Determinants of patient satisfaction: Lessons from large-scale inpatient interviews in Vietnam.

Patient satisfaction, a healthcare recipient's reaction to salient aspects of their service experience, has been considered an important metric of the overall quality of healthcare in both advanced and developing countries. Given the growing number of studies on patient satisfaction in developing and transitioning countries, publications using high-quality patient surveys in these countries remain scarce. This study examines factors associated with inpatient satisfaction levels using nationwide, large-scale interview data from 10,143 randomized and voluntary patients of 69 large and public hospitals in Vietnam during 2017-2018. We find that older patients, poor patients, female patients, patients with lower levels of education, patients not working for private enterprises (or foreign enterprises), and rural patients reported higher levels of overall satisfaction. Health insurance is found to have positive influence on overall patient satisfaction, primarily driven by limiting patient concerns about treatment costs, as well as increasing positive perceptions of hospital staff. We further find that patients who paid extra fees for their hospital admission expressed higher scores for hospital living arrangements and medical staff, but were mostly dissatisfied with treatment costs. These findings have important policy implications for current policy makers in Vietnam as well as for other countries with limited healthcare resources and ongoing healthcare reforms.

Introduction

Patient satisfaction is defined as “a healthcare recipient’s reaction to salient aspects of the context, process, and result of their service experience” [1]. In order to measure patient satisfaction, it is necessary to evaluate patient perceptions and to determine whether or not the patients perceived their medical needs were adequately met [2]. In advanced countries, patient-oriented outcomes and patient satisfaction surveys are central to designing and evaluating healthcare services and delivery system by reflecting the quality of services from the patients’ perspective and identifying patients who need additional attention [3]. For example, evaluation of patient satisfaction has been required in all French hospitals since 1996 and in Germany since 2005 [4]. In developing countries, recording patient views on healthcare delivery is becoming increasingly important as these countries shift from a doctor-to-patient relationship to the more modern provider-client attitude [5].

This study examines factors associated with inpatient satisfaction levels in Vietnam’s public hospitals during 2017-2018, using nationwide, large-scale interview data from randomized and voluntary responses to a survey provided to patients of 69 large and public hospitals. We look at how patient satisfaction levels respond to service-specific characteristics (length of stay and extra-paid services), patients’ financial conditions (employment status, health insurance, and income level), and demographic factors (age, gender, race, location and level of education). We further evaluate the associations of these factors with different salient aspects of their healthcare service experience: hospital staff, hospital facilities, and treatment costs.

The data show that elder patients, low-income patients, female patients, those with only primary level education, and rural patients tended to report higher scores on overall satisfaction of public hospitals’ services in Vietnam. More importantly, health insurance is positively associated with a higher level of patient satisfaction which may be primarily driven by reducing patient concerns about treatment costs. Patients who paid extra fees for their hospital admission reported higher scores for hospital living arrangements and medical staff, but were mostly dissatisfied with treatment costs.

This study attempts to close the existing gap in prior work on understanding patient satisfaction in Vietnam which is limited in specific health clinics and presents a major geographic limitation. Although studies on patient satisfaction have been increasing in developing countries including Vietnam, publications using high-quality and large-scale patient surveys are still scarce due to resource and time constraints. Despite a number of studies which have been conducted in Vietnam to evaluate patient satisfaction in a general inpatient setting or in specific healthcare clinics, these studies often present a major geographic limitation. For instance, one study explored psychological and socioeconomic factors associated with patients’ evaluation of healthcare quality, using data from over 2000 patients in a survey conducted in Hanoi in 2016 [6]. They found that the average score on health services quality was 3.55 (where 5 is the highest score), which was positively associated with younger age, higher-levels of income, and better health conditions. Three other studies were conducted only in the North of Vietnam [7-9]. Nguyen and Nguyen (2014) conducted a patient survey on 894 patients in 18 public hospitals in the North of Vietnam and suggested the average satisfaction score as 3.68 based on a 5-point Likert scale [7]. Vuong et al. (2018) also conducted a survey comprising of 900 patients in five provinces in Northern Vietnam from August 2014 to June 2015 [9]. The authors found a low satisfaction level from this group of patients: about 66% of participants perceived unsatisfactory levels about the health services. Other studies addressing some extent of geographical differences, however, were limited to some specific clinic setting [10, 11]. Tran et al. evaluated patient satisfaction with HIV care in Hanoi, Hai Phong (two provinces in Northern Vietnam), and Ho Chi Minh City (in Southern Vietnam) [10]. Interviewing 1016 patients in these clinics, the authors found that the percentage of respondents completely satisfied with overall service quality and treatment outcomes was 42.4% and 18.8%, respectively, which was considerably lower than the satisfaction level in general healthcare settings. To the best of our knowledge, our paper is the first study on patient satisfaction in Vietnam which addresses these geographic and specific clinic setting limitations by using a nationwide and large-scale interview data.

This paper contributes to the medical literature regarding determinants of patient satisfaction by providing analyses using a recent sample of data on 10,143 randomized and voluntary patients in Vietnam. Examining the correlation between demographic factors, health status, and multidimensional attributes of healthcare settings with patient satisfaction, prior studies provide mixed findings [4]. In a recent systematic review, Batbaatar et al. (2017) identified a total of 71 studies between 1980 and 2014 which often provided week and contradictory findings across samples on the relationship between various patient-related characteristics and patient satisfaction [12]. For instance, higher income group patients tended to be more satisfied with overall health services in a number of studies while another study concluded that lower-income patients were more satisfied with nursing care [12]. In addition, one study in Scotland found that patients with only primary education level reported higher patient satisfaction using data of 650 patients discharged from four acute care general hospitals during February and March 2002 [13]. One similar study showed that education level was not a significant predictor of overall patient satisfaction using a Norwegian national patient-experience survey of 10,912 patient [14]. These examples suggest that it is important to obtain more representative samples to attain better understanding of patient satisfaction.

Analyzing various demographic and economic characteristics of patients and services may help us better understand the quality of hospital services from the patients’ perspective and identify which sub-populations need additional attention. The focus of this study is Vietnam, the world’s fifteenth most populous country with a rapidly developing economy. The current healthcare system in Vietnam is a mixed public-private provider system under a number of initiatives and health finance reforms, which aims toward universal health coverage and improved general public health [15]. These findings have important policy implications for policy makers in Vietnam as well as for other countries with limited healthcare resources and ongoing healthcare reforms.

Materials and methods

Design and setting

This cross-sectional patient-level study is a secondary data analysis using a nationwide and large-scale interview data. In particular, this study used 10,143 randomized and voluntary responses of participants from 69 large and public hospitals located in 27 Northern, Central, and Southern provinces and cities in Vietnam in 2017-2018. Eligible participants were inpatients (or close relatives who accompanied patients during the treatment process and answered on behalf of the patients) satisfying: (1) at least 18 years old, (2) stayed at least one entire day in one of participated hospitals from April to July in the surveyed year, (3) survived after treatment.

