In-hospital cardiac arrest in a resource-limited emergency department in Somalia: Outcomes and priorities for resuscitation system development

Abstract

Background

In-hospital cardiac arrest (IHCA) is associated with high mortality, particularly in low- and middle-income countries where monitoring, emergency response systems, and post-resuscitation resources are limited. Evidence from Somalia remains scarce. This study aimed to describe the clinical characteristics, arrest circumstances, resuscitation-process documentation, short-term outcomes, and exploratory factors associated with return of spontaneous circulation (ROSC) among adult IHCA patients in a resource-limited emergency department in Mogadishu, Somalia.

Methods

This prospective observational study was conducted in the emergency department of Mogadishu Somalia–Turkey Recep Tayyip Erdoğan Training and Research Hospital between January 1, 2025 and April 30, 2026. Adult patients aged ≥18 years who experienced IHCA during the study period were consecutively enrolled. Data were collected prospectively using a standardized structured form. The primary outcome was ROSC, and the secondary outcome was survival beyond 24 h. Logistic regression was used only as an exploratory analysis because of the low number of ROSC events.

Results

A total of 240 patients were included. The mean age was 46.3 ± 15.2 years, and 136 patients (56.7%) were male. Most arrests occurred in the emergency department (200, 83.3%). Trauma was the leading admission diagnosis (139, 58.0%), predominantly severe traumatic brain injury (107, 76.9% of trauma cases), followed by sepsis (32, 13.3%). Witness status was documented in 190 patients; among these, 120 arrests (63.2%) were unwitnessed. Emergency response activation was documented in 43 patients (22.6%), and CPR initiation was documented in 35 patients (14.6%). ROSC was achieved in 14 patients (5.8%), and 3 patients (1.3%) survived beyond 24 h. In exploratory analysis, trauma admission, sepsis admission, HIV comorbidity, emergency department arrest, and weekend arrest were associated with reduced odds of ROSC.

Conclusion

Adult IHCA outcomes were extremely poor in this resource-limited emergency department, with low ROSC and minimal 24-h survival. The findings highlight a high burden of trauma-related arrest, frequent unwitnessed events, incomplete physiological and resuscitation-process documentation, low documented emergency response activation, and low documented CPR initiation. Context-specific strategies focusing on monitoring, early recognition, standardized emergency response activation, CPR training, structured documentation, and post-arrest review are needed to improve resuscitation systems in Somalia.

Introduction

In-hospital cardiac arrest (IHCA) remains a major cause of hospital mortality and is associated with substantial variation in survival across healthcare systems. Survival depends not only on cardiopulmonary resuscitation (CPR) quality, but also on early recognition of deterioration, timely activation of emergency response systems, rhythm recognition, defibrillation where indicated, and structured post-resuscitation care.¹

Contemporary resuscitation guidelines emphasize rapid identification of cardiac arrest, immediate initiation of high-quality CPR, coordinated team-based response, and standardized post-arrest stabilization as core components of the in-hospital chain of survival.² Adult basic and advanced life support algorithms further highlight the importance of organized resuscitation training, airway and ventilation management, medication use, and post-arrest care pathways.³ Where return of spontaneous circulation (ROSC) is achieved, access to structured post-resuscitation care, including hemodynamic optimization, oxygenation and ventilation control, neurological monitoring, and intensive care, is essential for improving outcomes.⁴

Large registry-based studies from high-income countries have shown that IHCA incidence and survival are strongly influenced by hospital systems, patient monitoring capacity, rapid response structures, and access to critical care.⁵ In low- and middle-income countries, these determinants are often constrained by overcrowding, delayed presentation, limited monitoring equipment, shortages of trained staff, inconsistent documentation, and restricted access to intensive care.⁶ Strengthening emergency care systems has therefore been identified as a key strategy for reducing avoidable mortality in resource-limited settings.⁷

