Admission pattern, Clinical outcomes and associated factors among patients admitted in medical intensive care unit at University of Gondar Comprehensive and specialized hospital, Northwest Ethiopia, 2019. A retrospective cross-sectional study.

Background: Intensive care unit (ICU) is a multidisciplinary staffed and specially equipped area of a hospital dedicated to providea care for patient with life-threatening illness. Provision of intensive care services to critically ill patients is a global enterprise. The care is advancing but in resource-limited settings, it is lagging far behind and ICU mortality is still higher due to various reasons. Objective: We aimed to determine the admission patterns, clinical outcomes and associated factors among patients admitted medical intensive care unit (MICU). Results: A total of 738 patients were admitted to MICU during September 2015- April 2019. Two hundred thirty four patients had incomplete data on the registries and their charts could not be located. So that, 504 (68%) of all ICU admissions had complete data and were analyzed. Out of the 504 patients, 268 (53.2%) patients were females. Cardiovascular disease 182(36.1%) was the commonest categorical admission diagnosis. The overall mortality rate of the MICU was 38.7 %. In the multivariate analysis, mortality was associated with need for mechanical ventilation (AOR=5.87, 95% CI: 3.24 -10.65) and abnormal mental status at admission (AOR = 2.8.8, 95% CI: 1.83-4.29). Patients who stayed less than four days in MICU are 5 times more likely to dies than who stayed(AOR= 5.58, 95% CI: 3.58- 8.69). Therefore, we recommend improving the acute critical care through the expansion of the care, supply emergency equipment’s and medications and implementation of admission criteria protocols and other local guidelines.

surgical procedures (2,3) Mortality in ICU is a global burden which results in a huge loss of productivity and financial costs. It varies across the world depend on ICU infrastructure, staff availability, and training, pattern, and cause of ICU admission. In developed contents like North America, Oceania, Asia and Europe in ICU mortality relatively low with the rate of 9.3%, 10.3, 13.7% and 18.7% respectively, while in the rest of the world such as South America, and the Middle East the mortality found to be 21.7% and 26.2% (4,5).
In Ethiopia,the first ICU was introduced more than 30 years ago in Addis Ababa Tikur Anbessa Hospital since then the number of intensive care service is increasing in both public and private hospitals of the country (6).Even though critical care services have been undergoing significant advancements for improvements due to technological progress, new scientific development in treatment outcome of the critically ill patientsglobally, the progress of critical care service provision in resource-limited settings is lagging far behind and the mortality is still higher (4,(7)(8)(9).
In Africa, the ICU mortality rate is high as it compared to the other developed continents. The mortality rate in Nigeria, Uganda, Tanzania, and Kenya were 32.9%, 40.1%, 41.1%, and 53.6% respectively (10)(11)(12)(13). Different studies showed that the mortality rate of ICU patients in Ethiopia is relatively similar to other Africa countries. Previous studies done in Jimma, Addis Ababa, and Mekelle showed that the mortality rate was 50.4%, 32% and 27% respectively (14)(15)(16).
In developing countries, the burden of critical illness is high. Severe infections such as pneumonia, malaria are endemic, road traffic accident, obstetric complications, and surgical emergencies are common. In addition, underlying malnutrition and HIV infection will make patients to have a poor physiologic function during admission thereby worsen the outcome of critically ill patients. Moreover, the burden is certain to extend with Growing urbanization and rising epidemics, (5,12,17).
Provision of ICU care is very challenging in developing countries like Ethiopia. The scarcity of drugs and medical equipment's, lack of well-trained staffs and poor infrastructures are the main challenges to provide optimal care to critically ill patients (3,18).
Apart from medical diagnosis of patients, most hospital in low income countries lack formal triage system and emergency department which forces critically ill patients to be seen in either in the ward or outpatient clinics on the bases of "first come first served" which will increases the burden of ICU and compromises patients outcome since triage system decrease waiting time and mortality through identification of critically ill patients with early warning score and care Delivery (1,19). There is scaricity of information regarding criticall ill patients admitted in the ICU in the study area. Therefore, the aim of the current study was to describe the admission patterns, clinical outcomes and associated factors among patients admitted in the medical intensive care unit of UOGCSH.

