In order to explore the situation of China’s community health delivery system before 2009 and provide a reference for China’s community health delivery system reform, the data that could comprehensively and accurately reflect the community patient’s life satisfaction, assessment of community medical service, and trust in community health delivery system in various regions of China was needed, so this study collaborated with the National Bureau of Statistics of China to carry out a large-scale 2008 national community resident household survey in 18 provinces, autonomous regions, and municipalities directly under the central government.

In this household survey, the National Bureau of Statistics of China adopted the most strict two-stage probability proportional to size (PPS) systematic sampling technique to select a probability sample of 3,600 community residents, and the advantage of this technique was that it could select a perfect stratified sample, but its disadvantage was that any deviation from the stratified sampling design could impact the effect of stratification. The face-to-face interviews for this household survey were conducted by the professional survey teams from the National Bureau of Statistics, local Bureaus of Statistics, and Tsinghua University. The professional investigator usually first invited the selected community resident to fill out the questionnaire of this household survey, no replacement was made if the selected community resident was away, refused to be interviewed, or failed to be interviewed after three attempts. If the selected community resident agreed to fill out the questionnaire of this household survey, but she/he was unavailable, or disabled in a way that impeded her/him from filling out the questionnaire, another family member that knew her/him best served as the respondent, and this family member was also asked to report her/his assessed values of questions in the questionnaire to check bias. A total of 3,306 valid responses were generated in the 2008 national community resident household survey, and the response rate was 91.83%.

The questionnaire of this household survey consisted of many sections, among them four sections had relation to this study. The first section inquired about the community resident’s certain personal characteristics (involving age, gender, marital status, education, income, employment status, occupation, health status, medical insurance, reimbursement percentage of medical cost, severity of disease, and stage of disease in the most recent community medical experience), and all these selected personal characteristics were considered as the possible influencing factors for the community resident’s life satisfaction in the most recent community medical experience. The second section inquired about the community resident’s life satisfaction in the most recent community medical experience. The third section inquired about the community resident’s assessment of community medical service in the most recent community medical experience, and the selection of five major aspects of community medical service was based on the top five aspects of community medical service that the community resident was most concerned about in the pre-survey for this study 4 5 , specifically speaking, the community resident’s assessments of five major aspects of community medical service consisted of assessment of medical treatment process, assessment of doctor-patient communication, assessment of waiting time for medical service, assessment of medical facility and hospital environment, and assessment of medical cost. The fourth section inquired about the community resident’s trust in community health delivery system in the most recent community medical experience, the selection of four major aspects of the community resident’s trust in community health delivery system was based on the interview materials in the pre-survey for this study 4 5 , and four major aspects of the community resident’s trust in community health delivery system were composed of trust in medical institution, trust in doctor, trust in prescription, and trust in recommended medical examination.

The personal characteristics of the study population were shown in Table 1. As the result of the stratified sampling design by the National Bureau of Statistics of China, the population distribution of each personal characteristic followed the natural distribution of community resident in China. The data in this household survey was collected anonymously, and the use of the dataset in this study was approved by the National Bureau of Statistics of China.

According to the regional per capita GDP and the richness of regional community medical resources in 2008, 18 provinces, autonomous regions, and municipalities directly under the central government in the 2008 national community resident household survey were divided into 3 groups by the National Bureau of Statistics of China:

1. Regions with the highest per capita GDP and the most abundant community medical resources (Group 1): Beijing, Shanghai;

2. Regions with the second highest per capita GDP and the second most abundant community medical resources (Group 2): Guangdong, Tianjin, Zhejiang, Fujian;

3. Regions with the lowest per capita GDP and the least abundant community medical resources (Group 3): Hubei, Hunan, Anhui, Heilongjiang, Jilin, Chongqing, Guangxi, Guizhou, Shaanxi, Ningxia, Qinghai, Xizang.

The 5-item self-reporting measure that assessed the community patient’s life satisfaction in the most recent community medical experience on a scale of 1 to 5 was employed, and the higher score reflected the community patient’s higher level of life satisfaction. In this study, the definition of satisfaction was a measure of how medical goods and medical services met or surpassed the community patient’s expectation, and then life satisfaction was operationalized as satisfaction with health-related quality of life and measured with the following question “If you compared your life after the most recent community medical experience with your life before the most recent community medical experience, was your health-related quality of life after the most recent community medical experience better than, equal to, or worse than that before the most recent community medical experience?”. The option “Much better” was assigned score 5; the option “Somewhat better” was assigned score 4; and the option “Equal” was assigned score 3; the option “Somewhat worse” was assigned score 2; and the option “Much worse” was assigned score 1.

