Evaluation of the implementation of the Critical Care Asia and Africa Intensive Care Unit registry in Ethiopia

Abstract

BACKGROUND

Intensive care unit (ICU) capacity is underdeveloped in sub-Saharan Africa, and outcomes for critical care lag behind higher resource settings. It is essential to understand local case-mix, processes of care and clinical outcomes in African ICUs to close these gaps. To this end, ICU registries are valuable tools for clinical research, quality improvement and capacity building. The Critical Care in Asia and Africa (CCAA) Network has specially developed a novel ICU registry for resource-limited settings.

AIM

To evaluate the feasibility, acceptability, and perceived sustainability of the CCAA ICU registry in the Ethiopian clinical context.

METHODS

Eight months following the pilot implementation of the CCAA ICU registry at two academic medical centers in Addis Ababa, Ethiopia, we conducted a qualitative evaluation. We conducted key informant interviews and focus group discussions with members of the ICU registry team, medical and nursing staff, and leadership team. We coded and analyzed transcripts deductively using a thematic content approach.

RESULTS

Emergent themes related to feasibility included data collection, data quality and factors necessary for success. Those related to acceptability included utility, accessibility and comparison to other methods. Those related to perceived sustainability included institutional future, ownership and expansion. Overall, respondents felt that the CCAA ICU registry was feasible and acceptable in their ICUs. They identified important threats to perceived sustainability including multiple channels of communication and infrastructure and human resource limitations, and proposed adaptations to address these threats.

CONCLUSION

The CCAA ICU registry is a promising tool for research and quality improvement in ICUs in sub-Saharan Africa, but successful implementation requires a clear understanding of regional and institutional influencing factors.

Key Words: Intensive care unit; Clinical registry; Implementation; Evaluation; Quality improvement; Ethiopia; Sub-Saharan Africa

Core Tip: In this study, implementation of the Critical Care Asia and Africa Intensive Care Unit registry was evaluated at two hospitals in Addis Ababa, Ethiopia. The implementation was considered an overall success by respondents, who commented positively on the registry’s feasibility and acceptability. This study identified key barriers and facilitators to implementation and threats to sustainability, including essential resources, communication, training, supervision and leadership. These findings can inform adaptations to the registry to optimize implementation and scale-up in the sub-Saharan Africa context.

INTRODUCTION

Intensive care units (ICUs) are organized systems that provide specialized medical and nursing care, continuous monitoring, and specialized physiologic organ support - termed critical care - to sustain life during a period of acute organ system insufficiency[1]. In low- and middle-income countries (LMICs), critical care is an essential part of the treatment of acute life-threatening conditions, and can be cost-effective when rescuing the critically ill from preventable deaths[2]. However, ICU capacity remains underdeveloped in LMICs because of inadequate resources including personnel, medications, and equipment, as well as poorly coordinated systems of care[3-6]. As a result, risk-adjusted ICU mortality in LMICs is over twice that of high-income countries (HICs)[7]. During the coronavirus disease 2019 (COVID-19) pandemic, reported ICU mortality for intubated patients approached 80% in African ICUs[8-10].

Quality improvement (QI) interventions can improve the health outcomes of ICU patients in LMICs, even within existing resource constraints[11,12]. However, QI efforts in LMICs are hampered by a paucity of published data about critical care capacity, case-mix, illness severity, processes of care, and clinical outcomes in those settings[13]. These data are essential for developing and implementing context-appropriate QI interventions that account for variations in demographics, disease patterns, and resource availability in LMICs[14].

ICU registries are databases that compile these data, providing a valuable tool to inform and evaluate QI interventions[15-17] and to improve outcomes of critically ill ICU patients by providing feedback to healthcare providers[18,19]. ICU registries developed in HICs often rely on significant funding, electronic medical records, human capital, and institutional support that are unavailable in LMICs[15-17]. The CCAA Network has attempted to overcome these challenges by developing a novel ICU registry adapted to LMICs, which was leveraged to enable standardized ICU reporting in LMICs during the COVID-19 pandemic to facilitate resource coordination and informed resource-appropriate treatment strategies[14,20].

Recently, significant efforts have been made in Ethiopia to improve quality and accessibility of acute and critical care services under the leadership of the Ethiopian Ministry of Health’s (MOH) Emergency, Injury, and Critical Care Directorate[21]. Although the CCAA ICU registry is a promising tool for research and QI efforts in LMICs in Asia, it has not yet been widely implemented in sub-Saharan Africa (SSA) and, more broadly, to date few studies have assessed the regional or institutional factors affecting implementation of ICU registries in SSA using a structured framework[22-24]. This study describes an evaluation of the feasibility, acceptability, and perceived sustainability of the CCAA ICU registry following its pilot implementation in Ethiopia. The purpose of this evaluation is to inform adaptations to the implementation strategy to optimize it for resource-limited contexts in Ethiopia and elsewhere in SSA.

