Statistical Analysis to Identify Factor Related with Patients Satisfaction on General Service in Assosa General Hospital, Western Ethiopia

Patient satisfaction is the extent to which patients are happy with their healthcare provided from the hospital/health center. It is considered as one of the desired outcomes of health care and it is directly related with utilization of health services. The aim of the study was to identify the factor related with patient satisfaction on general service in Assosa General hospital. To reach the aim, the data have been collected through questionnaire from 735 patients, selected using simple random sample of total 2509 population, in October 1 to November 30, 2021. Frequency distribution Table and pie chart were used for data description; binary logistic regression was used to identify the factors that affect patients satisfaction, using R version 4.1.3. The results of this study showed that 43.27% of patients were satisfied and the remaining 56.73% were not satisfied with the service they have received from the hospital. Patients’ satisfaction was associated with age, gender, health care quality of services, waiting time for treatment, fee-for-service, availability of drug, laboratory service, cleanness of hospital, doctor and nurse response. From the result of binary logistic regression it can be concluded that age, health care quality of services, waiting time for treatment, availability of drug, cleanness of hospital, doctor and nurse response has significant effect on the satisfaction of patients.


Introduction
Patient satisfaction is the extent to which patients are happy with their healthcare provided from the hospital/health center. It is considered as one of the desired outcomes of health care and it is directly related with utilization of health services. However, it is not brought to desired level due to different reasons through out the world. Asking the patients what they think about the care and treatment they received is an important step toward improving the quality of care and to insuring that local health services are meeting patient's needs (Prakash, 2010).
Globally, the patient satisfaction become an emerging health policy all over the world and a key determinant of quality of care (WHO, 2014). Several nations are developing innovations to improve the different aspects of quality. Many low-and middle-income countries have developed successful interventions, but require a global platform to share knowledge. This will allow nations to learn from successful interventions and adapt them to their local populations. It will also allow nations to avoid directing efforts towards unsuccessful interventions. Improving quality of care has proven challenging for all nations (Al-Abri and Al-Balushi, 2014). However, providing quality care to people everywhere remains the most important shared responsibility and opportunity to improve the health of people globally. With a deliberate emphasis on quality, nations will be able to make significant progress towards achieving the Sustainable Development Goals and attaining universal health coverage. Thus, there is a strong connection between health service quality and patients satisfaction (Busse, 2014

Study Area
The study was conducted at Assosa General hospital (AGH). Assosa General hospital is found in Benishangul-Gumuz Regional State, Assosa town; in which the region is one of the nine regional states established in 1994 by the constitution of Ethiopia. Assosa, capital city of the region which has 667 km distance from Addis Ababa. The hospitals gives services to patients with different cases and serves as biggest referral hospital in the region. It start to give service since 1958 E.C.

Study Population
The data for this study was obtained from the patient in AGH, Assosa town, Ethiopia. Patients who have been received care from AGH during October 1 to November 30, 2021 was considered as the target population of the study. The total number of patients who received care were estimated to be 2509 from patients goes to hospital for treatment between October 1 to November 30, 2020 year.

Sampling Techniques
Sampling technique was a method of selecting sample from entire population. For this study simple random sampling was used to select sample from a total population of 2509 patients received general service at AGH. A crosssectional study which deployed interviewer administered questionnaires has been carried out to assess the patient satisfaction with general services. For pediatric age group patients, adult care givers who accompanied them were used as respondents.

Sampling Design
The response of the patient satisfaction can be categorized in to two classes: satisfied and not satisfied. To find sample size for this study, the proportion of patient those who were satisfied and not satisfied, as well margin of error should be known. Thus, proportion of patients who are satisfied with hospital services was estimated to be 42% using Pilet Survey, margin of error is 3% and the desired confidence level of 95%. Based this information the sample size determination for this study was as follows:- , is the margin of error p=42%, is a proportion of patient those who are satisfied q=1-p, is a proportion of patient those who are not satisfied To select the sample size (n), the following two cases are considered:

Data Collection Procedure
Ethical permission has been obtained from the Research Ethics Review Board of Assosa University. Then primary data were taken from patient's in the hospitals by data collectors using structured questionnaire. The data collector in the hospital has been collect the data after they took one day training on the objective and relevance of the study, how to gather the appropriate information, procedures of data collection techniques and the whole contents of the interview. The researcher monitored the overall data collection process during the data collection period.