The 5-item self-reporting measures that assessed the medical treatment process, the doctor-patient communication, the waiting time for medical service, the medical facility and hospital environment, and the medical cost in the community patient’s most recent community medical experience on a scale of 1 to 5 were employed, and the higher score reflected the community patient’s higher rating for a certain aspect of community medical service: the option “Excellent performance in this aspect” was assigned score 5; the option “Good performance in this aspect” was assigned score 4; and the option “General performance in this aspect” was assigned score 3; the option “Bad performance in this aspect” was assigned score 2; and the option “Poor performance in this aspect” was assigned score 1. Here the definition of performance in this study was the extent of the execution or accomplishment of community medical service related work.

The 5-item self-reporting measures that assessed the community patient’s trust in medical institution, trust in doctor, trust in prescription, and trust in recommended medical examination in the most recent community medical experience on a scale of 1 to 5 were employed, in this study the definition of trust was the trait of believing in the honesty and reliability of the target object, and the higher score reflected the community patient’s higher degree of trust in a certain aspect of community health delivery system: the option “The high degree of trust in this aspect” was assigned score 5; the option “The relatively high degree of trust in this aspect” was assigned score 4; and the option “The medium degree of trust in this aspect” was assigned score 3; the option “The relatively low degree of trust in this aspect” was assigned score 2; and the option “The low degree of trust in this aspect” was assigned score 1.

Due to the fact that the ordered probit was the sole method that discerned unequal differences between ordinal categories in the dependent variable-the community patient’s life satisfaction, the ordered probit model was the sole solution to the research question of this study 36 37 38 . For example, the ordered probit model didn’t assume that the difference between choosing “Much better” and choosing “Somewhat better” was the same as the difference between choosing “Somewhat worse” and choosing “Much worse”. In fact the ordered probit in this study captured the qualitative differences between different levels of life satisfaction.

In the ordered probit model, the latent evaluation score y_i was a linear function of independent variables that were written as a vector x_i, here i was the sample number, and y_i=x_i*b+ε_i, b was a vector of coefficients, and ε_i was assumed to follow a standard normal distribution. When an ordered probit model with k cutoff points was established, p_j (j=1,2,…,k) were defined as the cutoff points of all y_i, then y_i≦p₁, p_j<y_i≦p_j+1 (j=1,2,…,k-1) or y_i>p_k. Following the notation, the ordered probit model was expressed as

Prob y i = y 0 | x i = Ф p 1 − x i * b

Prob y i = y j | x i = Ф p j + 1 − x i * b − Ф p j − x i * b j = 1 , 2 , … , k − 1

Prob y i = y k | x i = 1 − Ф p k − x i * b

here y_j (j=0,1,…k) were the discrete values of y_i and Ф was the standard normal cumulative distribution function 3 39 .

The marginal effect of x_i could be calculated according to this formula:

∂ Prob y i = y 0 | x i / ∂ x i = − b * φ p 1 − x i * b

∂ Prob y i = y j | x i / ∂ x i = − b * φ p j + 1 − x i * b − φ p j − x i * b j = 1 , 2 , … , k − 1

∂ Prob y i = y k | x i / ∂ x i = b * φ p k − x i * b

here was the standard normal density function, and based on (4), (5) and (6) the vector of coefficients b could be estimated 3 39 .

The first specified ordered probit model was estimated with respect to all samples/the samples in various groups in order to test whether and to what extent the community patient’s assessments of various major aspects of community medical service influenced life satisfaction in whole China/in various regions of China:

Model 1 : satisfaction i = ∑ l β l 1 assessment li + ∑ m β m 2 z mi + ε i

here i was the sample number; satisfaction_i was the community patient’s life satisfaction; assessment_li (l=1,2,…,5) were the community patient’s assessment of medical treatment process, assessment of doctor-patient communication, assessment of waiting time for medical service, assessment of medical facility and hospital environment, and assessment of medical cost; z_mi were control variables, since the community patient’s life satisfaction in the most recent community medical experience might be influenced by certain personal characteristics (involving age, gender, marital status, education, income, employment status, occupation, health status, medical insurance, reimbursement percentage of medical cost, severity of disease, and stage of disease in the most recent community medical experience), all selected personal characteristics were controlled as dummy variables in the regression model in order to address the interaction between social demographic figures and geographic variations; error term ε_i was assumed to follow a standard normal distribution.

The second specified ordered probit model was estimated with respect to all samples/the samples in various groups in order to test whether and to what extent various major aspects of the community patient’s trust in community health delivery system influenced life satisfaction in whole China/in various regions of China:

Model 2 : satisfaction i = ∑ n β n 1 trust ni + ∑ m β m 2 z mi + ε i

here trust_ni (n=1,2,…,4) were the community patient’s trust in medical institution, trust in doctor, trust in prescription, and trust in recommended medical examination; the descriptions of other variables were the same as those in the first specified ordered probit model.

