Speaker 1: Hello everybody, welcome to my channel. This is Ravi and I am going to explain you today clinical SAS topic 37 that is common statistical methods for clinical research. It is a part one and soon I will go and do the second part for this common statistical method for clinical research. And here you should know what is data. Data is nothing but responses collected from sample by the way of asking logical and unambiguous questions like you know unclear uncertain questions. How we are getting the data? We are asking something to subjects or any responses we are collecting from the subjects or sample that is called data. You should know what is data. And data is two types one is categorical data another one is continuous data. What is categorical data or classification data or qualitative data? The data which can be grouped is known as categorical data. If you see here something should be grouped like in sex, male and female. There are only two groups in case of race Asian, Hispanic, etc. In case of ethnicity non-Hispanic and you know other groups there should be a group right in ARMCD also A and B, C, D this should be group. This type of group data is known as categorical data. And you can see here the sex and ethnicity Hispanic, non-Hispanic, non-reported unknown. In case of race American, Indian or Alaskan native, Asian, black or African native, white unknown. These type of group races are present. So these variables we can consider as categorical data. What is continuous data or analysis data or quantitative data? The data which ranges from one value to another value. Like you know in case of age you can see this should be a numeric and it ranges from something 20 to 80 or any age should be happen to subject. In case of base, change from baseline these are the values, these are the variables we can consider as continuous data. Here you can see some example. Treatment variable comes under categorical data. Lab values are continuous data. In case of gender categorical data. Age is continuous data. Height, time, weight is continuous data. In case of race categorical data and change from baseline is a continuous data. What type of statistics we have in clinical research? There are descriptive statistics and inferential statistics. What is descriptive statistics? Nothing but it gives the details like you know n mean, median, mode, minimum, maximum, standard deviation. These type of things are summary descriptive statistics. And what are the inferential statistics? Nothing but the conclusion which we have taken based on the p value. We have available data. We should do some statistics and getting the p value. P value is nothing but probability value. Based on the probability value we can consider whether the drug is actively working or while we comparing with some standard two reference drug whether the two drugs have any significant difference or not. We can conclude based on the probability value. Those type of statistics we can consider as inferential statistics. And here what is hypothesis? Any statement which is under analysis called hypothesis. For example a statement which says there is no significant difference between the compared group that is called null hypothesis or fail to reject null hypothesis. Means if we are comparing two drugs if there is no difference between those drugs we can consider that is null hypothesis. A statement which says there is a significant difference that is reject null hypothesis. And here if the p value less than 0.05 then statistically significant difference is there. If it is greater than 0.05 there is no statistically significant difference is present. So based on the p value we can conclude whether the drug is significantly different or not significantly different. And while doing the data and assigning some statistical procedure what we have to do. First step is check whether the data is continuous data or categorical data. And second step is does we need to do the descriptive statistics or inferential statistics. And third one is third step is how many variables is there to analyze. So how many responses is there for each categorical data. And whether it is a normally distributed or not normally distributed. Whether it is a balanced or not balanced. So based upon these criteria we can do the we can assign the statistical procedure for the data analysis. In the categorical data analysis first four steps we have to do and continuous first three and last two. There is no responses consider for the continuous data statistical procedure assignment. And here what is the normally distributed not normally distributed data. If you see here normally distributed means mean median mode should be equal skewness and kurtosis should be zero or sample size should be less than three. Within this one criteria if meet the data we can call it as normally distributed data. And here these are the procedures. Statistics, categorical data analysis, continuous data analysis and proc or prokaryotic logistic. The categorical data analysis there is two types. Descriptive statistics analysis we use the proc freq procedure and inferential p-value statistics we use the same proc freq procedure with option. I have done a video for this proc freq with options. You can go and see my site SAS interview question number seven proc freq p-value video. And here for the continuous data analysis there is a inferential statistics and descriptive statistics. And for doing the descriptive statistics we can use proc means or proc summary for getting n mean median minimum and maximum related statistics. For case of inferential statistics for the continuous data you can check whether it is a normally distributed or not normally distributed. If it is not normally distributed we can directly go and use proc n par one way procedure. If it is a normally distributed then we can choose proc t-test or proc ANOVA proc GLM and proc mix. If in case if it is a balanced data we just go for the proc ANOVA procedure. If it is we don't know whether it is a balanced or unbalanced so we can choose proc GLM and proc mix procedure for analysis purpose. And what is the proc means procedure? In the proc means procedure it is used to getting the statistics like n mean standard deviation minimum and maximum. Here if you want to analysis this variable so you can carve in the where statement you can use that one and which requires statistics you can mention here. By using the proc means you will get the continuous data summary descriptive statistics. And in case of proc t-test procedure it is used for when we are using it is a continuous data and it is a normally distributed we can use the proc t-test procedure. The t-test procedure perform the t-test for one sample test two sample test and paid observation. The one sample t-test compare the mean of the sample given number two sample test compare the mean of first sample minus mean of the second sample and the paired observation t-test compare the mean of the difference in the observation of a given number. The one sample t-test you can see here this is the data only one variable time variable. By like this we can create this data set and writing the proc t-test alpha equal to 0.1 x y equal to 80 and where equal to time we will get all the statistics here. Here lower limit upper limit standard deviation etc statistics we are getting by using proc t-test. As well as here we can get PR greater than mod t 0.0329 this is the probability value we are getting proc t-test one way sample test. In case of two sample test are group t-test we can mention the class variable and where variable which is the classification variable here and where variable. We want to compare the values obtained from the two different group and if your group is independent to each other and the data is normally distributed in each group then group t-test can be used. If it is two variables and it should be a you know independent to each other so the group variable is categorical data type so we can use this gender variable in the class section and where variable in score is a where variable. By using you can get all these options like coherent and CI option you can use and you will get the difference between this male and female and this score difference versus 1 minus 1 2 means female versus male how much difference it is which method you want you can pick up in the data set level. If you are want to getting the data set you can write the ODS option and you will get all the data set name which one you want you can write it and take data from the data set and the p values PR greater than t value we are representing this p values and paid observation test for example this is BP before and after and parent before versus after it is a paired observation both are one depend on other and you will get this statistics like you know before and after difference and you will get the probability value here. The paired statement used to test whether the mean change in systolic blood pressure is significant difference 0 or not. If you want I will do second session for this remaining processes like PROC mixed and GLM and ANOVA it will take some time so will soon I will do this video also if you want if you like this video please like share comment thanks for watching this video for more interesting SAS interview topics please subscribe my channel thank you.