Juma A. Mfinanga

Background: Pediatric sepsis remains a leading cause of mortality in resource-limited settings. We evaluated the delivery of time-critical sepsis interventions and their association with in-hospital mortality among children in Tanzania. Methods: We conducted a secondary analysis of a prospective cohort of children (28 days to 14 years) presenting with sepsis to the Emergency Medicine Department at Muhimbili National Hospital, Dar es Salaam, Tanzania (July 2022 to November 2024). Sepsis was defined as fever and at least one World Health Organization warning sign. Local pediatric sepsis guidelines recommend antibiotics and intravenous fluids (IVF) for all patients with sepsis, regardless of the source. Primary exposures were receipt and timing of antibiotics and IVF. The primary outcome was in-hospital mortality. Multivariable logistic regression adjusted for age, nutritional status, referral status, and illness severity. Results: Among 755 children, 61.2% ( n = 462) received antibiotics, and 55.6% ( n = 420) received IVF. Median time to intervention was 2.4 h (IQR 1.6–3.7). 35.5% ( n = 268) received both interventions, and only 3.7% ( n = 28) received them within one hour. In-hospital mortality was 19.6% ( n = 148), with a median time to death of 4.1 days (IQR 1.6–9.5). In a multivariable analysis, time to initiation of both antibiotics and IVF was not significantly associated with in-hospital mortality (adjusted OR 1.29; 95% CI 0.83–2.0; p = 0.264). When modeled continuously, a longer time to IVF administration was associated with increased mortality risk (HR 1.05 per hour delay; 95% CI 1.01–1.10). Conclusions: Receipt of time-critical sepsis interventions (antibiotics and IVF) was delayed and inconsistent, with no observed association with decreased mortality, although residual confounding may remain. This both highlights gaps in early evidence-based care delivery and suggests additional, system-related influences on pediatric sepsis outcomes. Our findings underscore the need for context-specific strategies to strengthen pediatric sepsis management in resource-limited settings.

: Severe febrile illness (SFI), a major cause of child mortality in Sub-Saharan Africa (SSA), is frequently caused by bloodstream infections (BSI), increasingly caused by antimicrobial-resistant organisms. However, antimicrobial resistance (AMR) prevalence and patterns in this population are largely unknown. We conducted a prospective cohort study of children aged 28 days to 14 years with SFI at a referral hospital in Tanzania (July 2022–September 2023) to determine the prevalence, pathogen profile, and AMR patterns of BSIs. Blood cultures were performed on all participants. Isolates underwent AMR testing and whole genome sequencing (WGS) to identify AMR genes and assess genetic relatedness. The association between BSI and mortality was assessed using logistic regression. Among the 392 enrolled children (median age 17.8 months), 5.2% ( n = 20) had culture-confirmed BSI, with a case fatality rate of 45%. Gram-negative bacteria predominated ( Escherichia coli, Klebsiella pneumoniae ), with 79% of isolates exhibiting phenotypic resistance to ceftriaxone, a first-line antimicrobial. WGS revealed extended–spectrum β-lactamase genes ( CTX-M-15 , CTX-M-27 ) in most Enterobacterales isolates and identical isolates among 2/4 patients with Streptococcus pneumoniae and 2/3 Candida albicans infections. Children with BSI had significantly higher mortality than those without BSI (adjusted odds ratio 3.9; 95% CI: 1.5–10.1, P <0.01). These findings emphasize the urgent need for expanded AMR surveillance systems, empiric antibiotic regimens tailored to local AMR patterns, and culture-independent diagnostics in SSA. These data are critical for understanding the cause of severe infection and have the potential to guide policy aimed at reducing child mortality from BSI.

