Strengthening emergency care research quality: A commentary on a mid-study PDSA in a multisite project in Ghana

Abstract

Emergency medicine is rapidly evolving across Africa, yet the field continues to face significant gaps in research quality, particularly in study design and implementation. To generate actionable evidence that reflects the acute care needs of African populations, it is critical to embed quality management measures within clinical research processes. This commentary describes the use of the Plan-Do-Study-Act (PDSA), a simple and reproducible quality improvement (QI) tool, as a mid-study quality audit strategy during the implementation of the “Pathways to Care for Patients with Acute Infectious Illnesses in Ghana” (PATHCARE) study. Over two weeks, trained evaluators conducted structured observations and interviews with research assistants (RAs) across five study sites in Kumasi, Ghana. The audit revealed protocol deviations related to patient screening, recruitment, informed consent, and data collection that threatened data integrity and participant protection. Based on these findings, a targeted refresher training was conducted, leading to measurable improvements in RA knowledge of study protocol. This experience highlights the feasibility and impact of embedding low-cost, iterative quality audits within emergency care research. Especially in resource-constrained settings with limited research infrastructure, tools like the PDSA can serve as pragmatic solutions to enhance data reliability, safeguard ethical standards, and ultimately, improve the quality and utility of clinical research in African emergency medicine.

African relevance

• •This project, conducted in Ghana, highlights the importance of quality improvement (QI) in clinical research, particularly in resource-limited settings where trained research professionals may be lacking.
• •Our work demonstrates the feasibility and value of incorporating low-cost quality improvement tools like the PDSA cycle into the implementation of complex, multi-site clinical research projects in Africa.
• •This proactive approach ensured more accurate patient recruitment and data collection and enhanced the reliability of research findings. Such QI interventions are vital for generating high-quality evidence to strengthen emergency care across Africa.

Introduction and why QI in EM research matters

Emergency Medicine (EM) has grown substantially in Africa in the last two decades, [1,2] however, research in this space is still limited [3,4]. This can be attributed to a combination of factors, including a lack of a trained research workforce, inadequate research infrastructure, and conflicting priorities on the part of clinicians [[5], [6], [7]]. The consequence is the inability to conduct complex studies and a limited ability to make inferences, secure funding, and translate research to improve emergency care [7]. High-quality, patient-centred research is essential to comprehensively understand the acute health needs of African patients and translate findings to improve emergency care [7].

The International Council for Harmonization (ICH) E8(R1) guideline emphasizes that quality should be proactively designed into clinical studies, with a risk-proportionate approach focused on factors critical to participant protection and the reliability of study results [8]. Hence, effective quality management in clinical research is crucial for ensuring that data is accurate, participants are protected, and scientific objectives are achieved [9]. When the internal infrastructure is not available, often large studies will engage with contract research organizations (CROs) to provide external monitoring and QI. However, there are substantial costs associated with outsourced CROs for quality management, thus limiting their utility to industry-sponsored clinical studies, where per-enrollment costs may exceed ten thousand dollars [10,11]. Despite these constraints, quality control in research is non-negotiable. Without it, studies risk generating unreliable or non-generalizable findings. This is particularly true in emergency settings where patient acuity, rapid turnover, and complex care pathways pose additional challenges to data integrity and ethical conduct.

This commentary describes the design, implementation, and impact of a Plan-Do-Study-Act (PDSA) as a strategy for clinical research quality improvement (QI) in resource-constrained emergency care settings. The PATHCARE study, an ongoing investigation of care-seeking pathways for patients with acute infectious illnesses in Ghana, serves as an illustrative case. To strengthen the implementation and credibility of the study, we embedded a mid-study PDSA as a quality improvement intervention. This provided a structure for assessing the alignment between the protocol (“Plan”) and real-world execution (“Do”), analysing discrepancies (“Study”), and introducing corrective measures (“Act”). We highlight the utility of the PDSA to identify protocol deviations, inform study staff, and improve data reliability.

Adapting PDSA to the research context

The PDSA [12] is a widely used tool in clinical QI and offers a simple and iterative method to assess practices, test interventions, and refine processes [13,14]. The PDSA is commonly employed for continuous quality improvement in clinical care settings; its use in clinical research is less common [14,15]. In the PATHCARE study, we adapted this tool to audit how closely field research activities aligned with the study protocol and to gather insights for corrective action. Please see Supplemental Documents 1–4 for additional information on the PATHCARE study and the mid-study quality management strategy.

For our purpose, the PDSA tool was tailored to cover key research domains:

• •Participant screening and recruitment
• •Informed consent processes
• •Data collection procedures
• •Follow-up strategies
• •Ethics and protocol adherence

Key quality gaps identified

The mid-study quality improvement activity, employing the PDSA QI tool, uncovered several critical quality gaps that threatened the integrity of the research process (Fig. 1). Research assistants (RAs) inconsistently applied screening and inclusion criteria, often avoiding high-acuity or complex cases, which introduced potential sampling bias. Informed consent procedures varied widely between RAs, with deviations from the provided script and inadequate individual translation into local dialects. Clinical data collection was hindered by incomplete hospital records, reliance on convenience data, and challenges accessing information from previous health facilities. Some RAs engaged with the clinical team to fill gaps in clinical documentation, but others left the research data collection forms blank leading to incomplete data and small numbers for planned sub-analysis. Follow-up procedures were also unreliably implemented, with misunderstandings about eligibility and missed calls due to inaccurate contact details. Additionally, limited clinical training among part-time RAs contributed to variability in protocol adherence as they had a poor understanding of the clinical relevance of certain questions and difficulty with sensitive data collection. These gaps underscore the need for structured training, consistent mentorship, and embedded quality mechanisms throughout the research lifecycle.

