Questions: Part1: What is difference between informat and format A)Informat is used to read the variable from rawdata,

where as format is used to print the data read in declared format. How to use macro, global or local After I have defined a macro I can invoke it by adding the percent sign prefix to its name like this: % macro name. How to create a bunch of similar tables just has different conditions apply to How to read a excel file to a SAS dataset proc import datafile='C:\samp1.xls' out= Harshamanual dbms=excel ; run; How to do many to many merge
Many-many merge can be performed using PROC SQL. Though this many-many can be done with MERGE statement, it gives a message in the file saying MERGE HAS ONE OR MORE DATA SETS WITH REPEATS OF BY VARIABLES which means MANY-MANY merge has been performed using data step instead of proc sql. PROC SQL; CREATE TABLE PVR SELECT PVR1.ID PVR1.NAME PVR2.ST_CODE FROM PVR1 PVR2 WHERE PVR1.ID=PVR2.ID; QUIT;

Have you create any CDISC data I have created Cdisc Standard SDTM and ADaM datasets. What will you look at in log when doing merge If performing many to many merge I will look if any error occurs stating repeats of by variables. Or else will look at the total observations in datasets merged whether the merge occurred properly. What plots have been done, what procedure used I have done plot for lab parameters, Ihave used Box plot. Endpoints of oncology studies have been done Pfs, Over all survival, Duration of response, Best confirmed response CR,PR,SD,PD. Have you done ISS, what have been done I have Stacked Analysis Datasets from Individual Projects create an integrated analysis dataset. I followed pooling document with information on intergreating datasets. Part2: introduction....

1. Ur introduction, recent projects, oncology projects experience

2. about oncology project,and particular experience under each and every client on oncology

study.

3. How do u validate ADS and TLGs Validation we use proc compare to validate datasets. Other way by independent programming we write own code and check the output visual and see whether the dataset matches. With tables, by manual method we will check numbers etc in the output structure. 4. About CDISC experience, created any specs ?? hw much exp? I have experience in creating sdtm datasets from rawdata follwing mapping document. Creating adam datasets from sdtm following specs. Yes I created specs for each based on mock shells and sap. The specs contain Domainlist dataset, label, location, datastructure, keys,documentation. Codelist Each domain-variablename, label,length,core,format,origin,comments. 5. domains i have worked on list them Sdtm:dm,ae,vs,lb,ex,eg,ds,cm,ta,ti,tu,tr. ADam:adsl,adm,aae,avs,alb,aeg,aeff,atrsp. 6. They asked me what data you have in DS dataset In ds u have disposition data, subjects discontinued from study, reason of discontinuation. Variables like dedecod(completed,losttofollowup,withdraw,consent),dsspid,dscatm,epoch(screening,trt,phas e/followup),dsterm(obtained,randomized,completed). 7. Have you created any macros.Asked questions on macros based on the explanation. I have used macros, mostly in tables. To simplify the code and make work efficient I had macrotized proc sort-proc means-proc freq-proc transpose. I have created macros, one for imputing partial missing dates accroding to iso8601. Other macro was for validation, FOR CHECKING DATASET EXISTENCE,FOR CHECKING
PARAMETERS EXISTENCE, FOR CHECKING OUTPUT DATA EXISTENCE,FOR GENERATION OF STATS REQUIRED,FOR ORDERING STATS REQUIRED.

8. Asked if I worked on any Oncology studies.. asked me about the endpoints in it. Yes I worked on oncology studies, the endpoints were PFS, Overall Surivival, Duration of response, best confirmed response(CR,PR,SD,PD). 9. Asked me to explain about current study I am working on, what were the endpoints, what my

role was. My area in current study was on oncology, breast cancer. Most of my job was working on analysis datasets and generation of TLGs. The endpoints of my study was PFS, Overall survival and Duration of response rate. 10. What types of tables and graphs did u work on? Demographics Baseline charactersitics Summary of ae Incidence of ae by soc&pt Vitals- change from baseline Stats of ex by study drug Subject disposition Ecog performance status Efficacy tables: Summary of events & subject follow up Best valid confirmed response PFS summary of events Overall survival Graphs: KM plot of PFS KM plot for duration of response KM plot for over all survival KM plot duration of stable disease KM plot duration of CR Box plot for Lab test parameters 11. About ISS and ISE what domains did you work on? Integrating ADaM datasets 12. Questions about data step and proc SQL merging and differences. Repeats of by variable, using many to many merge. 13. did u create any domians. what are they SDTM Domains, DM AE VS LB 14. When you are given a spec and a dataset, you dont know anything about that data. what will you do ?

15. questions ON LAB data sets 16. KM graph? 17. What do you have on the x and y-axis for a Kaplan meier curve Y-Axis probability X-Axis -- Time

18. What do u have in a Efficacy table in oncology studies they framed questions from what I answered them. 19. what do u do ? programming or validation? Are you a primary programmer for that study? Programmer 20. How many patients are there in ur study ? Mine is a phase 2 study with 600subjects 21. particularly how to you to validation?steps? Independent programming/parallel programming 22. How do you create analysis data sets? From mock shells and sap specification is developed. Based on specification Analysis datasets are developed. 2) How many people are there in ur team and how many for programming and validation? 3) How many trials ure working on, which phases? phase 1 2 and 3 4) How much time you worked on oncology study? Like 70-80% or most of the time I am working on oncology study or phase3 studiesthat company mainly works on Oncology studies and phse3 trials. 5) How do you create analysis data sets? 7) What analysis datasets youre working on or created so far? 8) Progression and survival definitions (OS, PFS) - here you have to explain how to calculate the time to progression and survival time PFS- Progression free survival defines the time up until the patient progresses or death occurs from first dose date. OS-over all survival defines time from the first dose date till death due to any cause or if patient alive the last date the subject was censored alive. 9) Did you work on Tables, listings and plots.if yes you have to explain what

type of tables and plots you worked on and as programmer and validator

10) What is the validation process for tables, listings, datasets and plots?