In sub-Saharan Africa, emergency and prehospital care systems remain variably developed, and available evidence suggests that injury, sepsis, respiratory illness, cardiovascular disease, and other acute conditions contribute substantially to emergency care demand while system capacity remains limited.⁸ However, data on IHCA from Somalia are extremely scarce. This lack of local evidence limits the ability to evaluate resuscitation performance, identify modifiable system gaps, and develop context-specific interventions. This study aimed to describe the clinical characteristics, arrest circumstances, resuscitation-process documentation, short-term outcomes, and exploratory factors associated with ROSC among adult IHCA patients in a resource-limited emergency department in Mogadishu, Somalia, with the goal of informing context-specific resuscitation system development.⁹

Methods

Study design and setting

This prospective observational study was conducted in the emergency department of Mogadishu Somalia–Turkey Recep Tayyip Erdoğan Training and Research Hospital, a tertiary referral center in Mogadishu, Somalia. The study period extended from January 1, 2025 to April 30, 2026.

Hospital context and emergency response process

Mogadishu Somalia–Turkey Recep Tayyip Erdoğan Training and Research Hospital is a 250-bed tertiary teaching and referral hospital in Mogadishu, Somalia, serving Mogadishu and surrounding regions. Formerly known as Digfer Hospital, the facility was rebuilt and reopened in 2015 and provides emergency, medical, surgical, cardiovascular, diagnostic, and critical-care services. The adult intensive and coronary care network has a combined capacity of 42 beds. During the 16-month study period from January 1, 2025 to April 30, 2026, the hospital recorded 32,184 emergency department visits, 2204 ICU admissions, and 21,632 total hospital admissions.

During the same period, 240 adult IHCA cases were prospectively enrolled. This corresponds to a crude observed burden of 11.1 IHCA cases per 1000 total hospital admissions. However, this value was interpreted as a descriptive measure of observed IHCA burden rather than a formal hospital-wide IHCA incidence rate because the study surveillance pathway did not constitute a comprehensive hospital-wide cardiac-arrest registry covering all inpatient locations.

In-hospital cardiac arrest events were identified by clinical staff at the bedside and managed according to the hospital’s available emergency response process. Emergency response activation was defined as documented activation of the hospital emergency or resuscitation response process following recognition of cardiac arrest. The response generally involved emergency department clinicians and nurses, with involvement of anesthesia, intensive care, or other available specialist teams when clinically indicated and available. Resuscitation equipment, airway support, and defibrillation capacity were available according to local resources, although detailed Utstein variables such as initial rhythm, defibrillation timing, airway interventions, and medications were not consistently documented.

During the study period, there was no uniformly documented electronic code-blue registry capturing all Utstein variables for each IHCA event. Decisions to initiate or withhold resuscitation were made by the attending clinical team based on the patient’s clinical condition, perceived reversibility, and available resources. Formal treatment-limitation or futility decisions were not consistently captured as a separate structured variable. Therefore, the study could not reliably distinguish between non-initiation of CPR, undocumented CPR, delayed recognition after irreversible deterioration, and clinician judgement that resuscitation was futile.

Study population

All adult patients aged 18 years or older who experienced in-hospital cardiac arrest during the study period were eligible for inclusion. All eligible adult IHCA patients were consecutively enrolled during the study period. In-hospital cardiac arrest was defined as the sudden cessation of cardiac mechanical activity occurring after hospital arrival, confirmed clinically by unresponsiveness, apnea, and absence of a detectable pulse. Patients with documented do-not-resuscitate orders and those with missing primary outcome data, defined as unavailable ROSC status or unavailable 24-h survival status.

Data collection and variables

Data were collected prospectively using a standardized structured data collection form. Recorded variables included demographic characteristics, location of cardiac arrest, duration of illness before admission, primary admission diagnosis, comorbid conditions, arrest circumstances, emergency response activation, cardiopulmonary resuscitation (CPR) initiation, response time, CPR duration, return of spontaneous circulation (ROSC), post-arrest care location, and 24-h survival.

