Multisectoral coordination during the COVID-19 pandemic: practices, challenges and recommendations for future preparedness—a systematic literature review protocol

Abstract

Introduction

The COVID-19 pandemic amplified the need for robust multisectoral coordination; yet the specific mechanisms, benefits and challenges of such collaboration particularly in low- and middle-income countries (LMICs) remain poorly synthesised. This review aims to delineate the key elements, benefits, challenges and improvement strategies of multisectoral coordination during COVID-19 and to compare patterns between LMICs and high-income countries (HICs).

Methods and analysis

Eligible studies will include empirical qualitative, quantitative or mixed-methods research published in English between 1 January 2020 and 15 August 2024 that examines formal coordination mechanisms (eg, task forces, public-private partnerships, inter-agency committees) within the context of COVID-19. Searches will be conducted across PubMed, EBSCOhost, Emerald Insight, Google Scholar and selected grey-literature repositories. Citation chaining will be employed to identify additional sources.

Two reviewers will independently screen all records using Covidence, applying pre-piloted eligibility criteria to 5% of citations and proceeding only if inter-rater reliability achieves κ≥0.70. Data will be extracted into a Consolidated Framework for Implementation Research (CFIR)-informed template. Qualitative data will be analysed through framework synthesis, structured by the five CFIR domains. Quantitative data will be narratively summarised and, where outcomes are sufficiently similar across at least two studies, synthesised using a fixed-effect model.

Risk of bias will be assessed using Critical Appraisal Skills Programme for qualitative and Risk Of Bias In Non-randomised Studies of Interventions for non-randomised studies. Studies with serious or critical risk will be excluded from pooling. Subgroup analyses (LMIC vs HIC), sensitivity analyses (model and risk) and confidence grading using Confidence in the Evidence from Reviews of Qualitative Research and Grading of Recommendations, Assessment, Development and Evaluations will be conducted.

Ethics and dissemination

No primary data will be collected; thus additional Research Ethics Committee approval is unnecessary. The results will be disseminated via open-access publication, conference presentations and policy briefs for Nairobi County health stakeholders.

PROSPERO registration number

CRD42023466849.

STRENGTHS AND LIMITATIONS OF THIS STUDY

• This review will synthesise evidence across high-income and low-income settings, offering context-sensitive insights to inform the design of future pandemic coordination frameworks.

• The use of the Consolidated Framework for Implementation Research enables a structured, theory-informed integration of findings with direct relevance to implementation science and public sector coordination.

• Comprehensive multi-database and grey-literature searches will minimise retrieval bias and enhance the generalisability of recommendations.

• Independent dual-review processes for screening, extraction and appraisal will ensure methodological transparency and reproducibility.

• Restriction to English-language sources and the limited availability of comparable quantitative data may constrain generalisability and preclude meta-analysis.

Introduction and rationale

The COVID-19 pandemic exposed the fragility of global health systems and highlighted the critical role of multisectoral coordination in addressing complex public health crises. Effective coordination between stakeholders such as governments, health systems, private enterprises, civil society and academia has been pivotal in managing the pandemic’s far-reaching impacts.¹ However, despite its recognised importance, multisectoral coordination remains inadequately understood, particularly in low- and middle-income countries (LMICs), where systemic and structural barriers often hinder collaboration.²

During the COVID-19 pandemic, intergovernmental collaboration, global partnerships, decentralised humanitarian efforts, digital knowledge-sharing platforms, task forces and public-private partnerships were used as mechanisms for multisectoral coordination.³,⁶ Existing research highlights critical gaps in understanding how these mechanisms enable effective coordination. Fragmented governance, siloed decision-making and inconsistent communication have been cited as persistent challenges in LMICs, exacerbating the difficulties of resource mobilisation, equitable service delivery and policy implementation.^{7 8} While high-income countries (HICs) have benefited from robust emergency systems and centralised governance frameworks, LMICs such as Kenya face fragmented authority and limited digital infrastructure, which restrict the effectiveness of coordination efforts.^{9 10} Moreover, urban LMIC settings like Nairobi County present unique challenges, including high population density, resource inequities and socioeconomic diversity, which require tailored approaches to multisectoral collaboration.