Methods
Study design, area and period University of Gondar Comprehensive Specialized Hospital is a specialized teaching hospital located in Amhara National Regional state, North Gondar zone, which is about 738km away from Addis Ababa in Northwest of Ethiopia. It is one of the largest specialized hospitals providng service to about 5 million people. The critical care service of University of Gondar Comprehensive Specialized Hospital ( UoGCSH) was started in 2011 as a four-bed MICU capacity with two mechanical ventilators one defibrillator, four non-invasive hemodynamic monitoring devices, one ultrasound machine, and one ABG-analyzer machine. The MICU is an open ICU system run under the department of internal medicine.
Currently, the MICU has a team composed of one specialist in pulmonary and critical care medicine (PCCM), internal medicine residents, senior anesthetists, 2 critical care and 8 clinical nurses, anesthetists, physiotherapists, and other health professionals according to the demands of the individual cases. The MICU provides critical care service not only for internal medicine cases but also surgical and obstetrics cases.
A retrospective cross sectional study was conducted based on the ICU registration notes and charts.Patients who where admitted to the MICU from Septemper 2015 to April, 2019 and whose ICU registry notes and charts were available were included in the study. Missed Patients chart and incomplete data on the charts were the exclusion criteria ( Figure 1). D a t a p r o c e s s i n g a n d a n a l y s i s p r o c e d u r e s Data was entered, coded and cleaned using the Epi-data software and analyzed using SPSS version 20. Categorical variables were reported as frequencies and percentages whereas; median and interquartile range (IQR) were used for continous variables.

Study variables
Descriptive statistics was carried out and the results were presented using text, tables and graph. Model ftness was checked using a Hosmer-Lemeshow goodness-of-ftness test.
Crude odds ratios with their 95% confdence intervals were estimated in the bivariable logistic regression analysis to assess the association between each independent variable and outcome variable. In the bivariable logistic regression, variables with P-value < 0.2 were ftted into the multivariable logistic regression analysis. Adjusted odds ratios with their 95% confdence intervals were estimated to assess the strength of association, and variables with P-value < 0.05 were considered statistically signifcant factors.

