Patient satisfaction and associated factors among psychiatry outpatients of St Paulo’s Hospital, Ethiopia

Study design, period and area

An institutional based cross-sectional study was undertaken at SPHMMC, Addis Ababa, Ethiopia from May to June 2018. SPHMMC provides medical specialty services to an estimated 200 000 people annually. Around 17 330 were psychiatric patients who were referred from all over the country. An average of 1443 psychiatric patients visit the hospital on an outpatient or emergency basis monthly. All follow-up psychiatry outpatients were the source population. All sampled psychiatry outpatients aged 18 and above who had at least three visits comprised the study population, after exclusion of those who were severely ill, had cognitive impairment, or communication or hearing problems.

Sample size determination and technique

We determined the sample size by using the single population proportion formula, based on the assumption of 61.2% prevalence of patient satisfaction studies conducted in Dessie referral hospital psychiatric care services, north east Ethiopia28; with 95% CI and 4% marginal error, by adding a 5% non-response rate the sample size was determined to be 600. The study period was 1 month and the average monthly number of outpatients was 1443. The consecutive sampling technique was then used to select the study participants up to the required sample size of 600. The sampling procedure is depicted graphically in figure 1.

Request permission

Data collection procedures and instruments

Data were collected with a pretested structured questionnaire using face to face interview by five trained psychiatry nurses. The questionnaires were developed in English and translated to Amharic then back to English to maintain consistency. Data on patient satisfaction were collected by using a standardised Client Satisfaction Questionnaire (CSQ-8), developed for use in the mental health programme. The tool has eight items and the rank order is the same for all questions, with four options ranging from 1=“less favourable” to 4=“most favourable”.32 The CSQ-8 was validated in Egypt and has internal consistency (Cronbach’s α=0.88).26 The client satisfaction questionnaires were easily scored by summing the individual item scores depending on each answer. The results of the questionnaire range from a minimum score of 8 up to a maximum score of 32. A score of 8–20 is regarded as the patient being dissatisfied, 21–26 as satisfied, and 27–32 as highly satisfied.26 Information on the types of mental illness were collected by the reviewing patients’ charts. Sociodemographic and other variables were assessed using structured questionnaires developed from different studies in the literature. Internal consistency was demonstrated in this study by Cronbach's alpha (α=0.892).

Data processing and analysis

The collected data were edited, coded, and entered into Epi-Data version 3.1 and then exported into Statistical Package for Social Sciences (SPSS) version 20 for analysis. Ordinal logistic regression was used to test the associations between independent and dependent variables. Variables at bivariate analysis p<0.2 were candidates for multivariate analysis. The strength of associations was described by using OR and 95% CI. Variables with p value <0.05 at multivariate analysis were considered statistically significant. Descriptive analysis frequency table, mean, SD and graph were used to explore the data. Multicollinearity test was performed to the ordered logit model being fitted, by evaluating the variance inflation factor (VIF). In principal, variables have VIF values no more than 10. The parallel line test was performed to the ordered logit model being fitted for the general model, with χ²=38.44 and p=0.513, which is greater than the level of significance, failing to reject the null hypothesis. An ordinal logistic regression model to fit to the data was chosen.

Sociodemographic characteristics

Of the total of 600 outpatients invited to participate in the study, 589 (98.2%) completed the data; 11 (1.8%) provided incomplete data and were discarded. The mean age of the participants was 35.5 (12.8) years. Most of the participants (413, 70.1%) were between the age 18 and 39 years. The majority of the participants (367, 62.3%) were male, 374 (63.5%) were Orthodox, 233 (39.6%) were Amhara ethnic group, and 327 (55.5%) were single. The mean distance of travel to the hospital was 63.1 (128.0) km (minimum 2 km, maximum 785 km). More than four in five (483, 82%) were residents in urban areas and 486 (82.5%) were living ≤50 km from the health facility. Of the participants, 328 (55.7%) were charged for the services they received, 186 (31.6%) were unemployed, and 201 (34.1%) were in secondary school education (table 1).

Clinical characteristics of participants

In regard to the duration of illness, 215 (36.5%) had been ill for <12 months, 201 (34.1%) for 1–5 years, 86 (14.6%) for 6–10 years, and 87 (14.8%) for ≥11 years. Diagnoses of the participants included schizophrenia (186, 31.6%), depression (137, 23.3%) and bipolar disorders (125, 21.2%) (figure 2).