Physiological parameters at or immediately before cardiac arrest were recorded where available and included blood pressure, pulse rate, oxygen saturation, motor Glasgow Coma Scale (GCS), random blood glucose, and body temperature.

Variables and definitions

Cardiac arrest location was categorized as emergency department, intensive care unit, or operating theatre. These categories reflected the locations captured by the study surveillance and available emergency response documentation pathway during the study period. Ward arrests outside this pathway were not reliably captured and were therefore not included in the analytic cohort.

Return of spontaneous circulation was defined as restoration of a palpable pulse and effective circulation following resuscitation.

Witnessed cardiac arrest was defined as cardiac arrest directly observed by healthcare staff or patient attendants at the time of collapse. Unwitnessed cardiac arrest was defined as cardiac arrest not directly observed at the time of collapse. Cases in which witness status was not recorded in the clinical record or data collection form were classified as not documented.

Emergency response activation was defined as documented activation of the hospital emergency or resuscitation response process following recognition of cardiac arrest.

CPR initiation was defined as documented commencement of chest compressions and/or formal cardiopulmonary resuscitation following cardiac arrest. If CPR was not recorded in the clinical notes or data collection form, CPR initiation was classified as not documented rather than assumed to have been performed.

Response time was defined as the documented interval from recognition of cardiac arrest to initiation of resuscitation or arrival of the resuscitation response, according to the available clinical record. Because response time documentation was dependent on routine clinical recording, this variable was analyzed as a resuscitation process indicator rather than as a precise time-to-intervention measure.

Primary admission diagnosis was defined as the main clinical condition for which the patient was admitted before cardiac arrest. Admission diagnoses were categorized as trauma, sepsis, hypertensive emergency or hypertension-related acute presentation, central nervous system disorder, heart disease, diabetes mellitus-related presentation, or liver disease. Severe traumatic brain injury was recorded as a trauma subgroup rather than as a separate primary diagnosis.

Heart disease was recorded as a primary admission diagnosis when it represented the acute reason for hospital presentation. Chronic cardiac disease was not consistently documented as a separate comorbidity variable in the dataset and therefore was not reported among comorbid conditions.

The immediate cause of cardiac arrest according to full Utstein IHCA recommendations could not be reliably classified because initial arrest rhythm, defibrillation attempts, airway interventions, resuscitation medications, and presumed immediate arrest mechanism were not consistently documented. Therefore, this study reports primary admission diagnoses and clinical contexts associated with IHCA rather than definitive Utstein-classified immediate causes of cardiac arrest.

Comorbidities included documented hypertension, diabetes mellitus, HIV infection, malignancy, chronic kidney disease, chronic liver disease, and chronic central nervous system disease.

Case ascertainment

IHCA cases were identified prospectively through emergency department clinical records, resuscitation documentation, emergency response activation records, and study data collection forms completed during the study period. Eligible cases were reviewed by the study investigators to confirm that cardiac arrest occurred after hospital arrival and met the study definition of IHCA. Because the hospital did not have a comprehensive electronic cardiac arrest registry capturing all Utstein variables, case ascertainment depended on prospective clinical identification and available documentation.

Outcomes

The primary outcome was return of spontaneous circulation, defined as restoration of a palpable pulse and effective circulation following resuscitation. The secondary outcome was survival beyond 24 h after cardiac arrest. Additional process indicators included emergency response activation, CPR initiation, response time, CPR duration, and post-arrest care location.

Data quality and missing data

For variables with incomplete documentation, the number of patients with available data was recorded and reported. Percentages were calculated using the full study cohort as the denominator unless otherwise specified. For variables recorded only in a subset of patients, including witness status, physiological parameters, and post-arrest care location, the relevant denominator was explicitly stated.

Missing data were not imputed. Available-case analysis was used for variables with incomplete documentation. Absence of documentation was not assumed to indicate absence of the event, intervention, or physiological abnormality. In particular, lack of documented CPR initiation was treated as undocumented CPR initiation rather than definitive evidence that CPR was not performed, unless the clinical record explicitly stated that resuscitation was withheld or not initiated.

