Study setting and Period
The study was conducted in four public hospitals found in Harar and Dire Dawa Administration, Eastern Ethiopia. Harar town is the capital city of Harari region located at 526 Kms away from Addis Ababa. There are two public hospitals namely, Hiwot Fana Specialized University hospital and Jugal hospital. Dire Dawa Administration is located at 515 Kms from the capital city of Ethiopia, Addis Ababa. In the administration, there are two governmental and four private hospitals and eight health centers. For this study two public hospitals were selected: These are Dilchora referral hospital and Sabiyan primary level hospital. The study was conducted from March 1 to 30/2020.
Study Design and Population
Institutional based cross-sectional study design was employed among 413 Nurses working in four selected hospitals in Harar and Dire Dawa Administration.
All employed Nurses who are working at public hospitals were the source population and those Nurses working in the four public hospitals and available during the data collection period were study population. Nurses who were not available during the time of the study period due to annual, maternal, sick, study leave and newly employed in the last six months were excluded.
Sample Size Determination and Sampling Technique
The sample size was calculated by using single population proportion formula with the following assumptions: The proportion of intent to stay (P)was 57.75% (10), standard normal distribution (z=1.96), 5% margin of error(d)
Then by adding a 10% non-response rate the final sample size becomes 413.
Where: n=minimum sample size required for the study.
z= standard normal distribution (z=1.96) with confidence interval of 95%.
p=proportion of intent to stay among nurses (p=57.75%) (10).
d= is tolerable margin of error (d=0.05).
Four hospitals were purposively selected because they have higher numbers of nurses’ population. The sample size was proportionally allocated to selected hospitals. Finally, simple random sampling technique was employed with the lottery method to recruit the study participants from each hospital.
Data Collection Procedures
An interviewer guided self-administered structured questionnaire which was developed in English language was used to collect data. Participants were identified by using attendance sheets from respective wards then questionnaire was given to them after brief introduction of the aim of study. The questionnaire was collected back on the same day. The questionnaire was adopted and modified from previous different literatures (9, 10). The questionnaires address the nurses’ socio-demographic characteristics such as age, gender, marital status, work experience, job position, level of education, current area of practice, wards and having dependent family members, job satisfaction, intent to stay, autonomy, professional opportunity and organizational commitment. Intent to stay was measured by a scale created from six items. The items were answered on a five-point Likert scale with response options ranging from 1(strongly disagree) to 5(strongly agree). Similarly, job satisfaction was measured by 37-item scale with 5 subscales: autonomy and professional opportunities, scheduling, support, relationship and interaction and pay and benefits. These items were also answer on a 5-point Likert scale with response options ranging from 1(very dissatisfied) to 5(very satisfied). In the same manner, organizational commitment was measured by 22-item scale with 3 subscales. Affective, continuance and normative commitment. In the same way, the items were answer on a 5-point Likert scale with response options ranging from 1(strongly disagree) to 5(strongly agree).
Data quality was assured through Pre-test was done on 5% (21 participants) of the sample at Haramaya hospital, eastern Ethiopia with some time gap before the actual data collections and the amendment was made accordingly. One day training was given by principal investigators to the data collectors and supervisors prior to data collection. The training was focused on the objective of the study, understanding the meaning of each question, obtaining consent, keeping the confidentiality of the information they gathered and quality of data collection. Emphasis was given on the significance and the meanings of each question as well as how to explain for the participants incomprehensible manner if required. On top of this, supervisors were followed by data collectors and the principal investigator was also checked for the collected data clarity, accuracy and completeness on a daily basis. Double data entry was made.
Data Processing and Analysis
Cleaned data were coded and entered to EpiData 4.6 version and exported to SPSS windows version 20 for analysis. At the beginning of the analysis, scores for intent to stay; and each subscales of job satisfaction and organizational commitment were created by summation of the score’s statements include within each sub-scale. Descriptive analyses were done. Then the variables were re-coded and dichotomized. The descriptive findings were presented by tables and graphs.
The responses of the outcome variable were first added and overall mean were computed. Then, it was categorized as yes and no for the respective value of greater than and less than the mean. The job satisfaction and organizational commitment related factors were computed Then, the recoded score was added for each participant and an overall mean were computed. Finally, they were categorized and recoded as satisfied and dissatisfied for job satisfaction items and the rest high and low for organizational commitment items for a score of greater than and less than the mean respectively.
Bivariable analyses were computed to check association between dependent and each independent variable using binary logistic regression. All variables with p ≤ 0.25 in the bivariable analysis were included in the final model of multivariable analysis in order to control all possible confounders. Multi co-linearity test was carried out to see the correlation between independent variables by using standard error and co-linearity statistics (VIF >10 and standard error >2 was considered as suggestive of existence of multi co-linearity). The degree of association between independent and dependent variables were assessed by using adjusted odds ratio with 95% confidence interval. P-value less than or equal to 0.05 were considered as statistically significant. The goodness of fit was tested by Hosmer-Lemeshow statistic which was insignificant and omnibus test, which was significant.
Have intent to stay: nurses’ decision to stay in their profession which was measured by six items in nursing profession having a five-point Likert scale ranging from six to thirty with a mean score of 18.2 and higher than 18.2 score nurses considered as having intention to stay in nursing profession.
Satisfied: Nurses who scored > 60% of the sum of the satisfaction scales.
High commitment: Nurses who score > 60% of the sum of the commitment scales.