11) What is procedure to create plots ..Here you have to explain about Kaplan Maier

plots and others if you know any.

12) What kind of tables you worked on..safety and efficacy tables and give some

examples like pfs tables, ORR and overall survival tables.

13) Do you use any macros to generate to tables Yes ,ingeneral Proc freq ,procmean macros. Then to Mapppi index macro to map footnotes and titles. 14) How much % you worked on SDTM data domain deriving 15) Shell ask you each clients places and team members and what type of trials you worked on 16) What do you know about CDISC?
Clinical data interchange standard consortium, it CDISC is a global, open, multidisciplinary, non-profit organization that has established standards to support the acquisition, exchange, submission and archive of clinical research data and metadata.

17) What are the ADAM data sets you created?

18) How many analysis datasets you created shell ask you same question for each

19) Shell ask you without specification how would you work on datasets and tables. Assuming we have analysis datasets provided , though specifications are not given it means we have mock shells, following mock shells I would create tables using the datasets provided. 20) What therapeutic areas have you worked on

21) What phases have you worked on

22) Combining datasets.how do you doshell focus on merging

23) What is procedure for merging? Datastep Merge Proc sql many to many

24) Merging types and shell ask you how to do many to many merging Proc sql 25) How to create macro variables %let ,Callsymput,%macro%mend, proc sql into:, 26) How do you validate datasets Proc comapre 27) Witout specification how will you create analysis datasets.. question repeated thrice 28) Proc lifetest, ODS output detail Ods is used for output delivery system, Syntax: Ods pdf file= .pdf; Ods pdf close;

29) types of merge, how to do that, where proc sql is preferred? (many to many.. give

30) deatils about end points in adeff (efficacy data set) in onco project

Onyx Questions

1. Difference between Milestone and Epoch.

2. Merge using Data step and Proc SQL

3. How do u validate ADS and TLGs

4. About CDISC experience, created any specs ?? I said I have created SDTM mapping documents,

and worked on ADAM

5. What domains have you worked on list them (Special purpose, Interventions, Events, Findings)

and (domain names DM, DS, EX, LB, VS, CO, AE etc)

6. They asked me what data you have in DS dataset

Ans: Identifier variables, timing variables, reason for study discontinuation, gave some

examples on controlled terminology used for DSDECOD, and the issues I face resolving them in

my Integration study Tell it depending on what they ask you.

7. Have you created any macros.. Asked me to elaborate, what was the purpose to create them..

Asked questions on macros based on the explanation.

8. Asked if I worked on any Oncology studies.. asked me about the endpoints in it.

9. Asked me to explain about current study I am working on, what were the endpoints, what my

role was.

10. About ISS and ISE what domains did you work on, what were the issues you faced when pooling this data.

11. About Medra coding, how did I merge with AE dataset.

12. Questions about data step and proc SQL merging and differences.

Ans: Eg: If both datasets a and b have variables usubjid and race and by statement will not have

race in it how is it handled in data step and proc SQL

Data new;

Merge a b;

By usubjid;

Run;

My answer was the race from 1

st

dataset will be replaced by second one and the length

will be the length of RACE from second dataset (Not sure if I am correct on this.)

For Proc SQL I told if both have race on them and not merged on race (it will give variable

already exists on file warning). Have to cross check these answers from someone

13. Questions on Proc Transpose (What is reason behind the below error?)

Ans: Eg: proc transpose data=temp out=temp1;

Id datanumber;

This gave me error "The ID value " " occurs twice in the same BY group".

When you get that error it is telling you that you have multiple data points for one or more

variables that you are trying to create. SAS can force the transpose and delete the extra

datapoints if you add "let" to the proc transpose line. (Data should have unique values else we

get this error).

Eg: proc transpose data=temp out=temp1 let;

Id datanumber;

14. When you are given a spec and a dataset, you dont know anything about that data.. what will you do ?

Ans: Run a proc contents to know about the structure of the data, variables and their types and then check if the spec is mapped to source data correctly, and then do edit checks using proc means(continuous, numeric variables) and frequency(categorical, character variables)

15. Explain about CTC grades and recist criteria- How did u derive the CTC grades? (followed the spec to derive them)

16. How did u get the Low normal High ranges in LABS (Our raw data already had Low, Normal and High ranges, we dint derive the Low and High)

17. What do you have on the x and y-axis for a Kaplan meier curve (it is the time vs cumulative frequency)

18. What do u have in a Efficacy table in oncology studies (N, median, hazard ratio, proc phreg to calculate hazard ratio)

19. A dataset has 25 variables 1 suppose these numeric variables have some missing values and I want to convert those missing . values to 0 how can we do that ?

Ans: http://studysas.blogspot.com/2009/03/change-all-missing-values-of-all.html (refer this example) Using array function we can resolve this issue, We can do ;
data new; set old; array zero score1-score25; do over zero; if zero=. then zero=0; end; run;

20. How will a team be successful? 21. Time management? 22. Typical work day ?

23. How do you handle tight timelines ?

24. If at all you are missing a deadline, how do you handle and whom do you report ?

25. Mostly they tried to frame questions from what I explained them.

s 1 is usubjid which is a character, rest are all numeric (a,b,c..),