task sharingterm·dom as related but different entities, where task shiftingterm·dom concerns delegation, while task sharingterm·dom focuses on collaboration between different workers. Another eight panelists indicated that task shiftingterm·dom and task sharingterm·dom refer to the same phenomenon, but that task sharingterm·dom is the preferred term because it captures the intrinsically collaborative nature of these undertakings. The COATS Frameworkterm captures this distinction, and permits programme developers, scholars and other stakeholders to make deliberate decisions regarding their use of task shiftingterm·dom versus task sharingterm·dom terminology and approaches in their own work. The generalized purpose statement indicates that TS/Sterm·dom is intended to reduce morbidity, mortality, and burden of disease among populations where the inaccessibility of professionalized cadres limits access to effective care. Although health care tasks might be redistributed to providers with less training to deliver cost savings or improve operational efficiencies, this purpose statement reinforces that TS/Sterm·dom occurs when health care redistribution and delegation is driven by efforts to enhance health equity by meeting the health needs of underserved populations [ 1 , 18 , 19 ]. The general purpose statement also asserts that TS/Sterm·dom achieves its goals without compromising the standards of care, including maintaining safety and quality [ 20 – 23 ]. Opportunities arising from TS/S programmesterm·dom (Blue Box) The COATS Frameworkterm offers five opportunities for TS/Sterm·dom programmes to deliver health system improvements, depending on the implementation context. The Delphiterm·mth panelists identified that TS/Sterm·dom programmes are diverse and contextual, and that a given programme may therefore present opportunities that are particular to their specific settings and circumstances. These five opportunities demonstrate that TS/Sterm·dom programmes may have strategic functions that extend beyond reductions in morbidity and mortality. A given TS/Sterm·dom programme may build on one or more of these opportunities—or none at all. For example, a mental health TS/Sterm·dom programme may be designed specifically to deliver more culturally appropriate care by engaging community members who are more likely than professionals to share cultural, linguistic and social background with those receiving care [ 24 ]. Other TS/Sterm·dom initiatives, such as peer worker engagement in harm reduction programmes, may emphasize a change to conventional professional hierarchies by positioning peer workers to assume roles and responsibilities within the healthcare system that were previously restricted to professionals [ 25 ]. The rapid scale-up of essential interventions has been a priority for TS/Sterm·dom programmes related to HIV care in sub-Saharan Africa [ 21 , 26 ]. Practitioners, programme developers, policymakers and researchers may use these five opportunities to plan and appraise specific TS/Sterm·dom programmes within their specific contexts and programme goals. Implementation criteria for TS/Sterm·dom programmes (Red Box) Fig. 2 Concepts and Opportunities to Advance Task Shifting and Task Sharingterm (COATSterm) Framework The final section of the COATS Frameworkterm provides criteria for programme developers, policymakers, researchers and other stakeholders to assess if a given condition and context is suited to TS/Sterm·dom. Panelists identified that conditions that are critical to the success of TS/Sterm·dom programmes in one setting may be less relevant or absent in others.

The criteria for the implementation of a TS/Sterm·dom programme are divided into “necessary conditionsterm·dom” and “important considerationsterm·dom”. Necessary conditionsterm·dom refer to features of the available workers, the health problem, and the intervention that will generally need to be in place to make a TS/Sterm·dom programme successful. For example, workers willing to be trained to deliver the intervention and workers willing to provide that training will both be prerequisite for any successful TS/Sterm·dom programme, and programme developers should consider these needs as they conceive of a new TS/Sterm·dom initiative. Important considerationsterm·dom refer to concepts that will enable success in some settings but may be less relevant in others. For example, one panelist indicated that TS/Sterm·dom often requires resources for scale-up, but that “those resources don’t always need to be identified before the intervention is developed or tested.” Another panelist identified that the social acceptability of an intervention may not be a precondition to a successful TS/Sterm·dom programme because TS/Sterm·dom may be used to help improve the social acceptability of an intervention within a given community (Additional file 1 ).

