An institution-based comparative cross-sectional facility-based study was conducted from December 1–30, 2021 among those who are insured and non-insured under the CBHI scheme at Deder General Hospital, East Hararghe Zone, Oromia, Ethiopia. Deder Hospital is one of the eight hospitals’ in the East Hararghe zone and serves four populated woredas’ in East Hararghe (Malka Ballo, Goro Gutu, and Deder Woreda). According to the 2016 EDHS report, the hospital was serving about 357,904 males and 344,128 females, for a total of 702,032 populations being served by the hospital. Deder General Hospital was the first health facility where the federal government began a pilot study of community-based health insurance in 2011.

An institution-based comparative cross-sectional study was implemented.

All insured and noninsured patients who visited the outpatient department of the hospital during the study period were considered a source population. Whereas, all patients selected by a systematic random sampling method from outpatients who visited the hospital in the study period were the study populations. All insured and non-insured patients aged greater than 18 who visited the outpatient department of the hospital were included in the study, whereas patients who were critically ill and unable to communicate were not included in the study.

The sample size for the study was determined for both objectives by the double population formula using EPI Info version 7, and the largest number was selected. The sample size was calculated using the power formula of 80%.

The assumption were.

n = sample size.

P1 = proportion of satisfaction on availability of drugs in insured patients was taken as 60%, and P2 = proportion of satisfaction on availability of drugs in non-insured patients was taken as 72% from the study conducted in public hospitals in Addis Ababa, Ethiopia (25).

α = the level of significance (0.05).

β = Type II error 0.2; and none response rate 10%.

Depending on the above statistical assumption, the final sample size calculated for this study was.

n= 0.60 1 − 0.60 + 0.72 1 − 0.72 0.60 − 0.72 2 x 7.9 =532.

Therefore, the final sample size for the study was 532, and a one-to-one allocation method was used for both groups (insured and non-insured). Finally, the factor that gives the largest sample size was used to determine the final desired sample size for the study. Therefore, the study used a sample size of 266 for insured and non-insured patients. The study subjects were selected from the adult and pediatric outpatient departments of the hospital using systematic random sampling using the patient’s identification number (Figure 1).

A structured, interviewer-administered questionnaire, extracted from various pieces of literature, was used for data collection. The outcome variable of the study was patient satisfaction with health services. The independent variables of the study were socio-demographic characteristics (age, sex, marital status, educational levels, religion, and income). Health services-related factors (waiting times, availability of prescribed drugs, laboratory services, cost of services, staff services, privacy and confidentiality, information, accessibility, provider’s behavior and services, physical facilities, overall satisfaction with the quality of care received from the facility), and consultation and diagnosis (weighing the patient, measuring the temperature, using a stethoscope, examining the patient, asking about a history of past illness, asking about a history of present illness, asking if treatment was taken before arrival at the facility, and explaining the diagnosis).

Community-based health insurance (CBHI) is a prepayment method of financial contributions for healthcare that aims at risk pooling and the avoidance of catastrophic and impoverishing health expenditures (26).

Patient expectations were measured by the patient’s expectations, wants, and thoughts about what needed to be completed (25).

Not getting services: if the patient perceives he or she did not get all of the services (27).

Satisfaction with health services was measured using 16 questionnaires of five-point Likert scale items each rated from 1 (strongly disagree) to 5 (strongly agree), and the mean score was computed from the composite index score created from a summed score of 16 items. The level of satisfaction was satisfied when the participants scored above the mean score of the summed satisfaction index score and poor unless otherwise stated (28).

Insured Patient: A patient or their guardians who has paid an annual premium voluntarily to a local health care fund within a given period (2020/2021) and who had accessed care in the hospital at the time of interview.

A non-insured patient is a patient or caregiver who paid the full cost of each visit to the hospital but who had not accessed care in the hospital at the time of the interview.

Outpatient Department (OPD): the unit in the hospital where medical services are provided to clients.