Background: Critically ill patients have life-threatening conditions requiring immediate vital organ function intervention. But, critical illness in the emergency department (ED) has not been comprehensively described in resource-limited settings. Understanding the characteristics and dynamics of critical illness can help hospitals prepare for and ensure the continuum of care for critically ill patients. This study aimed to describe the pattern and outcomes of critically ill patients at the ED of the National Hospital in Tanzania from 2019 to 2021. Methodology: This hospital-records-based retrospective cohort study analyzed records of all patients who attended the ED of Muhimbili National Hospital between January 2019 and December 2021. Data extracted from the ED electronic database included clinical and demographic information, diagnoses, and outcome status at the ED. Critical illness in this study was defined as either a severe derangement of one or more vital signs measured at triage or the provision of critical care intervention. Data were analyzed using Stata 17 to examine critical illnesses’ burden, characteristics, first-listed diagnosis, and outcomes at the ED. Results: Among the 158,445 patients who visited the ED in the study period, 16,893 (10.7%) were critically ill. The burden of critical illness was 6,346 (10.3%) in 2019, 5,148 (10.9%) in 2020, and 5,400 (11.0%) in 2021. Respiratory (18.8%), cardiovascular (12.6%), infectious diseases (10.2%), and trauma (10.2%) were the leading causes of critical illness. Most (81.6%) of the critically ill patients presenting at the ED were admitted or transferred, of which 11% were admitted to the ICUs and 89% to general wards. Of the critically ill, 4.8% died at the ED. Conclusion: More than one in ten patients attending the Tanzanian National Hospital emergency department was critically ill. The number of critically ill patients did not increase during the pandemic. The majority were admitted to general hospital wards, and about one in twenty died at the ED. This study highlights the burden of critical illness faced by hospitals and the need to ensure the availability and quality of emergency and critical care throughout hospitals.

Hypertensive crisis is among the causes of morbidity and mortality in adult patients with hypertension in Sub-Saharan Africa. We aimed to determine the burden, risk factors and describe the management strategies of hypertensive crisis among adult patients seen at emergency departments of district and regional hospitals in Tanzania. This was a prospective multicenter longitudinal study which included all 162 district and regional hospitals in Tanzania. It was part of the Tanzania Emergency Care Capacity Survey (TECCS), a large assessment of burden of acute illness and emergency care capacity in Tanzania. Adult patients who presented to emergency departments with blood pressure ≥ 180/110mmHg were enrolled. Demographics, clinical presentation, management, and 24-hours outcomes were recorded using a structured case report form. Descriptive statistics were summarized in frequency and median, while logistic regression was used to evaluate the association between risk factors and presence of hypertensive crisis. We screened 2700 patients and enrolled 169 adults, henceforth proportion of adult patients with hypertensive crisis was 63 per 1000. Median age was 62 years (IQR 50–70 years) and predominantly females, 112 (66.3%). Majority 151(89.3%) were self-referred with two-wheel motorcycle being the commonest 46 (27.2%) mode of arrival to the hospital. Hypertensive emergency was found in over half 96 (56.8%) of the patients with hypertensive crisis, with oral medications administered in more than half of them, 71 (74%) as means to control the high blood pressure, and one-third 33 (34.4%) were discharged home. On multivariate analysis increasing age (AOR 4.53, p < 0.001), use of illicit drug (AOR 4.14, p-0.04) and pre-existing hypertension (AOR 8.1, p < 0.001) were independent risk factors for hypertensive crisis occurrence. Hypertensive crisis among adult patients attending district and regional hospitals is common (63 patients per every 1000 patients). Increasing age, use of illicit drug and pre-existing hypertension are independent associated factors for developing hypertensive crisis.