The descriptive statistics of the community patient’s life satisfaction, assessment of community medical service, and trust in community health delivery system were presented in Table 2. Generally speaking, the mean value of the community patient’s life satisfaction was between the assigned value for the option “Health-related quality of life after the most recent community medical experience was somewhat better than that before the most recent community medical experience” and the assigned value for the option “Health-related quality of life after the most recent community medical experience was equal to that before the most recent community medical experience”. Among the community patient’s assessments of various major aspects of community medical service, the mean value of the community patient’s assessment of medical treatment process was the highest, the mean values of the community patient’s assessment of waiting time for medical service and assessment of doctor-patient communication were the second highest, and the mean value of the community patient’s assessment of medical cost was in the medium level, while the mean value of the community patient’s assessment of medical facility and hospital environment was the lowest. The standard deviation of the community patient’s assessment of waiting time for medical service was significantly higher than the standard deviations of the community patient’s assessments of other aspects of community medical service. Among various major aspects of the community patient’s trust in community health delivery system, the mean values of the community patient’s trust in recommended medical examination and trust in doctor were the highest, and the mean value of the community patient’s trust in prescription was in the medium level, while the mean value of the community patient’s trust in medical institution was the lowest. The standard deviations of the community patient’s trust in doctor and trust in medical institution were significantly higher than the standard deviations of the community patient’s trust in prescription and trust in recommended medical examination.

The results of both model 1 and model 2 with respect to all samples were presented in Table 3. Among the community patient’s assessments of various major aspects of community medical service, assessment of medical cost and assessment of doctor-patient communication had the largest positive influences on the community patient’s life satisfaction, assessment of medical facility and hospital environment and assessment of medical treatment process had the second largest positive influences on the community patient’s life satisfaction, while assessment of waiting time for medical service had no significant influence on the community patient’s life satisfaction. Among various major aspects of the community patient’s trust in community health delivery system, trust in doctor had the largest positive influence on the community patient’s life satisfaction, trust in prescription had the second largest positive influence on the community patient’s life satisfaction, and trust in recommended medical examination had the third largest positive influence on the community patient’s life satisfaction, while trust in medical institution had no significant influence on the community patient’s life satisfaction.

t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01.

The results of model 1 with respect to the samples in various groups were presented in Table 4. The community patient’s assessment of medical cost, assessment of doctor-patient communication, assessment of medical facility and hospital environment, and assessment of medical treatment process had significant positive influences on life satisfaction in all groups, but the community patient’s assessment of waiting time for medical service had no significant influence on life satisfaction in any group. The positive influences of the community patient’s assessment of medical cost, assessment of doctor-patient communication, and assessment of medical facility and hospital environment on life satisfaction in regions with higher per capita GDP and more abundant community medical resources were larger than those in regions with lower per capita GDP and less abundant community medical resources. The positive influence of the community patient’s assessment of medical treatment process on life satisfaction in regions with lower per capita GDP and less abundant community medical resources was larger than that in regions with higher per capita GDP and more abundant community medical resources.

t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01.

The results of model 2 with respect to the samples in various groups were presented in Table 5. The community patient’s trust in doctor, trust in prescription, and trust in recommended medical examination had significant positive influences on life satisfaction in all groups, but the community patient’s trust in medical institution had no significant influence on life satisfaction in any group. The positive influences of the community patient’s trust in doctor, trust in prescription, and trust in recommended medical examination on life satisfaction in regions with lower per capita GDP and less abundant community medical resources were larger than those in regions with higher per capita GDP and more abundant community medical resources.

t statistics in parentheses, * p < 0.10, ** p < 0.05, *** p < 0.01.

In China’s community health delivery system, among five major aspects of community medical service, the medical cost (particularly in regions with higher per capita GDP and more abundant community medical resources), the doctor-patient communication (particularly in regions with higher per capita GDP and more abundant community medical resources), the medical facility and hospital environment (particularly in regions with higher per capita GDP and more abundant community medical resources), and the medical treatment process (particularly in regions with lower per capita GDP and less abundant community medical resources) were all key considerations in generating the community patient’s life satisfaction. Among four major aspects of the community patient’s trust in community health delivery system, trust in doctor (particularly in regions with lower per capita GDP and less abundant community medical resources), trust in prescription (particularly in regions with lower per capita GDP and less abundant community medical resources), and trust in recommended medical examination (particularly in regions with lower per capita GDP and less abundant community medical resources) were all important considerations in generating the community patient’s life satisfaction.