The participating hospitals are located nationwide in 27 provinces and cities of Vietnam. The locations of 29 participating hospitals in 2017 include: (1) 10 Northern provinces including Ha Giang, Son La, Bac Giang, Bac Ninh, Hanoi, Hung Yen, Nam Dinh, Thai Binh, and Ninh Binh; (2) 7 Central provinces including Nghe An, Quang Tri, Da Nang, Quang Ngai, Hue, Ninh Thuan, and Binh Dinh; (3) 4 Southern provinces including Ho Chi Minh City, Can Tho, Tien Giang, Ba Ria Vung Tau, and Ca Mau. The locations of the 60 participating hospitals in 2018 include: (1) 8 Northern provinces including Ha Giang, Son La, Bac Giang, Thai Nguyen, Quang Ninh, Hanoi, Ninh Binh, and Thanh Hoa; (2) 8 Central provinces including Nghe An, Quang Tri, Da Nang, Quang Ngai, Hue, Ninh Thuan, Binh Dinh, and Khanh Hoa; and (3) 7 Southern provinces including Ho Chi Minh City, Can Tho, Tien Giang, Ba Ria Vung Tau, Ca Mau, Vinh Long, and Hau Giang. Among these 60 participating hospitals in the 2018 survey, 21 hospitals were participated in the 2018 survey. The health system of Vietnam consists of the central level, provincial level, district level, and commune level. The surveys in this study covered three levels of hospitals in Vietnam: central, provincial, and district [16]. The lowest level of Vietnam’s healthcare system, commune health stations were not included in these surveys.

This secondary data analysis was determined to be exempt from review and informed consent by the institutional review board of the Indiana University Human Subject Office because the interviews were implemented by Vietnam Initiative for non-research purposes prior to this study. In addition, data of this paper were obtained from Vietnam Initiative in a fully anonymized and de-identified manner.

Data source

Data of this study comes from Vietnam Initiative. This institution designed the interview methods and contracted an independent interviewing organization to conduct these interviews for the project titled “Equitable Healthcare through Patient Satisfaction Index”. Fig 1 provides detailed steps of the sampling approach and data collection process implemented by Vietnam Initiative. In step 1, in collaboration with Vietnam Initiative, Medical Services Administration (MSA, operating under the Ministry of Health) of Vietnam sent the official letter to 80 public hospitals in 2018 and 50 hospitals in 2017. The letter called for voluntary participation of hospitals without any financial incentives. In step 2, the participating hospitals used the secured link provided in the letter to supply the lists of all eligible patients with cellphone numbers to Vietnam Initiative. For example, Vietnam Initiative received a list of 136,678 inpatients who stayed in one of the 29 participated public hospitals from April to July 2017.

Fig 1

Sampling approach and data collection process.

In step 3, Vietnam Initiative de-identified patient lists and randomly selected up to 250 patients from each hospital in 2017 (or up to 500 patients in 2018). For five small hospitals with the number eligible patients under 500, all eligible patients were included. The de-identified lists were sent to an independent survey contracter, the Development and Policies Research Center (DEPOCEN).

In step 4, DEPOCEN contacted provided cellphone numbers, gathered consent, and interviewed consented participants. DEPOCEN collected the data and sent the data files to Vietnam Initiative. Patients did not receive any financial incentives or training for these interviews. In step 5, Vietnam Initiative analyzed the data, estimated the patient satisfaction index per participating hospital, as well as disseminated the policy reports to stakeholders and participants.

Patients or their representatives (who accompanied the patients during their hospital treatment) were followed up with shortly after being discharged (2 to 5 months) and asked about their experiences and assessments of the overall satisfaction and specific attributes of their hospital services. The telephone survey completion rates are 95% of contacted patients in 2017 and 89% of contacted patients in 2018. These response rates are deemed acceptable for data collected from a large-scale and nationwide setting. On the other side, wrong telephone numbers and inability to reach out to the provided phone numbers accounted for 40% and 55.5% of all failed contacts in 2017 and in 2018, respectively. A typical phone call lasted from 7 to 10 minutes to collect responses on 12 key questions and some demographics of patients. Finally, data from 2,927 inpatients in 2017 and 7,562 inpatients in 2018 were successfully obtained by DEPOCEN. Among these 10,489 participants, 96.7% (10,143) had responded to all survey fields necessary for this study.

Measures

To examine the relationship between patient satisfaction levels and patients’ characteristics, two sets of outcome measures were utilized in in the first analysis: (1) continuous scores on overall patient satisfaction levels and (2) three constructed binary indicators based on these reported scores. The patients’ continuous scores reflect the degree to which patients’ expectation and needs were adequately met. The second measure of overall patient satisfaction levels is a constructed binary indicator to indicate whether a patient perceived at least a threshold percent of their expectations and needs were met. We constructed three binary indicators for patient satisfaction based on patients’ overall scores using three thresholds (80%, 90%, and 100%) based on the data distribution; i.e. whether a patient perceived at least 80% of their expectations and needs, at least 90% of their expectations and needs, or at least 100% of their expectations and needs. On average, 75% of patients in our sample reported at least 80% of their expectations and needs were met.

When examining patient satisfaction on three specific attributes of inpatient settings in the second analysis, we used patients’ ratings on several aspects of hospital staff, facilities, and service costs. Similar to prior work on patient satisfaction [6], a 5-point scale was used for most of these questions in the surveys (except for the question of informal costs), where 1 is the lowest level of satisfaction and 5 is the highest level of satisfaction. Patient satisfaction on hospital staff consists of scores on attitudes of medical staff, access to medical staff, and medical deliveries & instructions. Scores on hospital facilities include ratings on bed & associaries, hospital restrooms, and facility maps & directions. Patient satisfaction on treatment costs comprises scores on treatment costs, transparency in medicines & costs, and informal payments not requested.

Statistical analysis

Using data at the patient-level, we regress each patient satisfaction outcome on: (1) specific admission characteristics, including the length of stay and extra-paid services; (2) patients’ financial conditions including poverty status, employment status, and health insurance; and (3) demographic factors, such as age, gender, race, location, and level of education. For the first measure of the overall patient satisfaction level, the actual reported score, we used the Ordinary Least Square (OLS) models. Standard errors are robust to heteroskedasticity. For binary outcomes, we used logistic models with maximum likelihood estimation. For the 5-point scale ratings (ordered categorical variables) on three specific attributes of inpatient settings, we used ordered logistic regression, which preserves all the information in the 5-point scale, and avoids the need for dichotomization of patient responses.

Each regression includes hospital fixed effects, which address time-invariant unobserved factors that may differ across hospitals and communities. We also control for year fixed effects in order to capture nationwide changes. Standard errors are robust and clustered within hospitals. We performed regression analysis using Stata, version 16.0.

Limitations

Although these hospitals were selected carefully to cover all geographical regions and administrative tiers (central, province, and district), the surveys potentially could not reflect the patient satisfaction of the entirety of Vietnam’s patient populations. For instance, the insurance coverage in our data sample is 94%, which is slightly higher than the average insurance rate of Vietnam (89%). Among public hospitals, the number of involved hospitals is 69 which is fairly limited compared to the over 1,000 public hospitals in Vietnam. Furthermore, despite that private hospitals and clinics have been increasing in number in Vietnam (182 private hospitals vs. 1,183 public hospital) [17], they did not participate to the pilot surveys. Local health stations at the commune level—the third tier administrative country subdivision were not covered by the surveys of this research. These commune health stations provided 16.5% of all hospital beds in 2015 [17], however, they specialize in hygiene, vaccinations, antenatal care, and health education.

The second limitation of the current study is that all patient-reported variables were completed after discharge, creating potential measurement errors. In this context, patients were asked their experience from 2 to 5 months post treatment, which might affect the accuracy of their answers, especially for elderly patients.

Results

Participant characteristics

Table 1 provides basic statistics of the outcome and explanatory variables of 10,143 inpatients who had responded to all survey fields necessary for the regression analysis. The range of the observed overall patient satisfaction level is between 0% and 200%, a higher score indicates a higher level of patient satisfaction (first row). The median value and average score are 85% and 82.9%, respectively. The reported score’ range is usually 0% to 100%, only 10 patients of 10,143 respondents reported a score that exceeds 100% (7 respondents with a 110% and 3 respondents with a 200% score). It implies that these patients perceived that above 100% of their expectation and needs were met (exceedance of complete expectation). The data on three binary measures of overall patient satisfaction indicate that 75.5 percent of all patients reported at least 80% of needs met, 47.5 percent of patients reported at least 90% of their needs met, and only 15.8 percent of all patients perceived at least 100% of their needs met.