This systematic review is necessary to address these critical knowledge gaps. By synthesising evidence on the mechanisms, benefits and challenges of multisectoral coordination during the COVID-19 pandemic, the study aims to provide actionable recommendations to strengthen future pandemic preparedness. In order to achieve this, the review primarily seeks to:

1. Identify and compare the performance of key multisectoral coordination mechanisms specifically task forces, public-private partnerships and incident-management systems in terms of resource mobilisation and governance effectiveness across LMIC and HIC settings.

2. Determine which organisational structures, processes and contextual factors, as mapped to the Consolidated Framework for Implementation Research (CFIR), facilitate or hinder effective coordination during the COVID-19 response.

3. Synthesise and appraise the feasibility of stakeholder-proposed strategies for strengthening future multisectoral pandemic preparedness, thereby generating an evidence-informed blueprint for improved coordination in public health emergencies.

Methods of analysis

The protocol for this systematic review has been registered with the International Prospective Register of Systematic Reviews, and the full registration record is provided in Prospero number (CRD42023466849) for transparency and documentation of pre-specified methods. The review will follow the sequential stages recommended in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement and operationalised through the Cochrane Handbook: formulating a focused research question, systematically locating studies on COVID-19 coordination mechanisms, iteratively selecting eligible records, extracting quantitative and qualitative data into a CFIR-enriched template, independently appraising risk of bias with Critical Appraisal Skills Programme (CASP) and Risk Of Bias In Non-randomised Studies of Interventions (ROBINS-I), and finally collating, synthesising and reporting the findings. through a CFIR-anchored framework synthesis and conditional meta-analysis

Identifying the research question

What are the key elements, benefits, challenges and improvement strategies associated with multisectoral coordination during the COVID-19 pandemic?

The sub-questions are:

1. How did specific coordination mechanisms (task forces, public-private partnerships, incident-management systems) influence resource mobilisation and governance effectiveness in LMICs compared with HICs?

2. Which organisational structures, processes and contextual factors mapped to the CFIR domains facilitated or impeded effective coordination?

3. What actionable strategies have stakeholders proposed or implemented to strengthen future multisectoral pandemic preparedness, and what evidence exists for their feasibility?

Eligibility of the research question

The review will determine eligibility with two complementary frameworks: PICO (population, intervention, comparator, outcome) for quantitative evidence and SPIDER (sample, phenomenon of interest, design, evaluation, research type) for qualitative or mixed-methods evidence. The PICO framework (table 1) guides the appraisal of studies reporting quantitative indicators of coordination performance, while the SPIDER framework (table 2) supports the identification of qualitative and integrative evidence relevant to multisectoral coordination mechanisms. Each framework specifies the analytic lens (criteria) and the concrete determinants that a study must satisfy to be included.

Table 1. PICO framework for determination of quantitative study eligibility.

Criteria	Determinant
Participants	Stakeholders involved in multisectoral COVID-19 coordination (government officials, health agencies, private-sector actors, civil-society organisations, academic institutions).
Interventions	Formal coordination mechanisms such as national or sub-national task forces, public-private partnerships, incident-management systems or inter-agency committees.
Comparators	Either (1) the coordination structure that preceded COVID-19 in the same setting or (2) an alternative contemporaneous mechanism; findings will be contrasted between LMICs and HICs.
Outcomes	Quantitative indicators of coordination performance, eg, resource-mobilisation efficiency (time from request to disbursement), governance effectiveness (decision-making timeliness, inter-agency communication scores), plus any reported benefits, challenges or actionable recommendations.

HICs, high-income countries; LMICs, low- and middle-income countries; PICO, population, intervention, comparator, outcome.

Table 2. SPIDER framework for determination of qualitative/mixed-methods study eligibility.