Waiting time: the time interval between departures from registration for outpatient service and being seen by a physician at OPD.

The data was collected using an interviewer-administered structured questionnaire that was developed after a review of different literature that supports the objectives of the study (27, 29, 30). An exit interview data collection method was performed. The questionnaire, which consists of the Likert type, was first prepared in English, then translated into the native language (Afaan Oromo) for data collection purposes, and then back-translated to English to make sure the meaning was retained. The data was collected by two BSc nurses and two diploma clinical nurse professionals who were trained on the principles of data collection, components of tools, and ethical issues. Respondents were asked about their perceived satisfaction with the hospital services they received on the day of the interview, and their responses were scored using a five-level Likert scale of categories (very good =5, good = 4, somewhat good = 3, poor = 2, and very poor = 1).

The questionnaire was pretested on 5% of the sample size before the actual data collection at Chelenko Hospital, which provides similar services to the current study setting. Then the appropriateness of the checklist was evaluated, and vague questions were modified before the actual data collection. Besides, supervisors and the primary investigator regularly evaluated the obtained data for consistency, clarity, accuracy, and completeness. The internal reliability of 16 Likert scale perceived quality of health services measurement tools was estimated using Cronbach’s alpha value of 0.919, which was higher than the accepted value of 0.80 recommended by an author Kline (30).

All data were checked for completeness, cleaned, coded, and entered into Epi-data version 4.6 and exported to SPSS version 22.0 for further analysis. Descriptive statistics were used to summarize the data by using simple frequency tables and figures. The respondents were asked about their perceived satisfaction with the health care services they received and scored using a five-level Likert scale of categories. To examine the presence of a statistically significant association between insured and noninsured patients, a chi-square test was performed. Binary logistic regression was used to identify factors associated with satisfaction. All variables that were significant at a p value <0.25 in the bivariable analysis were considered for the multivariable analysis to control for all possible confounders. The model’s fitness was checked by the Hosmer-Lemeshow tests, and it was at a p-value of 0.314. Odd ratios along with the 95% CI were estimated to measure the strength of factors associated with the level of satisfaction with health services in both insured and non-insured patients. For all variables, the level of statistical significance was declared at p values less than 0.05.

Ethical approval was obtained from the Institutional Health Research Ethics Review Committee (IHRERC) of Haramaya University, College of Health and Medical Sciences. Then formal letters of cooperation were taken from East Hararghe zonal health offices to Deder General Hospital. Before obtaining informed consent from patients and guardians, they were given a clear description of the study objectives, procedure, duration, and possible risks and benefits of the study. Their right to participate or not and to withdraw in the middle of the interview was guaranteed. Informed, written, and signed consent was obtained from voluntary patients (guardians). The confidentiality of all study participants was maintained.

A total of 532 (266 insured and 266 uninsured) patients participated in this study, with a response rate of 100%. The total insured study participants, about half, 269 (50.6%), were female. Individuals who were insured and those who were not insured had a mean age with standard deviation (SD) of 43.10 ± 14.64 and 35.86 ± 13.39 years, respectively. Besides, seventy-four (27.8%) insured patients and sixty-six (24.8%) uninsured patients were aged equal to or greater than 50 years old. More than half of the study participants were urban dwellers 144 (54.1%) insured and 152 (57.1%) uninsured patients. Most of the study participants, insured 133 (50%) and uninsured 120 (45.1%), were Muslims. Regarding educational status, 39 (14.7%) of the respondents in the insured population and 45 (16.9%) uninsured patients attended higher educational levels, respectively. More than half of the study participants were urban dwellers, with 144 (54.1%) insured and 152 (57.1%) uninsured patients (Table 1).