Table 1

Summary statistics.

This table reports the descriptive statistics for variables of 9,961 patients (or their relatives) with complete data from 2017 and 2018 piloted interviews in Vietnam.

	Mean	Std. Dev	Median	Min	Max
Measures of Patient Satisfaction Level
Overall Patient Satisfaction Level
Percent of overall needs met [0-200]	82.9	(14.8)	85	0	200
Percent of patients with ≥80 of needs met [0-100]	75.5	(43.0)	100	0	100
Percent of patients with ≥90 of needs met [0-100]	47.5	(49.9)	0	0	100
Percent of patients with ≥100 of expectations [0-100]	15.8	(36.5)	0	0	100
Patient Satisfaction on Hospital Staff
Attitudes of medical staff [1-5]	4.14	(0.81)	4	1	5
Expertise of medical staff [1-5]	4.10	(0.74)	4	1	5
Access to medical staff [1-5]	4.08	(0.73)	4	1	5
Medical deliveries & instructions [1-5]	4.17	(0.66)	4	1	5
Patient Satisfaction on Hospital Facilities
Bed & associaries [1-5]	3.92	(0.85)	4	1	5
Hospital restrooms [1-5]	3.69	(0.99)	4	1	5
Facility maps/directions [1-5]	4.06	(0.70)	4	1	5
Patient Satisfaction on Treatment Costs
Treatment costs [1-5]	3.82	(0.77)	4	1	5
Transparency in medicines & costs [1-5]	4.01	(0.77)	4	1	5
Informal payments not requested [0-1]	0.91	(0.28)	1	0	1
Sociodemographic characteristics
Age [18-99]	41.1	(13.3)	38	18	99
Male [0-1]	0.46	(0.50)	0	0	1
Rural [0-1]	0.54	(0.50)	1	0	1
Race (Kinh) [0-1]	0.92	(0.27)	1	0	1
Postgraduate education [0-1]	0.018	(0.13)	0	0	1
College [0-1]	0.14	(0.35)	0	0	1
Vocational training [0-1]	0.13	(0.34)	0	0	1
High school [0-1]	0.22	(0.41)	0	0	1
Farmer/fisher/salt farmer [0-1]	0.17	(0.38)	0	0	1
Government sector [0-1]	0.15	(0.36)	0	0	1
Private companies/FDI [0-1]	0.20	(0.40)	0	0	1
Small business [0-1]	0.12	(0.33)	0	0	1
Hourly worker [0-1]	0.15	(0.36)	0	0	1
Retired, social beneficiaries [0-1]	0.078	(0.27)	0	0	1
Unemployed, students [0-1]	0.12	(0.32)	0	0	1
Health and financial conditions
Health insurance [0-1]	0.94	(0.24)	1	0	1
Poor [0-1]	0.13	(0.34)	0	0	1
Days of hospital stay, days/stay	10.7	(24.1)	7	0	1095
Extra service fees [0-1]	0.24	(0.43)	0	0	1
Observations (patient×year)	10,143

Regarding the specific attributes of hospital services, patients are least likely to be satisfied with the hospital facilities and treatment costs. Particularly, the average score for hospital restrooms is 3.69 (where 5 is the highest level of satisfaction), the average score for bed and associaries is 3.92, and the average score for treatment costs is 3.82. A typical patient is most happy with medical staff: 4.08 for access to medical staff, 4.10 for expertise of medical staff, and 4.17 for medical deliveries & instructions.

The average age of patients was 41.1 years old and patients from rural areas accounted for 54%. These demographic and patient characteristics are relatively representative of typical Vietnam patient populations. A typical patient stayed 7 days in hospitals, which is close to the average days per inpatient (6.7 days in 2015) [17]. Patients from minority groups accounted for only 8% of all participating patients while these people consisted for 13% of the country’s population [18]. Most respondents of the surveys had health insurance (94%) under their treatment, which were slightly higher than the insurance rate of Vietnam (89%). Poor patients accounted for 13% of respondents. Furthermore, 24% of patients voluntarily paid extra service fees during their hospital stays.

Factors associated with overall patient satisfaction

Two OLS models were used to examine the association between overall patient satisfaction scores and patient characteristics. Table 2 reports linear predictors of the overall patient satisfaction score using two OLS specifications: (1) health and financial conditions only and (2) health, financial conditions, and demographics. Without controlling for sociodemographic characteristics, it was found that length of stay and poverty level of income are positively associated with the overall patient satisfaction score. For instance, a typical poor patient tends to report 1.9 percentage points higher in the overall satisfaction score than non-poor patients (p<0.01, Model 1). Insurance coverage is not a significant predictor of the overall patient satisfaction score. When controlling for rurality, employment sector, education level, gender, race, and age, these health and financial conditions are no longer significant predictors of the overall satisfaction score at the 5% level. If possessing the same values on the other predictors, rural patients tend to have 2.43 percentage points higher in satisfaction score than urban patients (p<0.01, Model 2). On average, female patients reported 1.31 percentage points higher in satisfaction score than male counterparts (p<0.01, Model 2). Patients older than 50 years of age also reported 1.79 percentage points higher than younger patients (p<0.05, Model 2). In addition, higher level of education is negatively associated with the overall satisfaction score. We also found that patients, who work in the foreign-investment and enterprise sectors, tend to have lower satisfaction levels compared to employees in government sectors, students, and social beneficiaries.

Table 2

Factors associated with overall patient satisfaction rates.

OLS Estimates on Overall Satisfaction. Standard errors are robust and clustered within hospitals.* p<0.1 ** p<0.05 *** p<0.01.

Dep. Variable = overall patient satisfaction level	(1)	(2)
Health insurance [0-1]	1.28+	1.22
	(0.75)	(0.75)
Length of stay (4-7 days) [0-1]	0.43	0.28
	(0.41)	(0.40)
Length of stay (8-20 days) [0-1]	1.51**	1.12*
	(0.46)	(0.43)
Length of stay (≥21) [0-1]	0.79	0.50
	(0.65)	(0.64)
Extra service fees [0-1]	0.022	0.67+
	(0.36)	(0.36)
Poor [0-1]	1.90***	1.17*
	(0.49)	(0.49)
Rural [0-1]		2.43***
		(0.49)
Government sector [0-1]		2.93***
		(0.55)
Small business [0-1]		0.67
		(0.55)
Farmer/fisher/salt farmer [0-1]		1.93***
		(0.50)
Hourly worker [0-1]		1.28*
		(0.49)
Retired, social beneficiaries [0-1]		1.48+
		(0.80)
Unemployed, students [0-1]		1.80**
		(0.56)
Postgraduate education [0-1]		-5.15***
		(1.04)
College [0-1]		-3.75***
		(0.35)
Vocational training [0-1]		-1.39**
		(0.42)
Male [0-1]		-1.31***
		(0.36)
Age (≥50) [0-1]		1.79**
		(0.53)
Race (Kinh) [0-1]		-0.67
		(0.57)
Dep. Variable Mean	82.93	82.93
Dep. Variable SD	14.83	14.83
Observations	10,143	10,143
Adj R-squared	0.07	0.09

