|Home | About | Journals | Submit | Contact Us | Français|
Currently, China is in the process of medical and health care reform, and the establishment of primary medical and health services covering urban and rural residents is an important aspect of this process. Studying the satisfaction of residents of underdeveloped areas with their primary medical and health services and identifying the factors that can increase the satisfaction of different groups may improve patient compliance and ultimately improve health. Moreover, such research may provide a reference for the development of medical and health undertakings in similarly underdeveloped areas.
A face-to-face survey was conducted on a stratified random sample of 2200 residents in Gansu by using structured questionnaires. Demographic characteristics were collated, and questionnaires were factor-analysed and weighted using SPSS software to obtain scores for each factor, as well as total satisfaction scores. The characteristics of poorly satisfied populations were determined by a multiple linear regression analysis using SAS software. A cluster analysis was performed using SAS software for classification and a separate discussion of populations.
The hypertension self-awareness rate (11.29%) of the sampled population was lower than the average hypertension prevalence (23.85%), as recorded in the 2014 Health Statistical Yearbook of the region. The disease knowledge awareness factor was the lowest factor (2.857), whereas the policy awareness factor was the highest factor (4.772). The overall satisfaction was moderate (3.898).
The multivariate linear regression model was significant (p <0.05). The regression coefficients were -0.041 for minors; 0.065 for unemployed people; and 0.094 for people with an elementary school educational level, a value lower than that of other population groups.
A cluster analysis was used to divide the respondents into five groups. The overall satisfaction was lowest in the second population group (rural, middle-aged)(Fz=3.64) and was highest in the fourth population group(minors) (Fz=4.13). Different population groups showed different satisfaction rates in F1 to F6.
Hypertensive patients had low self-awareness, and residents had a poor grasp of disease and limited health knowledge. Their overall satisfaction was moderate. Residents expressed comparatively high satisfaction with the current policy. Minors, adults with low level of education, unemployed people and other vulnerable groups expressed low overall satisfaction. The degree of satisfaction varied greatly among the different groups. Targeted medical and health practices should be implemented for different groups; additionally, the public health practice should be strengthened.
The online version of this article (doi:10.1186/s12913-017-2200-9) contains supplementary material, which is available to authorized users.
Since the WHO introduced the Primary Health Care (PHC) strategy in the Alma-Ata Declaration of 1978, various countries have become committed to the development of health care. Primary Health Care is the basic health care that can generally be obtained by individuals and families in a community. The Declaration stated that the health inequalities between developed and developing countries and within countries should be of concern to all countries and that health care for all and a reduction of the gap in health status between developing and developed countries were the primary goals . Thirty years later, the WHO and member states reaffirmed the importance of the PHC strategy [2, 3] and emphasized the broad involvement of social factors, as well as the roles and functions of government in the health field . In those three decades, the PHC strategy has played a significant role, especially in low- and middle-income countries .
In 1988, the Chinese government made a commitment to this global health strategy and has since been devoted to medical and health care reform . Medical and health undertakings have developed tremendously, and the level of medical care has continually improved, thereby providing better medical and health services for all people in the country . With the improvement of the overall level of medical care, medical sub-specialties have been further divided, and general practitioners able to meet the primary care demands have become increasingly lacking, thus indirectly leading to difficulties and higher costs associated with obtaining medical services . From 1998 to 2013, the Chinese government conducted a total of five national health services surveys, which have revealed a continuously growing demand for health services in China and the disease spectrum gradually shifted towards chronic diseases. According to the fourth survey, conducted in 2008, health service utilization in China reached a record high; and according to the fifth survey, conducted after the new health reform, residents’ confidence in medical staff had increased. The difficulties and high costs of obtaining medical services have improved [9–11], but nevertheless, these two issues remain the most direct and practical concern of residents, especially for patients with common and frequently occurring diseases that need to be resolved in primary medical and health care institutions. Under the new health reform policy, the health resources shift towards the primary level, and primary medical and health services become the focus of medical and health work, which are considered to be key to solving the problem of medical services that are difficult to obtain and expensive [12, 13]. Only by constantly improving the capabilities of primary medical and health services and gradually enhancing patients’ satisfaction can residents’ health levels be improved.
