Management Outcome, Mortality and Factors Associated with Mortality Amongst Hospitalized Patients with Stroke. A Cross Sectional Study.

Background Outcomes, mortality due to stroke and factors related with mortality are not profoundly known in the northern Ethiopia. Hence, the main purpose of the study was to assess management outcome, mortality and factors associated with mortality amongst stroke patients. Methods A cross sectional study design was used to conduct the study. Patients medical chart were utilized to collect patient information. Patients medical registries who were diagnosed with stroke and treated in Ayder Comprehensive Specialized Hospital over the years were retrospectively seen. Sample size was estimated using single population formula. Medical charts have been assigned a number and patients were selected using systematic random sampling technique. The findings were analyzed using SPSS version 22. Using logistic regression analysis factors associated with mortality in hospital were identified. A P value less than 0.05 were deemed to be significant in all types of analyses. Results About 216 patients with the diagnosis of stroke were studied. Of the total patients studied, 126(58.3%) were females. From the total stroke patients analyzed, 102(47.2%) were discharged with improvement, 44(20.4%) left against medical advice, and 22(10.2%) patients discharged with no change. Moreover, about 48(22.2%) patients died in the hospital. The median time to death was 3 days ranging from 1 to 48 days and the average mean time to death was 6.2±8.7 days. One hundred thirty-six (62.9%) patients had affirmed to have at least one complication during their hospital stay. The total frequency of complication detected was 234 in number and aspiration pneumonia (n=92, 39.4%) was the frequently identified complication. Sever (3-8) record of Glasgow coma scale (GCS) at admission (Adjusted Odds Ratio=15.33, 95%CI:3.77-62.40, p=0.001) and being unconscious (AOR=2.61, 95%CI:1.06-6.40, p=0.037) during admission were positively associated with mortality.

Conclusion Substantially higher number of patients died at the hospital. Majority of the patients admitted to the hospital developed complication, aspiration pneumonia being the most frequently affirmed complication during their stay in the wards. Patients with severe GCS and who were unconscious during admission were more likely to die at the hospital.
Sketching out strategies to solve the problem will be conducive for the patients with stroke.

Background
Stroke constitutes and attributes for a major global health problem, and is the second most common cause of death and among the leading causes of disability in the world (1)(2)(3)(4). The prevalence's and incidences of stroke case in 2016 amounted to nearly 83 million and 13.6 million, respectively (3). About 15 million people suffer with stroke each year and nearly six million people worldwide die from stroke every year. Globally, it is speculated that one in six people will have a stroke in their lifetime (5). Studies has estimated a disability adjusted life years from a stroke to be over 87% in low and middle income countries (6). Moreover, stroke is attributable for more than 4% of direct healthcare costs in high-income countries (2). In Ethiopia, stroke was also attributed to premature deaths from non-communicable diseases (7).
In Africa, there is a paucity of comprehensive surveillance statistical data of stroke prevalence. Albeit, the extant reports showed that mortality and prevalence of disabling stroke in Africa are similar to or higher than the developed countries (8). In 2009, the estimate of new stroke case amongst people 15 years old and older was 483 thousand, which is equivalent to 81.2 (13.2-94.9)/100,000 person years. The incidence of new case scaled up by 10.8% which made the estimated new stroke case to be 535 thousand on 2013 (9). Moreover, some community-based research's illustrated that an agestandardised annual stroke incidence rate and prevalence accounted up to 316 per 100,000 population, 981 per 100,000, respectively (10). Despite, its higher prevalence and incidence, the outcome of stroke patients in Africa was troublesome. From a study done in south Africa, in the years between 2007 and 2011, the prevalence of crude stroke mortality was 114 per 100,000 person-years (11). In addition, from an estimate 75,000 per year stroke case prevalence's, 25 000 of the patients died within the first month. In that study, stroke attributed for 564 000 Disabled Adjusted Life Years(DALY) (12).
The prevalence of type of stroke in Ethiopia varies between studies. In many studies ratio of ischemic stroke outweighs haemorrhagic stroke (13)(14)(15)(16). In Ethiopia fatality rate of stroke patients was found to be 31.1% in one study. Besides, 45.4% of patients diagnosed with stroke discharged with neurological deficits (14). In Northwest Ethiopia, Gonder, 13% of the stroke patients died at the stroke unit with median of hospital stay being 6 days (9). Nevertheless, in the study area the trend of stroke mortality is not comprehensively and certainly known. Therefore, the main purpose of this study was to determine management outcome, mortality and factors associated with mortality amongst hospitalized patients with stroke.