Table 3 presents the predictors of the prevalence of whether a patient perceived at least a threshold percent of whether their expectations and needs were met. We obtain similar results using this maximum likelihood estimation method, although some results are sensitive to the selected threshold. We used 80% as a baseline threshold and also reported two alternative thresholds (90% in Model 2 and 100% in Model 3). The odds ratios indicate that health insurance is positively associated with the patient satisfaction level; the odds of insured patients of being satisfied (more than 80% of their needs met) are 31% higher than non-insured patients (p<0.05, Model 1). Patients who paid extra service fees are more likely to perceive higher levels of needs fulfilled (at least 80% of their needs, by 12%, p<0.05). Patients, who had a longer length of stay (8-20 days), had a higher level of satisfaction than their peers with shorter stays. When using a higher threshold, these factors, including health insurance, extra fees, and length of stay, are no longer statistically significant (Models 2 and 3). The odds of poor patients being satisfied, on two indicators of satisfaction with higher thresholds, are about 16-25% higher than their peers (two indicators statistically significant at least at the 5% level in Models 2 and 3). In all three models, the odds of a rural patient being satisfied are about 26-36% greater than for urban patients (p<0.01). In the same vein, the odds of being satisfied are positively associated with lower levels of education, female gender, and older age (statistically significant at least at the 5% in all three Models).

Table 3

Factors associated with Likelihood to have expectations met.

The odds ratios of logistic regressions, which include hospital fixed effects and year fixed effects, were reported. Standard errors are robust and clustered within hospitals.* p<0.1 ** p<0.05 *** p<0.01.

	(1)	(2)	(3)
Dep. Variable = overall patient satisfaction level	≥80%	≥90%	≥100%
Health insurance [0-1]	1.31**	1.06	1.16
	(0.16)	(0.11)	(0.14)
Length of stay (4-7 days) [0-1]	1.11	1.01	1.07
	(0.077)	(0.066)	(0.10)
Length of stay (8-20 days) [0-1]	1.18**	1.09	1.07
	(0.087)	(0.072)	(0.098)
Length of stay (≥21) [0-1]	1.10	1.11	1.02
	(0.11)	(0.11)	(0.15)
Extra service fees [0-1]	1.12**	1.03	1.13
	(0.061)	(0.057)	(0.093)
Poor [0-1]	1.17*	1.16**	1.25***
	(0.096)	(0.076)	(0.087)
Rural [0-1]	1.36***	1.29***	1.26***
	(0.090)	(0.073)	(0.091)
Government sector [0-1]	1.43***	1.39***	1.28**
	(0.12)	(0.10)	(0.16)
Small business [0-1]	1.17*	1.08	1.16
	(0.10)	(0.088)	(0.15)
Farmer/fisher/salt farmer [0-1]	1.25***	1.20**	1.65***
	(0.099)	(0.11)	(0.18)
Hourly worker [0-1]	1.11	1.12	1.44***
	(0.092)	(0.084)	(0.18)
Retired, social beneficiaries [0-1]	1.26*	1.12	1.54***
	(0.16)	(0.12)	(0.24)
Unemployed, students [0-1]	1.21**	1.11	1.57***
	(0.11)	(0.091)	(0.18)
Postgraduate education [0-1]	0.39***	0.43***	0.40***
	(0.068)	(0.088)	(0.10)
College [0-1]	0.62***	0.56***	0.43***
	(0.047)	(0.038)	(0.049)
Vocational training [0-1]	0.81***	0.81***	0.65***
	(0.057)	(0.052)	(0.062)
Male [0-1]	0.80***	0.80***	0.72***
	(0.045)	(0.040)	(0.051)
Age (≥50) [0-1]	1.19**	1.28***	1.47***
	(0.092)	(0.078)	(0.13)
Race (Kinh) [0-1]	0.87	0.86**	0.90
	(0.094)	(0.067)	(0.11)
Dep. Variable Mean	0.75	0.48	0.16
Dep. Variable SD	0.43	0.50	0.37
Observations	10,143	10,143	10,143
McFadden’s Adj R-squared	0.05	0.04	0.05

Perceptions on medical staff, facilities, and costs

We further examine the determinants of patient perceptions on whether or not their needs for medical practitioners were met in Table 4. The results suggest that health insurance is a positive determinant of patients’ perception of their hospital staff. Particularly, insured patients are significantly more likely to express higher scores for staff expertise and treatment instructions (29% higher in the odds of reporting higher levels of satisfaction and p<0.05 in Models 2; 27% higher in the odds and p<0.05 in Model 4). Long-stay patients also perceived higher levels of satisfaction regarding attitudes, accessibility, and instructions than short-stay patients (Models 1, 3, and 4). For instance, the odds of reporting higher levels of satisfaction on staff attitudes for patients who stayed 8-20 days and more than 3 weeks in hospitals are 25% and 33% higher than short-stay patients (less than 4 days), respectively (p<0.01 and p<0.05, Model 1). In addition, patients who paid extra service fees reported to have higher levels of needs fulfilled regarding hospital staff (staff attitudes and instructions) than other patients (p<0.05 in Models 1 and 4). Similar to the reaction to the overall satisfaction level, poor patients also expressed higher scores on the staff dimensions of their service experience (staff attitudes, staff expertise, and instruction). The demographic factors provide similar predictions on patient perceptions about medical staff as the reactions to the overall satisfaction level. Specifically, female patients, patients older than 50 years of age, rural patients, and those with only primary level education expressed higher scores for medical staff.

Table 4

Factors associated with patient satisfaction on hospital staff.

The odds ratios of ordered logistic regressions, which include hospital fixed effects and year fixed effects, were reported. Standard errors are robust and clustered within hospitals.* p<0.1 ** p<0.05 *** p<0.01.

	(1)	(2)	(3)	(4)
Dep. Variable = patient satisfaction regarding	Attitudes of staff	Expertise of staff	Access to staff	Medical Instructions
Health insurance [0-1]	1.11	1.29**	1.14	1.27**
	(0.082)	(0.13)	(0.097)	(0.15)
Length of stay (4-7 days) [0-1]	0.98	0.98	0.98	1.06
	(0.059)	(0.052)	(0.067)	(0.067)
Length of stay (8-20 days) [0-1]	1.25***	1.10	1.15*	1.18**
	(0.090)	(0.067)	(0.090)	(0.088)
Length of stay (≥21) [0-1]	1.33***	1.11	1.20**	1.29***
	(0.14)	(0.095)	(0.10)	(0.12)
Extra service fees [0-1]	1.11**	1.03	1.10*	1.11**
	(0.057)	(0.056)	(0.056)	(0.060)
Poor [0-1]	1.17**	1.19***	1.14*	1.18**
	(0.079)	(0.075)	(0.082)	(0.081)
Rural [0-1]	1.17***	1.20***	1.31***	1.18***
	(0.066)	(0.061)	(0.071)	(0.056)
Government sector [0-1]	1.23***	1.25***	1.42***	1.25***
	(0.080)	(0.066)	(0.087)	(0.081)
Small business [0-1]	0.95	0.95	0.97	1.04
	(0.072)	(0.063)	(0.067)	(0.066)
Farmer/fisher/salt farmer [0-1]	1.09	1.23***	1.27***	1.18**
	(0.075)	(0.085)	(0.10)	(0.088)
Hourly worker [0-1]	1.06	1.04	1.18**	1.10
	(0.071)	(0.078)	(0.084)	(0.081)
Retired, social beneficiaries [0-1]	1.04	1.01	1.21*	1.01
	(0.12)	(0.11)	(0.12)	(0.11)
Unemployed, students [0-1]	1.12	1.13	1.27***	1.10
	(0.081)	(0.082)	(0.098)	(0.10)
Postgraduate education [0-1]	0.69***	0.67***	0.68**	0.68**
	(0.098)	(0.091)	(0.12)	(0.13)
College [0-1]	0.68***	0.74***	0.71***	0.72***
	(0.039)	(0.050)	(0.043)	(0.047)
Vocational training [0-1]	0.78***	0.80***	0.76***	0.84***
	(0.048)	(0.053)	(0.045)	(0.052)
Male [0-1]	0.88***	0.88***	0.89***	0.94
	(0.032)	(0.041)	(0.040)	(0.039)
Age (≥50) [0-1]	1.44***	1.24***	1.58***	1.24***
	(0.085)	(0.071)	(0.11)	(0.084)
Race (Kinh) [0-1]	0.99	0.79***	0.93	0.84*
	(0.079)	(0.070)	(0.088)	(0.083)
Dep. Variable Mean	4.14	4.10	4.08	4.17
Dep. Variable SD	0.81	0.74	0.73	0.66
Observations	10,143	10,143	10,143	10,143