Results
Demographic and Admission Characteristics of among patients admitted to MICU A total of 738 patients were admitted from September 2015 to April 2019 .Two hundred thirty-four patients had incomplete data on the registries and their charts could not be located. So that, 504 (68%) patients were included in this study.
Patient age ranged from 15 to 98, median (IQR) was 32 (28 -40) years. Of the total admitted patients 268 (53.2%) were female with female to male ratio of (F: M) 1.1: 1 and most of the patients 264(52.4%) were urban residents.
Similarly, these disorders are the commonest categorical admission diagnosis in the specific year of the past five years ( Figure 2). All respiratory infections were under respiratory disease.
Infective heart disease and rheumatic heart disease were under cardiovascular disease Guillen Barrie Syndrome was under the category of Infectious disease The most common specific diagnosis at ICU admission was all types of myocardial infarction 96 (19 %) followed by Heart failure 56 (11.1%), ARDS 46(8.9%), septic shock 37 (7.3%) and HIV infection 25 (5%) of the all admissions (Table 3).   Patients who stayed in MICU for less than 4 days were 5 times more likely to die than who stayed four and above days. As illustrated in the following figure most of the patients died in the first 24 hours of admission (figure 5).
Bivariate logistic analysis shown that gender, diagnosis at admission, need for mechanical ventilation, duration of mechanical ventilation, length of ICU stay, mean arterial pressure at admission, and mental status at admission were significantly associated with the clinical out of MICU patients. However, in a multivariate analysis, the associated risk factors ofdeath were the need for mechanical ventilation, abnormal mental status at admission and length of ICU stay (Table 6).  (18,20,21).
However, the overall ICU mortality of the current study was higher than studies done in Mekelle, Ethiopia (27%) and Scandinavian countries (9.1%). (16,23). This discrepancy might be due to lack of necessary medical equipments (ABG-analyzer machine, portable dialysis machine and portable x-ray service), infrastructure, and training. In addition, the lack of high dependency unit in the study area might be one of the contributing factors for the higher rate of ICU mortality (18,22).
In our study, the mortality rate was lower than the previos studies done in other parts of Ethiopia Hosanna (46%), Jimma (50.4%), and Kenya (53.6%). This could be due to the difference in diagnosis of admission. Poly trauma and traumatic brain injury were the main diagnosis of admission in the previous studies.
In the current study, the median length of ICU stay found to be 4 days which resembles to the other African countries (13,22). In our study, Out of 504 patients, 118(23%) where stayed in ICU 24 hours or less. Out of whom 89 (75%) of patients died in the mentioned period of time which accounted for 45% of overall ICU mortality. In addition, patients who stayed in ICU for less than four days were 5 times more likely to die than patients who stayed four or more days with (AOR=5.58, p<0.001) which is comparable with the study done in Uganda (17). However, our result was different fom the study conducted in Hosanna, the length of ICU stay was more than 14 days,which was strongly associated with ICU mortality (OR=4.113,P <0.039) (24). The discrepancy on the median length of ICU stay might be explained due to late arrival to ICU and delay in intervention, shortage of crucial emergency drugs including inotropes, anti-arrhythmic drugs, and antibiotics, absence of airway management and cardiopulmonary resuscitation equipment's in the medical emergency ward. furthermore, vital signs and overall clinical conditions of the patients on admission to the MICU was found to very be poor, which may shown a gap in the continuity of care from emergency medicine department, medical and surgical wards to MICU. Early death might also be explained by a limited number of ICU bed since the World Federation of Societies of Intensive and Critical Care Medicine recommends the ICU need to have at least 5% of total hospital beds. Furthermore, shortageof functional mechanical ventilator which delays and denied admission of critically ill patients to MICU (18,25).
This study revealed that need for mechanical ventilation is independent risk factor of ICU mortality with (AOR: 5.578, P<0.001) which was similar with the studies done in Kenya and B r a z i l w e r e mechanical ventilator was independent risk factors for death (AOR= 10.7, p<0.001) and (AOR: 6.37, p<0.001) respectively (26,27). The possible explanation for this association could be related that mechanical ventilatoris initiated for the patients with respiratory failure, unable to protect the airway and hemodynamic instability.
Furthermore, patients who need intubation and mechanical ventilator more vulnerable for ventilator associated pneumonia and other nosocomial infection which further compromises the clinical outcome of critically ill patients (28,29). In addition, these patients who were on mechanical ventilator were presented with unstable vital sign and comorbid conditions which might be increased the odds of in ICU mortality.
Our study reveled that patients who presented with abnormal mental status were more likely to die in MICU than conscious patients (AOR: 2.741, P<0.001). This result may linked w i t h the severity of disease condition during admission. Disturbance level of consciousness is related with severe decompensated disease, cerebral hypo-perfusion due to sepsis, blood loss, poisoning, and neurological disorder. Beside to this patients with abnormal mental status might not protect their airways with further increases the probability of respiratory and infectious disease (30).
In the current study, majority of patients were adult with an average age group of 20-40 years and median age of 32 years. Since, most of young the patients were diagnosed with infectious disease and respiratory disorder which might increase the probability of MICU admission.Our result is in accordance with the studies done in other sub-Saharan counties as the median age of patients in Kenya (29 years) and Tanzania (34 years), years of age respectively (13,22). Contrarily, the patient's population in this study was younger than the developed nation as the mean age of ICU admission in Scandinavian countries 59.3 years, and Korea 64.7 years. The difference could be in the above-mentioned studies most of patients were admitted secondary to age-related degenerative and comorbid disease (3,20,31).
Despite the fact that the young patients took the majority proportion of admission the number of deaths increased with increasing age similar to the Addis Ababa and Mekelle (15,16). This might be the old patients were presented with Less organ physiologic preserve, unstable vital sign and more comorbid score.
In the present study, females were more likely to admit and diein MICU than males. Which was in contrast to the otherAfrican countrieswere males predominate the admission and mortality proportion of the ICU of Kenya, Uganda, Tanzania, and Nigeria (3,18,21,27).
The gender difference could be explained by the majority of females in the current study were admitted with comorbid illness and are aged with 26% of them were above the age of 60 years.
A disease of cardiovascular system accounted for 36% of all ICU admission followed by respiratory (17.9%). This is similar to the previous study done in Jimma and Mekelle (16,31). However, it is different from a study conducted in Uganda were infectious illness p r e d o m i n a n t (18)  the conception and design of the study, acquired, analyzed and interepted the data drafted and revised the manuscript. GF, NM, DY and NR partcipate in reviewing the design and methods of data collection, interpretation and preparation of the manuscript. All authors partcipate in preparation and critical review of the manuscripts. Inaddition, all authors read and approved the manuscript.
following admissions in the intensive care unit of a specialized hospital, in Ethiopia.     Length of ICU stay between survivor and non-survivor among patients admitted to This is a list of supplementary files associated with the primary manuscript. Click to download.