Aetiologies and outcomes of patients with abdominal pain presenting to an emergency department of a tertiary hospital in Tanzania: a prospective cohort study.

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.

Performance characteristics of a local triage tool and internationally validated tools among under-fives presenting to an urban emergency department in Tanzania.

BACKGROUND: A number of region-specific validated triage systems exist; however very little is known about their performance in resource limited settings. We compare the local triage tool and internationally validated tools among under-fives presenting to an urban emergency department in Tanzania. METHODOLOGY: Prospective descriptive study of consecutive under-fives seen at Muhimbili National Hospital (MNH), ED between November 2017 to April 2018. Patients were triaged according to Local Triage System (LTS), and the information collected were used to assign acuities in the other triage scales: Canadian Triage and Acuity Scale (CTAS), Australasian Triage Scale (ATS), Manchester Triage Scale (MTS) and South African Triage Scale (SATS). Patients were then followed up to determine disposition and 24 h outcome. Sensitivity, specificity, positive and negative predictive values for admission and mortality were then calculated. RESULTS: A total of 384 paediatric patients were enrolled, their median age was 17 months (IQR 7-36 months). Using LTS, 67(17.4%) patients were triaged in level one, 291(75.8%) level 2 and 26 (6.8%) in level 3 categories. Overall admission rate was 59.6% and at 24 h there were five deaths (1.3%). Using Level 1 in LTS, and Levels 1 and 2 in other systems, sensitivity and specificity for admission for all triage scales ranged between 27.1-28.4% and 95.4-98% respectively, (PPV 90.3-95.3%, NPV 47.1-47.4%). Sensitivity for mortality was 80% for LTS, and 100% for the other scales, while specificity was low, yielding a PPV for all scales between 6.9 and 8%. CONCLUSION: All triage scales showed poor ability to predict need for admission, however all triage scales except LTS predicted mortality. The test characteristics for the other scales were similar. Future studies should focus on determining the reliability and validity of each of these triage tools in our setting.