Assessment of residents’ satisfaction with primary medical and health services is a complex process. As early as the mid-20th century, many precedents had already been established for assessing health services on the basis of patients’ opinions and comments . Satisfaction is often used to evaluate the health services received by patients in health institutions [15–18], and it is influenced by numerous factors such as personal health status, demographic characteristics, socio-psychological factors, and characteristics of the health system. The improvement of the patient satisfaction can promote compliance and improve health . Measurement of the patient satisfaction also involves different indicators from multiple dimensions, such as satisfaction with the daily care received, satisfaction with the information provided, and satisfaction with the healthcare environment , but there is not yet a gold standard method or indicator for assessing patient satisfaction [21–25].
Li Jian-Jun and others have summarized the research on satisfaction published before 2000 and have found that interview methods (telephone and face-to-face)generate higher responses than mail surveys, and only 20% of studies consider the potential importance of expectations . Sara N Bleich and Emre Ozaltinb, in a 2009 study, have noted that participant satisfaction is critically influenced by their expectations . Some other studies have reached the same conclusion [28–31].
A study by Zhijian Li and colleagues has evaluated residents’ satisfaction with community health services after health care system reform in Shanghai, China. The authors have found that satisfaction across all dimensions improved since the reform was initiated, but disadvantaged groups had overall poorer satisfaction levels . A study by Sarah Atkinson and Dave Haran in one of Brazil’s poorest states, has found that user satisfaction can be a useful management tool. It was also found that the results of satisfaction assessment are influenced by the definition used and the method(s) used to measure satisfaction, so exploration from multiple perspectives is helpful in explaining the complexity of satisfaction .
In this paper, the methods and indicators used to assess the residents’ satisfaction with primary medical and health services were selected after analysis of numerous references and the results of China National Health Services Surveys, which cover public health, chronic diseases, infectious diseases and other aspects of healthcare. The assessment concentrates specifically on residents’ satisfaction with primary medical and health services under the new health reform in China.
Furthermore, the levels of satisfaction vary greatly among different countries and districts, owing to racial or ethnic and socio-economic factors . The area investigated in this paper was Gansu, which is located in the northwest of China and is an economically disadvantaged region of the country, with slow development of medical and health services  and a large rural population. In 2014, the total GDP of Gansu ranked fourth in China, but the per capita GDP ranked last in the country, largely because the economic structure is agriculture-based, with a few large concentrated industrial bases. Additionally, Gansu, which is in the process of urbanization, has typical features of developing countries.
Studying the satisfaction of residents of underdeveloped areas with their primary medical and health services and identifying the factors that can increase the satisfaction of different groups may improve patient compliance and ultimately improve health. Moreover, such research may provide a reference for the development of medical and health undertakings in similarly underdeveloped areas.
There is extensive literature regarding the satisfaction of residents and patients. Several studies have investigated residents’ satisfaction in less-developed areas under health care system reform. However, few studies have assessed patient satisfaction from multiple objective and subjective perspectives by using multiple linear regression with dummy variables, and studies focused on the various aspects of vulnerable members’ satisfaction have been especially rare.
According to the Gansu Provincial Health Department’s 2014 annual assessment of primary medical and health services, five counties/districts in Gansu (Chengguan District of Lanzhou, Gaolan County of Lanzhou, Zhuanglang County of Pingliang, Liangzhou District of Wuwei, Ganzhou County of Zhangye) were selected for this study after comprehensive consideration of their economic development status, health work foundation, and the ethnic and geographical features throughout the province. In each sampled county/district, three communities/villages were randomly selected, and the number of questionnaires was determined on the basis of the population of each community/village. A randomized household survey by trained investigators was used. The inclusion criteria for the participants were residents>9 years of age served by the local healthcare service providers. The exclusion criteria were residents who refused to provide informed consent and residents who were unable to answer the questionnaire (for psychiatric reasons). A total of 2200 questionnaires were distributed, and 2038 valid questionnaires were returned, with a response rate of 92.64%.
To avoid variations due to the questionnaire, the questionnaire was designed after examination of a large number of related studies and was improved according to pre-survey results. The questionnaire was divided into two parts: basic information about the residents and survey contents. The basic information included sex, age, occupation, household registration, work, educational level and prevalence of diabetes and hypertension, and the survey content involved knowledge of national public health programs, health knowledge, satisfaction with primary medical and health services and health policy, which was expected to comprehensively analyse residents’ satisfaction with both objective and subjective dimensions Additional file 1.
In this paper, the alpha reliability coefficient method was used to test the reliability of the questionnaire . The main consideration was whether there was high intrinsic consistency across the questionnaire items or indicators. The results of the analysis revealed high intrinsic consistency, because the inter-item reliability coefficient was 0.781, close to 0.8. Therefore, the questionnaire results were credible.