MATERIALS AND METHODS

The CCAA ICU registry was pilot implemented at two tertiary referral hospitals in Addis Ababa, Ethiopia starting in July 2021 with collaboration for research, implementation and training of the CCAA Network and Mahidol Oxford Research Unit. This qualitative evaluation of the pilot implementation was conducted in February 2022, using a hybrid analysis approach.

Study setting

St Paul’s Hospital Millennium Medical College (SPHMMC) was established in 1969, with the addition of a medical school in 2007 and a critical care fellowship in 2021. It has over 700 inpatient beds, 14 adult ICU beds, and provides a comprehensive array of medical and surgical subspecialty services. Addis Ababa Burn Emergency and Trauma Hospital (AaBET) is an affiliate of SPHMMC established in 2015 to improve emergency trauma and burn care. It has over 350 inpatient beds, 11 adult ICU beds, and is the largest trauma and emergency center in Ethiopia.

ICU registry

The CCAA ICU registry was initially developed and implemented in nine Asian countries with supported from a Wellcome Innovations Flagship award in 2020[14]. Under the supervision of a site lead, specially trained data collectors compile key variables regarding patient characteristics, clinical status and illness severity at the time of ICU admission. Additional data regarding processes of care, complications and clinical outcomes are compiled at the time of ICU discharge. Data are entered into tablet computers and uploaded in realtime to a secure, cloud-based platform for compilation. The site lead is responsible for ensuring data completeness and validity. The ICU registry platform is owned by CCAA, which funds data collection and hosts online data storage with grant support from the British National Institute for Health and Care Research and the Wellcome Trust. All collected data are the property of the local ICU registry team, who can access their data at any time through the cloud-based platform. The CCAA team oriented the site leads and data collectors to the registry platform prior to implementation, and continue to provide ongoing technical support. Preliminary analysis of ICU registry data from SPHMMC and AaBET has been published previously[25].

Respondent recruitment

Potential respondents were identified by SPHMMC and AaBET ICU leadership from staff rosters using a purposive sampling strategy and recruited in person by study team members who explained the reasons for doing the research. Recruitment continued until representation from medical, nursing, and leadership teams was achieved, and thematic saturation was reached.

Data collection

Three key-informant interviews (KIIs) and four focus group discussions (FGDs) were conducted in person in non-clinical spaces at SPHMMC and AaBET in February 2022, eight months after the start of pilot implementation, without non-participants present. KIIs and FGDs were conducted in English, the lingua franca of medical education and practice in Ethiopia, and facilitated by study team members with specialized training in qualitative methods and research ethics. Facilitators (Olani AB and Werkeshe G) were fluent in Amharic, the most widely spoken language in Ethiopia, and were able to clarify and interpret as necessary. Verbal informed consent was obtained for all respondents. KIIs and FGDs were recorded and transcribed and facilitators took written notes. KIIs and FGDs lasted 30-60 minutes each.

KIIs and FGDs were conducted using a semi-structured interview guide derived from the Diffusion of Innovations in Service Organizations (DISO) conceptual model, which was developed through a systematic literature review focusing primarily on healthcare research to guide critical thinking about the complex aspects of the innovation, the system in which it is introduced, and how these interact as implementation occurs[26]. We chose the DISO model because it comprehensively explores characteristics of the innovation and implementation context that were deemed relevant for this project, and has been adapted and used to guide QI projects in clinical settings including in SSA[22,27,28].

Respondents were prompted to discuss knowledge and beliefs about the CCAA ICU registry, structural characteristics, compatibility, adaptability, access to knowledge and information, relative advantage, relative priority, tension for change, self-efficacy, available resources, and leadership engagement. Questions were grouped under the predetermined topics of feasibility, acceptability, and perceived sustainability to guide discussion. The interview guide was developed and refined by study team members, piloted with Ethiopian ICU staff, and refined for clarity and completeness.

Statistical analysis

KII and FGD transcripts were coded using a thematic content analysis approach by members of the study team with specialized training in qualitative methods using a hybrid approach that incorporated both deductive coding based on the DISO domains of feasibility, acceptability, and sustainability[26,29,30] and inductive coding for emerging themes, sub-themes, and patterns. The first transcript was coded by all three coders to develop the codebook. Remaining transcripts were coded by two coders each using a consensus approach. Codes were aggregated in a matrix in Microsoft Excel according to topic, theme, and sub-theme and compared back to DISO domains.

