2.7. Statistical analysis

Revman5.3 software was used for statistical analysis of the test results. The binary variables were analyzed by relative risk (RR), and the continuous variables were analyzed by mean difference (MD) or standard mean difference (SMD), Both were expressed by 95% confidence interval (CI). If P > 0.1 and I² < 50%, the results are considered to be homogeneous, and the fixed effect model (FEM) is used for combined analysis; If P ≤ 0.1 and I² ≥ 50%, it means that there is heterogeneity among the research results. Try to find out the factors that cause the heterogeneity, and use subgroup analysis to calculate the statistics with heterogeneity. If there are no factors that cause clinical or methodological differences, or after sub-group analysis and processing, the results of the same type are still heterogeneous, use a random-effect model (REM). Combined statistics need to use the method of hypothesis test for Z test, P < .05 can be considered that the outcome index is statistically significant. The results of meta-analysis were shown by forest map, in which the vertical line of forest map was invalid, and its value was not statistically significant. Each horizontal line is the upper and lower limit of 95% confidence interval of each study, and its length represents the size of the confidence interval, in which the square represents the statistics of each test, and the square size represents the weight of the test. When the horizontal line falls on the left or right side of the vertical line, it indicates that the test has statistical significance; On the contrary, there was no statistical significance. We will contact the original author of the paper to obtain the complete data if there is definite data in the included literature.