Table 5 provides odds ratios of our ordered logistic regressions on patient perceptions on hospital living arrangements and amenities. Insured patients do not statistically differ from non-insured patients when rating their experience with hospital beds, restrooms, and maps/directions. Long-stay patients (8-20 days) perceived a significantly higher level of satisfaction regarding hospital beds than their short-stay peers (22% higher in the odds of reporting a higher level of satisfaction, p<0.05 in Model 1). Patients who paid extra service fees are much more likely to be satisfied with hospital beds and restrooms. In particular, the odds of these patients reporting higher levels of satisfaction with hospital beds and restrooms are 54% and 31% higher than patients who did not pay extra service fees, respectively (p<0.01 in both Models 1 and 2). Overall, poor patients did not express significantly higher scores on hospital facilities when using their healthcare service in these hospitals. Similar to patient’s reaction to the overall satisfaction level, patients older than 50 years of age, rural patients, and those with only primary level education expressed higher satisfaction with hospital facilities, although female patients are not more likely to report higher levels of satisfaction on hospital facilities than their male peers.

Table 5

Factors associated with patient satisfaction on hospital facilities.

The odds ratios of ordered logistic regressions, which include hospital fixed effects and year fixed effects, were reported. Standard errors are robust and clustered within hospitals.* p<0.1 ** p<0.05 *** p<0.01.

	(1)	(2)	(3)
Dep. Variable = patient satisfaction regarding	Hospital beds	Restrooms	Maps/directions
Health insurance [0-1]	1.10	1.01	1.16
	(0.093)	(0.073)	(0.12)
Length of stay (4-7 days) [0-1]	1.05	0.93	0.93
	(0.069)	(0.050)	(0.065)
Length of stay (8-20 days) [0-1]	1.22**	1.05	1.07
	(0.089)	(0.064)	(0.076)
Length of stay (≥21) [0-1]	1.23+	1.20*	1.12
	(0.13)	(0.11)	(0.099)
Extra service fees [0-1]	1.54***	1.31***	0.97
	(0.12)	(0.082)	(0.051)
Poor [0-1]	1.19*	1.14*	1.05
	(0.084)	(0.071)	(0.076)
Rural [0-1]	1.16**	1.22***	1.22***
	(0.062)	(0.069)	(0.073)
Government sector [0-1]	1.37***	1.42***	1.32**
	(0.093)	(0.094)	(0.13)
Small business [0-1]	1.02	1.11	1.12+
	(0.070)	(0.076)	(0.077)
Farmer/fisher/salt farmer [0-1]	1.32***	1.39***	1.32***
	(0.083)	(0.098)	(0.099)
Hourly worker [0-1]	1.15*	1.19*	1.09
	(0.069)	(0.089)	(0.084)
Retired, social beneficiaries [0-1]	1.08	1.13	1.22*
	(0.12)	(0.13)	(0.12)
Unemployed, students [0-1]	1.30***	1.22**	1.17*
	(0.095)	(0.089)	(0.092)
Postgraduate education [0-1]	0.65**	0.55***	0.40***
	(0.10)	(0.086)	(0.092)
College [0-1]	0.65***	0.61***	0.63***
	(0.036)	(0.034)	(0.044)
Vocational training [0-1]	0.79***	0.76***	0.90
	(0.041)	(0.038)	(0.063)
Male [0-1]	0.97	1.04	0.92+
	(0.039)	(0.049)	(0.043)
Age (≥50) [0-1]	1.28***	1.33***	1.19**
	(0.080)	(0.073)	(0.071)
Race (Kinh) [0-1]	0.92	1.00	0.96
	(0.081)	(0.098)	(0.090)
Dep. Variable Mean	3.92	3.69	4.06
Dep. Variable SD	0.85	0.99	0.70
Observations	10,143	10,143	10,143

Treatment costs and transparency in stated treatment costs are important aspects of patient satisfaction. We further explore the determinants of whether a patient felt satisfied with treatment costs in Table 6. Health insurance is a strong and positive predictor on patient satisfaction on treatment costs. The odds of insured patients being satisfied with treatment costs and transparency in cost information are 57% and 32% higher than uninsured patients, respectively (p<0.01 and p<0.05 in Models 1-2, respectively). Poor patients did not express significantly lower scores on hospital costs. As expected, patients who paid extra service fees are less likely to be satisfied with treatment costs (11% lower in the odds of reporting high levels of satisfaction on treatment costs, p<0.05 in Model 1). There is moderate evidence that these patients are more likely to be demanded to pay informal payments to hospital staff (p<0.10 in Model 3). Additionally, long-stay patients (only patients with 4-7 days and more than 21 days) are less likely to pay informal payments to hospital staff than their short-stay peers (31% lower in the odds of paying informal payments and p<0.01; 38% in the odds and p<0.05, respectively in Model 3). Patients older than 50 years of age are less likely to pay informal payments to hospital staff.

Table 6

Factors associated with patient satisfaction on treatment costs.

The odds ratios of ordered logistic regressions, which include hospital fixed effects and year fixed effects, were reported. Standard errors are robust and clustered within hospitals.* p<0.1 ** p<0.05 *** p<0.01.

	(1)	(2)	(3)
Dep. Variable = patient satisfaction regarding	Treatment costs	Transparency in medicines & costs	Informal Costs requested [0-1]
Health insurance [0-1]	1.57***	1.32**	0.98
	(0.17)	(0.14)	(0.20)
Length of stay (4-7 days) [0-1]	1.00	1.04	0.69***
	(0.064)	(0.066)	(0.083)
Length of stay (8-20 days) [0-1]	1.14*	1.15**	0.83
	(0.083)	(0.078)	(0.11)
Length of stay (≥21) [0-1]	1.04	1.30***	0.62**
	(0.096)	(0.13)	(0.12)
Extra service fees [0-1]	0.89**	1.08	0.79*
	(0.048)	(0.058)	(0.11)
Poor [0-1]	1.06	1.12*	1.02
	(0.069)	(0.072)	(0.12)
Rural [0-1]	0.96	1.17***	0.94
	(0.048)	(0.066)	(0.090)
Government sector [0-1]	1.06	1.23***	0.96
	(0.062)	(0.067)	(0.14)
Small business [0-1]	0.95	0.97	0.84
	(0.072)	(0.064)	(0.10)
Farmer/fisher/salt farmer [0-1]	1.09	1.09	1.03
	(0.085)	(0.085)	(0.17)
Hourly worker [0-1]	0.98	1.00	0.89
	(0.067)	(0.062)	(0.12)
Retired, social beneficiaries [0-1]	0.86	0.93	0.94
	(0.097)	(0.081)	(0.19)
Unemployed, students [0-1]	1.09	1.07	0.94
	(0.087)	(0.065)	(0.17)
Postgraduate education [0-1]	1.06	0.61***	0.80
	(0.15)	(0.092)	(0.17)
College [0-1]	0.86**	0.69***	1.07
	(0.057)	(0.041)	(0.098)
Vocational training [0-1]	1.02	0.75***	1.14
	(0.059)	(0.046)	(0.14)
Male [0-1]	0.90**	0.89***	1.13
	(0.042)	(0.037)	(0.097)
Age (≥50) [0-1]	0.96	1.17***	1.56***
	(0.060)	(0.067)	(0.18)
Race (Kinh) [0-1]	0.79**	0.99	0.95
	(0.079)	(0.066)	(0.16)
Dep. Variable Mean	3.82	4.01	0.91
Dep. Variable SD	0.77	0.77	0.28
Observations	10,143	10,143	10,143

