RESUMO
The purpose of this paper was to introduce the linear trend χ2 tests for the data of the two kinds of one-way ordinal two-dimensional contingency tables and the implementation of SAS and R software. The two one-way ordered two-dimensional contingency tables referred to the “R×2 contingency table” with R values ordinal variable as the cause variable and the “2×C contingency table”with C values ordinal variable as the outcome variable. Their linear trend tests required the use of the Cochran-Armitage’s linear trend χ2 test and the Lee’s linear trend χ2 test， respectively.
RESUMO
The purpose of this paper was to introduce the methods of the "layered person-time data" rate ratio analysis and the SAS implementation. The specific contents included the following four aspects： ① the point estimation and the confidence interval estimation of the layer i rate ratio in the stratified person-time data； ② the homogeneity test of the rate ratios among the different layers； ③ the point estimation and the confidence interval estimation of the common rate ratio； ④ the linear trend test of the incidence density in the stratified person-time data. Through two examples， it showed the statistical calculation process based on the SAS software to realize the aforementioned four kinds of the analysis contents， including providing the SAS program code， the explanation of the SAS output results， and giving the statistical and professional conclusions.
RESUMO
ObjectiveTo understand the status and relationship between social support level and Internet addiction among left-behind middle school students， so as to provide references for the targeted intervention and risk reduction of Internet addiction. MethodsStratified cluster random sampling method was used to recruit 3 532 middle school students from six middle schools in a county of Southeast Chongqing， including 1 945 （55.07%） left-behind students and 1 587 （44.93%） non-left behind students. All the selected students were assessed using self-designed general situation questionnaire， Diagnostic Scale of Internet Addiction for Middle School Students （DSIA-MSS） and Adolescent Social Support Rating Scale （ASSRS）， then the trend test was performed on relationships of social support to Internet addiction. Results① The detection rate of Internet addiction in the left behind group was higher than that in the non-left behind group （22.98% vs. 16.13%， χ2=25.732， P<0.01）. ② ASSRS score in the left behind group was lower than that in the non-left behind group ［（66.81±15.92） vs. （68.19±16.18）， t=2.544， P=0.011］. ③ Among three different living states of students （non-left behind， one-parent working outside， both-parents working outside）， the detection rate of Internet addiction demonstrated significant differences among students with different social support levels （χ2=35.220， 37.662， 31.434， P<0.01）. ④ A "dose-effect" relationship of social support to Internet addiction was found， indicating that students with higher social support levels suffered a lower risk of Internet addiction （P<0.01）. ConclusionLeft-behind middle school students develop a prominent problem of Internet addiction， which is closely related to their social support level， and the two aspects display a negative "dose-effect" relationship.
RESUMO
The purpose of the paper was to introduce the three special tests of the survival data and the SAS implementation. Specifically， it was the multiple comparisons， the trend test and the covariate test of the survival data. The multiple comparisons involved two situations： "the pairwise comparison" and "the comparison with control group". In the trend test， it involved two algorithms： "the log-rank test" and "the Wilcoxon test". In the covariate test， it involved "the single covariate test method" and "the multi-covariate test method of adding one covariate step by step". With the help of the SAS software and based on an example， this article implemented the three special tests mentioned above， explained the output results， and made statistical and professional conclusions.
RESUMO
Objective To assess the changing trend of serum uric acid (SUA) levles in health check-up subjects of Tianjin municipality 2007-2015.Method Data were collected from 266 043 subjects receiving annual health check-up from 2007 to 2015 in Tianjin municipality.Changing trends of SUA levels and incidence of hyperuricemia were analyzed during the 9 year period.Results From 2007 to 2015 SUA levels in males were increased from (347 ± 73) μmol/L to (373-± 78) μmol/L,while in females were increased from(253 ± 59) μmol/L to (267-± 61) μmol/L.The trend test showed that both for men and woinen the incidence of hypemricemia was increased in each year (men:x2 =270.30,women:x2 =15.10,both P =0.00).The incidence of hyperuricemia by each year was increased more markedly in men than that in women (x2 level:194.80-1 877.00,all P =0.00).The incidence of hyperuricemia in young adults (20-39) was higher than that in middle-ages (40-59),and the latter was higher than that in elderly (＞ 60).For the 7-year (2009-2015) incidence of hyperuricemia in young and middle-aged men,the trend test showed statistical siganificance (x2 level:14.11-233.90,all P =0.00).By contrast,the incidence hyperuricemia in women was increased with aging,and for the 9-year incidence of hyperuricemia in young women,the trend test showed statistical siganificance (x2 level:115.30-480.20,all P =0.00).Conclusion The incidence rates of hyperuricemia are high in Tianjin health check-up subjects,especiallly in young males and old females.There is a time trend of increasing incidence of hyperuricemia in the young and middle-aged subjects.The results indicate that the prevention and control of hyperuricemia in Tianjin are of clinical and social significance.
RESUMO
We described the time trend of acute myocardial infarction (AMI) from 1999 to 2013 in Tianjin incidence rate with Cochran-Armitage trend (CAT) test and linear regression analysis,and the results were compared.Based on actual population,CAT test had much stronger statistical power than linear regression analysis for both overall incidence trend and age specific incidence trend (Cochran-Armitage trend P value＜linear regression P value).The statistical power of CAT test decreased,while the result of linear regression analysis remained the same when population size was reduced by 100 times and AMI incidence rate remained unchanged.The two statistical methods have their advantages and disadvantages.It is necessary to choose statistical method according the fitting degree of data,or comprehensively analyze the results of two methods.
RESUMO
We described the time trend of acute myocardial infarction (AMI) from 1999 to 2013 in Tianjin incidence rate with Cochran-Armitage trend (CAT) test and linear regression analysis,and the results were compared.Based on actual population,CAT test had much stronger statistical power than linear regression analysis for both overall incidence trend and age specific incidence trend (Cochran-Armitage trend P value＜linear regression P value).The statistical power of CAT test decreased,while the result of linear regression analysis remained the same when population size was reduced by 100 times and AMI incidence rate remained unchanged.The two statistical methods have their advantages and disadvantages.It is necessary to choose statistical method according the fitting degree of data,or comprehensively analyze the results of two methods.
RESUMO
AIM:To introduce and evaluate two methods for sample size calculation in doseresponse research.METHODS:The methods used for sample size calculation in dose-response research include Cochran-Armitage trend(CAT) test proposed by Jun-mo Nam and the Unified contrast(CUC) method proposed by Chang. This paper takes the example of binary data,uses the two methods to calculate the required sample size and evaluate the power by simulation under the assumption that the response rate had a linear relationship with dosage under the logit scale.RESULTS:When the probability of response followed a linear trend on the logit scale, the sample sizes estimated from the two methods are approximately equal in various parameter's settings.The simulated power of CAT test was close to the expected one,however,the simulated power of CUC method was obviously affected by the values of contrast coefficient,if a contrast whose coefficients parallel the expected response, the CUC method was more powerful. CONCLUSION:When the probability of response follows a linear trend over dosage on the logit scale,the two methods reach almost the same result,in such circumstance,the Chang's unified contrast method will obtain an increasing power.