Fig. 1.

Mid‑Study PDSA Audit Findings Highlighting Protocol Deviations and Associated Risks.

The mid-study intervention: structured refresher training

For the Act category of the PDSA, we developed a training workshop targeting the identified gaps in adherence to the Plan or study protocol. Pre- and post- tests were utilized to gauge workshop efficacy. The test consisted of 10 multiple-choice questions covering the scope of RA activities in the research process. The statistical comparison is detailed in Supplement 4. The QI team visited all five research sites over two days, conducting 2-hour refresher training sessions at each site. The training sessions utilized adult learning methods highlighting aspects of data collection that were executed accurately and comprehensively addressed areas for improvement through five interactive case scenarios, each focusing on a specific aspect of the protocol requiring strict adherence. The themes of the workshop included:

• •The importance of adhering to inclusion criteria and avoiding selection bias
• •Standardizing the informed consent process using simplified and translated scripts
• •Enhancing clinical data collection skills, particularly for staff without a clinical background which included review of nomenclature and terminology in the protocol, clinical charts and data collection forms
• •Reinforcing the link between follow-up calls and the study’s primary outcome of 30-day mortality

Why this approach works: key enablers and insights

Several factors contributed to the success of the PDSA-based quality improvement approach implemented during the PATHCARE study. First, its low-cost, high-yield nature made it particularly well-suited to a resource-constrained research environment. The quality audit was conducted by two graduate-level public health trainees without the need for external consultants, specialized software, or additional infrastructure. All tools, including observation templates, interview guides, and follow-up logs, were developed using basic word processing and spreadsheet software, making the intervention both accessible and replicable.

Second, the intervention was designed with a deep understanding of the local research context. Rather than disrupting workflows, the audit was integrated into routine research activities and carried out with sensitivity to the clinical demands and availability of research staff. This facilitated RA cooperation and minimized resistance. Importantly, the approach prioritized mentorship over surveillance, fostering a learning environment in which research assistants felt supported rather than scrutinized. This contributed to greater engagement and ownership of the quality improvement process.

Third, the intervention was implemented at a strategically chosen time point—midway through the study—allowing for real-time identification and correction of problems before they could significantly undermine data quality or ethical standards. Immediate feedback and on-site coaching ensured that course corrections were relevant, actionable, and timely.

Finally, the model’s structured yet adaptable design lends itself to broader use in similar settings. The use of a standardized audit template enabled systematic assessment across multiple domains, while the findings directly informed a tailored refresher training that addressed site-specific gaps. The Act step of the PDSA should be tailored to the results of the other steps and may be more directly related to the operational steps of the protocol as well. These attributes make the approach scalable and reproducible for other emergency care research projects operating in low-resource environments. Collectively, these features demonstrate how a well-timed, context-sensitive, and resource-efficient intervention can drive meaningful improvements in research quality, even in the most constrained settings.

Limitations

While the PDSA proved valuable in this study for identifying areas for quality improvement, it is important to note that it has some inherent limitations [14]. The approach typically focuses on documenting what is happening in practice and does not provide a comprehensive framework to fully explore the underlying causes or reasons for the observed discrepancies or deviations from the planned protocol. We therefore recommend incorporating complementary methods, such as root cause analysis or Ishikawa Diagrams, to investigate significant deviations or errors identified during a PDSA-based quality audit.

The PDSA method is commonly employed for continuous quality improvement in healthcare and works best when it is used in iterative, repetitive cycles. While the PDSA activity only occurred once in this pilot QI study, this is not uncommon, as most public health and clinical programs only complete one cycle [16]. This may have limited the researchers' ability to fully evaluate the effectiveness of the mid-study training workshop in addressing the identified data quality issues and repeated PDSA cycles or a post-intervention observation phase may be necessary to achieve sustained and measurable improvements.

Conclusion: making quality research affordable and achievable

This commentary on the PATHCARE project highlights a practical, potentially cost-effective approach for improving clinical research quality in African emergency care settings. By embedding tailored PDSA-based quality management activities mid-study, we identified protocol deviations, designed an intervention to address those gaps, and improved RA knowlege of study procedures without requiring costly quality management infrastructure. In resource-constrained environments where research workforce is limited and data quality risks are high, tools like PDSA offer a powerful complement to traditional monitoring frameworks. They allow for proactive, real-time course correction and foster a culture of reflection and accountability among field staff.

We advocate for the broader adoption of structured quality audits as integral components of clinical research in LMICs. Funders, institutional review boards, and research leaders should support these practices and recognize that quality monitoring activities are not a luxury, but a prerequisite for ethical, impactful research. By lowering the barrier to QI implementation, this model has the potential to elevate the standard of emergency care research across Africa, ensuring that studies not only generate data, but do so with integrity, accuracy, and respect for participants.

Dissemination of results

This quality improvement project was presented at the 2024 Master of Public Health Capstone Symposium at the Johns Hopkins Bloomberg School of Public Health, Baltimore, USA. Also, an abstract of this paper was delivered as an oral presentation at the 2024 African Conference on Emergency Medicine in Botswana.

CRediT authorship contribution statement

Sarah Aba Osei-Frimpong: Conceptualization, Investigation, Writing – original draft, Writing – review & editing. Ama A. Essuman: Conceptualization, Investigation, Writing – original draft, Writing – review & editing. Teagan Lukacs: Conceptualization, Methodology, Writing – original draft, Writing – review & editing. Amanda Varnauskas: Writing – original draft. Joseph Bonney: Conceptualization, Methodology, Writing – original draft. Bhakti Hansoti: Conceptualization, Supervision, Writing – review & editing.

Declaration of competing interest

The authors declared no conflicts of interest.