Discussion

Using novel, nationwide, and large-scale patient interview data, this paper examines the associations of the patient satisfaction level and various patient characteristics, including service-specific characteristics, patients’ financial conditions, and demographic factors. Compared to prior work on patient satisfaction in Vietnam, we found a higher level of patient satisfaction using a more representative sample of patients and close relatives. The average patient satisfaction score was 82.9 percent of patients’ expectations and needs met. The average scores on different attributes of healthcare settings varied from 3.69/5 (lowest for hospital restrooms) to 4.17/5 (highest for medical deliveries and instruction). Using a smaller sample of patients in the North of Vietnam, Nguyen and Nguyen (2014) suggested the average satisfaction score as 3.68/5 while another study reported a lower score of 3.55/5 from a Hanoi-based group of patients [6, 7].

Our findings suggest that older patients, low-income patients, female patients, those with only primary level education, patients not working for private enterprises (or foreign enterprises), and rural patients reported higher levels of overall satisfaction. Some of these findings are contradictory to that of similar studies in Vietnam. For example, Vuong et al. (2017) found that patient satisfaction level is positively associated with patient characteristics such as younger age and higher-level of income [6]. Interestingly, our findings are consistent with some findings from prior studies in some developed countries regarding age and levels of education. For instance, two studies, one conducted in Scotland looking at 650 patients discharged from four acute care general hospitals during February and March 2002, and the second study conducted in 32 different large tertiary hospitals in the USA; both showed that patients older than 50 years of age, patients who had a shorter length of stay or better health status, and those with only primary level education reported higher satisfaction scores related to variable health service-related domains [13, 19]. Batbaatar et al. (2017) also suggested that the majority of their reviewed studies concluded that older patients were more satisfied with health services than younger patients [12]. In addition, our study suggested that female patients reported a higher level of satisfaction, which is different from the results in several studies [13, 19].

Health insurance is found to have positive influence on overall patient satisfaction (the low threshold—80 percent of expectations and needs met), primarily driven by limiting patient concerns about treatment costs, as well as increasing positive perceptions of hospital staff. This result is consistent with the finding of positive influence of insurance coverage on improving patient satisfaction in another study in Vietnam [9]. We further find that patients who paid extra fees for their hospital admission expressed higher scores for hospital living arrangements (hospital beds and restrooms) and medical staff (attitudes and instructions), but were mostly dissatisfied with treatment costs.

Similar to a study conducted in a public hospital in France [20], patients in our study reported that hospital living arrangements and amenities were their major dissatisfaction. For instance, the average score for hospital restrooms is 3.92 and the score for hospital beds is 3.92 (where 5 is the highest level of satisfaction), respectively. This exploration suggests that there is room for improvement in the public hospital environment to improve patient satisfaction and quality of care.

These findings suggest that Vietnam is achieving important steps moving forward toward universal health coverage while improving patient satisfaction. Particularly, prior to the 1980s, the government provided basic healthcare for everyone in Vietnam, until a policy shift toward a mixed market economy, leading to the inclusion of a private healthcare sector and the introduction of fee-for-service in public hospitals [21]. Vietnam has utilized a single-scheme social health insurance since 1992, first covering low-income patients and formal-sector workers (workers in jobs with normal hours and regular pay), then extending to other populations. As of May 2019, 89% of Vietnamese populations (84 million people) were covered by this social health insurance system [22]. Private households pay for their health insurance premiums, with varying government subsidies based on the enrollee’s economic status. Understanding patient satisfaction in public hospitals, the dominant healthcare clinics in Vietnam, is becoming increasingly important in this mixed public-private healthcare system. This finding suggests that Vietnam is achieving important steps moving forward toward universal health coverage while improving patient satisfaction. It is worth noting that our sample is slightly over-represented by insured patients (94% compared to the overall 89% insurance rate of Vietnam by May of 2019). Although most patients in this sample were insured, the average satisfaction score for treatment costs is 3.82, one of the lowest average scores compared to other multidimensional attributes of healthcare settings. Vietnam’s policy shift toward a mixed market economy and the introduction of fee-for-service in public hospitals may play an important role in explaining this major dissatisfaction.

Patient satisfaction has been considered a mandatory barometer to evaluate how well a healthcare system is working in a number of advanced countries such as France and Germany, however, it receives limited attention in developing countries including Vietnam, where few interventions include patient satisfaction as a component. Remarkably, the most comprehensive patient surveys in Vietnam were conducted by an independent, non-governmental, non-political, and academic institution in 2017 and 2018, as the first attempt to measure patient satisfaction at a nationwide scale. Using data from these surveys, this study provides better understanding of patient satisfaction in Vietnam and identifies which sub-populations of patients and specific aspects of healthcare settings require additional attention.

2 Dec 2019

PONE-D-19-21890

Determinants of Patient Satisfaction: Lessons from Large-Scaled Patient Interviews in Vietnam

PLOS ONE

Dear Dr Nguyen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Jan 16 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

James M. Lightwood

Academic Editor

PLOS ONE

Additional Editor Comments:

Thank you for your submission. Both reviewers think that the manuscript is a potentially interesting and valuable contribution and worthy of publication, but also think it requires a major revision. I agree, and have requested a major revision.

Reviewer 1 presents a clear and concise list of comments and suggestions on methodological issues, and on format and presentation. I suggest each one of Reviewer 1's points be carefully addressed. Reviewer 2 agrees with some of Reviewer 1's points, and where both reviewers agree, those points should be given very careful attention. Reviewer 1 brings up an issue that escaped my notice, which is that there are no line numbers in the submission. Please ensure line numbers are in the revised submission and if you have difficulty with the PLoS ONE submission system in producing them in the revised manuscript, please contact the production staff for assistance.

Reviewer 2 raises two important concerns. The first concern is that the manuscript is vague on the primary source of the data. I interpreted the manuscript as saying that the primary data reported in the main results are from the MSA, but Reviewer 2 interprets it as saying that the data come from several sources. Please clarify this issue. Reviewer 2's second concern, based on their experience with survey research and data collection in Vietnam, is that MSA data may be biased due to influence of data collection personnel on patient responses. Please provide a careful response to Reviewer 2 on this point. If you agree that bias in the MSA data might be an important issue, I wonder if you have enough data from other sources to perform a sensitivity analysis or check on the primary source of data.

