OUP user menu

Patient experiences with inpatient care in rural China

Heather Sipsma, Yu Liu, Hong Wang, Yan Zhu, Lei Xue, Rachelle Alpern, Martha Dale, Elizabeth Bradley
DOI: http://dx.doi.org/10.1093/intqhc/mzt046 452-458 First published online: 27 June 2013

Abstract

Objective To describe patient experiences with hospital inpatient care among participants living in rural China and to examine their associations with sociodemographic characteristics, hospital type and province.

Design Cross-sectional study using data from questionnaires administered to members of randomly selected households in 2010. We used linear and logistic regression to determine associations between patient ratings of care and key components of their experience and between patient ratings of care and sociodemographic characteristics, hospital type and province.

Setting Households located in seven provinces in rural China.

Participants Household members >15 years who reported being admitted to the hospital within the last 365 days with valid data on our outcome measures (n = 443).

Main Outcome Measures Patient evaluations of health care experiences.

Results Approximately 31% of participants rated their experiences 5.0 out of 5.0 (best), but 22% rated their experiences ≤3.0. Fifteen percent would not recommend the facility to family and friends. Five factors emerged, of which, ‘communication with nurses’ was most strongly and consistently associated with overall patient ratings. Multivariable models showed that ratings for township and county-level hospitals were significantly lower than above county-level hospitals. Variation also existed across the seven provinces.

Conclusions Findings suggest that patients on average have high ratings of hospital care, but a notable proportion of participants, particularly those receiving care in county-level hospitals, continue to be less than fully satisfied. As China further develops its health system, establishing routine monitoring of patients' experiences will be important to ensure the system is responsive to the population needs.

  • quality improvement
  • hospital care
  • patient experiences
  • China

Introduction

During the last decade, the health care system in China has undergone substantial reform with the introduction of the New Cooperative Medical Scheme (NCMS) in 2003. The NCMS sought to improve access to health care services for rural China by providing a community risk-pooling health protection system supported by central, provincial, county and township governments [1]. The NCMS program now covers close to 100% of the rural population, and the total amount that government spent on health care more than doubled between 2006 and 2010 [2]. Evaluations of the NCMS reform have shown notable improvements in access and utilization [35] and high satisfaction with the program [6, 7]; however, little is known about patients' experiences with these services.

Previous research assessing patient experiences with health services in China has focused on urban populations and has not adequately explored patient experiences in rural China. A study of urban patients who had been admitted to a hospital in Hong Kong reported that nearly 22% of patients rated their experience poorly using the Picker Patient Experience Questionnaire-15 (PPE-15), with clinicians showing respect for patients as particularly important to overall ratings [8]. Another study from Hong Kong, China, reported that almost 98% of respondents who had received outpatient care reported a score of 5 or higher out of 10 to describe their overall healthcare experiences [9]. In a third study from the City of Dalian, China, overall patient ratings were not provided, but significant differences in ratings were documented by facility type and service provider [10]. Other studies have developed and validated satisfaction measures among specialized populations, including patients with liver cancer [11, 12], nursing home residents [13] and patients receiving nursing care at home [14]. Two of these studies reported associations between satisfaction with care and quality of life [12, 13]. Other studies have reported that 90% of patients receiving rehabilitation services were satisfied with the care they received [15] and that inpatients perceived high quality of nursing care [16, 17]. To date, only one study on patient experiences of health care services has been conducted among the rural population in China [18], which represents more than 70% of China's population and is the target for NCMS reforms. This study, however, was conducted among only county-level hospitals in two provinces [18]. We seek to build upon this previous work with a broader sample in seven provinces and by specifically examining differences in experiences for different levels of hospitals and across provinces in rural China.

We therefore sought to understand patients' experiences with hospital inpatient services among rural households in China who had enrolled in the NCMS. Specifically, we aimed to (1) estimate the overall ratings of hospital inpatient services among participants living in rural China, (2) determine the strength of associations between patients' overall ratings of services and key components of their experience (i.e. communication with doctors and nurses, physical environment, pain management and medication management) and (3) examine associations between overall ratings and participant sociodemographic characteristics, facility type and province. Our findings can be useful as baseline data to evaluate current and subsequent changes in patient experiences with health services in rural China in the context of the NCMS reform.