RESULTS

The study included 23 respondents, three in KIIs and 20 in FGDs. included ICU medical and nursing directors, director of the hospital quality office, ICU staff physicians and nurses, and the ICU registry data collectors. No respondents dropped out of the study. Years of ICU experience ranged from 1-10 (median 3). Topics, themes and subthemes are summarized in Table 1.

Table 1 Feasibility, acceptability and perceived sustainability of the Critical Care Asia and Africa Intensive Care Unit registry in Ethiopia.

Topic	Theme	Subtheme
Feasibility	Data collection	Online ICU registry platform
		Data collectors
		Integration into clinical activities
	Data quality	Accuracy and completeness
		Missed data
		Monitoring
	Needs for success	Resources
		Training
		Communication
Acceptability	Utility	Research
		Quality improvement
		Clinical care
		Current use
	Accessibility	Ease
		Who has access
	Comparison to other methods	Organization
		Comprehensiveness
		Availability
Perceived sustainability	Institutional future	Expectations
		Threats
		Adaptations
	Ownership	Distrust
		Data loss
	Expansion	Stakeholders
		Challenges
		Solutions

ICU: Intensive care unit.

Feasibility

Overall, respondents indicated that the ICU registry implementation was feasible to implement in their hospitals. Four key themes emerged during the qualitative analysis - data collection, data quality, and needs for success.

Data collection: Emergent subthemes included the ICU registry’s online platform, data collectors and integration into clinical care. Multiple respondents commented on the novel nature of the online platform, allowing data collection using tablet computers. Several commented that the online format was easy to use and facilitated ready availability of data for future analysis. However, several respondents mentioned potential technical problems because the platform requires functioning computers with continuous electricity and a reliable internet connection, which are not always available. Several respondents stressed the importance of an option to record registry data offline and upload later to adapt to intermittent internet connection, which did not exist at the time of this evaluation.

Respondents differed in their data collection preferences. Some preferred a highly trained data collector who is present every day, while others wanted to share data collection responsibilities between multiple staff members. One data collector commented, “The ICU registry is filled on a daily basis and it is tiresome for someone with other duties to fill in all the required sections. Hence, it needs a designated, trained, and experienced data encoder”. On the other hand, one physician suggested, “If the entire staff gets access to the application on their mobile phone, it will be easier... Documentation shouldn’t be the responsibility of one person”.

Concerns about balancing clinical roles of registry data collectors were part of broader comments about registry integration into ICU clinical activities. Some respondents felt that the ICU registry should be completely separate from clinical activities, so that bedside providers can focus on caring for their patients and not be distracted by data collection. Others noted that the success of the registry required support from the entire ICU staff to ensure accurate and comprehensive data collection. One physician commented, “We need to convince the entire staff to be cooperative and willing. We need to include it in the culture of the ICU”.

Data quality: Emergent subthemes included accuracy and completeness, missed data, and monitoring. Most respondents believed data being collected were accurate and complete, reflecting the clinical reality in the ICU. An ICU director noted, “A lot of important information is being collected including the primary diagnosis, daily complications, the severity level and outcome of patients admitted to our ICU and much more important information in our day-to-day activities”. A nurse commented, “We are recording what we see on the mechanical ventilator, monitor, and the patient. So, it is real and reflects the real situation of the patient”. A few respondents noted that the accuracy of the data being collected was dependent on the training and attention to detail of the data collector.

However, several respondents identified instances where data might be missed. The data collector worked on weekdays during the day, so patient admissions and clinical events occurring overnight could be missed because the data collector would gather those data retrospectively from paper charts. Respondents also mentioned data points that were not included in the ICU registry, such as nutritional information and procedures like chest tube insertion. One physician mentioned that diagnoses could change over time and worried that multiple diagnoses might not be captured. Another respondent stressed the tension between comprehensive data collection and overburdening the data collectors, which could cause data loss and threatening project feasibility.

Multiple respondents discussed the importance of monitoring of the data collection process. Several mentioned the role of the ICU registry site lead in supervising data collectors and ensuring thoroughness. A data collector noted, “Supervision is also important. Sometimes we can be busy and forget to register. So, the supervisor is needed to check whether the data is being registered or not”. A physician pointed out the importance of regularly reviewing collected data to ensure completeness, saying, “We need further evaluation by our physicians and nurses to know the information being missed”.