Finally, there are two important issues regarding the statistical methodology used, and for which set of data. The manuscript reports use of data that use two different scales for two different outcome variables. The first is a patient assessment of the degrees to which their expectations were met, which can exceed 100 percent; the second is a 5 point rating scale of patient satisfaction, which is converted to a 0-1 binary scale to indicated 'unsatisfied' versus 'satisfied' (Methods section page 6). However, the Results section appears to only report the second outcome measure. This issue need to be clarified. I don't see where the results on first set of outcome measures are reported. Were the measures of degree to which patient expectations were met converted to a 0-1 binary scale as well, and reported with the other type of outcome variable? Reviewer 2 seems to say that the first set of outcomes is reported, and wonders how patient satisfaction can exceed 100 percent. But I can't find where analysis for this first set of outcome is reported, and I admit I am guessing that what you mean for this set of outcomes is patient reports of the degree to which their expectations are met (so it can exceed 100 percent). Please provide more detail on the definition of this set of outcomes, and make sure it is clear to reader which set of outcomes is reported in each section of the Results.

Second, I interpret the manuscript as saying the primary results reported use logistic regression, and ordinary least squares (OLS) was used for a sensitivity analysis that is reported in a supplementary table (Results section, page 9) . However I don't see the supplementary table with the submission. If the supplementary table is to remain unpublished, then the authors should state that. If not, please ensure that any supplementary material is included in the revised submission. In any case, the results of the OLS sensitivity analysis should be described in the main text, and I can't find where it is. Reviewer 2 interprets the manuscript as reporting some main results are from the OLS regression, and says that these results should be omitted, or the analysis changed to use only logistic regression, since logistic regression is the only correct regression method. However, all the main results seem to be from logistic regression to me. So, please clarify this issue.

Regarding the issue of logistic regression versus OLS for binary data, I don't completely agree with Reviewer 2. In some cases OLS is an acceptable regression approach as well as logistic regression, and can be more robust. All the explanatory variables used for your analysis appear to be 0-1, and you have a relatively large number of observations, so some objections to OLS don't hold. The fact that all the explanatory variables are 0-1 should keep the predicted probabilities from the estimated model between 0 and 1, for example.

The question is how to compare the results from logistic and OLS regressions, since logistic regression and OLS report the coefficients on a different scale. One method is to pick a reference slope coefficient for the logistic and OLS regressions, and find the ratio of the other coefficients to the reference coefficient for the logistic and OLS regressions, respectively. The ratios should be approximately the same for each regression. The second approach is to compare the predictions of the logistic and OLS regressions. If both methods are appropriate for your data set, both should produce similar predictions, and the correctly and incorrectly classified observations should have substantial overlap. If the two methods produce substantially different predictions, the method with the best overall predictive performance should be preferred. Also, you may want to try ordered logit regression, which preserves all the information in the 5 point scale, and avoid the need for dichotomization of patient responses. Standard errors that are robust to heteroskedasticitiy should be used for the OLS regression. A routine for ordered logit regression should be provided in major statistical packages, such as SAS and Stata. These issues are discussed at length in the text by Jeffrey Wooldridge, Econometric Analysis of Cross Section and Panel Data, MIT Press, second edition, chapters 15 and 16 (I think the information is in the first edition too, but not sure in which chapters).

The authors should state what statistical package was used for the estimation (SAS, SPSS, Stata, something else?).

I also agree with both reviewers that much of the content in the manuscript is in the wrong section. For example, describing the OLS sensitivity analysis should be in the Methods section, not the Results section. Also, the descriptive statistics of the data are presented in a 'Descriptive information' section. I believe that unless there was something unexpected in the data that lead to a change in analysis, the descriptive statistics are customarily reported in the Results section.

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. Please upload a copy of Supporting Information S1 Table which you refer to in your text on page 16.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: General comments: This study is designed to explore determinants of Vietnamese patient satisfaction of inpatient treatment. It used a large scale of sample, that could be representative for Vietnamese population. I would recommend publishing this manuscript after revising. I have a few suggestions for improving clarity in a few places, highlighted below:

- Format:

o There is no line number as requested by the journal. It is a bit difficult to review and cite where are the issues.

o The section Conclusion should be renamed as Discussion.

- Page 3, paragraph 1: Relevant literature about correlation between patient characteristics with patient satisfaction is not sufficient, the author just focused on educational level.

- Page 5: this content is repeated at the methods section and should be in the methods section

- Page 7, 8:

o Providing a diagram could be helpful to understand the sampling approach and data.

o How long for the interview? Did participants receive any thing for their participation? What is the relationship between the research team and the surveyors? Did the surveyors receive training for interview?

o What is the software used for data analysis?

- Page 9:

o The insurance rate of Vietnam should be updated as it has been increasing significantly recent years

o Text in result should not repeat figures that mentioned in the tables

o Table 1: binary variables should be presented with percentages, not MEAN. I am a bit confused about the value of the over patient satisfaction level. Please explain how you have the overall patient satisfaction level: overall needs met with min/ max value are 0/200. This should state in the method section.

- Page 10:

o From line 19: should be presented in the discussion section

- The outcomes should be compared with similar studies in Vietnam and in other countries.

- Page 15: “Female, poor, rural patients, students, and social beneficiaries tend to have higher levels of patient satisfaction in public hospitals, the primary healthcare providers in Vietnam. These findings suggest that the public healthcare system in Vietnam still meets the medical needs for relatively vulnerable patient populations” � the language here is quite personal view, lack of evidence from this part of result to contribute for this statement. Those factors could affect to medical literacy. Patient expectations and patient satisfaction is then impacted by medical literacy level.

Reviewer #2: 15 Nov 2019

To: Authors

RE: Manuscript # PONE-D-19-21890, entitled "Determinants of Patient Satisfaction: Lessons from Large-Scaled Patient Interviews in Vietnam"

The manuscript was developed based on a recent large-scale survey in recent times, it has some merit that may be considered for publication. However, while I was reading the manuscript, I had some following comments as suggestions for improvement:

1. Introduction:

- much information referenced in this part or even in other parts of the manuscript is not updated (many cited materials were published around 2000-2004)

- in page 3 (line 10-13), I was confused about a statement "Although studies on patient satisfaction have been increasing in developing countries, publications using high-quality and large-scale patient surveys are still scarce due to resource and time constraints". When reading this, I expected that I could have an opportunity to see some discussion or supporting evidence from developing countries, yet, all of a sudden, only Turkey (developed country) and Bangladesh case studies were presented. I do believe that's not coherent and convincing as we know so many studies on patients satisfaction on healthcare were conducted and reported in both advanced and less advanced nations.

- in page 3 (line 3-11 from the bottom), instead of just listing prior studies in Vietnam, it's expected to discuss some brief results from such studies (prevalence of satisfaction, associated factors, limitations) to see the existing gaps.

- page 5 (the last paragraph), interviewing 10,489 patients is huge, I just wonder how data collectors handled this: Is it just telephone- or face-to-face-based interview or combined? More descriptions provided in methods section will help readers understand.

2. Materials and Methods:

- The info/descriptions in this part are really mixed up, it's not really clear. I suggest that authors start with describing the design and setting, then methods, data collection and analysis. It's not logical to present the procedure of regression at the beginning of this section.