CPR initiation was recorded only when commencement of chest compressions and/or formal cardiopulmonary resuscitation was documented in the clinical record or data collection form. Absence of documented CPR initiation was not assumed to mean that CPR was definitely not performed unless the clinical record explicitly stated that resuscitation was withheld or not initiated. Therefore, documented CPR initiation was interpreted as a resuscitation-process documentation variable rather than a definitive measure of all CPR attempts.

Statistical analysis

Data were analyzed using IBM SPSS Statistics version 26.0. Categorical variables were summarized as frequencies and percentages. Continuous variables were summarized as mean with standard deviation or median with range/interquartile range, according to distribution and available reporting format.

Associations between clinical variables and ROSC were assessed using logistic regression analysis. Univariable logistic regression was first performed for demographic, clinical, arrest-related, and resuscitation process variables. Variables with p < 0.05 in univariable analysis were entered into a multivariable logistic regression model to identify factors associated with ROSC. Results were reported as odds ratios (ORs) with 95% confidence intervals (CIs). A two-sided p value < 0.05 was considered statistically significant.

Because of the small number of ROSC events, the regression model was limited to variables identified in univariable analysis and was not intended for development of a predictive model. Formal model diagnostics and collinearity assessment were limited by the low event count; therefore, results were interpreted descriptively and cautiously. Given the low number of ROSC events, regression findings were interpreted as exploratory and hypothesis-generating rather than definitive evidence of independent prediction.

Ethical considerations

The study was approved by the Research Ethics Committee of Mogadishu Somali Turkey Training and Research Hospital (Ethics Protocol No.: MSTH22674). The study was conducted in accordance with the principles of the Declaration of Helsinki. As part of routine hospital admission procedures, patients or caregivers provide consent for anonymized clinical data from hospital medical records to be used for research purposes. No patient names or directly identifiable information were included in the study dataset or manuscript.

Results

Patient characteristics and arrest location

A total of 240 adult patients with in-hospital cardiac arrest (IHCA) were prospectively enrolled between January 1, 2025 and April 30, 2026. The mean age was 46.3 ± 15.2 years, and 136 patients (56.7%) were male. Most patients were young adults aged 18–44 years (159, 66.5%), while 63 (26.3%) were aged 45–64 years and 18 (7.2%) were aged 65 years or older.

Most arrests occurred in the emergency department (200, 83.3%), followed by the intensive care unit (ICU) in 30 patients (12.5%) and the operating theatre in 10 patients (4.2%). The duration of illness before admission was less than one week in 178 patients (74.3%) and one week or longer in 62 patients (25.7%) (Table 1).

Table 1.

Baseline clinical characteristics, arrest location, admission diagnoses, and comorbidities.

Characteristic	Frequency, n (%)
Total cohort	240 (100.0)
Sex
Male	136 (56.7)
Female	104 (43.3)
Age group, years
18–44	159 (66.5)
45–64	63 (26.3)
≥65	18 (7.2)
Location of cardiac arrest
Emergency department	200 (83.3)
Intensive care unit	30 (12.5)
Operating theatre	10 (4.2)
Duration of illness before admission
<1 week	178 (74.3)
≥1 week	62 (25.7)
Primary admission diagnosis
Trauma	139 (58.0)
Severe head injury among trauma cases	107 (76.9)
Sepsis	32 (13.3)
Hypertension-related presentation	22 (9.2)
Central nervous system disorder	16 (6.7)
Heart disease	14 (5.8)
Diabetes mellitus	11 (4.6)
Liver disease	6 (2.5)
Comorbidities among patients with documented comorbidity, n = 62
HIV infection	28 (45.0)
Hypertension	14 (22.0)
Malignancy	9 (14.0)
Diabetes mellitus	8 (13.0)
Kidney disease	1 (2.0)
Liver disease	1 (2.0)
Central nervous system disease	1 (2.0)

Abbreviations: HIV, human immunodeficiency virus; n, number.