Needs for success: Emergent subthemes included resources, training, and communication. Regarding resources, respondents highlighted the needs for funding for staff, training, and lines of communication. In particular, respondents stressed the importance of funding to cover salaries for data collectors and the site supervisor to ensure complete and accurate data collection but indicated that the necessary resources were not prohibitive. An ICU director said, “It doesn’t require many resources… Considering the importance of the database, I don’t think this fee is expensive for any teaching hospitals in the country”.

Respondents mentioned the importance of training for both the data collection team and those in research and leadership roles. Respondents noted that the data collection team needed initial orientation and iterative feedback addressing data collection problems. One nurse commented, “The registered data… must be accurate and the person that fills the data should get appropriate training”. For those in research and leadership roles, respondents pointed out that guidance about data access and use would be helpful. An ICU director stated, “Physicians, leaders, including ICU department heads and head nurses, and others need training from the developers of the database to utilize it, they have to know the data they can extract from the database”.

Respondents stressed the need for good communication between the hospital’s ICU registry team and international collaborators, the site leads and the data collectors, the ICU registry team and clinical staff, and the ICU registry team and hospital leadership. Regarding communication between the hospital’s ICU registry team and international collaborators, respondents reported that it was going well with responsiveness to feedback about the data collection process. One data collector commented, “We communicate frequently about the issues related to amendments and difficulties, through these means of communications, we have improved a lot of issues from the registry”. Respondents also reported a positive impression of communication between the site leads and data collectors. One data collector commented, “There is serious follow-up in our hospital. The site leads follow us on a regular basis. Last time, I forgot to register one patient and the site lead reminded me why did I forget to register that particular patient. This shows how the follow-up is strict”. However, several respondents mentioned the need to improve communication between the ICU registry team and clinical staff, saying that the clinical staff could support the data collection process better if they received a formal introduction to the registry from the registry team, for example informing data collectors about important events occurring overnight. Finally, multiple respondents commented on the need to improve communication between the ICU registry team and hospital leadership. One physician suggested, “I am not sure whether the leaders have awareness of the registry. But I believe that if the leaders have awareness of the registry, they will provide the necessary support”.

Acceptability

Overall, respondents reported that the ICU registry was acceptable and beneficial. One physician said, “I think it is a good system. We should start in other ICUs and continue in our hospital as well”. An ICU director commented, “It has a positive value… It is accepted by the hospital”. Comments on acceptability fell under three themes - utility, accessibility, and comparison to other methods.

Utility: Emergent subthemes included research, QI, clinical care and current use. Multiple respondents commented that the ICU registry could be used to research case-mix, care processes, and clinical outcomes. One nurse stated, “To understand whether the provided treatment is being successful or not, the ICU registry could help... It is also helpful for the hospital to understand the number of cases treated in the ICU”.

Regarding research, several respondents noted that standardizing ICU data collection across facilities could allow for multicenter studies. One nurse commented, “If we have such a system in different hospitals, it is possible to share a clinical experience… The benefit is not only for the patient but also for the country at large”. An ICU director described the utility of these data for informing local evidence-based practice, saying, “We are relying on evidence from Western nations to guide our activities... We don’t have pieces of evidence generated by ourselves. We have to use the database for such purposes and modify our protocols”.

Multiple respondents noted that the ICU registry could inform much-needed QI efforts. One physician mentioned that the ICU registry’s longitudinal data collection was well suited for evaluating QI projects, saying, “After the registry is implemented sometime later, then we can compare it to see the impact and changes over time”. An ICU director also suggested that QI efforts guided by the registry could improve staff morale, commenting, “This will help to improve your clinical care and also patient outcome. If you have a better outcome of the patient, then the staff satisfaction will increase. You will also learn from your mistakes”.

Several respondents said that they had used the ICU registry to retrieve clinical data including culture results and trends in vital signs and laboratory values for use in patient care. Others voiced ambiguity about whether ICU registry data should be considered for clinical use, and said they ultimately hoped for one integrated electronic medical record for both clinical documentation and research.

Some respondents voiced concern about how the compiled ICU registry data were currently being used. For example, an ICU director said, “We are just collecting the data and we are not using it adequately... The goal of the database shouldn’t be data collection only, but they should use it”.

Accessibility: Emergent subthemes included ease and who has access. Most respondents thought the innovation of an electronic ICU registry platform made it easy both to input and to extract data, noting that it could be accessed from different locations and could generate reports summarizing data. One ICU director said, “It is great to have electronic data. You can access it anytime and anywhere whenever you need it”.