- There are many pieces of information that confused readers such as pilot, why is what is pilot and why is pilot as I thought the pilot is just something the researchers did before the official study. The process of data collection is not well described such as it was said that the surveys were conducted by the Medical Services Administration (MSA) under VMOH (Vietnam Ministry of Health), Vietnam Initiative of Indiana University and Oxfam Vietnam. I think collecting data in this way could introduced a lot of bias to data. I have been working in Vietnam for more than 20 years (with focusing on research and teaching on health system, health policy and management and with a lot of interactions with policy makers in VMOH at the national and sub-national meetings and conferences; most of people raised the issue of bias about the way data were collected using the hospital system itself (such as nurses and/or doctors wearing blue/uniform clothes can not give patients freedom to report their true responses or even with data collected by MSA may not be a good way to produce valid data because MSA is just one of divisions of MOH, patients and people would be afraid of not reporting positive results in clients/satisfaction surveys because of the intervention or presence of MOH officials or staffs. Over the past years, VMOH and hospital leaders in Vietnam suggested that there be a national independent agent/organization totally independently of VMOH and the healthcare system to conduct the survey on patients satisfaction, but so far it's not feasible due to public budget limitation. In your current surveys, even you had Vietnam Initiative of Indiana Uni and Oxfarm Vietnam involved, but the main contact was still MSA of MOH.

- Also, how to have a patient's contact info is an issue. Did every single hospital asked for patients' informed consent to provide to MSA (MOH) so the surveys were conducted or without any informed consents?If this process was actually done? How was it undertaken?...

- What was the fund of the study: Authors mentioned that there was not fund for this research. However, I guessed that although there was no direct fund for incentives for patients, MOH(VMA) and/or Vietnam Initiative of Indiana Uni and Oxfarm Vietnam may actually provide some financial support for some basis logistic stuff such as contact, travel, data analysis and management, etc) because this study was really costly with alot of surveys and admin processes?

- May be there was a mistake in reporting Cronbach's alpha, as I know that this coefficient is not analysed for each item, but it is done for a group of items (say at least two items or more). If so, please correct this across the methods section as well as the whole paper.

- When authors said 95% and 89% in 2017 and 2018 participated, it's not necessary to remention "Remarkably, only 5% and 10.6% of all contacted patients refused to participate to the 2017 interviews and the 2018 interviews, respectively".

- As this study was a focus on inpatients, may be authors should use inpatients in the title because the experience and satisfaction between inpatients and outpatients are quite different

- 16% of all patients perceived over 100% of their needs met seems impossible?

3. Results:

- As this study, logistic regression was applied for inferential analysis, the OLS is not appropriate because it is only used for linear regression, while logistic regression works with the method of maximum likelihood estimation (MLE).

- As there is no discussion section, it's really difficult to distinguish between results and discussion as the authors mixed them up in one section "results'

- As this study was Vietnam-based context, it's also better to compare with other studies in Vietnam (I could not see much this comparison in this section)

- There is a quite interesting result "female patients, patients older than 50 years of age, rural patients, and those with only primary level education expressed higher scores for medical staff", please explain why or provide some possible assumptions for this as this results was just put forward in the text.

4. Conclusions:

- The first sentence was too lengthy

- The result or conclusion "Female, poor, rural patients, students, and social beneficiaries tend to have higher levels of patient satisfaction in public hospitals, the primary healthcare providers in Vietnam." could be misunderstood if no explanations are provided; perhaps these groups have low expectations and demands.

5. Limitations:

- Two other limitations are 1) cross-sectional design can not determine causality; 2) the nature of surveys conducted by MOH introduces some bias.

- Another limitation could be due to the nature of data collection conducted by MOH/MSA.

6. References:

- Not updated because just 6 among 24 items have been published within 5 recent years.

Overall, I think this manuscript might be published with major revisions required.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

22 Jan 2020

Please find the attached document.

Submitted filename: PlosOne-Response letter - 2019-01-16.docx

Click here for additional data file.

16 Jun 2020

PONE-D-19-21890R1

Determinants of Patient Satisfaction: Lessons from Large-Scaled Inpatient Interviews in Vietnam

PLOS ONE

Dear Dr. Nguyen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

While the manuscript is much improved, Reviewer 1 is of the view that additional revisions are needed to address his/her comments. I look forward to seeing a revised version of the manuscript along with a point by point response.

Please submit your revised manuscript by Jul 31 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

David Hotchkiss

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors tried to address all the reviewers’comments. The revision solved some of the issues that reviewers and editors pointed out and help to improve the manuscript’s quality.

However, in my opinion, this manuscript still needs more work to meet the criteria for publication. Objectives of the study were not presented clearly and concise (line 149-169). The method section was not well organized and scientific. It is very nice that the authors added a flow diagram to help visualize the sampling approach. If the authors put a bit more notice on presenting the study design, data collection and also describing major variables with their value, it would be easier for readers to follow. Such as the overall patient satisfaction level, in line 261 -265 p10, was stated that: “The score' range is usually 0% to 100%. It can exceed 100% if a patient perceived that above 100% of their expectations and needs were met (exceedance of complete expectation).” But in Table 1 p34, the range is [0-200]. Maybe I missed something…

The authors should consider about how to report the dependent variables as the previous comments and then present the results following the major objectives order. I feel confused and have to read several times to understand.

This manuscript has some merit, but I hope the authors could consider carefully about the above points for this manuscript to be ready for publication.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

8 Jul 2020

In response to comments from Reviewer 1:

R1-1. (We will number these comments consecutively in order to refer back to them as needed)

“Objectives of the study were not presented clearly and concise (line 149-169).”

Thank you. We have revised the introduction to present the objectives in a more concise fashion in lines 48-69 and lines 97-103 (the version with track changes). One example is the following:

“This study examines factors associated with inpatient satisfaction levels in Vietnam's public hospitals during 2017-2018, using nationwide, large-scale interview data from randomized and voluntary responses to a survey provided to 69 large and public hospitals. We look at how patient satisfaction levels respond to service-specific characteristics (length of stay and extra-paid services), patients' financial conditions (employment status, health insurance, and income level), and demographic factors (age, gender, race, location and level of education). We further evaluate the associations of these factors with different salient aspects of their healthcare service experience: hospital staff, hospital facilities, and treatment costs.”

R1-2. “The method section was not well organized and scientific. It is very nice that the authors added a flow diagram to help visualize the sampling approach. If the authors put a bit more notice on presenting the study design, data collection and also describing major variables with their value, it would be easier for readers to follow.”

Thank you for this comment. We have made a number of edits in the method section to improve the presentations of our study design and measures. The changes are color-coded in blue (lines 187-190, 236-341, and 266-271).

R1-3. “Such as the overall patient satisfaction level, in line 261 -265 p10, was stated that: “The score' range is usually 0% to 100%. It can exceed 100% if a patient perceived that above 100% of their expectations and needs were met (exceedance of complete expectation).” But in Table 1 p34, the range is [0-200]. Maybe I missed something…” The authors should consider about how to report the dependent variables as the previous comments and then present the results following the major objectives order. I feel confused and have to read several times to understand.

Thank you for raising this point. We made some edits to provide some details of this variable in lines 316 –323, following:

“The range of the observed overall patient satisfaction level is between 0% and 200%, a higher score indicates a higher level of patient satisfaction (first row). The median value and average score are 85% and 82.9%, respectively. The reported score' range is usually 0% to 100%, only 10 patients of 10,143 respondents reported a score that exceeds 100% (7 respondents with a 110% and 3 respondents with a 200% score). It implies that these patients perceived that above 100% of their expectation and needs were met (exceedance of complete expectation).”

Submitted filename: PlosOne-Response letter - 2020-07-05.docx

Click here for additional data file.

4 Sep 2020

Determinants of Patient Satisfaction: Lessons from Large-Scaled Inpatient Interviews in Vietnam

PONE-D-19-21890R2

Dear Dr. Nguyen,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

David Hotchkiss

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

9 Sep 2020

PONE-D-19-21890R2

Determinants of Patient Satisfaction: Lessons from Large-Scale Inpatient Interviews in Vietnam

Dear Dr. Nguyen:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. David Hotchkiss

Academic Editor

PLOS ONE