Request permission

Patient satisfaction related factors

Regarding patient satisfaction related factors, 528 (89.6%) of the participants were satisfied with the level of confidentiality and discretion, 495 (84%) were highly satisfied with the cleanliness of the hospital in general, 465 (78.9%) were satisfied with the information received about treatment, 457 (77.6%) were satisfied with the relationship with the healthcare staff members, and 397 (67.4%) were satisfied with the opening hours of the treatment centre; however, 339 (57.6%) were dissatisfied with the waiting time for the visit (table 2).

Magnitude of patient satisfaction

Concerningt the magnitude of patient satisfaction, 50.3% (95% CI 46.0 to 54.2) were highly satisfied, 31.03% (95% CI 27.2 to 34.8) were satisfied, and 18.7% (95% CI 15.4 to 22.1) were dissatisfied.

Patient satisfaction with service according to the CSQ-8 responses

The two items for which more than half the patients were highly satisfied were: recommending the service to others (346, 58.7%); and coming back to the service again if further help was needed (table 3).

Factors associated with patient satisfaction

In bivariate analysis, factors including male sex, age, religion, ethnicity, living condition, education, occupation, type of mental illness, service, and medication availability were candidates for multivariate analysis with a value of p<0.2. In multivariate analysis, male sex, inability to read and write, being unemployed, free service, and availability of medication were significantly associated with patient satisfaction at p<0.05.

The odds of highly satisfied versus satisfied and dissatisfied for males was 2.3 times more likely as compared with females; likewise the odds of highly satisfied and satisfied versus dissatisfied for males was 2.3 times more likely as compared with females (adjusted OR (AOR) 2.30, 95% CI 1.57 to 3.36). The odds of highly satisfied versus satisfied and dissatisfied for respondents unable to read and write were 2.2 times greater as compared with those with a degree level of education and above; likewise the odds of highly satisfied and satisfied versus dissatisfied for participants unable to read and write were 2.2 times greater as compared with those with a degree level of education and above (AOR 2.23, 95% CI 1.10 to 4.66).

The odds of highly satisfied versus satisfied and dissatisfied for unemployed participants was 1.69 times more likely as compared with employed patients; likewise the odds of highly satisfied and satisfied versus dissatisfied for unemployed participants was 1.69 times more likely as compared with employed subjects (AOR 1.69, 95% CI 1.15 to 2.47).

The odds of highly satisfied versus satisfied and dissatisfied for receiving the service for free were 1.57 times greater as compared with respondents being charged a fee; likewise the odds of highly satisfied and satisfied versus dissatisfied for receiving the service for free were 1.57 times greater as compared with respondents being charged a fee (AOR 1.57, 95% CI 1.11 to 2.22).

The odds of highly satisfied versus satisfied and dissatisfied for availability of medicine were 1.62 times more likely as compared with no availability; likewise the odds of highly satisfied and satisfied versus dissatisfied for availability of medicine were 1.62 times more likely as compared with no availability (AOR 1.62, 95% CI 1.13 to 2.23) (table 4).

Main findings

This study examined the level of patient satisfaction in a psychiatry outpatient setting. Of the study subjects, 50.3% were highly satisfied, 31.03% were satisfied, and 18.7% were dissatisfied. The proportion of patients being highly satisfied with the general mental healthcare services was lower in our study than in those in psychiatry outpatient settings elsewhere, such as 61.2% in Ethiopia, 60% in Tanzania, 60% in Austria, and 58% and 72% in two different studies in Pakistan.19 24 25 27 28 The discrepancy might be due to the use of different measurement tools (locally developed tools were used in Austria, and the Patient Satisfaction Questionnaire III in Pakistan), different patient characteristics or study populations, different data collection methods (a self-administered questionnaire was used in Tanzania), different study design (a prospective study was undertaken in one of the Pakistan reports), differences in mental health literacy and mental health services, and sociocultural differences. Conversely, our finding was higher than the 31.2% noted in Egypt among psychiatry outpatients, and 44.8% among anxiety disorder outpatients in Los Angeles, California.15 26 The possible reasons for the difference might be due to the type of mental disorder of the patients and the use of different instruments and setting (the California study was done at a primary healthcare location), study design (a prospective study design was used in California), and sociocultural differences. On the other hand, our findings are comparable with the studies done in Mexico and Ireland, 53% and 52%, respectively.16 17