Several respondents raised concern about who has access to the data, which they viewed as unfair. One physician noted, “At the moment, few numbers of people have access to the platform. But that has to be changed… I don’t think it is fair to deny the access for the physicians from whom we are collecting the data”.

Comparison to other methods

Emergent subthemes included organization, comprehensiveness, and availability. Most respondents reported that the ICU registry was better organized and thus more reliable than other research methods, including retrospective chart reviews and clinical logbooks, in which charts could be missed or lost and data omitted.

Several respondents noted that data collection with the registry was more comprehensive and standardized than other methods. An ICU director stated, “There was no good data registry in our ICU previously. We were registering the patient’s chart number, admission diagnosis, date of admission, the outcome at discharge, and discharge date. These were the only pieces of information we were collecting”. Another physician commented that prospectively collected data was at lower risk for bias, saying, “Collecting data is very difficult… It's usually retrospective… (which) could be biased with memory, biased with keeping the chart, and so forth”.

Several respondents noted that using registry data for research was easier, faster, and cheaper to access than a retrospective chart review. One physician said, “To get clinical data, you should review... hundreds of charts and investigation sheets, which is time-consuming. But the database registry system is easier for everyone to access”.

Perceived sustainability

Respondents had varying confidence in the perceived sustainability of the ICU registry in Ethiopia. Comments on perceived sustainability fell under three themes - institutional future, ownership and expansion.

Institutional future: Emergent subthemes included expectations, threats and adaptations. Many respondents voiced optimism that the ICU registry would be sustained at their hospital, though some were skeptical. One physician noted, “We have witnessed several great pilot projects with great impact but for some reasons they haven’t continued… We don’t have problem in starting a new thing but sustaining it”.

Respondents identified threats to perceived sustainability at their hospital, including added workload, need for ongoing training and support from leadership and clinicians, and funding. A nurse stated, “It will be sustainable if the management supports it… if the staff get training on the software and work appropriately… and if it is well-financed”. Others pointed out that ICU staffing is subject to periodic turnover that can make it difficult to sustain systems like the registry. One physician noted, “We used to have a permanent staff at the ICU station… so it was very easy to start something… But now we have staff rotations… every six months”. At the hospital level, respondents identified key stakeholders whose support is vital, including the ICU director, medical director, ICU nurses and clinicians, finance department, and CEO. One physician mentioned the need for champions among clinical staff, “you need representatives from the ICU like physicians’ representatives and nursing representatives to take the torch and make sure that the staff understands the registry”. Several respondents noted that the pilot implementation was grant-funded and raised concern that the hospital might not provide ongoing funding or resources.

Respondents suggested adaptations to the registry data elements to improve perceived sustainability, either narrowing to reduce data collection burden or expanding to improve data quality. Suggested additions included nutritional support, occurrence of specific procedures, locally common complications, and treatment cost.

Ownership: Several respondents raised concerns about ICU registry ownership, with emergent subthemes including distrust and data loss. Although the registry platform is provided free of charge by the CCAA Network, some respondents worried if that was sustainable. One ICU director commented, “We don’t know the owner of the software. We don’t know whether they will transfer the software ownership to us or not, or if they have the plan to sell it… Maybe for the future, the department needs to hold the registry”. The respondent suggested that this uncertainty could lead to cynicism and undermine support for the ICU registry among staff. Another physician commented, “Last time there was an application in our hospital. After using it for a while, we were requested to buy the application. If there were no paper-based documentation as backup, we would lose all information because we had no budget to buy the application”. A few other respondents voiced skepticism about relying on an electronic ICU registry that could be corrupted or susceptible to data loss, with one physician noting “The application might be crashed as well, so it is not good to fully 100% rely on the digital registry system”.

Expansion: Emergent subthemes included stakeholders, challenges, and solutions. Respondents were hopeful about expansion, with one ICU director saying, “Hopefully, it will be adopted by the MOH and implemented throughout Ethiopia and will be the standard of data collection”. Key stakeholders for expansion included regional health bureaus, and the Emergency and Critical Care Directorate of the MOH. Respondents stressed the importance of communication with leaders to win support. One ICU director suggested, “We have to show them the outcome, feasibility and other things to convince them... the way it is changing our day-to-day activities, how easy it is, and solving problems”. Another physician suggested that data from the current pilot could demonstrate the value of the ICU registry, “We should identify the strength of the registry because it is helpful to convince other hospitals... If it is failed in our setting, it would likely fail in other settings as well”. Several respondents noted that if the MOH valued the ICU registry, it could play an important role in providing necessary resources or mandating expanded implementation.