Criteria	Determinant
Sample	Stakeholders from LMICs and HICs who participated in multisectoral coordination during the COVID-19 pandemic (government, health sector, private sector, civil society, academia).
Phenomenon of interest	Multisectoral coordination processes, mechanisms and contextual factors influencing COVID-19 response, with explicit comparison or contrast between LMICs and HICs.
Design	Qualitative designs (eg, interviews, focus groups, case studies) and mixed-methods studies that permit extraction of qualitative data relevant to coordination.
Evaluation	Descriptions of key elements, perceived benefits and experienced challenges of coordination; assessments of strategy effectiveness; and contextual factors influencing outcomes.
Research type	Descriptive or explanatory empirical studies that explore how and why coordination mechanisms functioned, including facilitators, barriers and recommended improvements.

HICs, high-income countries; LMICs, low- and middle-income countries; SPIDER, sample, phenomenon of interest, design, evaluation, research type.

Identifying relevant studies

Eligible studies will be located through comprehensive keyword searches in PubMed, EBSCOhost (CINAHL, MEDLINE, PsycINFO), Emerald Insight and Google Scholar. To minimise publication bias, grey-literature sources such as WHO technical documents, government white papers and pre-print servers will also be searched. Search strings will combine controlled vocabulary (MeSH/Thesaurus) and free-text terms and will be piloted for sensitivity and specificity before full execution. All searches will be limited to records published in English between 1 January 2020 and 15 August 2024.

The following thematic keyword blocks will be used for the search:

• COVID-19 terminology (eg, “COVID-19”, “SARS-CoV-2”, “coronavirus pandemic”).

• Coordination mechanisms (eg, “task force”, “public-private partnership”, “incident-management system”, “inter-agency committee”).

• Implementation constructs (eg, “resource mobilisation”, “governance effectiveness”, “barriers”, “facilitators”, “CFIR”).

• Context comparators (eg, “low- and middle-income countries”, “high-income countries”).

A librarian at the University of KwaZulu-Natal will peer-review each strategy using the Peer Review of Electronic Search Strategies (PRESS) checklist. Line-by-line search strings, run dates and record yields will be archived in online supplemental table S1. A sample overview of the electronic search strategy, including database, search terms and retrieved records, is provided in table 3. Retrieved citations will be exported to EndNote V.20 for de-duplication and then uploaded to Covidence.

Table 3. Electronic search record (to be completed during the audit process).

Database	Date searched	Search line example	Records retrieved
PubMed		(“COVID-19”[Mesh] AND (“task force” OR “public-private partnership”) …)	#

Study selection

Eligibility criteria

Empirical studies will be assessed for inclusion based on study design, relevance to multisectoral coordination during COVID-19, stakeholder focus, outcome measures and publication characteristics, as outlined in table 4.

Table 4. Inclusion and exclusion criteria.

Inclusion	Exclusion
Empirical studies including quantitative, qualitative or mixed-methods that examine multisectoral coordination mechanisms (task forces, PPPs, incident-management systems, inter-agency committees) during the COVID-19 pandemic	Editorials, commentaries, opinion pieces, protocols without results
Stakeholders: government, health agencies, private sector, civil society, academia	Studies unrelated to COVID-19 or not addressing coordination
Outcomes: resource-mobilisation efficiency, governance timeliness, barriers/facilitators, CFIR-mapped determinants or recommended strategies	Publications in languages other than English
Published 1 Jan 2020 to 15 Aug 2024 in peer-reviewed outlets or credible grey literature	Duplicate publications of the same dataset (the most complete version retained)

CFIR, Consolidated Framework for Implementation Research; PPPs, public-private partnerships.

Two reviewers will independently screen titles/abstracts; any record judged potentially eligible by either reviewer will proceed to full-text review. Full texts will likewise be screened in duplicate, with disagreements resolved by an adjudicating third reviewer. Prior to full screening, a 5% pilot sample will test the criteria; Cohen’s κ≥0.70 will be required before progressing. Screening decisions and reasons for exclusion will be logged in Covidence, and the overall process will be depicted in a PRISMA 2020 flow diagram.