The previous studies separately showed that the community patient’s assessment of community medical service and trust in community health delivery system were important considerations when the community patient comprehensively evaluated community medical service to generate life satisfaction 5 6 23 24 25 26 27 35 . But the community patient’s assessment of community medical service/the community patient’s trust in community health delivery system was usually taken as a whole in most of these previous studies, while few studies subdivided the community patient’s assessment of community medical service/the community patient’s trust in community health delivery system into various aspects and further explored the differences of their influencing effects on the community patient’s life satisfaction 5 6 23 24 25 26 27 35 . And there were few studies that explored the regional differences of the influences of both the community patient’s assessment of community medical service and trust in community health delivery system on life satisfaction in China.

Most previous studies obtained their conclusions only through the qualitative analysis or the authors’ experiences, while only a small number of previous studies adopted the strict quantitative analysis, but their conclusions were usually drawn on the basis of a small range of community patients, then the robustness and universality of their conclusions remained controversial.

The most important contribution of this study was that the influences of the community patient’s assessments of various major aspects of community medical service/various major aspects of the community patient’s trust in community health delivery system on life satisfaction in whole China/in various regions of China before 2009 were studied under a systematic and comprehensive framework for the first time. The second most important contribution of this study was that in order to perform the strict quantitative analysis, this study collaborated with the National Bureau of Statistics of China to carry out a large-scale 2008 national community resident household survey for the first time in China. The third most important contribution of this study was that the regional differences of the key influencing factors for the community patient’s life satisfaction in China before 2009 were found for the first time. The fourth most important contribution of this study was that the robustness and universality of the major findings on the situation of China’s community health delivery system before 2009 in this study were much better than those in previous studies.

The growing community health delivery systems in most countries faced the challenge to meet the growing need for primary care in all communities, especially in medically underserved communities. Since the expectation of the whole society forced CHC to provide community medical service for the public more efficiently, more economically, more effectively, and more equally in China, the challenge for China’s community health delivery system was much more severe than the challenges for most other countries’ community health delivery systems, and then the holistic and systematic approaches to promote the community patient’s life satisfaction were more urgently needed in China than in most other countries.

In China’s future community health delivery system reform, on the basis of the major findings on the situation of China’s community health delivery system before 2009, the following inspirations on the effective approaches to promote the community patient’s life satisfaction in whole China/in various regions of China were found. The reduction of medical cost (particularly in regions with higher per capita GDP and more abundant community medical resources), the improvement of doctor-patient communication (particularly in regions with higher per capita GDP and more abundant community medical resources), the promotion of medical facility and hospital environment (particularly in regions with higher per capita GDP and more abundant community medical resources), and the improvement of medical treatment process (particularly in regions with lower per capita GDP and less abundant community medical resources) could help promote the community patient’s life satisfaction. The promotion of trust in doctor (particularly in regions with lower per capita GDP and less abundant community medical resources), the promotion of trust in prescription (particularly in regions with lower per capita GDP and less abundant community medical resources), and the promotion of trust in recommended medical examination (particularly in regions with lower per capita GDP and less abundant community medical resources) could be beneficial to the promotion of the community patient’s life satisfaction.

Several limitations of this study should be noted. First, the response rate of the large-scale 2008 national community resident household survey was 91.83%, and then the slight deviation from the stratified sampling design might have the slight influence on the major findings of this study. Second, there might be other potential influencing factors for the community patient’s life satisfaction, although these potential influencing factors were not contained or controlled in this study, they might affect the influences of the community patient’s assessments of various major aspects of community medical service/various major aspects of the community patient’s trust in community health delivery system on life satisfaction. Third, in fact the community patient’s assessments of various major aspects of community medical service/various major aspects of the community patient’s trust in community health delivery system interrelated with each other, and this situation might affect the relative influences of the community patient’s assessments of various major aspects of community medical service/various major aspects of the community patient’s trust in community health delivery system on life satisfaction. Fourth, this study didn’t include the impact of the community health delivery system reform after 2008.

The findings in this study could prove useful for both practitioners in CHC and policy makers in health administration departments. Practitioners in different regions of China could find the different effective approaches to promote the community patient’s life satisfaction (through promoting the corresponding key influencing factors among major aspects of community medical service/major aspects of the community patient’s trust in community health delivery system). Policy makers in different regions of China could implement the different effective interventions for the promotion of the community patient’s life satisfaction. For example, in order to effectively promote the community patient’s life satisfaction, the improvement of medical treatment process through targeted laws, targeted policies, targeted regulations, and targeted measures should be emphasized particularly in regions with lower per capita GDP and less abundant community medical resources, while the reduction of medical cost, the improvement of doctor-patient communication, and the promotion of medical facility and hospital environment through targeted laws, targeted policies, targeted regulations, and targeted measures should be emphasized particularly in regions with higher per capita GDP and more abundant community medical resources.