Methods

Study design and sample

We conducted a cross-sectional study in 2010 of 15 698 individuals in 4209 households in seven provinces that were representative of the provinces along the north side of the Yellow River in China. We selected 2 or 3 representative counties within each province and collaborated with the county health administrator to identify 5 to 13 townships within each county, including at least 1 high income, 1 medium income and 1 low-income township. Research teams visited three villages within each township, including one high, one medium and one low-income village for that township, as determined by the township official in the local health bureau. Households were randomly selected within each township. Each of the 12 research teams included 14 interviewers (undergraduates), 1 supervisor (MPH or PhD student) and 1 faculty member. Research teams approached the head of each household in person and explained the study. After obtaining informed consent, they interviewed the head of household. If he or she were not available, they began by interviewing the next most senior person in the home. Interviews were conducted in Mandarin with all available members of the household. Individuals younger than 18 years were consented to participate in the survey by their parents or guardians. We limited our analytic sample to self-respondents who were older than 15 years (n = 4663). Our sample was further limited to participants who reported having been admitted to a hospital in the past 365 days (n = 490). Individuals who were admitted to a hospital in the last 365 days were older, in worse health, less educated and more likely to be female than individuals who had not been admitted to a hospital in the last 365 days (P-values <0.05). We additionally excluded from this sample people who were not the first respondent from each household (n = 13) and who did not have complete outcome data (n = 34). Our total sample size was 443 participants.

Measures

Patients' experiences with their health care were measured using a modified 10-item version of the Inpatient Assessment of Health Care (I-PAHC), which used several items from the CAHPS instrument [1922] and had been validated for use in low-income settings [23]. The I-PAHC was developed by first examining existing instruments used to assess patient healthcare experiences in low-income countries and then conducting focus groups in Ethiopia to learn about aspects of care that were most important. Interviews among stakeholders in the Ministry of Health, physicians and hospital administrators were used to review and modify items for their relevance and comprehensibility. Items were then pilot tested and further modified as appropriate. The final set of items was validated among 230 patients from five hospitals and three health centers in Ethiopia [23] and was re-validated for use in China (see Supplementary material, Appendix). The five factors derived from the I-PAHC included communication with nurses (two items; Cronbach's alpha = 0.92), communication with doctors (three items; Cronbach's alpha = 0.84), physical environment (two items; Cronbach's alpha = 0.61), pain management (one item) and medication management (two items; Cronbach's alpha = 0.70). On the survey, patients' overall rating of health services was measured with an additional two items. First, participants were asked to indicate the quality of care they received on a scale from 1 to 5 where 1 = ‘worst visit’ and 5 = ‘best visit’. This variable was retained as ordinal for analysis. Second, participants were asked the likelihood with which they would recommend the facility to friends and family. Responses were on a 4-point scale, ranging from ‘definitely no’ to ‘definitely yes’. For analysis, we dichotomized responses as ‘definitely yes’ or ‘probably yes’ versus ‘definitely no’ or ‘probably no’.

We also gathered data on participant sociodemographic characteristics including age, gender, education (below primary, primary school graduate, middle school graduate or high school graduate), marital status (currently married or not currently married), household size, residential location (province), self-rated health (1 = excellent to 5 = very poor) and the total household income for the past year. Participants also indicated the type of hospital at which they received their care (township hospital, county hospital, above county-level hospital and ‘other’ hospitals).

Statistical analysis

We generated means and frequencies to describe patient experiences with care as well as distributions of participant sociodemographic characteristics. We used linear regression models to determine the associations between five factors derived from the I-PAHC (communication with nurses, communication with doctors, physical environment, pain management and medication management) and overall ratings of care. Last, to determine the associations between overall ratings of care and participant sociodemographic characteristics, hospital type and province, we used multivariable linear and logistic regression models, as appropriate. Participant sociodemographic characteristics included patient age, gender, education, marital status, household size, self-reported health and household income level as has been done in similar research [8, 9]. Specifically, we hypothesized that patients' reported experiences would vary significantly by region and type of facility, with patients in more rural (versus urban) regions and treated at lower- (versus higher-) level health facilities rating their experiences more poorly. A complete case analysis was conducted, because frequency of missing data was low (<4%). All analyses were completed with IBM SPSS Statistics 19 (2010).