The survey used a sub-questionnaire from the third-party evaluation projects of the Gansu National Public Health Equalization Program; the sub-questionnaire has been revised by several experts and professors after extensive analysis of references and has been applied in various projects for many years. As a result, the validity of the questionnaire was high.
After collecting the questionnaires, data were entered by double-entry using EPIDATA.
There are several limitations of this study. First, we were unable to cover the entire province, owing to the limited personnel available. Second, face-to-face research was assisted by the staff members from the health services institution, which may have influenced the responses regarding satisfaction. Third, the sample size was too small to allow the participants to be divided into different groups when the data were analysed.
From the basic information portion of the questionnaire, the demographic characteristics and the sources of health knowledge (from multiple choice question) for the respondents are summarized in Table 1
The prevalence of hypertension in the patient sample was 11.29%. According to the provincial statistical department’s documents, the prevalence of hypertension among Gansu’s residents was 23.85%.
Because the main part of the questionnaire involved many questions, a factor analysis was performed on 26 questions in the main questionnaire using SPSS 20.0 software to reduce dimensionality. The KMO coefficient, a measure of sampling adequacy, was 0.862, which was greater than 0.5. The approximate χ2 in Bartlett’s test was 14288.411, with 325° of freedom and p=0.000, thus indicating that the 26 questions selected were suitable for the factor analysis.
We assessed the variance contributions of various factors, and those with eigenvalues greater than 1 were extracted. Kaiser standardized orthogonal rotation was performed, which converged after six iterations. On the basis of the component matrix after rotation, six factors were extracted.
The answers to each questionnaire question were rated on a Likert scale: very satisfied-5, moderately satisfied-3, and very dissatisfied-1. Awareness objective questions were scored as follows: correct-5, incorrect-1. The higher the score, the higher the respondents’ objective and subjective satisfaction, and vice versa. Six factors and questions for each factor were weighted using the Delphi method  to obtain scores for six factors F1, F2, F3, F4, F5, F6, as well as overall score Fz. The details of the six factors and the distribution of various factors are shown in Table 2.
The residents’ overall satisfaction was 3.90, which was moderate. The score for disease knowledge awareness was the lowest factor, at only 2.86. The residents’ subjective satisfaction with primary medical and health services was low at 3.42. The health policy awareness score was the highest factor; that is, residents were very satisfied with the primary medical and health policies.
A multiple linear regression analysis was used to explore the influence of each feature on the total score Fz. The overall satisfaction score Fz was set as the dependent variable, and the age group, household registration, education and occupation were set as independent variables, but the independent variables were polytomous variables. In most of the recent satisfaction studies, the independent variables have been considered polytomous variables in the logistic regression analysis, which has failed to reflect the differences in satisfaction among various feature subgroups. Therefore, in this paper, n-1 dummy variables were introduced for each independent variable, where n was the number of categories in the polytomous variables, and the dummy variables were (0,1) variables. The purpose of introducing dummy variables was to quantify the unquantifiable variables. The value of the dummy variables was usually 0 or 1. The introduction of dummy variables made the linear regression model more complex, but the problem description became more concise, and the satisfaction of different subgroups within each group was visible, which was closer to reality.
To simplify the analysis, the last item in each polytomous variable was set as the control dummy variable; i.e., the elderly subgroup in the age group;the alien permanent population in the household registration; master’s degree or above in the educational level; and other occupations in the occupation category. Because the dummy variables in one element were an indivisible whole, they had to be entered at the same time in one block during variable screening. Stepwise variable selection, with dummy variables simultaneously entered in one block, was set for the independent variables and Fz was the dependent variable. The results showed that the model could be used to analyse and predict the findings (p=0.000<0.05), as shown in Tables 3.
The regression coefficient referred to the disparity between the dummy variable group and control groups. Because p was largely influenced by the sample size, and variables were entered into the model gradually, the regression equation in case the model was significant (p <0.05 in Table 4.1) could be obtained as Yz=3.461-0.041*X1+0.018*X2+0.036*X3+……+0.189*X15. Positive and negative regression coefficients indicated positive and negative correlations. The greater the absolute value, the higher the degree of influence.
According to the regression coefficients of the regression equation, the influence of each polytomous variable on the overall satisfaction could be obtained, i.e., the influence of the presence or absence of the dummy variable on Yz. The PE of the control dummy variables in each term was 0.
