Survival and predictors of adult patients admitted to intensive care unit in Yekatit 12 Hospitals, Addis Ababa, Ethiopia.

Ethical approval

An official letter was obtained from Africa Medical College department of Public Health in order to obtain ethical clearance from Addis Ababa Health Bureau and to get permission and cooperation from Addis Ababa Health Bureau and Yekatit 12 hospital. In order to conduct research in Yekatit 12 hospital, the hospital administrator requested the authors obtain ethical clearance from Addis Ababa Health Bureau. After the review of the proposal of this research, ethical clearance was obtained from Addis Ababa Health Bureau research review and ethics committee (reference number: AAHB/04/345/14). To ensure confidentiality, name and any other personal identities were not included during data collection in the data extraction sheet/data collection tools, any document in the raw data, and in writing the research findings. A consent form was signed by the registrar office of Yekatit 12 Hospital to use patient medical records as necessary. Before data extraction from the patient registration sheet, oral informed consent was obtained from all patients/healthy participants (or their legal guardians for minors and concerning deceased participants, consent were obtained from next of kin of the participants via phone communication). This is because data were extracted from the patient medical chart after patients were discharged. This made it difficult to access written consent from each patient. Oral informed consent was accepted by Addis Ababa Health Bureau research review and ethics committee.

Study design, settings, and participants

A retrospective cohort study of survival analysis was conducted by reviewing all eligible medical charts of ICU patients. There were 45 number of beds in the ICU units. A total of 438 patients admitted at Yekatit 12 Hospital Medical College ICU from March 15th, 2020 to March 15th, 2022 participated in this study. The follow-up with patients were done starting from admission up to discharge.

Sample size and sampling procedure

The sample size was determined by a double population proportion formula using Epi-Info version 7.0 statistical software with the following assumptions: Confidence Interval 95%, study Power=80% and one-to-to-one ratio based on a study conducted at Nigist Eleni Mohammed Memorial Hospital, Hosanna, Southern Ethiopia.¹¹

Data collection procedure

Data was collected by well-trained nurse data collectors with a Bachelor of Science using a data extraction format. The data extraction format was developed by the authors by referring to the Global Health Care of the Critically Ill in Low-Resource Settings (recommended research instrument for low-resource setting) and edited as a local context to fit the research objective.^3,7,17,20 The data extraction format was pretested on 5% of the samples. The inclusion criteria were all patients 18 years and above (because the clinical spectrum of disease for children is significantly different from adult illnesses) admitted to the ICU during March 15^th, 2020 to March 15^th, 2022. There were 1234 adult patients were admitted in total to the ICU in the same period of time. We excluded 5 adult patient cards with incomplete medical record information. The study participants were selected using systematic random sampling techniques. For each patient, potential risk factors such as presence of comorbidities (diabetic mellitus, cardiovascular disorder, cancer, dyslipidemia, and neurologic disorder) and reason for admission (Sepsis, shock, trauma, brain injury) were extracted from the charts. The authors analyzed the first admission in a given year for patients who may have multiple ICU admissions within a year. The exit time was set as the earliest of either the death date or the ICU discharge date. The data extraction form Ref. 23 had three sections that cover socio-demographic characteristics, admission and discharge information, and survival of adult patients admitted in the ICU.

Definition of outcomes, exposures, and predictors variables

Data analysis

The data was analyzed using SPSS version 23. The main outcome of this study was the survival status of patients admitted to ICU. The primary outcome measure was mortality during the ICU stay. To minimize information bias, the checklist was cross checked with the registers and cards. Baseline characteristics of patients at the time of ICU admission was reported as number and percentage for categorical variables and as mean and standard deviation (SD) for normally distributed continuous variables, or as the median and inter-quartile range (IQR) if not. A Kaplan-Meier curve was used to describe the probability of survival in ICU stay, thus the exit time was set as the earliest of either the death date or the ICU discharge date. Kaplan-Meier (KM) method was used to estimate overall survival (OS), while the Cox proportional hazard model was used for multivariate analysis to determine the possible associations of predictor variables, to control possible confounders and to obtain an adjusted hazard ratio at a p-value less than 0.05.

Socio-demographic characteristics

A total of 438 adult ICU patients were included at all stage completing follow-up and analyzed in the study. The underlying data are publicly available.²² The mean age of the study participants was 44.34 (SD±17.7). More than half (237, (54.1%) of the study participants of the study were female. Most of the study participants (233, (53%) attended higher education. Married participants of this study constitute more than half (261, (61%) of the study population. The employed and unemployed participants of the study were 251 and 90, respectively. The mean time of follow-up was 4.15 days with minimum 1 and maximum 10 days respectively (these days were the patients admitted and follow until discharged).

Clinical characteristics

The main reasons for admission were sepsis (35, 8%), shock (63, 14.4%), trauma (23, 5.3%), brain injury (40, 9.1%), cardiovascular disorder (86, 19.6%), respiratory failure (79, 18%) and other cases (112, 25.6%). Some participants of the study were admitted with co-morbidity. Among the co-morbidities, the number of participants with a respiratory disorder was (89, 20.3%). A total of 128 participants (29.3%) suffered cardiovascular disorders. The rest of the co-morbidities included dyslipidaemia, cancer, neurological diseases, and diabetes mellitus (Table 1).

Survival time of ICU patients

The total 438 ICU patients were followed for a total of 1816 person days (PD) of observation. The mean time of survival was 4.15 days with a minimum of one and maximum of 10 days.

Survival and mortality rate of ICU patients

The cumulative survival and mortality were 69.7% and 30.3%, respectively. The survival rate of ICU patients was 168 persons per 1000 person days (survival=168/1000PD, 95% CI: 0.15, 0.89). While the incidence of mortality or mortality rate of patients in the same hospital was 73 per 1000 person days (mortality= 73/1000PD, 95% CI: 0.073, 0.086) (Figure 1).

Figure 1. KM Survival curve of adult ICU patients of Yekati 12 hospital, Ethiopia 2022.

Predictors for survival of ICU patients

The finding of a multivariate analysis results confirmed that the risk of mortality among ICU patients attending primary education was twice that of patients attending higher education with an adjusted hazard ratio (AHR) (AHR=2.097, 95% CI:1.081,4.067). Patients who attended secondary education were more than one time at risk of suffering from mortality (AHR=1.79, 95% CI: 1.11, 2.88). Patients admitted to ICU because of shock were more than four times at risk of death compared to other causes (AHR=4.51, 95% CI: 2.41, 8.45). The risk of mortality among ICU patients admitted because of brain injury was more than two times compared to other patients (AHR=2.77, 95% CI: 1.18, 6.47). Patients with respiratory failure were more than two times at risk of mortality (AHR=2.42, 95% CI: 1.19, 4.87) (Table 2).