During the study period, the hospital recorded 21,632 total admissions, 2204 ICU admissions, and 32,184 emergency department visits. The 240 enrolled IHCA cases corresponded to an observed IHCA burden of 11.1 cases per 1000 total hospital admissions; this was not interpreted as a formal hospital-wide IHCA incidence rate because case ascertainment did not constitute a comprehensive hospital-wide cardiac-arrest registry.

Admission diagnoses and comorbid conditions

Trauma was the leading admission diagnosis, accounting for 139 cases (58.0%). Among trauma-related admissions, severe traumatic brain injury was documented in 107 patients (76.9% of trauma cases). Other admission diagnoses included sepsis (32, 13.3%), hypertension-related presentations (22, 9.2%), central nervous system disorders (16, 6.7%), heart disease (14, 5.8%), diabetes mellitus (11, 4.6%), and liver disease (6, 2.5%).

Comorbidities were documented in 62 patients. HIV infection was the most frequently reported comorbidity (28, 45.0%), followed by hypertension (14, 22.0%), malignancy (9, 14.0%), and diabetes mellitus (8, 13.0%). Chronic kidney disease, chronic liver disease, and chronic central nervous system disease were each documented in one patient (2.0%) (Table 1).

Arrest circumstances, monitoring, and resuscitation process

Witness status was documented in 190 patients. Among these, 120 arrests were unwitnessed and 70 were witnessed. Witness status was not documented in 50 patients. Among witnessed arrests, 68 (97.1%) were witnessed by healthcare staff and 2 (2.9%) by patient attendants.

Cardiac arrests occurred during daytime hours in 86 patients (36.0%), nighttime hours in 66 patients (27.4%), and weekends in 88 patients (36.6%). Emergency response activation was documented in 43 patients (22.6%), and CPR initiation was documented in 35 patients (14.6%). Response time was recorded as ≤2 min in 128 patients (53.6%) and >2 min in 112 patients (46.4%). The mean duration of CPR was 24.4 min, with a range of 11–86 min (Table 2).

Table 2.

Arrest circumstances, resuscitation process, physiological documentation, and outcomes.

Variable	Frequency, n (%)
Witness status
Witnessed	70 (29.2)
Unwitnessed	120 (50.0)
Not documented	50 (20.8)
Among witnessed arrests, n = 70
Witnessed by healthcare staff	68 (97.1)
Witnessed by patient attendants	2 (2.9)
Time of arrest
Daytime	86 (36.0)
Nighttime	66 (27.4)
Weekend	88 (36.6)
Resuscitation process indicators
Emergency response activated	43 (22.6)
CPR initiation documented	35 (14.6)
Response time
≤2 min	128 (53.6)
>2 min	112 (46.4)
Mean CPR duration, min	24.4 (range, 11–86)
Physiological measurements recorded
Blood pressure recorded	167 (69.5)
Motor GCS recorded	167 (69.5)
Pulse rate recorded	136 (56.8)
Oxygen saturation recorded	95 (39.5)
Random blood glucose recorded	67 (27.9)
Body temperature recorded	30 (12.6)
Short-term outcomes
ROSC achieved	14 (5.8)
Survival beyond 24 h	3 (1.3)
Post-arrest management among patients achieving ROSC, n = 14
ICU	7 (50.0)
General ward	4 (28.6)
Emergency department	3 (21.4)
Postmortem findings
Postmortem performed	75 (31.3)
Postmortem not incorporated into clinical diagnosis	18/75 (24.0)

Abbreviations: CPR, cardiopulmonary resuscitation; GCS, Glasgow Coma Scale; ICU, intensive care emergency department; ROSC, return of spontaneous circulation.