Results

Sample characteristics

The average age of participants was 49 years (Table 1). Slightly more than half (52%) were male, and 29% had less than a primary level education. Almost 95% were currently married. Approximately 8% of participants reported having ‘excellent’ health, 33% reported having ‘good’ health, 26% reported having ‘fair’ health, 28% reported having ‘poor’ health and 3% reported having ‘very poor’ health. Twenty-one percent had been admitted to a township hospital, 51% to a county hospital, 25% to an above county-level hospital and 3% to another type of facility.

View this table:
Table 1

Overall sample characteristics and patient satisfaction (n = 443)

Overall n (%)
Age (mean ± SD)49.3 ± 14.80
Gender
 Male229 (51.7%)
 Female214 (48.3%)
Education
 Less than primary128 (28.9%)
 Primary school graduate103 (23.3%)
 Middle school graduate157 (35.4%)
 High school graduate55 (12.4%)
Marital status
 Currently married417 (94.3%)
 Other (including never married)25 (5.7%)
Household size (mean ± SD)3.8 ± 1.73
General health
 Excellent37 (8.4%)
 Good146 (33.2%)
 Fair116 (26.4%)
 Poor126 (28.6%)
 Very poor15 (3.4%)
Total household income (Yuan; mean ± SD)22 710 ± 33 583
Total household expenditure (Yuan; mean ± SD)20 047 ± 34 052
Type of facility
 Township hospital94 (21.2%)
 County hospital227 (51.2%)
 Above county-level hospital109 (24.6%)
 Other13 (2.9%)
Province
 Shandong30 (6.8%)
 Henan92 (20.8%)
 Shanxi40 (9.0%)
 Neimenggu68 (15.3%)
 Shannxi68 (15.3%)
 Gansu84 (19.0%)
 Ningxia61 (13.8%)
Overall rating of quality of care (mean ± SD)4.0 ± 0.83
 1 (worst)4 (0.9%)
 214 (3.2%)
 378 (17.6%)
 4211 (47.6%)
 5 (best)136 (30.7%)
Would recommend facility
 Yes377 (85.1%)
 No66 (14.9%)

Patient experiences with hospital inpatient care

The average patient rating of their experience with inpatient services was 4.0 out of 5.0 (the best stay), with 31% rating services 5.0 (Table 1). An additional 22% rated their experience as 3.0 or less. Average ratings for specific items ranged from 2.4 to 3.6, with ratings for 8 of the 10 items averaging between 3 (‘usually’) and 4 (‘always’) (Table 2). Despite high average ratings, a notable proportion of participants (15%) would either definitely not or probably not recommend the hospital to their family and friends (Table 1).

View this table:
Table 2

Patient experiences with inpatient service utilization (I-PAHC) (n = 443)

Mean ± SD
During this hospital stay, how often did nurses treat you with courtesy and respect?3.4 ± 0.72
During this hospital stay, how often did nurses listen carefully to you?3.5 ± 0.72
During this hospital stay, how often did doctors treat you with courtesy and respect?3.6 ± 0.66
During this hospital stay, how often did doctors listen carefully to you?3.5 ± 0.67
During this hospital stay, how often did doctors explain things in a way you could understand?3.4 ± 0.69
During this hospital stay, how often was the area around your bed kept clean?3.5 ± 0.75
Is it quiet at night around your place of stay in the hospital?3.4 ± 0.82
During this hospital stay, how often was your pain well controlled?3.3 ± 0.80
Before giving you any new medicine, how often did hospital staff tell you what the medicine was for?a2.8 ± 1.24
Before giving you any new medicine, how often did hospital staff describe possible side effects in a way you could understand?a2.4 ± 1.25
  • aDescriptive statistics presented only among patients who reported receiving new medication during their hospital stay (n = 244).