Methods
The study was conducted in Ayder Comprehensive Specialized Hospital (ACSH) which is found in Mekelle city, Tigray region, Ethiopia. ACSH provides a broad range of medical services to both inpatients and outpatients of all age groups. A hospital based crosssectional study was conducted to run the study. Patient who were admitted to the ACSH since September 2013 to September 2018 were deployed for study selection. Patients who were diagnosed with all types of stroke, age of 18-years and older were included in the study. Patients medical registry and chart was the main source of data. Patients with incomplete medical registry and chart were excluded from the study.
Sample size was calculated with assumptions of confidence level at 95% CI, degree of precision at 0.05 and z=1.96. Taking prevalence of 15.1% mortality in to consideration and adding contingency of 10%, the calculated sample size was 216. patients who were admitted to medical and intensive care unit of the hospital were identified from the hospital management information system patient registration book. Moreover, patients chart stroke cases were identified through review of discharge logbooks from medical and intensive care unit wards. Each patient's medical chart was assigned a number and patients were selected and included in the study using systematic random sampling technique.
Data collection tool which enables data collector to extract the data from patient's chart was used to collect patients' medical and sociodemographic information. The data collection tool was developed by reviewing reputable literatures and adopting patients medical record format in the hospital. The data collection tool was written in English. The main contents of the data collection tool entailed sociodemographic characteristics, medications given for patients, type of diagnosis, type of stroke, comorbidities, complications recorded, treatment out come and Glasgow coma scale score. Data were collected by three pharmacy technicians who were employed out of ACSH. All the information's were retrospectively collected from patient's chart.
Data collectors have been trained prior to data collection process and procedure. In addition, to that the data collection tool was pretested in 10 stroke patients which were excluded from analysis. Based on the pretests findings, the data collection tool was amended. During data collection the principal investigator supervised closely the data. At the end of each data collection day the principal investigator was also checking the completeness and quality of the data.
Ethical clearance was obtained from the office of Health Research Ethics Review Committee of college of health sciences, Mekelle University. The study was conducted after getting official permission from the ACSH. To ensure confidentiality, the names of patients and address were not recorded.

Statistical Analysis
Before data entry, data was cleaned and coded. After checking completeness and categorization of the data, it was entered in to SPSS version 21 software for analysis.
Logistic regression was used to identify factors associated with mortality. Predictors of mortality were identified using multivariable logistic regression analysis. Level of significance was declared at p value less than 0.05 for all types of analysis.

Operational definitions
Stroke: defined as "rapidly developing clinical signs of focal (or global) disturbance of cerebral function lasting longer than 24 hours, unless interrupted by death, with no apparent cause other than that of vascular origin". By applying this definition transient ischemic attack, which is defined to last less than 24 hours, and patients with stroke symptoms caused by subdural hemorrhage, tumors, poisoning, or trauma, are excluded (17).

Sociodemographic and clinical characteristics
A total of 216 patients diagnosed with all types of stroke were recruited and analyzed.
Amongst the patients studied, 126(58.3%) were females in gender. The mean age of the patients was 61.2±15. 6 year ranging from minimum of 20 to 95 years. Concerning residence of the patients, when 115(53.2%) were from rural, the rest 101(46.8%) patients were urban residents. The median length of hospital stay of the patients were 8 days, whereas the median time from sign and symptom onset to hospital admission was 24 hours. Ischemic stroke was the most commonly diagnosed type of stroke, which was affirmed in (n=111), 51.4% of the patients followed by hemorrhagic stroke (n=96, 44.4%) ( Table 1).
Regarding to the recorded complications, about 136 patients had diagnosed with at least one complication during their hospital stay. A total of 234 complication records were identified in the total 216 stroke patients studied. Aspiration pneumonia (n=92, 39.4%) were the frequently identified complication. Increased intracranial pressure (ICP) (n=40, 17%) and cardiorespiratory failure (n=38, 16.2%) followed aspiration pneumonia ( Figure   2). who were with sever record of GCS (3-8) during admission were 15 times more likely to die than those patients who were with mild score of GCS (13)(14)(15). In addition to that, stroke patients who were unconscious at admission were more than two times more likely to die than those of conscious stroke patients at admission (Table 3).