Variables in the Study
The variable that has been used as dependent for this study was satisfaction of patients having categories: not satisfaction and satisfied(Y=0 not satisfied and Y=1 satisfied).

Descriptive Statistics
The frequency distribution table and pie chart were used to summarize patient satisfaction data obtained from the hospital using structured questionnaire.

Chi-squared Test of Independence
The chi-squared test of independence is one of the most basic and common hypothesis tests in the statistical analysis of categorical data. It will be used to test the association between binary response variable and explanatory variables. In addition it is used to identify which predictor variable are taken in the model and which are not. The predictor variable with p-value greater than 25% are not included in the model (Agresti, 1997).

Binary Logistic Regression Model
Binary logistic regression is used when the dependent variables are either binary (dichotomous) variable. It can be used to predict a dependent variable on the bases of continues or categorical independent variables and rank the relative importance of independent variable, to assess the interaction effect or to understand the impact of covariate(s) (Agresti, 1997). Unlike the discriminate analysis, the logistic regression does not have the requirements of the independent variables to be normally distributed, linearly related, nor equal variance with in each group (Tabachnick and Fidel, 1996).

Parameter Estimation for the Model
The maximum likelihood and non-iterative weighted least squares are the two most computing estimation methods used in fitting logistic regression model. When the assumption of normality of the predictors does not hold, the non-iterative weighted least squares method is less efficient. In contrast, the maximum likelihood estimation (MLE) method is appropriate for estimating the logistic model parameters due to this less restrictive nature of the underlying assumptions (Hosmer and Lemeshow, 1989). Hence, the MLE technique would be applied to estimate parameters of the model.

Model Building
The methods of selecting a subset of covariates in Binary logistic regression are essentially similar to those used in any other regression models. Thus, the model was built using Hosmer and Lemeshow (1998) recommendation.

Model Diagnostics
Assessing goodness of fit involves investigating how close values predicted by the model with that of observed values. The question is to test whether the fitted model fit the data or not. The deviance is used for this purpose (Agresti, 1997).

Descriptive Analysis
The data for this study has been taken from 735 patients treated at AGH, Ethiopia during October 01 to November 30, 2021. As shown in Table 1 below, from the total of 735 patients in the hospital, 417(56.73%) were not satisfied by service the have been obtained from the hospital and the remaining 318(43.27%) were satisfied. Most of patients in the hospital were female with age category of 15-50 years old. The treated patients in the hospital were mostly from rural area.
Looking the response of patients on the health quality care of service in the hospital, 56.79% were responded as there is no service quality and the remaining 43.30% were responded as there is service quality care in the hospital. Patients in the hospital receive treatments after staying more than 2 hours in most cases.
By considering fee for service in the hospital about 52.90% response of patient were as there as there is no acceptable fee for service they provided and the remaining 47.10% patients respond as there is acceptable fee for service they provided. The majority of respondent respond as there is poor availability of drug, laboratory service and cleanness of hospital. The respondents are also highly respond as there is fair doctor and nurse response while giving treatment in the hospital.

Availability of Data
The data will be given upon request on behalf of the corresponding author

Ethical Consideration
The Research Ethics Review Board of Assosa University would provide an ethical clearance for the study. The data has been collected after written permission was given to AGH and Assosa University Research Directors write an official cooperation letter to the hospital for the permission. The data obtained from the hospital were kept confidentially.

Not applicable
Software Used SPSS version 16 was used for coding and R version 4.1.3 was used for data analysis