Background: Mortality among under-five children in Tanzania remains high. While early presentation for treatment increases likelihood of survival, delays to care are common and factors causing delay to presentation among critically ill children are unknown. In this study delay was defined as presentation to the emergency department of tertially hospital i.e. Muhimbili National Hospital, more than 48 h from the onset of the index illness. Methodology: This was a prospective cohort study of critically ill children aged 28 days to 14 years attending emergency department at Muhimbili National Hospital in Tanzania from September 2019 to January 2020. We documented demographics, time to ED presentation, ED interventions and 30-day outcome. The primary outcome was the association of delay with mortality and secondary outcomes were predictors of delay among critically ill paediatric patients. Logistic regression and relative risk were calculated to measure the strength of the predictor and the relationship between delay and mortality respectively. Results: We enrolled 440 (59.1%) critically ill children, their median age was 12 [IQR = 9–60] months and 63.9% were males. The median time to Emergency Department arrival was 3 days [IQR = 1–5] and more than half (56.6%) of critically ill children presented to Emergency Department in > 48 h whereby being an infant, self-referral and belonging to poor family were independent predictors of delay. Infants and those referred from other facilities had 2.4(95% CI 1.4–4.0) and 1.8(95% CI 1.1–2.8) times increased odds of presenting late to the Emergency Department respectively. The overall 30-day in-hospital mortality was 26.5% in which those who presented late were 1.3 more likely to die than those who presented early (RR = 1.3, CI: 0.9–1.9). Majority died > 24 h of Emergency Department arrival (P-value = 0.021). Conclusion: The risk of in-hospital mortality among children who presented to the ED later than 48 h after onset of illness was 1.3 times higher than for children who presented earlier than 48 h. It could be anywhere from 10% lower to 90% higher than the point estimate. However, the effect size was statistically not significant since the confidence interval included the null value Qualitative and time-motion studies are needed to evaluate the care pathway of critically ill pediatric patients to identify preventable delays in care.

Background: The survival of children who suffer cardiac arrest is poor. This study aimed to determine the predictors and outcome of cardiac arrest in paediatric patients presenting to an emergency department of a tertiary hospital in Tanzania. Methodology: This was a prospective cohort study of paediatric patients > 1 month to ≤ 14 years presenting to Emergency Medicine Department of Muhimbili National Hospital (EMD) in Tanzania from September 2019 to January 2020 and triaged as Emergency and Priority. We enrolled consecutive patients during study periods where patients’ demographic and clinical presentation, emergency interventions and outcome were recorded. Logistic regression analysis was performed to identify the predictors of cardiac arrest. Results: We enrolled 481 patients, 294 (61.1%) were males, and the median age was 2 years [IQR 1–5 years]. Among studied patients, 38 (7.9%) developed cardiac arrest in the EMD, of whom 84.2% were ≤ 5 years. Referred patients were over-represented among those who had an arrest (84.2%). The majority 33 (86.8%) of those who developed cardiac arrest died. Compromised circulation on primary survey (OR 5.9 (95% CI 2.1–16.6)), bradycardia for age on arrival (OR 20.0 (CI 1.6–249.3)), hyperkalemia (OR 8.2 (95% CI 1.4–47.7)), elevated lactate levels > 2 mmol/L (OR 5.2 (95% CI 1.4–19.7)), oxygen therapy requirement (OR 5.9 (95% CI 1.3–26.1)) and intubation within the EMD (OR 4.8 (95% CI 1.3–17.6)) were independent predictors of cardiac arrest. Conclusion: Thirty-eight children developed cardiac arrest in the EMD, with a very high mortality. Those who arrested were more likely to present with signs of hypoxia, shock and acidosis, which suggest they were at later stage in their illness. Outcomes can be improved by strengthening the pre-referral care and providing timely critical management to prevent cardiac arrest.