The framework does not prescribe a measure or threshold to establish if the criteria have been met. These criteria are intended for interpretation and adaption to their specific context so that practitioners and programme developers can appraise whether and how the criteria have been addressed in their particular circumstances.

Discussion Task shiftingterm·dom and task sharingterm·dom is a leading and promising health systems strategy to address health workforceterm·key shortages, transform health care delivery, and improve health outcomes and inequities [ 1 , 2 ]. However, TS/Sterm·dom is not a universal solution to insufficient or inequitable access to care. Conceptual frameworks can guide and refine a range of public health policies and decision-making, including for example, the implementation of screening programmes, immunization programmes, action on the social determinants of health, or public health ethics [ 4 , 27 – 30 ]. The COATS Frameworkterm offers a conceptual framework to guide and refine TS/Sterm·dom programmes.

Other researchers have examined the various context-specific facilitators and barriers to successful TS/Sterm·dom implementation [ 31 , 32 ]. This study builds on this research with a consensusterm·key from international TS/Sterm·dom scholars to advance a conceptual framework for TS/Sterm·dom programmes. A central strength of our study is the use of a high quality international Delphiterm·mth process, with reproducible panel selection processes, prespecified consensusterm·key criteria, and a prespecified number of Delphiterm·mth rounds [ 13 ]. This process supported the iterative theorization processes involved in the development of a conceptual framework, achieved consensusterm·key across the COATSterm concepts, and integrated the breadth and diversity of TS/Sterm·dom experience and expertise worldwide [ 3 ]. The framework was built based on the prevailing consensusterm·key of the panelists, but also incorporated dissenting views to ensure that the resulting framework incorporated varied terminology (task shiftingterm·dom and task sharingterm·dom) and was adaptable to a wide range of contexts and circumstances.

The COATS Frameworkterm offers a practical definition of TS/Sterm·dom (including the distinction between task shiftingterm·dom and task sharingterm·dom approaches), and guides programme development and implementation with a central purpose statement for TS/Sterm·dom. The necessary conditionsterm·dom and important considerationsterm·dom provides criteria that stakeholders can use to support TS/Sterm·dom programme implementation. The COATS Frameworkterm is both simple and adaptable: it offers a singular purpose and conditions suitable to all TS/Sterm·dom programmes, but also incorporates opportunities and criteria that attend to the diversity of circumstances or contexts where TS/Sterm·dom may be implemented. The COATS Frameworkterm is intended to prompt and support adaptations to local circumstances and contexts.

Our study and resulting framework have limitations. Our Delphiterm·mth panel was convened deliberately to engage panelists with scholarly experience in conducting syntheses across TS/Sterm·dom programs and elicit consensusterm·key among these practitioners. The findings are therefore a consensusterm·key among participating panelists and may not be representative of other TS/Sterm·dom experts. One-third of panelists were from low- and middle-income countries and the majority conducted their work concerning TS/Sterm·dom in lower-income populations. Therefore, our findings and framework may require adaptation for other populations and settings. Panelists’ expertise may differ from the voices of policymakers, providers, or patients, and the resulting framework may well be adapted by other stakeholders. Our definition of TS/Sterm·dom did not explicitly include task redistribution directly to patients or artificial intelligence systems [ 6 ]. However, patient self-care and artificial intelligence may be sufficiently distinct from conventional TS/Sterm·dom to merit separate conceptual frameworks.

Conclusion The COATS Frameworkterm delivers an essential theorization of TS/Sterm·dom. Better theorization of TS/Sterm·dom can drive more deliberate, strategic, and effective programmes. The COATS Frameworkterm is both comprehensive and adaptable, and suited to support refinements across a range of TS/Sterm·dom policy and programmes, including planning, decision-making, implementation, evaluation and research worldwide.

Supplementary Information Additional file 1. TS/Sterm·dom Delphiterm·mth Questionnaires and Analysis of Dissenting Views.