Two-step clustering analysis uses BIRCH(balance Iterative Reducing and Clustering using Hierarchies, which is focus on the large sample clustering in hierarchical clustering. This algorithm summarizes the information of cluster through the clustering characteristics, and then cluster. Centroid distance between two cluster is indicated by Euclidean Distance and Manhattan Distance. The distance between each sample point in two cluster is indicated by average inter-cluster distance, average intra-cluster distance and variance increase distance. In the first step, the measurement of changes in distances is logarithm Likelihood function. In the second step, the result in step 1 is involved in, and each sample is cluster analyzed.
Using SAS 9.4 software, various characteristic values and factors were classified by a two-step cluster analysis, which yielded good results. The 2038 respondents were divided into five groups after individual case exclusion as shown in Table 4.
The mean factor scores for the above five groups of people are shown in Table 5.
As can be clearly seen from the results, for the overall satisfaction Fz, the fourth category of people (rural, middle-aged) had the highest score (Fz=4.13), the second category (minors) had the lowest score (Fz=3.64), and the other three groups exhibited almost no differences. The middle-aged population stood out among groups regarding dietary health knowledge and healthy lifestyle knowledge (F5=4.51, 4.52; F6=4.86, 4.74), thus indicating that the middle-aged were more concerned about health knowledge and received better health education. For F1, the subjective satisfaction with primary medical and health services, the middle-aged rural population and elderly population scored higher (F1=3.68, 3.66), whereas the middle-aged urban population scored lower (F1=3.12). The elderly group scored low in factors related to the public health programme and disease knowledge awareness (F2=3.43, F3=2.01), but they expressed high subjective satisfaction with the medical and health services. The minors had a low awareness of health policies and low levels of various health knowledge (F2, F4, F5, F6).
Because the expectations and demands for health services may affect the satisfaction of residents, concrete implementation of graded medical care is crucial to improve residents’ satisfaction while avoiding unnecessary waste of medical resources. During the Eighteenth Session of the Standing Committee of the Twelfth National People’s Congress held on December 22, 2015, the Chinese government proposed a health reform plan for accelerating the construction of a graded medical care system . The progress of this graded medical care system will be examined in further studies.
In general, hypertensive patients had low self-awareness, and residents had a poor grasp of disease and health knowledge. The overall satisfaction level was moderate. Residents expressed comparatively high satisfaction with policy. Minors, adults with low levels of education, unemployed people and other vulnerable groups expressed low overall satisfaction. The degree of satisfaction varied greatly among different groups. Targeted medical and health practices should be implemented for different groups, and public health practice should be strengthened.
Questionnaire about residents’ satisfaction with primary medical and health services in Gansu Province. This questionnaire is about the satisfaction of residents, which is the way to collect the data in the research. (DOCX 16 kb)
Data about residents’ satisfaction in Gansu Province. This raw data was imported with epidata without any management. (XLSX 526 kb)
I would like to express my gratitude to all those who helped me during the writing of this paper especially Suqin Lian. I gratefully acknowledge the help of my friends, Tom Lake and Mark Liu, who has offered me a significant revision on the manuscript.
I also owe a special debt of gratitude to all the professors and editors in BMC., from whose hard work benefited a lot and academically revise for the paper.
The funding is from Commission of Health and Family Planning in Gansu Province, I really appreciate your support.
Commission of Health and Family Planning in Gansu Province which is a government sector in charge of medical and health work, funded this study.
The questionnaire was printed in paper. The data was imported into computer with EPIDATA, and the data is availability in machine-readable format upon request from the corresponding author whenever possible, but not available online.
WND and QYZ participated in the design, the acquisition, analysis and interpretation of data, and drafted the manuscript. CSY is responsible for the project and reviewed the manuscript. WLF and LLX reviewed the manuscript and provided some comments, as well as participating in the finalization and interpretation of the paper. All authors have read and approved the final manuscript. SQL provided some technical help about the SAS software in the process of data analyze.
The authors declare that they have no competing interests.
Ethics committee of School of Public Health, Lanzhou University approved this study. Informed consent to participate was obtained from each participant from whom data was collected in this study. Three hundred twenty nine participants were below the age of 16, parental/guardian consent was all instead obtained.
All authors have agreed to the submission to the journal and that the manuscript is not currently under submission in any other journal.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Weinan Dong and Qingyu Zhang are co-first authors.
Weinan Dong, Email: moc.qq@910814823.
Qingyu Zhang, Email: moc.qq@8679151201.
Chunsheng Yan, Email: nc.ude.uzl@hshcnay.
Wanling Fu, Email: moc.liamg@125tratswen.
Linlin Xu, Email: moc.qq@2878813751.