Background: Due to the high prevalence of human immunodeficiency virus (HIV) in Tanzania, provider-initiated HIV testing for patients attending any health care setting is recommended. However, follow-up and linkage to care by those tested remain poor. We determined the feasibility and efficacy of text messaging to promote follow-up among otherwise healthy trauma patients who underwent provider-initiated HIV testing and counseling at an emergency department (ED) in Tanzania. Material and methods: This randomized controlled trial (RCT) was conducted at Muhimbili National Hospital (MNH) ED between September 2019 and February 2020. Adult trauma patients consenting to HIV testing and follow-up text messaging were randomized to standard care (pre-test and post-test counseling) or standard care plus a series of three short message service (SMS) text message reminders for follow-up in an HIV clinic, if positive, or for retesting, if negative. Investigators blinded to the study assignment called participants 2 months after the ED visit if HIV-positive or 4 months if HIV-negative. We compared the proportion of people in the intervention and control groups completing recommended follow-up. Secondary outcomes were the proportion of patients agreeing to testing, proportion of patients agreeing to receiving text messages, and the proportion of HIV-positive and HIV-negative patients in each study arm who followed up. Results: Of the 290 patients approached, 255 (87.9%) opted-in for testing and agreed to receive a text message. The median age of the study population was 29 [IQR 24–40] years. There were 127 patients randomized to the intervention group and 128 to the control group. The automated SMS system verified that 381 text messages in total were successfully sent. We traced 242 (94.9%) participants: 124 (51.2%) in the intervention group and 18 (488%) in the control group. A total of 100 (39.2%) subjects reported completing a follow-up visit, of which 77 (60.6%) were from the intervention group and 23 (17.9%) were from the control group (RR = 3.4, 95% CI 2.3–5.0). This resulted in a number needed to treat (NNT) of 2.3. Of the 246 HIV-negative participants, 37% underwent repeat screening: 59% of those in the intervention group and 16% in the control group (RR = 3.7, P = < 0.0001, NNT 2.3). Among the nine positive patients, all five in the intervention group and only three in the controls had follow-up visits. Conclusion: Automated text message is a feasible and effective way to increase follow-up in HIV-tested individuals in a limited income country.

Background: Abdominal pain in adults represents a wide range of illnesses, often warranting immediate intervention. This study is to fill the gap in the knowledge about incidence, presentation, causes and mortality from abdominal pain in an established emergency department of a tertiary hospital in Tanzania. Methods: This was a prospective cohort study of adult (age ≥ 18 years) patients presenting to the Emergency Medicine Department of Muhimbili National Hospital (EMD-MNH) in Dar Es Salaam, Tanzania with non-traumatic abdominal pain from September 2017 to October 2017. A case report form was used to record data on demographics, clinical presentation, management, diagnosis, outcomes and patient follow-up. The primary outcome of mortality was summarized using descriptive statistics; secondary outcome was, risks for mortality. Results: Among 3381 adult patients present during the study period, 288 (8.5%) presented with abdominal pain, and of these 199 (69%) patients were enrolled in our study. Median age was 47 years (IQR 35–60 years), 126 (63%) were female, and 118 (59%) were referred from another hospital. Most common final diagnoses were malignancies 71 (36%), intestinal obstruction 11 (6%) and peptic ulcer disease 9 (5%). Most common EMD interventions given were intravenous fluids 57 (21%), analgesia 49 (25%) and antibiotics 40 (20%). 160 (80%) were admitted of which 15 (8%) underwent surgery directly from EMD. 24-h and 7-day mortality were 4 (2%) and 7 (4%) respectively, while overall in hospital-mortality was 16 (8%). Among the risk factors for mortality were male sex Relative Risk (RR) 2.88 (p = 0.03), hypoglycemia (RR) 5.7 (p = 0.004), ICU admission (RR) 14 (p < 0.0001), receipt of IV fluids (RR) 3.2 (p = 0.0151) and need for surgery (RR) 6.6 (p = 0.0001). Conclusion: Abdominal pain was associated with significant morbidity and mortality as evidenced by a very high admission rate, need for surgical intervention and a high in-hospital mortality rate. Future studies and quality improvement efforts should focus on identifying why such differences exist and how to reduce the mortality.