Out of all the participants in this study, only 98 (36.8%) insured and 149 (56.0%) uninsured patients reported that pharmacy staff gave explanations on the use of medications. In similar circumstances, 219 (82.3%) insured and 200 (75.2%) uninsured patients were reported as having their privacy and confidentiality maintained. Out of all the participants, only 81 (30.4%) of the insured and 84 (31.6%) of the uninsured patients received all of the prescribed medication. Similarly, only 32 (12.0%) of insured patients and fifty-seven (21.4%) of uninsured patients were seen by doctors within 30 min after arrival. Regarding the cost of services, 38 (14.3%) of insured patients and 47 (17.7%) of uninsured patients paid less than 100 Ethiopian birr (Table 2).

Regarding the objective quality of care measures used for consultations and diagnosis, 90 (33.8%) of insured and 136 (51.1%) uninsured patients, 84 (31.6%) of insured and 128 (48.1%) of uninsured patients, and 86 (32.3%) of insured and 104 (39.1%) of uninsured patients reported that providers measured the temperature of the patient, used a stethoscope, and weighed the patients, respectively. The diagnosis was explained only for 95 (35.7%) of insured patients and 143 (53.8%) of uninsured patients who were seen in outpatient services. Only 97 (36.5%) insured and 144 (54.1%) uninsured patients were asked if the treatment was taken before arrival (Table 3).

The level of household satisfaction with the CBHI scheme was rated using five questions, each worth five points on a Likert scale. The CBHI plan satisfaction score ranged from 5 to a maximum of 25 points for respondents. The average level of satisfaction was 4.178. After that, families were classified as satisfied if their score was higher than the mean and unsatisfied if it was below the mean. Overall patient satisfaction at Deder General Hospital was 65.6% (95% CI: 61.5–69.50) based on 349 out of 532 respondents rating the hospital’s medical services as satisfactory. Overall, 183 (68.8%; 95% CI: 62.8–74.4) insured patients and 166 (62.4%; 95% CI: 56.8–68.0) uninsured patients expressed pleasure (Figure 2).

In the bi-variable model, factors such as education level, how patients visited the hospital, use of a stethoscope, measure of temperature, how patients visited the hospital, waiting time, whether pharmacy staff explained the use of drugs or not, availability of drugs, and cost of service were significant at a 0.25 value of p and considered for the final multivariable analysis to assess factors associated with satisfaction among insured patients. In the final multivariable analysis, three variables (level of education, availability of drugs, and cost of service) were found to have a statistically significant association with satisfaction.

The odds of being satisfied with health care services were 4.90 (AOR = 4.90; 95% CI: 2.05–11.75) and 3.08 (AOR = 3.08; 95% CI: 1.05–9.03) times higher among patients who attended secondary school and had a higher level of education compared to individuals without a formal education. Similarly, patients who received all of the prescription medication were 3.49 (AOR: 3.49; 95% CI: 1.43–8.54) times more satisfied than those who received only a portion of the ordered medication, who were 3.34 (AOR: 3.34; 95% CI: 1.61–6.94) times more satisfied than those who received none of the prescribed medication at all. Moreover, an insured patient who paid less than 100 Ethiopian birr for treatment and follow-up services was 4.85 (AOR: 4.85; 1.35–17.40) times more satisfied than those who were paid between 100 and 500 Ethiopian birr and those treated with an annual payment (Table 4).

Factors such as education level, use of a stethoscope, measuring of temperature, how patients visited the hospital, waiting time, whether pharmacy staff explained the use of drugs or not, availability of prescribed drugs, and amount of payment were candidates for multivariable analysis at a p-value less than 0.25 in bivariable mode. In multivariable analysis, only the availability of prescribed drugs was significantly associated with satisfaction in uninsured patients. Patients without insurance who received all of their doctor’s prescriptions were 6.28 times (AOR = 6.28; 95% CI: 3.26–12.05) more satisfied than those who received part of their prescriptions, and 3.40 times (AOR = 3.40; 95% CI: 1.70–6.78) more satisfied than those who did not receive all of their prescriptions (Table 5).

It may be challenging to provide more meaningful information regarding participant perception and the time sequence of the associations because the study only used a quantitative method and a cross-sectional design.