Physiological documentation at the time of arrest was incomplete and varied by parameter. Blood pressure and motor GCS were recorded in 167 patients (69.5%), pulse rate in 136 (56.8%), oxygen saturation in 95 (39.5%), random blood glucose in 67 (27.9%), and body temperature in 30 (12.6%) (Table 2). Because physiological abnormalities were not consistently documented across the full cohort, these variables were interpreted as indicators of documentation and monitoring capacity rather than as definitive predictors of outcome.

Resuscitation outcomes

ROSC was achieved in 14 patients. Using the full cohort denominator, this corresponds to 5.8% of all included IHCA patients. Three patients survived beyond 24 h, corresponding to 1.3% of the full cohort. Among patients achieving ROSC, post-arrest care was provided in the ICU for 7 patients (50.0%), the general ward for 4 patients (28.6%), and the emergency department for 3 patients (21.4%).

Postmortem examination was performed in 75 patients (31.3%). Among postmortem cases, findings were not incorporated into the clinical diagnosis in 18 cases (24.0%) (Table 2). The resuscitation process and short-term outcome cascade are presented in Fig. 1.

Fig. 1.

Utstein-style patient flow and short-term outcomes among adult in-hospital cardiac arrest patients.

Exploratory factors associated with return of spontaneous circulation

In exploratory logistic regression analysis, trauma admission, sepsis admission, HIV comorbidity, arrest occurring in the emergency department, and weekend arrest were associated with reduced odds of ROSC.

Trauma admission was associated with lower odds of ROSC (OR 0.12; 95% CI 0.03–0.45; p = 0.002), as was sepsis admission (OR 0.14; 95% CI 0.04–0.52; p = 0.003). HIV comorbidity was also associated with reduced odds of ROSC (OR 0.10; 95% CI 0.01–0.72; p = 0.027). Arrest occurring in the emergency department was associated with reduced odds of ROSC (OR 0.10; 95% CI 0.02–0.46; p = 0.003). Weekend arrest was similarly associated with reduced odds of ROSC (OR 0.20; 95% CI 0.04–0.90; p = 0.042) (Table 3). No statistically significant associations were observed between ROSC and age, sex, or time of day. A forest plot of exploratory regression findings is provided as Supplementary Fig. 1.

Table 3.

Exploratory factors associated with reduced odds of return of spontaneous circulation.

Factor	Odds ratio (95% CI)	p value
Trauma admission	0.12 (0.03–0.45)	0.002
Sepsis admission	0.14 (0.04–0.52)	0.003
HIV comorbidity	0.10 (0.01–0.72)	0.027
Arrest in emergency department	0.10 (0.02–0.46)	0.003
Weekend arrest	0.20 (0.04–0.90)	0.042

Abbreviations: CI, confidence interval; HIV, human immunodeficiency virus; ROSC, return of spontaneous circulation.

Discussion

This prospective observational study provides one of the first detailed descriptions of adult IHCA in a resource-limited emergency department in Somalia. The findings demonstrate extremely poor short-term outcomes, with ROSC achieved in only 14 patients and 24-h survival documented in only 3 patients. The study also identifies a high burden of trauma-related arrest, frequent unwitnessed events, incomplete physiological documentation, low documented emergency response activation, low documented CPR initiation, and exploratory associations between reduced odds of ROSC and trauma admission, sepsis admission, HIV infection, emergency department arrest, and weekend arrest.

The low ROSC rate observed in this cohort is consistent with reports from some low-resource hospital settings. A Ugandan tertiary hospital study reported ROSC in 14 patients (7.4%) and 24-h survival in three patients, demonstrating similarly poor outcomes in a resource-constrained sub-Saharan African setting.¹⁰ The similarity between these findings and the present study suggests that poor IHCA outcomes in the region may reflect shared structural challenges, including delayed recognition of deterioration, limited continuous monitoring, inconsistent CPR initiation, and restricted access to post-arrest critical care.