Corresponding authors will be contacted (one email plus one reminder) for missing information; non-response after 21 days will be coded as ‘unavailable’. The search, selection and inclusion criteria align with the PICO framework for quantitative evidence and the SPIDER framework for qualitative and mixed-methods evidence, ensuring comprehensive yet transparent study identification.

Data extraction

A standardised, pilot-tested extraction form will be used to ensure consistent capture of meta-data across all included studies. The form will be implemented in Excel/Google Sheets and pre-populated by Python/LLM scripts containing the variables shown in table 5. However, the full data extraction template used for standardising study information is attached in online supplemental file S2. Two reviewers will complete the form independently; a second reviewer will verify all critical fields, and discrepancies will be resolved by a third reviewer. A 5% pilot on early-included papers will test the form; modifications will be logged in the Amendment Log.

Table 5. Data extraction form (full template provided in online supplemental file S2).

Section	Variables captured
1. Bibliographic information	Study ID; authors & year; title; country & World-Bank income group; journal/grey-literature source; study period; data source; publication type; primary unit of analysis; stakeholder group(s).
2. Aims	Stated study objectives and/or research questions.
3. Methodology	Study design; sampling frame; data-collection methods; coordination mechanism examined; CFIR domain(s) addressed; risk-of-bias tool applied.
4. Results	Quantitative outcomes (resource-mobilisation efficiency, governance timeliness, etc); qualitative benefit themes; challenge themes; enabler themes; illustrative quotes; effect direction; statistical estimate if available.
5. Discussion	Authors’ interpretation, reported facilitators/barriers, noted limitations, modelling or implementation gaps.
6. Conclusions and recommendations	Actionable strategies proposed; feasibility notes; reviewer-coded CFIR mapping of each recommendation.
7. Risk of bias	CASP domain scores (qualitative) or ROBINS-I overall rating (quantitative).

CASP, Critical Appraisal Skills Programme; CFIR, Consolidated Framework for Implementation Research; ROBINS-I, Risk Of Bias In Non-randomised Studies of Interventions.

All verbatim qualitative excerpts linked to coordination determinants will be imported into NVivo V.14 and coded deductively to the five CFIR domains; inductive codes will be added as needed. Extracted files, appraisal forms and logs will reside on an encrypted local drive mirrored to Google Drive with weekly automated back-ups.

Collating, summarising and reporting results

This stage will follow three sequential steps, namely analysis, results reporting and interpretation.

Data synthesis

A descriptive numerical summary will record the total number of included studies, study design, publication year, country income group, coordination mechanism type, stakeholder composition and risk-of-bias profile. Qualitative findings will be integrated through a CFIR-anchored framework synthesis that will group evidence by domain (intervention characteristics, outer setting, inner setting, characteristics of individuals, process) and subconstructs. Quantitative results will be narratively summarised; where ≥2 studies report commensurate outcomes, a fixed-effect or random-effect meta-analysis will be undertaken after assessing heterogeneity (χ², I²).

Subgroup and sensitivity analyses

To illuminate contextual contingencies, a subgroup comparison will contrast effect directions between LMICs and HICs, explicitly noting how CFIR domains manifest differently across resource settings. Sensitivity analyses will (1) exclude studies at serious or critical risk of bias, (2) re-run descriptive syntheses after omitting studies with outlying sample sizes or non-comparable outcome definitions and (3) test the influence of analytic choices such as fixed-effect versus narrative reporting.

Results reporting

Outputs will include:

• A PRISMA 2020 flow diagram.

• A study-characteristics table.

• A domain-by-study CFIR matrix.

• A benefits–challenges–enablers table with illustrative quotes.

• Forest plots for any pooled quantitative outcomes.

• Summary-of-findings tables displaying Grading of Recommendations, Assessment, Development and Evaluations (GRADE) (quantitative) and Confidence in the Evidence from Reviews of Qualitative Research (CERQual) (qualitative) certainty ratings.