Discussions
In this study, an attempt had been made to assess the outcome of stroke patients after they have been hospitalized in ACSH. After the patients had admitted to hospital, 62.9% of the patients developed at least one stroke complication making the total frequency of complications to be 234. The most frequent complication identified was aspiration pneumonia with 39.4% of the frequencies, followed by ICP increment, which attributed for 17% of the total frequencies calculated. After medical and therapeutic treatments, 42.7% of the patients discharged with improvement. Nevertheless, when 20.4% of the patients left against medical advice, 10.2% of the patients have been discharged with no change.
Moreover, 22.2% of the stroke patients died at the hospital. The median time to death was 3 days ranging from a minimum of 1 day to the maximum of 48 days. The average mean time to death was 6.2 ± 8.7 days. Sever (3)(4)(5)(6)(7)(8) score of GCS at admission and being unconscious at admission were the two predictors of mortality in the hospital.
More than 60% of the stroke patients in the hospital developed at least one and /or above type of complication. Aspiration pneumonia was the most frequent type of stroke complication detected in our study. From a study conducted in Addis Ababa, the prevalence of complications was 71.8% which was relatively higher than our finding. In this study aspiration pneumonia was the most frequent complication among the stroke patients which is similar to our finding (18). Moreover, in a study from Gonder aspiration pneumonia was also the commonest type of complication discovered in stroke patients (19). The number of patients in the study conducted in Addis Ababa were few in number and it was a prospective study design. The discrepancy between the two studies might be originated from that study design difference.
The mortality proportion in our hospital was 22.2%. This is relatively similar with a study conducted in Addis Ababa, which reported a prevalence of 23% (18). Locally from our country, a relatively lower mortality rate with, a mortality prevalence of 13% from Gonder, Northwest Ethiopia (19) and 15.1% from Shashemene (20), 16.2% in Ambo (21) were respectively reported. Moreover, in neighbor country Kenya a mortality prevalence of 5% has also been found (22). The difference might be due to difference in study setting. ACSH is a comprehensive specialized hospital. It serves for referred patients with exaggerated complications from the region and other peripheral areas of Afar region and Amhara region. Therefore, relatively patients in the hospital are not similar to those patients who served in the other regions hospitals. Besides, there is time delay between symptom onset and admission that could result in treatment delay predisposing to mortality which might be due to issue of transportation and lack of symptom awareness of stroke. On the contrary, in a study conducted in Tikur Anbessa Hospital, capital of Ethiopia, 44.5% of the patients died at the hospital (13). Similarly, in a study from Gambia 41% of the stroke patients died at hospital (23). Our study setting is comprehensive specialized hospital which has Computed Tomography scan and Magnetic Resonance Imaging diagnosing materials. This might enable patients to easily confirm and detect stroke as fast as possible and intervene as quick as much.
Patients who were unconscious were more likely to die in hospital than stroke patients who were mentally alert. This finding concurred with the report of Tikur Anbessa hospital.
In a study conducted in Addis Ababa, altered mental state were found an independent determinant of mortality. In that study, patients with alert mental state were less likely to die (13). When patients are unconscious they are imperatively at risk of infectious disease and other lethal complications. As far as they have swallowing problem, they will have risk of death. Swallowing difficulties and level of consciousness at admission were a predictor of fatal out come in stroke patients like our findings (23). Sever (3-8) score of GCS was also a predictor of mortality in our study. Many studies also found to be advocative of our report. Patients at who were with lowest score of GSC were at risk of stroke complications which ultimately precipitates patients death (23, 24).
The mean hospital stay of the patients on the wards was 9.91 days. In contrast to this, a higher hospital stay was reported from a study at university of Gondar hospital Ethiopia, and Nairobi hospital, Kenya, which found a mean length of hospital stay of stroke patient to be 13 days and 12.5 days respectively (19,22). This difference might be due to difference in patient load and number of availability of beds. The median time from stroke symptoms onset to hospital admission was 24 hours in our setting which is similar to retrospective cross-sectional study in Shashemene with a median time from symptom onset to hospital admission of 23.3 hours (20). In contrary, a hospital-based multicenter prospective study in Brazil illustrated that the median time from symptoms onset to hospital admission was 12.9 hour (25). This disparity might be due to better infrastructure and greater awareness of people about the ways of management and risks of stroke in hospital admission in Brazil.
This study is the only study done in our hospital. It conceals relatively larger population.
There is no immaculate study globally. Hence this study has also its own flaws. For instance, this is hospital based study and as such the results cannot be generalized to the general population. As this study is based on retrospective chart review, the data obtained might be affected by the documentation culture of the hospital and health care professional. Significant numbers of patients were discharged against medical advice which made the in hospital mortality non representative of the short term mortality. The study hasn't assessed the rate of disability and functional limitation; those were not documented in the patient chart record. Moreover, the management intervention done was not also studied.

Conclusion
Significantly higher number of patients have been died at the hospital. The stay of the patients in the hospital was also longer. Majority of the patients had a diagnosis of, aspiration pneumonia being the most frequent type of complication. Unconsciousness of patients at admission and patients with lower GSC scores were risk factors for stroke to be lethal. Strategies that could decrease mortality prevalence should be streamed at the hospital setup level to improve stroke patients further better outcomes.

Ethics approval and consent to participate
The study was approved by the Institutional Review Board of Collage of Health Sciences, Mekelle University. An informed written consent was obtained from all patients.

Consent to publish
Not applicable.

Availability of data and materials
The datasets supporting the conclusions of the study are included in the article. Any additional data will be available on request.

Competing interests
The authors declare that they have no competing interests.

Funding
The study was not supported by any funding agent.

Authors' Contributions
All authors (SWA, KG, KG, YLN, DMD, TMA) contributed toward data analysis, drafting and revising the paper and agree to be accountable for all aspects of the work.