In contrast, IHCA outcomes reported from higher-resource settings are substantially better, although still associated with high mortality. A major review of IHCA highlighted that survival is influenced by patient characteristics, arrest rhythm, location of arrest, time to CPR, time to defibrillation, institutional response systems, and quality of post-resuscitation care.¹¹ The markedly lower survival in the present cohort therefore likely reflects not only patient-level severity, but also system-level constraints in monitoring, response activation, resuscitation delivery, and post-arrest care.

Trauma was the leading admission diagnosis in this cohort, accounting for more than half of all arrests, with severe traumatic brain injury representing the dominant trauma subgroup. This pattern differs from many high-income IHCA cohorts, where cardiac, respiratory, and medical causes often predominate, and reflects the broader emergency care burden in low-resource settings where injury contributes substantially to preventable mortality.¹² The predominance of severe head injury is clinically important because these patients may deteriorate rapidly due to intracranial hypertension, hypoxia, hypotension, hemorrhage, aspiration, or delayed neurosurgical intervention.

The association between trauma admission and reduced odds of ROSC is biologically plausible. Traumatic cardiac arrest is frequently preceded by severe hypoxia, hemorrhagic shock, obstructive physiology, or catastrophic neurological injury, and outcomes are poor when reversible causes are not rapidly identified and corrected.¹³ In the present setting, delayed presentation, limited prehospital care, restricted imaging or neurosurgical capacity, and constrained blood product availability may further reduce the likelihood of successful resuscitation after trauma-related arrest.

Sepsis was also associated with reduced odds of ROSC. This is consistent with the recognized global burden of sepsis and its disproportionate impact in low-resource regions, where delayed presentation, limited diagnostic capacity, antimicrobial access challenges, and restricted intensive care resources contribute to high mortality.¹⁴ Sepsis-associated cardiac arrest is often the terminal manifestation of profound circulatory failure, metabolic acidosis, hypoxia, and multi-organ dysfunction, which may explain the poor reversibility observed in this cohort.

HIV infection was the most frequently documented comorbidity and was associated with reduced odds of ROSC. This finding should be interpreted cautiously because the study did not capture immune status, antiretroviral therapy use, opportunistic infections, or HIV disease stage. However, HIV may act as a marker of underlying immunosuppression, advanced infection, malnutrition, delayed presentation, or comorbidity burden, all of which may worsen physiological reserve during critical illness and reduce the probability of successful resuscitation.¹⁵

A major system-level finding was the high proportion of unwitnessed arrests among patients with documented witness status. Unwitnessed IHCA is clinically important because it usually reflects delayed recognition, delayed CPR initiation, and prolonged no-flow or low-flow time. Prior evidence has shown that time to CPR is a critical determinant of survival after cardiac arrest, and any delay in recognition or intervention substantially reduces the probability of ROSC and meaningful recovery.¹⁶

The low documented rate of CPR initiation is a major finding and should be interpreted cautiously. CPR initiation was documented in only 35 patients, but absence of documentation was not assumed to mean that CPR was definitely not performed. This finding may reflect true non-initiation of resuscitation, incomplete documentation during high-acuity events, delayed recognition after irreversible deterioration, or clinical judgement that resuscitation was futile, particularly among patients with catastrophic trauma or severe traumatic brain injury. Because formal treatment-limitation or futility decisions were not consistently captured as a structured variable, the study cannot determine the relative contribution of these explanations. Standardized cardiac arrest documentation, post-event review, and clearer recording of resuscitation decisions may therefore be essential first steps for improving resuscitation quality and interpretability of future IHCA data.¹⁷

Arrest occurring in the emergency department was associated with reduced odds of ROSC. This likely reflects a combination of high illness severity at presentation, overcrowding, limited monitoring, delayed escalation, and competition for staff attention in a high-throughput emergency environment. Emergency departments in LMICs often function as both acute receiving emergency departments and holding areas for critically ill patients, which can lead to prolonged stays without the monitoring intensity available in ICU or operating theatre settings.¹⁸

