2.1 Survey respondents
This is a comparative study based on cross-sectional data, and 1,250 patients who were selected in 18 medical alliances in Beijing from October to December 2016.
In the "Beijing Main Functional Area Plan" published by the Beijing government on September 17, 2012, the 16 districts in Beijing were divided into four different functional areas. The A and B districts were classified as the "capital functional core area." The C, D, E, and F districts were classified as the “urban functional development area.” The G, H, I, J, and K districts were classified as the “new area of urban development,” and the L, M, N, O, and P districts were classified as the “ecological conservation development area.” Therefore, this classification is adopted for the current addresses of the respondents in the questionnaire.
2.2 Interview and questionnaire
First, this study included a 48-person interview with doctors using the qualitative interview method to explore their attitudes toward medical alliances. We did not set the number of interviewees in advance. Instead, according to the principle of "information saturation" in qualitative research, we stopped the interview process after the 48th interview when we thought the collected materials were sufficient. Ten personal in-depth interviews and 7 focus groups with 5-6 persons in each group were conducted. We used the interview guidelines to conduct the interviews. We prepared three questions: 1. What is the current status of the medical alliance?; 2.What are the current problems of the medical alliance?; and 3. What are the results and effects of the medical alliance? These questions are related to patient satisfaction.
We used the thematic framework analysis method to sort out and summarize the collected text data and formed different dimensions and questions of the questionnaire. After the first version of the questionnaire was formulated, we conducted two rounds of expert discussion, inviting doctors, hospital administrators and staff of health administration departments to repeatedly discuss and modify the questionnaire. The experts agreed that the final version of the questionnaire was effective, reliable and scientific and could be used for questionnaire surveys. Finally, we conducted a test of reliability and validity. The Cronbach coefficient was 0.67, the KMO value was 0.60, the sample met the reasonable requirements for the data structure, p<0.001 passed Bartlett’s test of sphericity, and the cumulative variance interpretation rate value was 54.04%. In summary, this questionnaire has good reliability and validity [27-28].
The questionnaire includes 24 questions across four parts. The first part includes the characteristics of the patients (including gender, age, household registration, length of time at residence, current address, chronic disease status, patient type, medical insurance type, average monthly medical expenses), the second part is the patients’ degree of understanding and method of the medical alliance policy. The third part is the degree of utilization of the medical alliance (including patients’ willingness and reasons for choosing cooperative hospitals or core hospitals for treatment and the construction effect of the medical alliance), and the fourth part is the patients’ satisfaction with the medical alliance.
2.3 Survey methods
The study used a stratified random sampling method to collect patients from 18 core hospitals and 80 cooperative hospitals of 18 medical alliances in 16 districts of Beijing. Because the number of medical institutions and health personnel in district C and district D is much higher than in other districts, two medical alliances were randomly selected from district C and district D, and one medical alliance was randomly selected from other areas to ensure the representativeness of the survey subjects.
The inclusion criteria of this study were patients aged between 18 and 85 years who had received medical services in the medical alliance. The exclusion criteria in this study were patients who had no autonomous behavior ability or awareness. To improve the valid response rate of the questionnaire, we distributed the questionnaire through government channels of the Beijing Municipal Health Commission and obtained informed consent from the respondents. We distributed 540 questionnaires in core hospitals. A total of 457 valid questionnaires were collected, and the valid response rate was 84.63%. We distributed 900 questionnaires in cooperative hospitals. A total of 793 valid questionnaires were collected, and the valid response rate was 88.11%. In total, 1,250 questionnaires were collected, and the valid response rate was 86.81%.2.4 Patient satisfaction calculation method
The calculation method used for the overall satisfaction score is very typical in the international literature. Patients’ satisfaction score = (number of very dissatisfied patients * 1 + number of quite dissatisfied patients * 2 + number of neither satisfied nor dissatisfied patients * 3 + number of quite satisfied patients * 4 + number of very satisfied patients * 5)/total number of participants in the evaluation [29-31].