Background: In Tanzania, there is no national trauma registry. The World Health Organization (WHO) has developed a data set for injury that specifies the variables necessary for documenting the burden of injury and patient-related clinical processes. As a first step in developing and implementing a national Trauma Registry, we determined how well hospitals currently capture the variables that are specified in the WHO injury set. Methods: This was a prospective, observational cross-sectional study of all trauma patients conducted in the Emergency Units of five regional referral hospitals in Tanzania from February 2018 to July 2018. Research assistants observed the provision of clinical care in the EU for all patients, and documented performed assessment, clinical interventions and final disposition. Research assistants used a purposefully designed case report form to audit the injury variable capture rate, and to review Ministry of Health (MoH) issued facility Register book recording the documentation of variables. We present descriptive statistics for hospital characteristics, patient volume, facility infrastructure, and capture rate of trauma variables. Results: During the study period, 2891 (9.3%) patients presented with trauma-related complaints, 70.7% were male. Overall, the capture rate of all variables was 33.6%. Documentation was most complete for demographics 71.6%, while initial clinical condition, and details of injury were documented in 20.5 and 20.8% respectively. There was no documentation for the care prior to Emergency Unit arrival in all hospitals. 1430 (49.5%) of all trauma-related visits seen were documented in the facility Health Management Information System register submitted to the MoH. Among the cases reported in the register book, the date of EU care was correctly documented in 77% cases, age 43.6%, diagnosis 66.7%, and outcome in 38.9% cases. Among the observed procedures, initial clinical condition (28.7%), interventions at Emergency Unit (52.1%), investigations (49.0%), and disposition (62.9%) were documented in the clinical charts. Conclusions: In the regional hospitals of Tanzania, there is inadequate documentation of the minimum trauma variables specified in the WHO injury data set. Reasons for this are unclear, but will need to be addressed in order to improve documentation to inform a national injury registry.

Background: Elevated serum lactate levels have been shown in numerous studies to be associated with serious adverse events, including mortality. Point of care lactate level is increasingly available in resource-limited emergency department (ED) settings. However, little is known about the predictive ability of for serious adverse events. Aim of the study: We aimed to describe the utility of serum lactate level as a predictor of 24-h serious adverse events among adult patients presenting at the Emergency Medicine Department (EMD) of Muhimbili National Hospital (MNH) and MUHAS Academic Medical Center (MAMC). Methods: This was a prospective observational study involving adult patients presenting to the EMD-MNH and MAMC from 1 September 2018 and 31 October 2018. Eligible patients with at least one lactate level test drawn while in the ED were examined in terms of their demographics, relevant clinical characteristics, and any serious adverse event (SAE) within 24 h of arrival. The sensitivity and specificity of lactate level to predict outcomes of interest were determined using the best cut-off point constructed from AUROC to see how well lactate level could discriminate which patients would have adverse events in the next 24 h. Categorical and continuous variables were compared with the chi-square test and two-sample t test, respectively. Results: We screened 2057 (20.9%) out of 9828 patients who presented during study period, and enrolled 387 (18.8%). The overall median age was 54 years (interquartile range 40–68 years) and 206 (53.2%) were male. Using local triaging system, a total of 322 (83.2%) was triaged as an emergency category. The mean lactate level was 3.2 ± 3.6 mmol/L, 65 (16.8%) patients developed at least one SAE, with 42 (11%) who required ICU/HDU, 37 (10%) needed ventilator support, 10 (3%) required inotropes, and 9 (2%) developed cardiac arrest. The overall 24-h mortality was 28 (7%). The AUC of serum lactate level for overall 24-h mortality was 0.801 (95%CI, 0.7–0.9, P ≤ 0.001). At the optimal cutoff value (3.8 mmol/L), lactate level had a sensitivity and specificity for 24-h mortality of 64% and 85%, respectively. Mortality of the high-lactate level group (³3.8 mmol/L) was significantly higher than that of the low-lactate level group (< 3.8 mmol/L), 23.8% vs. 2.9%, respectively (95%CI 3.8–17.2, p < 0.001), with the relative risk of mortality in the high-lactate level group being 8.1 times higher compared to the low-lactate level group. Conclusion: The utility of lactate level in predicting mortality was similar to that seen in high-resource settings. A serum lactate level of ³3.8 mmol/L predicted 24-h serious adverse events in unselected patients seen in the high-acuity area of our ED. Incorporating serum lactate level in ED in lower- and middle-income countries (LMICs) can help identify patients at risk of developing serious adverse events.