Factors associated with overall ratings of care

Four of the five factors from the I-PAHC were significantly (P-values <0.05) associated with the overall rating of care in multivariable analysis (Table 3). Every one-unit increase in scores for ‘communication with nurses’ was associated with a 0.35 (95%CI = 0.21–0.49) increase in overall ratings; a one-unit increase in ‘communication with doctors’ and ‘physical environment’ was associated with a 0.30 (95%CI = 0.13–0.46) and 0.14 (95%CI = 0.03, 0.25) increase in overall rating scores, respectively. ‘Pain management’ was not significantly associated with overall rating scores (B = 0.03, 95%CI = −0.07, 0.12). Among patients who received new medication during their hospital stay, a separate multivariable model suggested that ‘medication management’ was significantly associated with a 0.11 (95%CI = 0.02, 0.20) increase in overall rating scores.

View this table:
Table 3

Regression models examining independent associations of patient experience factors on overall patient ratings and recommendations

 Overall rating B (95% CI)
All patients (n = 432)Patients with new medications (n = 241)
Communication with nurses0.35 (0.21, 0.49)**0.37 (0.16, 0.58)**
Communication with doctors0.30 (0.13, 0.46)**0.25 (−0.00, 0.50)***
Physical environment0.14 (0.03, 0.25)*0.10 (−0.05, 0.26)
Pain management0.03 (−0.07, 0.12)0.03 (−0.10, 0.17)
Medication managementNot included0.11 (0.02, 0.20)*
  • *P < 0.05; **P < 0.01; ***P = 0.054.

Differences in patient overall ratings of care by hospital type and province

Multivariable models showed that overall ratings for township hospitals and county-level hospitals were significantly lower than ratings for above county-level hospitals (P-value <0.05 and P-value <0.01, respectively; Table 4). Additionally, ratings among participants from Henan, Shannxi, Gansu and Ningxia Provinces were significantly higher than among participants from Shandong Province (all P-values <0.05). Similarly, odds of recommending the facility among county hospitals were significantly lower than above county-level hospitals (P-value <0.05). Odds of recommending the facility were also greater among participants from Shannxi Province compared with participants from Shandong Province (P-value <0.05) and among participants from Shannxi and Ningxia Provinces compared with Shanxi Province (P-value <0.05).

View this table:
Table 4

Regression models examining correlates of the overall rating of quality of care received and recommending the facility (n = 427)

Overall Ratinga B (95% CI)Would recommendb OR (95% CI)
Age0.01 (0.00, 0.02)*1.00 (0.98, 1.03)
Gender
 Male−0.05 (−0.23, 0.12)1.46 (0.79, 2.72)
 FemaleREF1.00
Education
 Less than primaryREF1.00
 Primary school graduate0.12 (−0.11, 0.36)0.86 (0.35, 2.12)
 Middle school graduate0.08 (−0.15, 0.31)0.51 (0.22, 1.16)
 High school graduate0.16 (−0.14, 0.45)0.78 (0.27, 2.27)
Marital status
 Currently married0.27 (−0.09, 0.64)2.89 (0.96, 8.72)
 Not currently marriedREF1.00
Household size0.01 (−0.05, 0.06)0.96 (0.80, 1.15)
Self-reported health−0.04 (−0.12, 0.04)0.94 (0.71, 1.25)
Total household income (per 1000)0.00 (−0.00, 0.00)1.00 (0.99, 1.01)
Type of facility
 Township hospital−0.25 (−0.49, −0.01)*0.88 (0.33, 2.30)
 County hospital−0.33 (−0.53, −0.13)**0.43 (0.20, 0.94)*
 Above county-level hospitalREF1.00
 Other−0.28 (−0.79, 0.22)0.49 (0.09, 2.79)
Province
 ShandongREFREF
 Henan0.51 (0.15, 0.87)**1.82 (0.61, 5.47)
 Shanxic0.22 (−0.19, 0.63)0.91 (0.26, 3.19)
 Neimenggu0.27 (−0.11, 0.65)1.84 (0.59, 5.73)
 Shannxid0.52 (0.15, 0.90)**3.52 (1.01, 12.28)*
 Gansu0.41 (0.05, 0.77)*2.48 (0.81, 7.63)
 Ningxiac,d0.51 (0.12, 0.89)**3.24 (0.86, 12.25)
  • aLinear regression model.

  • bLogistic regression model.

  • cStatistically significant differences (P < 0.05) for recommending the facility.

  • dStatistically significant differences (P < 0.05) for recommending the facility.