Others identified potential challenges, including lack of manpower, poor infrastructure, and need for institutional support. As one physician explained, “My concern is that most ICUs in Ethiopia are in an infantile state”. Several respondents noted that district ICUs are rarely staffed by physicians and nurses with critical care expertise and that this problem is exacerbated by staff rotations. One nurse noted, “The system needs trained manpower. Unless the person filling the data is trained, it will affect the data quality”. Regarding the poor infrastructure in district hospitals, one nurse stated, “There are resource-related challenges everywhere... It would be worst at periphery hospitals. There could be… an absence of computers and internet access... Without these resources, implementation of the registry could be impossible”. Several respondents also anticipated resistance to implementation of the ICU registry by ICU staff or hospital leadership. One nurse suggested, “They may resist if they think the registry system takes additional time. So, awareness creation on the benefit of the registry could be important”.

Regarding solutions to those challenges, several respondents suggested focusing on expanding to regional hospitals with more resources instead of more remote district hospitals. Some respondents suggested performing a baseline assessment of resources and needs to inform implementation. One ICU director suggested adapting the registry dataset for more resource-constrained settings, saying, “Since ICU level varies… having a core and an extended version of the registry will also be an asset... they need to decide based on their needs”. A data collector commented that ICUs with different levels of specialization might have different needs, saying, “Some ICUs might be surgical, medical, pediatrics, or neonatal... The information that is drawn from such ICUs might be totally different. The tool also has to be modified to better align with the level and type of ICUs”.

DISCUSSION

There is a dearth of clinical data on critical care in SSA, which poses a critical challenge to QI and systems strengthening efforts in the region[31,32]. The creation of ICU registries is an important step to aid evaluation of the impact of interventions to improve ICU performance[11,33]. Regional and national ICU registries are widely used in clinical research to describe case-mix, processes of care and clinical outcomes - including Australian and New Zealand Intensive Care Society in Australia and New Zealand, the Brazilian Research in Intensive Care Network in Brazil, the Intensive Care National Audit and Research Centre in the United Kingdom, and the Japanese Intensive Care Patient Database in Japan[15,34-36]. ICU registries also commonly track quality indicators including processes of care and healthcare associated events that make them powerful tools for QI, although very little of these data come from SSA[37,38]. The COVID-19 pandemic resulted in a rapid increase in ICU capacity in Ethiopia and other parts of SSA, although not always with adequate resources, training and coordinated systems of care to achieve optimal outcomes[10]. This trend has lent increased urgency to creation of ICU registries in SSA[23].

Implementation of clinical registries and other health information systems is a complex socio-technical process, and evaluations need to incorporate organizational, contextual, human, social and technological factors[39]. However, little research has examined how organizational cultures, health system priorities, and infrastructure influence registry implementation in LMICs[40]. In Asia, two previous studies been conducted to evaluate the implementation of the CCAA ICU registry. A quantitative analysis of the structure, coverage, reliability and validity of the registry in seven Asian countries found that it achieved a high quality of data collection, with structural and organizational performance comparable to ICU registries in HICs[41]. A qualitative phenomenological study addressing diffusion, dissemination, and perceived sustainability of the registry in four South Asian countries highlighted the importance of efforts to increase the innovation-system fit, the influence of motivated champions, and access to resources and expertise, while reliance on individuals and the priorities of other health care actors posed risks to perceived sustainability[40]. In South Africa, an ICU registry was developed and evaluated using a field audit for data completeness that found that incorporating data collection into the daily flow of clinical work and refining the design and user interface led to successful implementation and good compliance[24]. The decision between integrating data collection into routine clinical documentation and employing designated data collectors operating parallel to clinical care must be made in light of the implementation context including expertise and stability of the workforce and the approach to clinical documentation. As Ethiopian hospitals transition from paper to electronic medical records, a hybrid platform that with an embedded clinical registry may be advantageous[42].

This evaluation of the pilot implementation of the CCAA ICU registry at two tertiary care hospitals in Ethiopia adds to this burgeoning body of data about ICU registry implementation in LMICs. The implementation was considered an overall success by respondents, who commented positively on the registry’s feasibility and acceptability. Respondents felt strongly that the ICU registry was an improvement over previous data collection methods and provided vital information for clinical research and QI efforts in their facilities. This qualitative study identified key barriers and facilitators to implementation as well as threats to perceived sustainability, which can inform adaptations to optimize implementation and scale-up in the SSA context (Table 2). Important insights related to developing communication strategies with ICU staff, hospital leadership and governmental health bodies, and defining the minimal resources necessary for successful implementation. These findings correspond with the previous qualitative evaluation of the CCAA ICU registry in South Asia, which highlighted innovation-system fit, influence of champions, and access to resources and expertise as key themes, and identified data sharing, communication and adaptability as determinants of implementation success[40].