Weekend arrest was also associated with reduced odds of ROSC, suggesting a possible “weekend effect.” This may reflect reduced staffing levels, lower availability of senior decision-makers, slower escalation, limited access to diagnostics or critical care beds, and reduced operational readiness outside routine working hours. While this finding requires confirmation in larger studies, it identifies a potentially modifiable service-delivery problem that could be addressed through weekend staffing models, resuscitation team coverage, and standardized emergency response protocols.¹⁹

Physiological documentation was incomplete and varied substantially by parameter. Although blood pressure and motor GCS were documented in approximately two-thirds of patients, oxygen saturation, random blood glucose, and body temperature were recorded less consistently. These findings should therefore be interpreted primarily as evidence of gaps in monitoring and documentation rather than as definitive evidence of specific pre-arrest physiological patterns. Nevertheless, incomplete physiological monitoring in a high-risk emergency department population highlights the need for structured reassessment, early warning systems, continuous monitoring of critically ill patients where feasible, and clear escalation protocols.²⁰

The findings have direct implications for the development of context-specific resuscitation strategies in Somalia. A feasible improvement package should prioritize structured monitoring of high-risk patients, standardized emergency response activation, regular CPR and basic life support training, clear role allocation during resuscitation, reliable availability of airway and defibrillation equipment, standardized cardiac arrest documentation, and post-event review. In resource-limited settings, improving recognition, response activation, documentation, and basic resuscitation processes may be more practical and immediately impactful than complex technologies that require intensive infrastructure.²¹

This study has several strengths. It was prospective, included a relatively large cohort for a single-center IHCA study in Somalia, and captured clinically relevant patient-level, arrest-level, and system-level variables. However, limitations must be acknowledged. The study was conducted at a single tertiary referral hospital, which may limit generalizability. Several Utstein-recommended variables, including initial arrest rhythm, defibrillation attempts and timing, presumed immediate cause of arrest, airway interventions, medications administered during resuscitation, survival to hospital discharge, and neurological outcome, were not consistently available. This limits comparison with international IHCA registries and restricts causal interpretation of resuscitation outcomes. Documentation was incomplete for some physiological and process variables, including witness status and CPR initiation, which may have introduced documentation bias. Regression findings should also be interpreted as exploratory because the number of ROSC events was low relative to the number of variables examined. Nevertheless, the study provides important baseline evidence for resuscitation system development and future multicenter research in Somalia.²²

Conclusion

In this prospective observational study of adult in-hospital cardiac arrest in a resource-limited emergency department in Mogadishu, Somalia, outcomes were extremely poor, with low ROSC and minimal 24-h survival. Trauma, particularly severe traumatic brain injury, was the leading admission diagnosis, and many arrests were unwitnessed or incompletely documented. Trauma admission, sepsis admission, HIV comorbidity, emergency department arrest, and weekend arrest were associated with reduced odds of ROSC, although these findings should be interpreted as exploratory.

These findings highlight critical gaps in early recognition, monitoring, emergency response activation, CPR delivery, documentation, and post-arrest care. Context-specific resuscitation strategies focusing on early warning systems, structured emergency response pathways, regular CPR training, improved documentation, and access to essential post-arrest care are urgently needed. Future multicenter studies should evaluate IHCA epidemiology across Somalia and assess whether targeted resuscitation system interventions can improve survival outcomes.

CRediT authorship contribution statement

Mohamed Mukhtar Mohamed: Investigation, Data curation. Said Abdirahman Ahmed: Visualization, Supervision. Nasteho Mohamed Sheikh Omar: Writing – review & editing, Methodology. Ömer Metin: Methodology, Investigation, Formal analysis. Yusuf Çokay: Writing – review & editing, Formal analysis, Conceptualization. Hassan Adan Ali Adan: Writing – original draft, Supervision, Resources.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

No external funding was received for this study.

All authors have reviewed and approved this declaration.

Appendix A. Supplementary material

The following are the Supplementary material to this article:

Supplementary Fig. 1.

Forest Plot of exploratory factors associated with return of spontaneous circulation (ROSC).