  • *P ≤ 0.05; **P ≤ 0.01.

Age was the only sociodemographic characteristic associated with overall ratings; every year increase in age was associated with significantly higher ratings (P-value <0.05). No sociodemographic characteristics were associated with odds of recommending the facility to family or friends (all P-values >0.05).

Discussion

Our findings indicate that overall patient ratings of hospital inpatient care were generally high for many patients in rural China, although a notable minority would not recommend the service to friends or family and rated their experiences unfavorably. One study from urban China found that nearly 22% of inpatients rated their experience poorly (less than 7 out of 10) using the PPE-15 [8], which may be findings similar to ours, although comparisons are difficult because the studies use different scales. Overall patient ratings in rural China appear lower than in higher income countries, such as the US [24] but are generally higher than those found among studies from rural areas in other low-income countries. For instance, in Nepal, approximately 33% of inpatients were dissatisfied with their care because of improper treatment [25]; in rural Mozambique, 45% of outpatients reported their care as fair or poor [26], and 37% of patients in rural Ethiopia reported being less than fully satisfied [27], but again, it is difficult to make comparisons given the different instruments used.

Participants who received care at county hospitals compared with above county-level hospitals had significantly lower overall ratings of care and were less likely to recommend the facility to friends and family. Additionally, regional differences in patient experience were apparent; participants from less wealthy provinces reported greater overall ratings than participants from the wealthiest province (Shandong Province), possibly due to lower expectations among people in poorer areas. The lower ratings among participants from Shandong Province, however, emphasize the need to address the quality of care throughout this region. Last, younger participants tended to report lower overall ratings of care, which is consistent with other literature [9, 18]. This finding persisted after adjusting for hospital type and province and suggests that younger patients may have expectations for higher quality health services.

Although the NCMS coverage has reached nearly 100% throughout rural China [2], our findings suggest that the variation in patient experiences across different levels of hospitals and different provinces is substantial. Furthermore, despite increased financial protection, a notable proportion of participants, particularly those receiving care in county-level hospitals, continue to be less than fully satisfied with their care.

In this study, the factors derived from the modified I-PAHC instrument were strongly associated with overall ratings of care. ‘Communication with nurses’ was the factor most strongly and consistently related to overall ratings of care and thus identified an important component of care that might be targeted in order to improve overall patient experience. Communication with physicians and physical environment, and for those receiving medications, the discussions about purpose and side effects of medications, were also important components of patients' overall rating of their experience. As China seeks to build a patient-centered model of health care, feedback on patient experience can be an important part of evaluating progress. Policies and practices that promote enhanced communication between patients and clinicians, rather than enhanced technical skills for clinicians, may be most central to improving patient experiences.

Our findings should be interpreted in the light of our study limitations. First, the data are self-reported through face-to-face interviews and therefore may be influenced by social desirability bias, resulting in artificially high ratings. Even with this potential bias, however, a sizable proportion of responses suggested some level of dissatisfaction. Nevertheless, these proportions may be underestimates. Second, it is not clear to what extent lower patient ratings were related to objective problems in quality of care. Nevertheless, our data do provide insight into the subjective views of care, an important element in patient-centered quality of care. Last, we did not examine the influence of patient experiences on the future use of health care or the health of the patient. Such longitudinal follow-up is an important topic for future research.

Previous literature that has evaluated the NCMS reform has shown notable improvements in access and utilization [35] but has not adequately explored patient experiences in rural China. Our study contributes to this literature by demonstrating substantial variation across facility types and regions and suggests that patient experience may be an important indicator for future monitoring of the NCMS and its impact on patients. Although improving financial access to health care is an important policy objective, ensuring high quality care from the patients' perspectives is equally important and sometimes overlooked in lower-income settings. Our findings suggest that, although patients on average have relatively high ratings of hospital care, a sizable proportion report less positive experiences with county-level hospital care. As China further develops its health system, establishing routine monitoring of patients' experiences will be important to ensure the system is responsive to the population needs.

Supplementary material

Supplementary material is available at INTQHC online.

Funding

This work was supported by the Research Center for Healthcare Management, Tsinghua University and Tsinghua University School of Economics and Management.

References

View Abstract