Table 2 Facilitators and barriers to implementation of the Critical Care Asia and Africa Intensive Care Unit registry in Ethiopia.

Characteristics	Explanation
The innovation
Relative advantage	Data stored online is safe and easily accessible
Compatibility	Easy to collect data using online platform with tablets
Observability	Pilot implementation at SPHMMC and AaBET and successes in Asia can demonstrate utility to other hospitals in Ethiopia
Adaptability	Would be helpful to have a means of offline data collection and asynchronous upload
	Desire for alterations in dataset to suit lower resource or specialized hospital
Communication and influence
Champions	Site leads who can advocate to hospital leadership
	Need to prioritize generating awareness and buy-in from hospital leadership and quality office
Social networks	Lack of awareness and support from ICU clinical staff
Outer context
Incentives and mandates	Will be helpful to get buy-in from the Ministry of Health to support broader implementation and sustained use
Interorganizational norm-setting and networks	Potential to develop a regional ICU registry network for research and QI
System readiness for innovation
Tension for change	High perceived value of the collected data among ICU registry team
Assessment of implications	Need to ensure that ICU clinical staff are aware of the utility and accessibility of the ICU registry data
Dedicated resources	Designated data collectors without concurrent clinical responsibilities, not all hospitals can afford to take staff out of clinical roles for data collection
	Continuous electricity and internet access, not universally available
	Equipment and financial support from CCAA
Adopter
Needs/motivation	Can fulfill hospital reporting mandates
	Desire to conduct research and quality improvement projects
	Skepticism about if or how data are being used
	Distrust about who has ownership of collected data
Skills	Staff with specialized knowledge needed for data collection and supervision, not available at all hospitals
Implementation process
Human resources	Importance of initial orientation and ongoing training for site leads and data collectors
	High workload for data collectors
	Current staffing model leads to gaps at night and on weekends
	High turnover of ICU staff and inexperience providers could be a challenge in some hospitals
Hands-on approach by leaders	Engaged site leads needed to monitor quality of collected data
External collaboration	Training and technical support from CCAA team

SPHMMC: St Paul’s Hospital Millennium Medical College; AaBET: Addis Ababa burn, Emergency and Trauma Hospital; ICU: Intensive care unit; CCAA: Critical Care Asia and Africa.

Effective communication was a major determinant of the feasibility and perceived sustainability of the ICU registry identified by respondents, aligning with the domain of Communication and Influence in the DISO conceptual model[26]. This study provided nuanced insights into multiple channels of communication between stakeholders including international collaborators, local registry team members, ICU staff, hospital leadership and governmental health bodies. Effective communication and responsiveness within the local registry team and between them and international collaborators were identified as important factors contributing to the project’s feasibility. In contrast, improving communication with the ICU staff, hospital leadership and governmental health bodies were identified as key opportunities to improve perceived sustainability. For example, respondents from the local registry team and ICU staff had differing impressions about whether data collection should the responsibility of one data collector or a responsibility shared among staff and how data collection should be integrated into or separated from clinical activities. These misunderstandings highlighted the importance of orienting staff to the ICU registry to ensure they understand its intended use, and of maintaining open channels of communication to address concerns that arise in the local implementation environment. Roles and responsibilities around data collection, ownership of the ICU registry, and access to collected data emerged as important topics that should be addressed through such channels. Although CCAA ensured that contributing sites maintained ownership and full access to all data collected in their ICUs with the option to participate in multi-institutional studies on a voluntary basis, not all clinical staff were aware of this agreement, highlighting an important point for clarification in orientation and communication with the clinical staff.

Respondents also emphasized the importance of familiarizing hospital leadership and governmental health bodies with the ICU registry, including its purpose and relevance, and providing routine updates on implementation successes and clinical trends. These forms of communication were identified as essential for establishing buy-in for ICU registry implementation, maintaining ongoing institutional support and ensuring resource availability. Improved communication with ICU staff, hospital leadership and governmental health bodies could also inform adaptations to fit local needs and resource limitations. This was especially important because lack of infrastructure and human resources were identified as important potential barriers to the feasibility and perceived sustainability of this ICU registry, and are also commonly identified barriers to implementing trauma registries in LMICs[43].