2.5 Statistical analysis
The data were double-entered using Epidata 3.1 software to establish a database, and SPSS 20.0 was used for statistical analysis. A nonparametric test was used for the age and the average monthly medical expenses. A chi-square test was used for the type of hospital, gender, household registration, length of time at residence, current address, patient type, medical insurance type, chronic disease status, level of understanding of policy and method of understanding. P<0.05 was considered statistically significant.
In logistic regression analysis, this study took the patients' satisfaction evaluation of the medical alliance services as the dependent variable and reduced the dimensionality of the ordered dependent variable as a binary variable (very dissatisfied, quite dissatisfied, and neither satisfied nor dissatisfied were classified as dissatisfied; quite satisfied and very satisfied were classified as satisfied) [25]. The independent variables included the type of hospitals the patients visited, the basic information of the patients (gender, age, household registration, current address, length of time at residence, chronic disease status, patient type, medical insurance type, and average monthly medical expenses), and the patients’ cognition of medical alliances (levels of understanding and policy) (see Table 1). Model 1 included the hospital type. Then, we introduced the confounding factor of hospital type in the next two steps. Model 2 included the type of hospital and patients’ basic information (gender, age, household registration, current address, length of time at residence, chronic disease status, patient type, medical insurance type, and average monthly medical expenses), and Model 3 included the type of hospital, patients’ basic information, and patients’ cognition of medical alliances (level of understanding of policy and method of understanding).
Table 1 Variable assignment
Characteristics
|
Variable
|
Assignment
|
Patients’ satisfaction
|
Y
|
1=Dissatisfied; 2=Satisfied
|
Type of hospital
|
X1
|
0=Core hospital; 1=Cooperative hospital
|
Gender
|
X2
|
0=Male; 1=Female
|
Age (years)
|
X3
|
1=0-20; 2=21-40; 3=41-60; 4=61-80; 5=81-100
|
Household registration
|
Urban area
|
X4
|
0=No; 1=Yes
|
Suburbs
|
X5
|
0=No; 1=Yes
|
Nonnative
|
X6
|
0=No; 1=Yes
|
Residence time
|
X7
|
1=Within half a year; 2=Half a year to one year; 3=One to two years; 4=More than two years
|
Current address
|
Capital functional core area
|
X8
|
0=No; 1=Yes
|
Urban functional development area
|
X9
|
0=No; 1=Yes
|
New area of urban development
|
X10
|
0=No; 1=Yes
|
Ecological conservation development area
|
X11
|
0=No; 1=Yes
|
Nonnative
|
X12
|
0=No; 1=Yes
|
Chronic disease status
|
X13
|
0= No chronic diseases; 1= Any chronic diseases
|
Patient type
|
X14
|
0= Inpatient; 1= Outpatient
|
Medical insurance type
|
UEBMI (Urban employee-based medical insurance)
|
X15
|
0=No; 1=Yes
|
URBMI (Urban resident-based medical insurance)
|
X16
|
0=No; 1=Yes
|
NMI (National medical insurance)
|
X17
|
0=No; 1=Yes
|
NRCMS (New rural cooperative medical scheme)
|
X18
|
0=No; 1=Yes
|
CI (Commercial insurance)
|
X19
|
0=No; 1=Yes
|
Out of pocket
|
X20
|
0=No; 1=Yes
|
Average monthly medical
expenses (yuan)
|
X21
|
1=Less than 300; 2=301-500; 3=501-800; 4=801-1000; 5=More than 1001
|
Level of understanding of medical alliance policy
|
X22
|
1= Very little understanding; 2= Little understanding; 3= General understanding; 4= Some understanding; 5= High level of understanding
|
Method of understanding
|
Media reports
|
X23
|
0=No; 1=Yes
|
Community promotion
|
X24
|
0=No; 1=Yes
|
Hospital promotion
|
X25
|
0=No; 1=Yes
|
Recommendations from relatives and friends
|
X26
|
0=No; 1=Yes
|
Others
|
X27
|
0=No; 1=Yes
|