Some 31.0% of the participants were satisfied with the general mental health service based on CSQ-8. This is consistent with reports of 25% in Tanzania, 34.3% in Mexico, 24% in Austria, and 25.5% and 18.7% in Pakistan.16 19 24 25 27 In contrast, this finding is lower than that in the studies done in Egypt (50%) and Ireland (38%).17 26 The possible reason for the variance in the finding might be the differences in the characteristics of the study population, sociocultural differences and the data collection methods (a self-administered questionnaire was used in Ireland). On the other hand, the magnitude of satisfaction is higher in our study than that in the study in Austria (24%)19 and the two studies in Pakistan (25% and 18.5%).24 25 The possible reason for the difference in the finding might be the use of different measurement tools (hospitalised patients’ satisfaction questionnaire in psychiatry in Austria, and patient satisfaction questionnaire III in Pakistan), the study population, and sociocultural, mental health literacy and mental health service differences.

In this study, the magnitude of dissatisfaction is consistent with the studies done in Egypt (18.6%), Saudi Arabia (18.3%), China (22.1%), Germany (18.5%), Austria (16%), Romania (22%) and Pakistan (16.5%).18–21 23 24 26 Conversely, the level of dissatisfaction in our study was found to be higher than in other studies done in Tanzania (5%), Ireland (10%), Pakistan (9.3%) and Mexico (12.7%).16 17 25 27 The possible reason for the variance in the finding might be the characteristics of the study populations, sociocultural differences and the different data collection methods (a self-administered questionnaire was used in Ireland). The level of dissatisfaction found in our study was also lower than in studies done in north east Ethiopia (38.8%), India (43%) and Los Angeles (55.2%).15 22 28 This discrepancy might be due to the different measurement tools (Charleston Psychiatric Outpatient Satisfaction Scale, a 15-item scale assessing patient satisfaction, was used in the north east Ethiopia study), and differences in the characteristics of the study population.

Regarding the factors associated with patient satisfaction, male sex was significantly associated with patient satisfaction in our study. The results are supported by studies done in Ethiopia, Nigeria, Qatar and Ireland.9 17 28 31 In our study the inability to read and write was significantly associated with patient satisfaction. This finding coincides with studies done in India, Saudi Arabia, Sir Lanka and Ireland.17 23 29 30 Being unemployed is also significantly associated with patient satisfaction and is supported by studies done in India, Qatar and Sir-Lanka.22 30 31 In our study, obtaining services free of charge was significantly associated with patient satisfaction. However, this is not supported by findings from other studies. The possible reason could be that, in our study, individuals who received services for free were poor and could not afford medical costs, so they might be grateful for fulfilment of their needs without charge.

In this study the availability of medicine was statistically associated with patient satisfaction. This finding is supported by a study done in Tanzania.27 The possible reason might be that the insufficient supply of psychiatric drugs in hospital could create dissatisfaction because patients are unable to afford and access the prescribed medications at a private pharmacy.

According to our study, factors such as marital status, distance from the health institution, residency, infrastructure, excessive reliance on laboratory and physical examinations, relationship with healthcare staff members, and information received about treatment were not associated with patient satisfaction. The possible reasons might be the use of different tools, data collection methods, study designs and the study subjects.

Limitations

Despite providing invaluable baseline data, this study has some limitations. Firstly, social desirability bias; the data were collected by an interviewer administered method which might lead to either over reporting or under reporting. Secondly, recall bias; participants may be forgetful and there was no cross checking of the information they gave us. In this study, only adult psychiatry patients were included, so it is difficult to generalise our findings to all psychiatry patients because inpatients, children and adolescent psychiatry patients were not included in the study.

Implications

This study has many strengths. The extent of patient satisfaction was assessed by using a reliable standard tool and the satisfaction level was quantified on ordinal scale descriptions. The study showed that further improvement of patient satisfaction is required. Male sex, an inability to read and write, being unemployed, receiving services free of charge and availability of medication were significantly associated with patient satisfaction. More than half of the participants were dissatisfied with the waiting time to receive services. The timely provision of services and meeting patients’ expectations are helpful for achieving good health outcomes.