The novel online format of the CCAA ICU registry was enthusiastically received. However, adaptations to enable offline data collection with asynchronous upload to the cloud-based ICU registry platform could improve feasibility at sites with intermittent electricity and internet access. Similar challenges have been noted previously when implementing novel medical information systems in Ethiopia and other LMICs[44,45]. This is an ongoing barrier in Ethiopia - a recent report estimated that only 70% of Ethiopian hospitals have access to electricity, although this number may be higher among hospitals with ICUs[46]. The success of the ICU registry also relies on engaging data collectors and site leads with sufficient medical expertise, training on the ICU registry, time to complete their duties, and commitment to the success of the project. A similar challenge was noted in the performance evaluation of the CCAA ICU registry in Asia[41]. Notably, a recent national review of ICU capacity in Ethiopian hospitals highlighted manpower shortages, with only 29% of Ethiopian ICUs continually staffed by physicians with formal ICU training, 16% continually staffed by nurses with specialist ICU training, and this expertise concentrated in Addis Ababa[47].

The findings of this study should be considered in light of its strengths and limitations. This study is novel in addressing a critical knowledge gap about implementing ICU registries in SSA, providing key insights about how to adapt the CCAA registry and other novel clinical registries to the regional and institutional factors of this resource-limited context to optimize feasibility, acceptability, and perceived sustainability. By including respondents from multiple disciplines and leadership roles, as well as two different ICUs with distinctive patient populations and institutional cultures, we were able to capture diverse experiences, perspectives and insights. Nevertheless, both SPHMMC are AaBET are relatively high-resource institutions within Ethiopia, with specialized medical and nursing staff that are not widely found elsewhere in the country. Though many of our respondents have worked in other, lower-resource institutions, there may be challenges in disseminating the CCAA ICU registry that were not sufficiency anticipated here. We were also unable to elicit input from other stakeholders including public health workers in the MOH or regional health bureaus. Social desirability bias may have motivated some respondents to omit or minimize challenges that they faced. We attempted to mitigate this by utilizing facilitators who were not part of the ICUs’ clinical or leadership teams and protecting respondents’ anonymity. Given grant-sponsored donor funding of this pilot implementation, we were unable to assess true sustainability including institutionalization with cost absorption and independence from external support, but only stakeholders’ perceptions of that possibility. It is possible that additional threats to perceived sustainability of the ICU registry will emerge over time, so in practice the evaluation process should not be constrained to a single time point. Nevertheless, the results of this early evaluation were useful in informing efforts to strengthen the ICU registry implementation process.

Since we conducted this evaluation, the ICU registries at SPHMMC and AaBET have continued to prospectively collect data on ICU patients, with modifications proposed by this evaluation including efforts to strengthen communication between the ICU registry team and the ICU clinical staff and hospital leadership, with particular focus on building understanding of the purpose of the registry and ownership of the collected data by the ICU teams at SPHMMC and AaBET. A study published in 2024 analyzed ICU registry data to describe the case-mix, processes of care and clinical outcomes in the ICUs at SPHMMC and AaBET, identifying ICU-acquired infections as a priority for QI efforts[25]. Since then, several ongoing research and QI projects have utilized ICU registry data. The ICU registry was also adapted to generate reports on key performance indicators that are shared with hospital management, the quality directorate and the MOH. The CCAA ICU registry also has been implemented subsequently in multiple other hospitals in Ethiopia and SSA. Future scale-up of the CCAA ICU registry and development of a regional ICU registry network would allow for systematic evaluation of the burden of illness and processes of care in ICUs in SSA using a common dataset, as CCAA has accomplished in Asia[14]. This registry network would facilitate clinical outcomes and implementation science research to determine how to best use the limited resources available for critical care, informing locally-derived, context-appropriate guidelines, standards of care and QI initiatives[48].

CONCLUSION

The CCAA ICU registry is a promising tool for research and QI in resource-limited settings. However, this pilot demonstrates that successful implementation in SSA requires a clear understanding of regional and institutional influencing factors. This implementation evaluation provides a nuanced picture of the myriad factors affecting the feasibility, acceptability, and perceived sustainability of the CCAA ICU registry in two well-resourced hospitals in Addis Ababa, Ethiopia. It thus offers insights to inform scalable implementation across LMICs, contingent on resource alignment, stakeholder engagement, and contextual adaptation. Successful adaptation of the CCAA ICU registry, including formalizing communication, training, supervision and leadership approaches, can enhance its implementation and dissemination processes in order to address a crucial data gap about critical illness and critical care throughout SSA.