Quality appraisal

Two reviewers will independently assess the study quality. Qualitative studies will be appraised using the CASP Qualitative Checklist (10 items), with item-level ratings synthesised into overall low, moderate or high risk using pre-defined thresholds. Non-randomised quantitative studies will be assessed with ROBINS-I across its seven domains, assigning an overall risk level based on the worst-rated domain. Discrepancies will be resolved by consensus or through third-reviewer adjudication.

Risk-of-bias ratings will directly inform both sensitivity analyses and confidence assessments. For quantitative outcomes, the GRADE approach will be used to classify the overall certainty as high, moderate, low or very low, based on sequential consideration of risk of bias, inconsistency (eg, I² or narrative divergence), indirectness, imprecision and publication bias. Observational studies will begin at ‘Low’ and be upgraded based on factors such as large effect sizes, dose-response gradients or absence of plausible confounding. Summary-of-findings tables will be produced using GRADEpro guideline development tool.

For qualitative findings, the CERQual framework will guide confidence ratings across four domains: methodological limitations, coherence, adequacy of data and relevance. Each thematic CFIR-coded finding will be rated as high, moderate, low or very low certainty, and explanatory footnotes will be used to justify any downgrades. A CERQual evidence profile will accompany the results, reinforcing the transparency and interpretability of the synthesis.

Meta bias assessment

To address meta-bias, we will evaluate publication bias whenever at least ten studies report the same quantitative outcome, visually inspecting funnel plots and applying Egger’s test (two-tailed, p<0.10) in R; any asymmetry will be interpreted in light of study size and heterogeneity, while domains represented solely by peer-reviewed qualitative sources will be noted for potential transferability concerns. Selective-reporting bias will be checked by comparing prespecified outcomes with those actually reported, logging discrepancies via an adapted Outcome Reporting Bias In Trials decision tree and incorporating them into overall risk-of-bias and certainty ratings. For missing data, corresponding authors will be contacted once with a 21-day follow-up; quantitative studies lacking dispersion statistics will have SD imputed where feasible, otherwise remaining in narrative synthesis, and qualitative gaps will be flagged as ‘data absent’ in NVivo and reflected in CERQual adequacy. All decisions will be recorded in the Amendment Log, and sensitivity analyses will exclude studies reliant on imputed or unrecoverable data.

Dissemination plan

The protocol will be disseminated to key stakeholders in public health, governance and research to foster engagement and collaboration. Primary audiences include researchers, academics and policymakers involved in pandemic preparedness and multisectoral coordination. Special emphasis will be placed on reaching stakeholders in LMICs, including Nairobi County officials, public health practitioners and regional bodies such as the African Union. International organisations such as WHO and other global health actors will also be targeted to encourage alignment with broader pandemic preparedness strategies.

Patient and public involvement

Patients and members of the public were not involved in the design, conduct, reporting or dissemination plans of this systematic-review protocol.

Protocol amendment plan

Any methodological deviations from the original protocol, including refinements to eligibility criteria, adjustments to data extraction procedures or analytic approach modifications, will be transparently documented in the Protocol Amendments Plan (online supplemental file 3). This file provides a version-controlled audit trail in alignment with best practices for protocol fidelity and reproducibility in evidence synthesis.

Discussion

Multisectoral coordination has emerged as a cornerstone of resilient public health systems, particularly during complex emergencies like the COVID-19 pandemic. However, despite global recognition of its importance, empirical clarity on the structures, processes and contextual enablers that underpin effective coordination remains limited, especially in LMICs, where systemic fragmentation and resource constraints often hamper joint responses. This systematic review seeks to bridge this knowledge gap by systematically synthesising evidence across both LMIC and high-income contexts, thereby generating comparative insights that are not only descriptive but analytically transferable.

By employing a CFIR-guided framework synthesis, the study offers a theory-informed lens through which to distill implementation-relevant findings on what works, for whom and under what conditions. Conditional meta-analytic strategies will supplement these findings where quantitative evidence permits. Importantly, the review is oriented not just toward academic synthesis but toward practice-relevant guidance. The findings aim to inform policymakers, planners and emergency response coordinators on how to institutionalise coordination mechanisms that are inclusive, agile and contextually grounded.